PART 2 Coran Pediatric Surgery, 7th Ed PDF

CHAPTER 28
Principles of
Pediatric Oncology,
Genetics of Cancer,
and Radiation
Therapy
Matthew J. Krasin and Andrew M. Davidoff
A number of milestones in the evolution of cancer therapy

have come from the field of pediatric oncology. The first clear
evidence that chemotherapy could provide effective treatment for childhood malignancy occurred in 1950 when
Farber reported temporary cancer remission in children with
acute lymphoblastic leukemia (ALL) treated with the folic
acid antagonist aminopterin.1 The first successful use of
a multidisciplinary approach to cancer treatment occurred
in the 1960s and 1970s through the collaborative efforts of
pediatric surgeons, radiation therapists, and pediatric oncologists aiming to improve the treatment of Wilms tumor in
children.2 Such a multidisciplinary approach is now used
throughout the field of oncology. The successful use of a
combination of chemotherapeutic agents to cure Hodgkin
disease and ALL during the 1960s led to the widespread
use of combination chemotherapy to treat virtually all types
of cancers. Since the late 1980s, neuroblastoma has been the

paradigm for the use of therapies of variable intensity,
depending on risk stratification determined by clinical and
biological variables, including molecular markers. Other advances in pediatric oncology have included the development
of interdisciplinary, national cooperative clinical research
groups to critically evaluate new therapies, the efficacy of
dose-intensive chemotherapy programs in improving the
outcome of advanced-stage solid tumors, and the supportive
care necessary to make the latter approach possible. The
development and application of these principles and advances have led to substantially increased survival rates for
children with cancer and profound improvements in their
quality of life.
Additionally, advances in molecular genetic research in
the past 3 decades have led to an increased understanding
of the genetic events in the pathogenesis and progression
of human malignancies, including those of childhood.
A number of pediatric malignancies have served as models
for molecular genetic research. Chromosomal structural
changes, activating or inactivating mutations of relevant
genes or their regulatory elements, gene amplification, and
gene imprinting may each play a role in different tumor
types. In some instances, these genetic events occur early
in tumorigenesis and are specific for a particular tumor type,
such as the chromosomal translocation t(11;22)(q24;q12) in
Ewing sarcoma; other aberrations occur in a variety of different tumor types and are almost always associated with additional genetic changes, such as chromosome 1p deletion in
neuroblastoma and Wilms tumor. Some alterations involve
oncogenesgenes that, when activated, lead directly to
cancerwhereas others involve tumor suppressor genes,
whose inactivation allows tumor progression. The result of
alterations in these genetic elements, regardless of the mechanism, is disruption of the normal balance between proliferation and death of individual cells. These discoveries have
highlighted the utility of molecular analysis for a variety of
purposes, including diagnosis, risk stratification, and treatment planning; the understanding of syndromes associated
with cancer; genetic screening and genetic counseling; and
prophylactic treatment, including surgical intervention.
Soon, treatment regimens are likely to be individualized on
the basis of the molecular biological profile of a patients tumor. In addition, molecular profiling will lead to the development of new drugs designed to induce differentiation of
tumor cells, block dysregulated growth pathways, or reactivate silenced apoptotic pathways.
Epidemiology and Survival

Statistics
------------------------------------------------------------------------------------------------------------------------------------------------
Cancer in children is uncommon; it represents only about

2% of all cancer cases. Nevertheless, after trauma, it is the
second most common cause of death in children older than
1 year. Each year, approximately 130 new cases of cancer are
identified per million children younger than 15 years (or
about 1 in 7000). This means that in the United States, about
9,000 children younger than 15 years are diagnosed with
cancer each year, in addition to 4,000 patients aged 15 to
19 years.3 Leukemia is the most common form of cancer
397
398
PART III
MAJOR TUMORS OF CHILDHOOD
TABLE 28-1
Frequency of Cancer Diagnoses in Childhood
Molecular Biology of Cancer

------------------------------------------------------------------------------------------------------------------------------------------------
Type of Cancer
Percentage of Total
Leukemia
Brain tumors
Lymphoma
Neuroblastoma
Sarcoma
Wilms tumor
Osteosarcoma
Retinoblastoma
Liver tumors
30
25
15
8
7
6
5
3
1
in children, and brain tumors are the most common solid

tumor of childhood (Table 28-1). Lymphomas are the
next most common malignancy in children, followed by
neuroblastoma, soft tissue sarcomas, Wilms tumor, germ cell
tumors, osteosarcoma, and retinoblastoma. A slightly different distribution is seen among 15- to 19-year-olds, in whom
Hodgkin disease and germ cell tumors are the most frequently diagnosed malignancies; non-Hodgkin lymphoma,
nonrhabdomyosarcoma soft tissue sarcoma, osteosarcoma,
Ewing sarcoma, thyroid cancer, and melanoma also occur
with an increased incidence.
In general, the incidence of childhood cancer is greatest
during the first year of life, peaks again in children aged 2
to 3 years, and then slowly declines until age 9. The incidence then steadily increases again through adolescence.
Each tumor type shows a different age distribution pattern,
however. Variations by gender are also seen. For example,
Hodgkin disease, ALL, brain tumors, neuroblastoma, hepatoblastoma, Ewing sarcoma, and rhabdomyosarcoma are
more common in boys than in girls younger than 15 years,
whereas only osteosarcoma and Ewing sarcoma are more
common in boys than in girls older than 15 years. However,
girls in the older age group have Hodgkin disease and thyroid
cancer more frequently than boys do. Distribution also varies
by race: White children generally have a 30% greater incidence of cancer than do black children. This difference is
particularly notable for ALL, Ewing sarcoma, and testicular
germ cell tumors. The probability of surviving childhood
cancer has improved greatly since Farber induced the first remissions in patients with ALL. In the early 1960s, approximately 30% of children with cancer survived their disease.
By the mid-1980s, about 65% of children with cancer were
cured, and by the mid-1990s, the cure rate had increased to
nearly 75%.4 Currently, greater than 80% are cured. These
great strides have resulted from three important factors:
(1) the sensitivity of childhood cancer, at least initially, to
available chemotherapeutic agents; (2) the treatment of
childhood cancer in a multidisciplinary fashion; and (3)
the treatment of most children in major pediatric treatment
centers in the context of a clinical research protocol using
the most current and promising therapy. Although progress
in the treatment of some tumor types, such as ALL and
Wilms tumor, has been outstanding, progress in the treatment of others, such as metastatic neuroblastoma and rhabdomyosarcoma, has been modest. Therefore there is still
a need for significant improvement in the treatment of
childhood cancer.
During normal cellular development and renewal, cells evolve

to perform highly specialized functions to meet the physiologic needs of the organism. Development and renewal
involve tightly regulated processes that include continued cell
proliferation, differentiation to specialized cell types, and
programmed cell death (apoptosis). An intricate system of
checks and balances ensures proper control over these physiologic processes. The genetic composition (genotype) of a cell
determines which pathway(s) will be followed in exerting that
control. In addition, the environment plays a crucial role in
influencing cell fate: Cells use complex signal transduction
pathways to sense and respond to neighboring cells and their
extracellular milieu.
Cancer is a genetic disease whose progression is driven by a
series of accumulating genetic and epigenetic changes influenced by hereditary factors and the somatic environment.
These changes result in individual cells acquiring a phenotype
that provides them with a survival advantage compared with
surrounding normal cells. Our understanding of the processes
that occur in malignant cell transformation is increasing; many
discoveries in cancer cell biology have been made by using
childhood tumors as models. This greater understanding of
the molecular biology of cancer has also contributed significantly to our understanding of normal cell physiology.
NORMAL CELL PHYSIOLOGY

Cell Cycle
Genetic information is stored in cells and transmitted to subsequent generations of cells through nucleic acids organized
as genes on chromosomes. A gene is a functional unit of heredity that exists on a specific site or locus on a chromosome,
is capable of reproducing itself exactly at each cell division,
and is capable of directing the synthesis of an enzyme or other
protein. The genetic material is maintained as DNA formed
into a double helix of complementary strands. The cell must
ensure that replicated DNA is accurately copied with each cell
division or cycle. DNA replication errors that go uncorrected
potentially alter the function of normal cell regulatory proteins. The molecular machinery used to control the cell cycle
is highly organized and tightly regulated.5 Signals that stimulate or inhibit cellular growth converge on a set of evolutionarily conserved enzymes that drive cell-cycle progression.
Various checkpoints exist to halt progression through the
cell cycle during certain environmental situations or times
of genetic error resulting from inaccurate synthesis or damage.
Two of the most well-studied participants in the cell-cycle checkpoint system are TP53 and retinoblastoma (RB) proteins.6
In normal circumstances, cells divide and terminally differentiate, thereby leaving the cell cycle, or they enter a resting state.
Inactivation of the effectors of cell-cycle regulation or the bypassing of cell-cycle checkpoints can result in dysregulation of the
cell cycle, a hallmark of malignancy.
Signal Transduction
Signal transduction pathways regulate all aspects of cell function, including metabolism, cell division, death, differentiation, and movement. Multiple extracellular and intracellular
CHAPTER 28
PRINCIPLES OF PEDIATRIC ONCOLOGY, GENETICS OF CANCER, AND RADIATION THERAPY
signals for proliferation or quiescence must be integrated by

the cell, and it is this integration of signals from multiple pathways that determines the response of a cell to competing and
complementary signals. Extracellular signals include growth
factors, cytokines, and hormones; the presence or absence
of adequate nutrients and oxygen; and contact with other cells
or an extracellular matrix. Signaling mediators often bind
to membrane-bound receptors on the outside of the cell,
but they may also diffuse into the cell and bind receptors in
the cytoplasm or on the nuclear membrane. Binding of a
ligand to a receptor stimulates the activities of small-molecule
second messengersproteins necessary to continue the transmission of the signal. Signaling pathways ultimately effect the
activation of nuclear transcription factors that are responsible
for the expression or silencing of genes encoding proteins
involved in all aspects of cellular physiology.
Receptors with tyrosine kinase activity are among the most
important transmembrane receptors. Several important transmembrane receptors with protein kinase activity have been
identified and grouped in families on the basis of structural
similarities.7 These families include the epidermal growth
factor receptors (EGFRs), fibroblast growth factor receptors,
insulin-like growth factor receptors (IGFRs), platelet-derived
growth factor receptors (PDGFRs), transforming growth factor
receptors, and neurotrophin receptors (TRKs). Abnormalities
of members of each of these families are often found in pediatric malignancies and therefore are thought to play a role in
their pathogenesis. Characteristic abnormalities of these receptors often form the basis of both diagnostic identification
of certain tumor types and, more recently, targeted therapy
for tumors with these specific abnormalities.
Programmed Cell Death
Multicellular organisms have developed a highly organized
and carefully regulated mechanism of cell suicide to maintain
cellular homeostasis. Normal development and morphogenesis are often associated with the production of excess cells,
which are removed by the genetically programmed process
of cell death called apoptosis. Apoptosis limits cellular expansion and counters cell proliferation. Apoptosis is initiated
by the interaction of death ligands, such as tumor necrosis
factor-a (TNF-a), FAS, and TNF-related apoptosis-inducing
ligand (TRAIL), with their respective receptors. This interaction is followed by aggregation of the receptors and recruitment of adapter proteins to the plasma membrane, which
activate caspases.8 Thus the fate of a cell is determined by
the balance between death signals and survival signals.9
An alternative to cell death mediated by receptorligand
binding is cellular senescence, which is initiated when chromosomes reach a critical length. Eukaryotic chromosomes
have DNA strands of unequal length, and their ends, called
telomeres, are characterized by species-specific nucleotide
repeat sequences. Telomeres stabilize the ends of chromosomes, which are otherwise sites of significant instability.10
With time and with each successive cycle of replication, chromosomes are shortened by failure to complete replication of
their telomeres. Thus telomere shortening acts as a biological
clock, limiting the life span of a cell. Germ cells, however,
avoid telomere shortening by using telomerase, an enzyme
capable of adding telomeric sequences to the ends of chromosomes. This enzyme is normally inactivated early in the
growth and development of an organism. Persistent activation
399
or the reactivation of telomerase in somatic cells appears to

contribute to the immortality of transformed cells.
Malignant Transformation
Alteration or inactivation of any of the components of normal
cell regulatory pathways may lead to the dysregulated growth
that characterizes neoplastic cells. Malignant transformation
may be characterized by cellular de-differentiation or failure
to differentiate, cellular invasiveness and metastatic capacity,
or decreased drug sensitivity. Tumorigenesis reflects the
accumulation of excess cells that results from increased cell
proliferation and decreased apoptosis or senescence. Cancer
cells do not replicate more rapidly than normal cells, but they
show diminished responsiveness to regulatory signals. Positive
growth signals are generated by proto-oncogenes, so named
because their dysregulated expression or activity can promote
malignant transformation. These proto-oncogenes may encode growth factors or their receptors, intracellular signaling
molecules, and nuclear transcription factors (Table 28-2).
Conversely, tumor suppressor genes, as their name implies,
control or restrict cell growth and proliferation. Their inactivation, through various mechanisms, permits the dysregulated growth of cancer cells. Also important are the genes
that regulate cell death. Their inactivation leads to resistance
to apoptosis and allows the accumulation of additional genetic
aberrations.
Cancer cells carry DNA that has point mutations, viral
insertions, or chromosomal or gene amplifications, deletions,
or rearrangements. Each of these aberrations can alter the
context and process of normal cellular growth and differentiation. Although genomic instability is an inherent property of
the evolutionary process and normal development, it is
through genomic instability that the malignant transformation
of a cell may arise. This inherent instability may be altered by
inheritance or exposure to destabilizing factors in the environment. Point mutations may terminate protein translation,
alter protein function, or change the regulatory target sequences that control gene expression. Chromosomal alterations create new genetic contexts within the genome and
lead to the formation of novel proteins or to the dysregulation
of genes displaced by aberrant events.
Genetic abnormalities associated with cancer may be detected
in every cell in the body or only in the tumor cells. Constitutional
or germline abnormalities either are inherited or occur de novo
in the germ cells (sperm or oocyte). Interestingly, despite the
presence of a genetic abnormality that might affect growth
regulatory pathways in all cells, people are generally predisposed
to the development of only certain tumor types. This selectivity
highlights the observation that gene function contributes to
growth or development only within a particular milieu or physiologic context. Specific tumors occur earlier and are more often
bilateral when they result from germline mutations than when
they result from sporadic or somatic alterations. Such is often
the case in two pediatric malignancies, Wilms tumor and retinoblastoma. These observations led Knudson11 to propose a twohit mechanism of carcinogenesis in which the first genetic
defect, already present in the germline, must be complemented
by an additional spontaneous mutation before a tumor can arise.
In sporadic cancer, cellular transformation occurs only when
two (or more) spontaneous mutations take place in the same cell.
Much more common, however, are somatically acquired chromosomal aberrations, which are confined to the malignant cells.
400
PART III
TABLE 28-2
Proto-oncogenes and Tumor Suppressor Genes in Pediatric Malignancies
Oncogene Family
Proto-oncogene
Chromosome Location
Tumors
Growth factors and receptors
ERBB2
TRK
SRC
17q21
9q22
7p11
Signal transducers
Transcription factors
H-RAS
c-MYC
MYCN
11p15.1
18q24
2p24
Glioblastoma
Neuroblastoma
Rhabdomyosarcoma,
Osteosarcoma, Ewing sarcoma
Neuroblastoma
Burkitt lymphoma
Neuroblastoma
Syndrome
Tumor Suppressor Gene
Chromosome Location
Tumors
Familial polyposis coli
APC
5q21
Familial retinoblastoma
RB
13q24
WAGR*
Denys-Drash{
Beckwith-Weidemann{
WT1
WT1
WT2 (?)
11p13
11p13
11p15
Li-Fraumeni
Neurofibromatosis type 1
Neurofibromatosis type 2
TP53
NF1
NF2
17q13
17q11.2
22q12
von Hippel-Lindau
VHL
3p25-26
Intestinal polyposis, colorectal

cancer
Retinoblastoma,
osteosarcoma
Wilms tumor
Wilms tumor
Wilms tumor,
hepatoblastoma, adrenal
Multiple (see text)
Sarcomas, breast cancer
Neurofibroma,
neurofibrosarcoma, brain
tumor
Renal cell cancer,
pheochromocytoma, retinal
angioma, hemangioblastoma
Protein kinase
*WAGR: Wilms tumor, aniridia, genitourinary abnormalities, and mental retardation.

{
Denys-Drash: Wilms tumor, pseudohermaphroditism, mesangial sclerosis, renal failure.
{
Beckwith-Weidemann: multiple tumors, hemihypertrophy, macroglossia, hyperinsulinism.
These aberrations affect growth factors and their receptors,

signal transducers, and transcription factors. The general types
of chromosomal alterations associated with malignant transformation are shown in Figure 28-1. Although a low level
of chromosomal instability exists in a normal population of
cells, neoplastic transformation occurs only if these alterations affect a growth-regulating pathway and confer a growth
advantage.
Abnormal DNA Content
Normal human cells contain two copies of each of 23 chromosomes; a normal diploid cell therefore has 46 chromosomes.
Although cellular DNA content, or ploidy, is accurately determined by karyotypic analysis, it can be estimated by the much
simpler method of flow cytometric analysis. Diploid cells have
a DNA index of 1.0, whereas near-triploid (also termed hyperdiploid) cells have a DNA index ranging from 1.26 to 1.76. The
majority (55%) of primary neuroblastoma cells are triploid or
near triploid (e.g., having between 58 and 80 chromosomes),
whereas the remainder are near diploid (35 to 57 chromosomes) or near tetraploid (81 to 103 chromosomes).12 Neuroblastomas consisting of near-diploid or near-tetraploid cells
usually have structural genetic abnormalities (e.g., chromosome 1p deletion and amplification of the MYCN oncogene),
whereas those consisting of near-triploid cells are characterized by three almost complete haploid sets of chromosomes
with few structural abnormalities.13 Of importance, patients
with near-triploid tumors typically have favorable clinical
and biological prognostic factors and excellent survival
rates compared with those who have near-diploid or neartetraploid tumors.14

Chromosomal Translocations
Many pediatric cancers, specifically hematologic malignancies
and soft tissue neoplasms, have recurrent, nonrandom abnormalities in chromosomal structure, typically chromosomal
translocations (Table 28-3). The most common result of a
nonrandom translocation is the fusion of two distinct genes
from different chromosomes. The genes are typically fused
within the reading frame and express a functional, chimeric
protein product that has transcription factor or protein kinase
activity. These fusion proteins contribute to tumorigenesis by
activating genes or proteins involved in cell proliferation. For
example, in Ewing sarcoma the consequence of the t(11;22)
(q24;q12) translocation is a fusion of EWS, a transcription
factor gene on chromosome 22, and FLI-1, a gene encoding
a member of the ETS family of transcription factors on chromosome 11.15 The resultant chimeric protein, which contains
the DNA binding region of FLI-1 and the transcription activation region of EWS, has greater transcriptional activity than
does EWS alone.16 The EWSFLI-1 fusion transcript is detectable in approximately 90% of Ewing sarcomas. At least
four other EWS fusions have been identified in Ewing sarcoma; fusion of EWS with ERG (another ETS family member)
accounts for an additional 5% of cases.17 Alveolar rhabdomyosarcomas have characteristic translocations between the
long arm of chromosome 2 (75% of cases) or the short arm
of chromosome 1 (10% of cases) and the long arm of chromosome 13. These translocations result in the fusion of PAX3
CHAPTER 28
401
TABLE 28-3
Common, Recurrent Translocations in Soft Tissue Tumors
Tumor
Ewing sarcoma/primitive
neuroectodermal tumor
Desmoplastic small round

cell tumor
Synovial sarcoma
Alveolar rhabdomyosarcoma
Malignant melanoma of
soft part (clear cell sarcoma)
Myxoid liposarcoma
Extraskeletal myxoid
chondrosarcoma
Dermatofibrosarcoma
protuberans and giant cell
fibroblastoma
Congenital fibrosarcoma and
mesoblastic nephroma
Lipoblastoma
Genetic
Abnormality
Fusion
Transcript
t(11;22)(q24;q12)
t(21;22)(q22;q12)
t(7;22)(p22;q12)
t(17;22)(q12;q12)
t(2;22)(q33;q12)
t(11;22)(p13;q12)
t(11;22)(q24;q12)
t(X;18)(p11.23;
q11)
t(X;18)(p11.21;
q11)
t(2;13)(q35;q14)
t(1;13)(p36;q14)
t(12;22)(q13;q12)
FLI1-EWS
ERG-EWS
ETV1-EWS
E1AF-EWS
FEV-EWS
WT1-EWS
FLI1-EWS
SSX1-SYT
SSX2-SYT
t(12;16)(q13;p11)
t(12;22)(q13;q12)
t(9;22)(q22;q12)
CHOP-TLS(FUS)
CHOP-EWS
CHN-EWS
t(17;22)(q22;q13)
COL1A1-PDGFB
t(12;15)(p13;q25)
ETV6-NTRK3
t(3;8)(q12;q11.2)
t(7;8)(q31;q13)
?
?
PAX3-FKHR
PAX7-FKHR
ATF1-EWS
From Davidoff AM, Hill DA: Molecular genetic aspects of solid tumors in
childhood. Semin Pediatr Surg 2001;10:106-118.
FIGURE 28-1 Spectrum of gross chromosomal aberrations using

chromosomes 1 and 14 as examples. HSR, homogeneously staining
regions. (From Look AT, Kirsch IR: Molecular basis of childhood cancer.
In Pizzo PA, Poplack DG [eds]: Principles and Practices of Pediatric
Oncology. Philadelphia, Lippincott-Raven, 1997, p 38.)
(at 2q35) or PAX7 (at 1p36) with FKHR, a gene encoding a

member of the forkhead family of transcription factors.18
The EWS-FLI-1 and PAX7-FKHR fusions appear to confer a
better prognosis for patients with Ewing sarcoma and alveolar
rhabdomyosarcoma, respectively.19,20 Translocations that generate chimeric proteins with increased transcriptional activity
also characterize desmoplastic small round cell tumor,21 myxoid
liposarcoma,22 extraskeletal myxoid chrondrosarcoma,23 malignant melanoma of soft parts,24 synovial sarcoma,25 congenital
fibrosarcoma,26 cellular mesoblastic nephroma,27 and dermatofibrosarcoma protuberans.28
Proto-oncogene Activation
Proto-oncogenes are commonly activated in transformed cells
by point mutations or gene amplification. The classical example of proto-oncogene activation by a point mutation involves
the cellular proto-oncogene RAS. RAS-family proteins are
associated with the inner, cytoplasmic surface of the plasma
membrane and function as intermediates in signal transduction pathways that regulate cell proliferation. Point mutations
in RAS result in constitutive activation of the RAS protein
and therefore the continuous activation of the RAS signal
transduction pathway. Activation of RAS appears to be
involved in the pathogenesis of a small percentage of pediatric

malignancies, including leukemia and a variety of solid
tumors.
Gene amplification (i.e., selective replication of DNA
sequences) enables a tumor cell to increase the expression
of crucial genes whose products are ordinarily tightly controlled. The amplified DNA sequences, or amplicons, may be
maintained episomally (i.e., extrachromosomally) as double
minutes-paired chromatin bodies lacking a centromere or as
intrachromosomal, homogeneously staining regions. In about
one third of neuroblastomas, for example, the transcription
factor and proto-oncogene MYCN is amplified. The MYCN
copy number in neuroblastoma cells can be amplified 5-fold to
500-fold and is usually consistent among primary and
metastatic sites and at different times during tumor evolution
and treatment.29 This consistency suggests that MYCN amplification is an early event in the pathogenesis of neuroblastoma.
Because gene amplification is usually associated with advanced stages of disease, rapid tumor progression, and poor
outcome, it is a powerful prognostic indicator.30,31 The cell
surface receptor gene ERBB2 is another proto-oncogene that
is commonly overexpressed because of gene amplification,
an event that occurs in breast cancer, osteosarcoma, and
Wilms tumor.32
Inactivation of Tumor Suppressor Genes
Tumor suppressor genes, or antioncogenes, provide negative
control of cell proliferation. Loss of function of the proteins
encoded by these genes, through deletion or mutational inactivation of the gene, liberates the cell from growth constraints
402
PART III
and contributes to malignant transformation. The cumulative

effect of genetic lesions that activate proto-oncogenes or inactivate tumor suppressor genes is a breakdown in the balance
between cell proliferation and cell loss because of differentiation or apoptosis. Such imbalance results in clonal overgrowth
of a specific cell lineage. The first tumor suppressor gene to
be recognized was the retinoblastoma susceptibility gene RB.
This gene encodes a nuclear phosphoprotein that acts as a
gatekeeper of the cell cycle. RB normally permits cell-cycleprogression through the G1 phase when it is phosphorylated,
but it prevents cell division when it is unphosphorylated.
Inactivating deletions or point mutations of RB cause the protein to lose its regulatory capacity. The nuclear phosphoprotein gene TP53 has also been recognized as an important
tumor suppressor gene, perhaps the most commonly altered
gene in all human cancers. Inactivating mutations of the
TP53 gene also cause the TP53 protein to lose its ability to regulate the cell cycle. The TP53 gene is frequently inactivated in
solid tumors of childhood, including osteosarcoma, rhabdomyosarcoma, brain tumors, anaplastic Wilms tumor, and a
subset of chemotherapy-resistant neuroblastoma.3335 In addition, heritable cancer-associated changes in the TP53 tumor
suppressor gene occur in families with Li-Fraumeni syndrome, an autosomal dominant predisposition for rhabdomyosarcoma, other soft tissue and bone sarcomas,
premenopausal breast cancer, brain tumors, and adrenocortical carcinomas.36 Other tumor suppressor genes include
Wilms tumor 1 (WT1), neurofibromatosis 1 (NF1), and von
Hippel-Lindau (VHL). Additional tumor suppressor genes
are presumed to exist but have not been definitively identified.
Epigenetic Alterations
As stated previously, the hallmark of cancer is dysregulated
gene expression. However, not only do genetic factors influence gene expression but epigenetic factors do as well, with
these factors being at least as important as genetic changes
in their contribution to the pathogenesis of cancer. Epigenetic alterations are defined as those heritable changes in
gene expression that do not result from direct changes in
DNA sequence. Mechanisms of epigenetic regulation most
commonly include DNA methylation and modification of
histones, although the contribution of microRNAs (miRNA),
a class of noncoding RNAs, is becoming increasingly
recognized.
DNA Methylation DNA methylation is a reversible process
that involves methylation of the fifth position of cytosine within
CpG dinucleotides present in DNA. These dinucleotides
are usually in the promoter regions of genes; methylation of
these sites typically causes gene silencing, thereby preventing
expression of the encoded proteins. This process is part
of the normal mechanism for imprinting, X-chromosome
inactivation, and generally keeping large areas of genomic
DNA silent, but it may also contribute to the pathogenesis of
cancer by silencing tumor suppressor genes. However, both
abnormal hypomethylation and hypermethylation states exist
in human tumors, resulting in both dysregulated expression
and silencing, respectively, of affected genes. These modifications of the nucleotide backbone of human DNA are
becoming increasingly recognized in human cancer, both for
their frequency and importance. For example, promoter
methylation resulting in silencing of caspase 8, a protein involved in apoptosis, likely contributes to the pathogenesis of
MYCN-amplified neuroblastoma37 as well as Ewing sarcoma.23
Histone Modification Histones are the proteins that give
structure to DNA and, together with the DNA, form the major
components of chromatin. The functions of histones are
to package DNA into a smaller volume to fit in the cell, to
strengthen the DNA to allow replication, and to serve as a
mechanism to control gene expression. Alterations in histones
can mediate changes in chromatin structure. The compacted
form of DNA, termed heterochromatin, is largely inaccessible
to transcription factors and therefore genes in the affected
regions are silent. Other modifications of histones can cause
DNA to take a more open or extended configuration (euchromatin), allowing for gene transcription. The N-terminal tails of
histones can be modified by a number of different processes
including methylation and acetylation, mediated by histone
acetyl transferases (HAT) and deacetylases (HDAC), and
histone methyltransferases (HMT). Each of these processes
alters histone function, which, in turn, alters the structure
of chromatin and therefore the accessibility of DNA to transcription factors. Methylation of the DNA itself can also effect
changes in chromatin structure.
MicroRNA As stated above, miRNAs are a group of small,
regulatory noncoding RNAs that appear to function in gene
regulation. These miRNAs are single-stranded RNA fragments
of 21 to 23 nucleotides that are complementary to encoding
mRNAs.25 Their function is to down-regulate expression of
target mRNAs; it is estimated that miRNAs regulate the expression of about 30% of all human genes.38 These miRNAs regulate gene expression primarily by incorporating into
silencing machinery called RNA-induced silencing complexes
(RISC). MiRNAs are involved in a number of fundamental
biological processes, including development, differentiation,
cell-cycle regulation, and senescence. However, broad analyses of miRNA expression levels have demonstrated that
many miRNAs are dysregulated in a variety of different cancer
types, including neuroblastoma and other pediatric tumors,39
frequently losing their function as gene silencers/tumor suppressors. The activity of miRNAs, like gene expression, is also
under epigenetic regulation.
METASTASIS
Metastasis is the spread of cancer cells from a primary tumor to
distant sites and is the hallmark of malignancy. The development of tumor metastases is the main cause of treatment
failure and a significant contributing factor to morbidity
and mortality resulting from cancer. Although the dissemination of tumor cells through the circulation is probably a
frequent occurrence, the establishment of metastatic disease
is a very inefficient process. It requires several events, including the entry of the neoplastic cells into the blood or lymphatic
system, the survival of those cells in the circulation, their
avoidance of immune surveillance, their invasion of foreign
(heterotopic) tissues, and the establishment of a blood supply
to permit expansion of the tumor at the distant site. Simple,
dysregulated cell growth is not sufficient for tumor invasion
and metastasis. Many tumors progress through distinct stages
CHAPTER 28
403
that can be identified by histopathologic examination, including hyperplasia, dysplasia, carcinoma in situ, invasive cancer,
and disseminated cancer. Genetic analysis of these different
stages of tumor progression suggests that uncontrolled growth
results from progressive alteration in cellular oncogenes and
inactivation of tumor suppressor genes, but these genetic
changes driving tumorigenicity are clearly distinct from those
that determine the metastatic phenotype.
Histologically, invasive carcinoma is characterized by a lack
of basement membrane around an expanding mass of tumor
cells. Matrix proteolysis appears to be a key part of the mechanism of invasion by tumor cells, which must be able to move
through connective tissue barriers, such as the basement
membrane, to spread from their site of origin. The proteases
involved in this process include the matrix metalloproteinases
and their tissue inhibitors. The local environment of the target
organ may profoundly influence the growth potential of
extravasated tumor cells.40 The various cell surface receptors
that mediate interactions between tumor cells and between
tumor cells and the extracellular matrix include cadherins,
integrins (transmembrane proteins formed by the noncovalent
association of alpha and beta subunits), and CD44, a transmembrane glycoprotein involved in cell adhesion to hyaluronan.41 Tumor cells must decrease their adhesiveness to escape
from the primary tumor, but at later stages of metastasis, the
same tumor cells need to increase their adhesiveness during
arrest and intravasation to distant sites.
new blood vessels have developed) can grow to only a limited

size, approximately 2 to 3 mm3. At this point, rapid cell
proliferation is balanced by equally rapid cell death by apoptosis, and a nonexpanding tumor mass results. The switch to an
angiogenic phenotype with tumor neovascularization results in
a decrease in the rate of apoptosis, thereby shifting the balance
to cell proliferation and tumor growth.50,51 This decrease in
apoptosis occurs, in part, because the increased perfusion
resulting from neovascularization permits improved nutrient
and metabolite exchange. In addition, the proliferating endothelium may supply, in a paracrine manner, a variety of factors
that promote tumor growth, such as IGF-I and IGF-II.52
In experimental models, increased tumor vascularization
correlates with increased tumor growth, whereas restriction
of neovascularization limits tumor growth. Clinically, the
onset of neovascularization in many human tumors is temporally associated with increased tumor growth,53 and high
levels of angiogenic factors are commonly detected in blood
and urine from patients with advanced malignancies.107
In addition, the number and density of new microvessels
within primary tumors have been shown to correlate with
the likelihood of metastasis, as well as the overall prognosis
for patients with a wide variety of neoplasms, including pediatric tumors such as neuroblastoma and Wilms tumor.54,55
ANGIOGENESIS
------------------------------------------------------------------------------------------------------------------------------------------------
Angiogenesis is the biological process of new blood vessel

formation. This complex, invasive process involves multiple
steps, including proteolytic degradation of the extracellular
matrix surrounding existing blood vessels, chemotactic migration and proliferation of endothelial cells, the organization of
these endothelial cells into tubules, the establishment of a
lumen that serves as a conduit between the circulation and
an expanding mass of tumor cells, and functional maturation
of the newly formed blood vessel.42,43 Angiogenesis involves
the coordinated activity of a wide variety of molecules, including growth factors, extracellular matrix proteins, adhesion
receptors, and proteolytic enzymes. Under physiologic conditions, the vascular endothelium is quiescent and has a very
low rate of cell division, such that only 0.01% of endothelial
cells are dividing.4244 However, in response to hormonal cues
or hypoxic or ischemic conditions, the endothelial cells can be
activated to migrate, proliferate rapidly, and create tubules
with lumens.
Angiogenesis occurs as part of such normal physiologic activities as wound healing, inflammation, the female reproductive cycle, and embryonic development. In these processes,
angiogenesis is tightly and predictably regulated. However,
angiogenesis can also be involved in the progression of several
pathologic processes in which there is a loss of regulatory
control, resulting in persistent growth of new blood vessels.
Such unabated neovascularization occurs in rheumatoid
arthritis, inflammatory bowel disease, hemangiomas of childhood, ocular neovascularization, and the growth and spread
of tumors.45
Compelling data indicate that tumor-associated neovascularization is required for tumor growth, invasion, and
metastasis.4649 A tumor in the prevascular phase (i.e., before
Molecular Diagnostics
The explosion of information about the human genome has led
not only to an improved understanding of the molecular
genetic basis of tumorigenesis but also to the development of
a new discipline: the translation of these molecular events
into diagnostic assays. The field of molecular diagnostics has
developed from the need to identify abnormalities of gene or
chromosome structure in patient tissues and as a means of
supporting standard histopathologic and immunohistochemical diagnostic methods. In most instances, the result
of genetic testing confirms light microscopic- and immunohistochemistry-based diagnosis. In some instances, however
(e.g., primitive, malignant, small round cell tumor; poorly
differentiated synovial sarcoma; lipoblastic tumor), molecular
analysis is required to make a definitive diagnosis.
The molecular genetic methods most commonly used to
analyze patient tumor material include direct metaphase
cytogenetics or karyotyping, fluorescence in situ hybridization
(FISH), and reverse transcriptase polymerase chain reaction
(RT-PCR). Additional methods, such as comparative genomic
hybridization, loss of heterozygosity analysis, and complementary DNA (cDNA) microarray analysis, may eventually
become part of the routine diagnostic repertoire but are
currently used as research tools at referral centers and academic institutions. Each standard method is summarized in
Table 28-4. As with any method, molecular genetic assays
have advantages and disadvantages, and it is important to
understand and recognize their limitations.
The value of molecular genetic analysis of patient tissue
is not limited to aiding histopathologic diagnosis. Many of
the most important markers provide prognostic information
as well. MYCN amplification in neuroblastomas,13 for example, is strongly associated with biologically aggressive behavior. Amplification of this gene can be detected by routine
404
PART III
TABLE 28-4
Comparison of the Cytogenetic and Molecular Methods Routinely Used as Aids in Pathologic Diagnosis of Soft Tissue Tumors
Method
Purpose
Advantages
Disadvantages
Cytogenetics
Low resolution analysis of metaphase

chromosomes of cells grown in culture
Does not require a priori knowledge

of genetic abnormalities
Available in most diagnostic centers
In situ
hybridization
Detection of translocations, amplifications,

and gene deletions by hybridization of
nucleic acid probes to specific DNA or
mRNA sequences
PCR and RTPCR
Extremely sensitive detection of DNA

sequences and mRNA transcripts for the
demonstration of fusion genes, point
mutations, and polymorphisms
Can be applied to chromosomal

preparations as well as cytologic
specimens, touch preparations, and
paraffin sections
Morphologic correlation is possible
Multiple probes can be assayed at the
same time
Rapid (usually only requires 2 days)
Highest sensitivity and specificity of
all the molecular diagnostic
techniques
DNA sequencing of PCR products can
confirm result and provide additional
information
Requires minimal tissue
Versatile; can be applied to fresh
tissue as well as formalin-fixed,
paraffin-embedded tissue
Morphologic correlation is possible
The presence of normal tissue will
usually not affect test results
Rapid (usually requires 3-5 days)
Requires fresh, sterile tumor tissue for

growth in culture
Low sensitivity; will only detect large
structural abnormalities
No histologic correlation
Slow and technically demanding (takes
up to several weeks to perform)
Cannot detect small deletions or point
mutations
Interpretation can be difficult, especially
with formalin-fixed, paraffin-embedded
material
Only a limited number of specific nucleic
acid probes are available commercially
Formalin-fixation diminishes sensitivity
Combinatorial variability within fusion
gene partners requires appropriate
redundant primer design to avoid falsenegative test results
Extreme sensitivity requires exacting
laboratory technique to avoid falsepositive test results
From Davidoff AM, Hill DA: Molecular genetic aspects of solid tumors in childhood. Semin Pediatr Surg 10: 2001;106-118.
PCR, polymerase chain reaction; RT-PCR, reverse transcriptase polymerase chain reaction.
metaphase cytogenetics or by FISH, and current neuroblastoma protocols include the presence or absence of MYCN
amplification in their stratification schema. Some fusion gene
variants are also thought to influence prognosis. In initial
studies, two examples noted to confer relatively favorable
prognoses are the type 1 variant fusion of EWS-FLI1 in Ewing
sarcoma or primitive neuroectodermal tumor20 and the
PAX7-FKHR fusion in alveolar rhabdomyosarcoma.19
New technologies are emerging that permit accurate,
high-throughput analysis or profiling of tumor tissue: Gene
expression can be analyzed by using RNA microarrays, and
proteins by using proteomics. These approaches identify a
unique fingerprint of a given tumor that can provide diagnostic or prognostic information. Proteomic analysis can also
identify unique proteins in patients serum or urine; such a
profile can be used for early tumor detection, to distinguish
risk categories, and to monitor for recurrence. Additional
types of omics that are currently being used to evaluate
tumor or patient specimens include transcriptomics (RNA
and gene expression), metabolomics (metabolites and metabolic networks), and pharmacogenomics (how genetics affects
host drug responses). Information from each of these areas of
investigation provides an increasingly precise and unique
perspective on the biology, clinical behavior, and responsiveness to specific therapeutic interventions of individual patient
tumors. It is through these analyses that personalized therapy
is likely to be realized. In addition, it is anticipated that with
the identification of new, critical components of oncogenesis
and tumor progression will come new druggable targets
for cancer therapy. Drugs that act on these targets will not
only be effective anticancer agents but, because of their

specificity, will also have a broader therapeutic window,
thereby improving safety and minimizing toxicity.
Childhood Cancer and Heredity

------------------------------------------------------------------------------------------------------------------------------------------------
Advances in molecular genetic techniques have also improved

our understanding of cancer predisposition syndromes.
Constitutional gene mutations that are hereditary (i.e., passed
from parent to child) or nonhereditary (i.e., de novo mutations in the sperm or oocyte before fertilization) contribute
to an estimated 10% to 15% of pediatric cancers.56
Constitutional chromosomal abnormalities are the result of
an abnormal number or structural rearrangement of the
normal 46 chromosomes and may be associated with a predisposition to cancer. Examples are the predisposition to leukemia seen with trisomy 21 (Down syndrome) and to germ
cell tumors with Klinefelter syndrome (47XXY). Structural
chromosomal abnormalities include interstitial deletions
resulting in the constitutional loss of one or more genes.
Wilms tumors may be sporadic, familial, or associated with
specific genetic disorders or recognizable syndromes. A better
understanding of the molecular basis of Wilms tumor has
been achieved largely through the study of the latter two types
of tumors. The WAGR syndrome (Wilms tumor, aniridia, genitourinary abnormalities, and mental retardation) provides an
easily recognizable phenotype for grouping children likely to
have a common genetic abnormality. Constitutional deletions
from chromosome 11p13 are consistent in children with
CHAPTER 28
WAGR syndrome57 and also occur in approximately 35% of

those with sporadic Wilms tumor.58 A study of a large series
of patients identified the gene deleted from chromosome
11p13 as WT1.59 This gene encodes a nuclear transcription
factor that is essential for normal kidney and gonadal development60 and appears to act as a tumor suppressor, but its
precise role is unclear at this time. Aniridia in patients with
WAGR syndrome is thought to occur after the loss of one copy
of the PAX6 gene located close to WT1 on chromosome 11.61
Denys-Drash syndrome, which is characterized by a very high
risk of Wilms tumor, pseudohermaphroditism, and mesangial
sclerosis leading to early renal failure, is associated with germline mutations in the DNA binding domain of WT1.62 The mutated WT1 protein appears to function by a dominant negative
effect. Only 6% to 18% of sporadic Wilms tumors have WT1
mutations.62,63
In another subset of patients with Wilms tumor, there is
loss of genetic material in a region distal to the WT1 locus
toward the telomeric end of chromosome 11 (11p15).39
It has therefore been suggested that there is a second Wilms
tumor susceptibility gene, tentatively named WT2, in 11p15.
Loss of heterozygosity at this locus has also been described in
patients with Beckwith-Wiedemann syndrome, a congenital
overgrowth syndrome characterized by numerous growth
abnormalities as well as a predisposition to a variety of malignancies, including Wilms tumor.64
Neurofibromatosis type 1 (NF1) is one of the most common
genetic disorders. The NF1 protein normally inhibits the
proto-oncogene RAS, but in patients with NF1, mutation of
one copy of the gene combined with deletion of the other permits uncontrolled RAS pathway activation. These patients are
then susceptible to myelogenous disorders, benign tumors,
gliomas, and malignant peripheral nerve sheath tumors. An
inherited predisposition to pediatric cancers is also associated
with Li-Fraumeni syndrome (which results from mutations
which inactivate the TP53 gene and put patients at risk for
osteosarcoma, rhabdomyosarcoma, adrenocortical carcinoma,
and brain tumors, among other tumors), familial retinoblastoma (which results from mutations that inactivate the RB gene
and put patients at risk for osteosarcoma as well as retinoblastoma), familial adenomatous polyposis, and multiple endocrine neoplasia syndromes. Another set of inherited risk
factors is represented by mutations of DNA repair genes
(so-called caretaker genes), as seen in xeroderma pigmentosa
and ataxia-telangiectasia.65 Understanding these complex
syndromes and their pathogenesis is important in efforts
to screen for early detection and, possibly, for prophylactic
therapy.
Recently, the germline mutation associated with hereditary
neuroblastoma has been identified as activating mutations in
the tyrosine kinase domain of the anaplastic lymphoma kinase
(ALK) oncogene on the long arm of chromosome 2 (2p23).66
Further molecular studies have revealed that common genetic
variation at chromosome bands 6p221 and 2q3567 are associated with susceptibility to, and likely contribute to the etiology of, high-risk neuroblastoma, providing the first evidence
that childhood cancers also arise because of complex interactions of polymorphic variants. Finally, the same group has
also shown that inherited copy number variation at chromosome 1q21.1 is associated with neuroblastoma, implicating
a neuroblastoma breakpoint family gene in early neuroblastoma genesis.68
405
Genetic Screening
------------------------------------------------------------------------------------------------------------------------------------------------
Along with an increased understanding of the molecular basis

of hereditary childhood cancer has come the opportunity to
identify children who are at high risk of malignancy and, in
some cases, to intervene before the cancer develops or when
it is still curable. Two examples include familial adenomatous
polyposis and familial thyroid cancer.
Familial adenomatous polyposis is an autosomal dominant inherited disease in which hundreds to thousands of adenomatous intestinal polyps develop during the second and
third decades of life. Mutations of the adenomatous
polyposis coli (APC) gene on chromosome 5q21 occur in
approximately 80% of kindreds of persons who have the
disease.69,70 These mutations initiate the adenomatous process by allowing clonal expansion of individual cells that, over
time, acquire additional genetic abnormalities that lead to the
development of invasive colorectal carcinoma.71 Prophylactic
colectomy is recommended for patients with this germline
mutation, although the most appropriate timing for this
intervention in children with familial adenomatous polyposis
is controversial. These patients are also at increased risk of
hepatoblastoma.72
Medullary thyroid carcinoma (MTC) is a rare malignancy
that may occur sporadically or as part of two syndromes: multiple endocrine neoplasia (type 2A or 2B) syndrome or familial
MTC syndrome. In children, MTC is much more likely to
occur in association with a familial syndrome. An apparently
100% association between germline RET mutations73 and
MTC guides the recommendation for prophylactic thyroidectomy in affected patients. There is no effective adjuvant treatment other than surgery for MTC, highlighting the need for
early intervention. Patients with germline RET mutations
should also be screened for pheochromocytoma, which occurs in 50% of patients with multiple endocrine neoplasia
type 2A, and hyperparathyroidism, which occurs in 35% of
such patients, although these entities generally arise in older
patients beyond the pediatric age range.74 In addition,
patients who are at risk for MTC or have newly diagnosed
MTC, as well as their relatives, should be screened for the
germline RET mutation so that appropriate surgical and
genetic counseling can be given.
General Principles
of Chemotherapy
------------------------------------------------------------------------------------------------------------------------------------------------
Cytotoxic agents were first noted to be effective in the treatment of cancer in the 1960s, after alkylating agents, such as
nitrogen mustard gas, used during World War II, were observed to cause bone marrow hypoplasia. Chemotherapy is
now an integral part of nearly all cancer treatment regimens.
The overriding goal of cancer chemotherapy is to maximize
the tumoricidal effect (efficacy) while minimizing adverse side
effects (toxicity). This goal can be difficult to achieve, however,
because the dose at which tumor cells are affected is often similar to the dose that affects normal proliferating cells, such as
those in the bone marrow and gastrointestinal tract. Despite
the early promise of chemotherapy and the observation that
most tumor types are initially sensitive to chemotherapy, often
406
PART III
exquisitely so, the successful use of chemotherapy is often

thwarted by two factors: the development of resistance to
the agent and the agents toxicity to normal tissues. Nevertheless, chemotherapy remains an integral part of therapy when
used as an adjunct to treat localized disease or as the main
component to treat disseminated or advanced disease.
A number of principles and terms are essential to the understanding of chemotherapy as a therapeutic anticancer modality.
Adjuvant chemotherapy refers to the use of chemotherapy for
systemic treatment following local control generally by surgical
resection or radiation therapy of a clinically localized primary
tumor. The goal in this setting is to eliminate disease that is not
detectable by standard investigative means at or beyond the
primary tumors site. Neoadjuvant chemotherapy refers to
chemotherapy delivered before local therapeutic modalities,
generally in an effort to improve their efficacy; to treat micrometastatic disease as early as possible, when distant tumors
are smallest; or to achieve both of these aims. Induction chemotherapy refers to the use of chemotherapeutic agents as the primary treatment for advanced disease. In general, chemotherapy
given to children with solid tumors and metastatic disease at
the time of first examination has a less than 40% chance of
effecting long-term, disease-free survival. Exceptions include
Wilms tumor with favorable histologic features, germ cell
tumors, and paratesticular rhabdomyosarcoma, but most
children with metastatic disease are at high risk of disease
recurrence or progression. Combination chemotherapy refers
to the use of multiple agents, which generally have different
mechanisms of action and nonoverlapping toxicities, that
provide effective, synergistic antitumor activity and minimal
side effects.
The mechanisms of action and side effects of commonly
used agents are listed in Table 28-5. Alkylating agents interfere
with cell growth by covalently cross-linking DNA and are
not cell-cycle specific. Antitumor antibiotics intercalate
into the double helix of DNA and break the DNA strands.
Antimetabolites are truly cell-cycle specific, because they interfere with the use of normal substrates for DNA and RNA
synthesis, such as purines and thymidine. The plant alkaloids
can inhibit microtubule function (vinca alkaloids, taxanes) or
DNA topoisomerases (camptothecins inhibit topoisomerase I;
epipodophyllotoxins inhibit topoisomerase II), and these actions also lead to breaks in DNA strands. Topoisomerases
are a class of enzymes that alter the supercoiling of doublestranded DNA. They act by transiently cutting one (topoisomerase I) or both (topoisomerase II) strands of the DNA to relax
the DNA coil and extend the molecule. The regulation of DNA
supercoiling is essential to DNA transcription and replication,
when the DNA helix must unwind to permit the proper
function of the enzymatic machinery involved in these
processes. Thus topoisomerases maintain the transcription
and replication of DNA.
The common toxic effects of these agents are also listed in
Table 28-5. Most toxicity associated with chemotherapy is reversible and resolves with cessation of treatment. However, some
chemotherapeutic agents may have lifelong effects. Of particular
concern is that certain drugs can lead to a second malignancy.
Most notable is the development of leukemia after the administration of the epipodophyllotoxins and cyclophosphamide.75
Finally, understanding the metabolism of chemotherapeutic
agents is important. Certain agents require metabolism at a
specific site or organ for their activation or are eliminated from
the body by a specific organ (see Table 28-5). The processes of

activation and elimination require normal organ function
(e.g., the liver for cyclophosphamide); therefore children with
liver or kidney failure may not be able to receive certain agents.
RISK STRATIFICATION
Major advances in the variety of chemotherapeutic agents and
dosing strategies used to treat pediatric cancers in the past
30 years are reflected in improved patient survival rates.
Regimen toxicity (including late effects, which are particularly
important in the pediatric population) and therapeutic
resistance are the two main hurdles preventing further advancement. As more information about diagnostically and
prognostically useful genetic markers becomes available,
therapeutic strategies will change accordingly. With molecular
profiling, patients can be categorized to receive a particular
treatment on the basis of not only the tumors histopathologic
and staging characteristics but also its genetic composition.
Some patients whose tumors show a more aggressive biological profile may require dose intensification to increase their
chances of survival. Patients whose tumors do not have an
aggressive biological profile may benefit from the lower
toxicity of less intensive therapy. Such an approach may allow
the maintenance of high survival rates while minimizing longterm complications of therapy in these patient populations.
The paradigm for the use of different therapeutic intensities
on the basis of risk stratification drives the management of
pediatric neuroblastoma. There is increasing evidence that
the molecular features of neuroblastoma are highly predictive
of its clinical behavior. Most current studies of the treatment of
neuroblastoma are based on risk groups that take into account
both clinical and biological variables. The most important
clinical variables appear to be age and stage at diagnosis,
and the most powerful biological factors appear to be MYCN
status, ploidy (for patients younger than 1 year), and histopathologic classification. These variables currently define
the Childrens Oncology Group risk strata and therapeutic
approach, which are further refined by determining whether
there is 1p/11q LOH. At one extreme, patients with low-risk
disease are treated with surgery alone; at the other extreme,
patients at high risk for relapse are treated with intensive
multimodality therapy that includes multiagent doseintensive chemotherapy, radiation therapy, and stem cell
transplantation. Other factors, such as 17q gain, caspase
8 inactivation, and TRKA/B expression, are currently being
evaluated and may help further refine risk assessment in the
future. The management of other solid pediatric tumors is also
shifting to risk-defined treatment. For example, the current
protocol for the management of patients with Wilms tumor
includes risk stratification and therapy adjustment based
on molecular analysis of the primary tumor for 16q and 1p
deletions.
TARGETED THERAPY
Another major change in the approach to the treatment of cancer has been the concept of targeted therapy. Until recently, the
development of anticancer agents was based on the empirical
screening of a large variety of cytotoxic compounds without
particular regard to disease specificity or mechanism of action.
Now, one of the most exciting prospects for improving the
TABLE 28-5
Common Chemotherapeutic Agents
Synonyms
Alkylating
agents
Carboplatin
CBCDCA
Paraplatin
Cisplatin
CDDC
Platinol
Cyclophosphamide
CTX
Cytoxan
Ifosfamide
IFOS
Ifex
Dacarbazine
Temozolomide
Nitrogen Mustard
TMZ
Mechlorethamine
DTIC
Temodar
Mustargen
Melphalan
L-PAM
Alkeran
Busulfan
Antimetabolites
Busulfex
Cytarabine
Ara-C
Cytosar
Fluorouracil
5-FU
(Several)
Method of
Elimination
Susceptible
Solid
Tumors
Mechanism of Action
Common Toxic Effects
Platination,
intrastrand and
interstrand DNA
cross-linking
Platination,
intrastrand and
interstrand DNA
cross-linking
Alkylation,
intrastrand and
interstrand DNA
cross-linking
Alkylation,
intrastrand and
interstrand DNA
cross-linking
Methylation
Methylation
Alkylation,
intrastrand and
interstrand DNA
cross-linking
Alkylation,
intrastrand and
interstrand DNA,
cross-linking
Alkylation,
intrastrand and
interstrand DNA
cross-linking
Inhibits DNA
polymerase,
incorporated into
DNA
Inhibits thymidine
synthesis,
incorporated into
DNA/RNA
A, H, M, (esp. thrombocytopenia), N/V
BT, GCT,
NBL, STS
A, N/V, R (significant), ototoxicity,

neuropathy
BT, GCT,
NBL, OS
A, N/V, SIADH, M, R, cardiac, cystitis
Liver
H, R (minor)
Broad, BMT
A, CNS, N/V, M, R, cardiac, cystitis
Liver
H, R (minor)
Broad
H, N/V, M, hepatic vein thrombosis

CNS, N/V, M
A, M (significant), N/V, mucositis,
vesicant, phlebitis, diarrhea
Liver
Spontaneous
R
R
Spontaneous
hydrolysis
NBL, STS
BT
BT
M, N/V, mucositis, diarrhea
Spontaneous
hydrolysis
NBL, RMS,
BMT
A, H, M, N/V, P, mucositis
BMT
M, N/V, diarrhea, CNS
Target cell
Biotransformation
Limited
CNS, N/V, M, cardiac, diarrhea,

mucositis, skin, ocular
Target cell
Biotransformation,
renal (minor)
GI
carcinomas,
liver tumors
Continued
Agent
Site of
Activation
CHAPTER 28
Class of Drug
Brand
Name
407
408
PART III
TABLE 28-5
Common Chemotherapeutic Agentscontd
Class of Drug
Antibiotics
Plant Alkaloids
Agent
Synonyms
Brand
Name
Mercaptopurine
6-MP
Purinethol
Methotrexate
MTX
Trexall
Dactinomycin
Actinomycin-D
Cosmegen
Bleomycin
BLEO
Blenoxane
Anthracyclines
Daunomycin
Daunorubicin
Cerubidine
Adriamycin
Doxorubicin
Adriamycin
Epipodophyllotoxins
Etoposide
VP-16
VePesid
Teniposide
VM-26
Vumon
Vinca alkaloids
Vincristine
VCR
Oncovin
Mechanism of Action
Common Toxic Effects
Site of
Activation
Inhibits thymidine
synthesis,
incorporated into
DNA/RNA
Blocks folate
metabolism, inhibits
purine synthesis
DNA intercalation,
strand breaks
DNA intercalation,
strand breaks
H, M, mucositis
Target cell
Susceptible
Solid
Tumors
Method of
Elimination
Biotransformation,
renal (minor)
Limited
CNS, H, M, R, mucositis, skin
R, H (minor)
OS
A, H, M, N/V, mucositis, vesicant
RMS, Wilms
P, skin, mucositis
H, R
GCT
A, M, N/V, cardiac, diarrhea, vesicant,

potentiate XRT reaction
Limited
A, M, N/V, cardiac, diarrhea, mucositis,

vesicant, potentiate XRT reaction
Broad
Topoisomerase II
inhibitor, DNA strand
breaks
Topoisomerase II
breaks
A, M, N/V, mucositis, neuropathy,

diarrhea
Broad
A, M, N/V, mucositis, neuropathy,

diarrhea
Degraded
Broad
Inhibits tubulin
polymerization,
blocks mitosis
A, SIADH, neuropathy, vesicant
H
H
Broad
DNA intercalation,
strand breaks, free
radical formation
DNA intercalation,
strand breaks, free
radical formation
Vinblastine
Velban
Inhibits tubulin
polymerization,
blocks mitosis
A, M, mucositis, vesicant
Taxanes
Paclitaxel
Taxol
A, M, cardiac, mucositis, CNS,

neuropathy
Docetaxel
Taxotere
Interferes with
microtubule
formation
Interferes with
microtubule
formation
Topoisomerase I
breaks
Topoisomerase I
breaks
L-Asparagine
depletion, inhibits
protein synthesis
Nuclear receptor
mediated apoptosis
A, H, M, N/V, mucositis, diarrhea, skin
TPT
Hycamtin
Irinotecan
CPT-11
Camptosar
L-Asparaginase
Erwinia
Elspar
Corticosteroids
GCT
NBL, RMS
H, R (minor)
NBL, RMS
degraded
Limited
H, R (minor)
BT
A, neutropenia, cardiac, mucositis, CNS,

neuropathy
A, H, M, N/V, diarrhea
H, GI
CNS, H, coagulopathy, pancreatitis,

anaphylaxis
avascular necrosis, hyperglycemia,
hypertension, myopathy, pancreatitis,
peptic ulcers, psychosis, salt
imbalance, weight gain
Toxic effects: A, alopecia; CNS, central nervous system toxicity; H, hepatotoxicity; M, myelosuppression; N/V, nausea and vomiting; P, pulmonary toxicity;
R, renal toxicity; SIADH, syndrome of inappropriate antidiuretic hormone; XRT, x-ray therapy.
Solid tumors: BMT, conditioning for bone marrow transplantation; BT, brain tumor; EWS, Ewing sarcoma; GCT, germ cell tumors; NBL, neuroblastoma; OS, osteosarcoma; RMS, rhabdomyosarcoma; STS, soft tissue
sarcoma; W, Wilms tumor.
Camptothecins
Topotecan
CHAPTER 28
Miscellaneous
VLB
409
410
PART III
therapeutic index of anticancer agents, as well as overcoming

the problem of therapy resistance, involves targeted therapy.
As the molecular bases for the phenotypes of specific malignancies are being elucidated, potential new targets for therapy
are becoming more clearly defined. The characterization of
pathways that define malignant transformation and progression has focused new agent development on key pathways involved in the crucial processes of cell-cycle regulation,
receptor signaling, differentiation, apoptosis, invasion, migration, and angiogenesis, which may be perturbed in malignant
tissues. Information about the molecular profile of a given tumor type can be assembled from a variety of emerging
methods, including immunohistochemistry, FISH, RT-PCR,
cDNA microarray analysis, and proteomics. This information
can then be used to develop new drugs designed to counter
the molecular abnormalities of the neoplastic cells. For example, blocking oncogene function or restoring suppressor
gene activity may provide tumor-specific therapy. In addition,
molecular profiling may lead to the development of drugs
designed to induce differentiation of tumor cells, block dysregulated growth pathways, or reactivate silenced apoptotic
pathways.
Some agents target alterations in the regulation of cell proliferation. Trastuzumab (Herceptin) is a monoclonal antibody
that binds to the cell surface growth factor receptor ERBB2
with high affinity and acts as an antiproliferative agent when
used to treat ERBB2-overexpressing cancer cells.76 Pediatric
high-grade gliomas that overexpress EGFR may be amenable
to a similar therapeutic agent, gefitinib (Iressa), a smallmolecule inhibitor of EGFR (ERBB1).77 In addition, smallmolecule tyrosine kinase inhibitors, such as imatinib
(Gleevec), designed to block aberrantly expressed growthpromoting tyrosine kinasesABL in chronic myelogenous
leukemia78 and c-KIT in gastrointestinal stromal tumors79
are being evaluated in clinical trials. Imatinib may also be useful in treating pediatric tumors in which PDGF signaling plays
a role in tumor cell survival and growth. Also of potential
therapeutic utility are small-molecule inhibitors that recognize
antigenic determinants on unique fusion peptides or one of
the fusion peptide partners in tumors that have chromosomal
translocations (e.g., sarcomas). Tumors that depend on autocrine pathways for growth (e.g., overproduction of IGF-II
in rhabdomyosarcoma or PDGF in dermatofibrosarcoma
protuberans) may be sensitive to receptor blocking mediators
(e.g., antibodies to the IGF-II or PDGFR).
Other agents target alteration of the cell death and differentiation pathways. Caspase 8 is a cysteine protease that
regulates programmed cell death, but in tumors such as neuroblastoma, DNA methylation and gene deletion combine to
mediate the complete inactivation of caspase 8, almost always
in association with MYCN amplification.80 Caspase 8-deficient
tumor cells are resistant to apoptosis mediated by death receptors and doxorubicin; this resistance suggests that caspase
8 may be acting as a tumor suppressor. However, brief exposure of caspase 8deficient cells to demethylating agents, such
as decitabine, or to low levels of interferon gamma can lead to
the reexpression of caspase 8 and the resensitization of the
cells to chemotherapeutic drug-induced apoptosis. Histone
deacetylase also seems to have a role in gene silencing associated with resistance to apoptosis81; therefore histone deacetylase inhibitors, such as suberoylanilide hydroxamic acid
(SAHA), are also being tested for the treatment of certain
pediatric malignancies. Finally, cells with alterations in programmed cell death as a result of the persistence or reactivation of telomerase activity, which somatic cells normally lose
after birth, can be targeted by various telomerase inhibitors.
Several methods of targeting tumor cell differentiation are
being used for the treatment of neuroblastoma. Treatment
with 13-cis-retinoic acid, a vitamin A derivative that signals
through receptors that mediate transcription of different sets
of genes of cell differentiation, including HOX genes, is now
standard of care for maintenance therapy in patients with
high-risk neuroblastoma.82,83 Also, different neurotrophin receptor pathways appear to mediate the signal for both cellular
differentiation and malignant transformation of sympathetic
neuroblasts to neuroblastoma cells. Neurotrophins are
expressed in a wide variety of neuronal tissues and other
tissues that require innervation. They stimulate the survival,
maturation, and differentiation of neurons and exhibit a
developmentally regulated pattern of expression.84,85 Neurotrophins and their TRK tyrosine kinase receptors are particularly important in the development of the sympathetic nervous
system and have been implicated in the pathogenesis of neuroblastoma. Three receptorligand pairs have been identified:
TRKA, TRKB, and TRKC, which are the primary receptors for
nerve growth factor, brain-derived neurotrophic factor
(BDNF), and neurotrophin 3 (NT-3), respectively.84 TRKA
appears to mediate the differentiation of developing neurons
or neuroblastoma in the presence of nerve growth factor ligand and to mediate apoptosis in the absence of nerve growth
factor.85 Conversely, the TRKB-BDNF pathway appears to
promote neuroblastoma cell survival through autocrine or
paracrine signaling, especially in MYCN-amplified tumors.86
TRKC is expressed in approximately 25% of neuroblastomas
and is strongly associated with TRKA expression.87 Studies are
ongoing to test agonists of TRKA in an attempt to induce
cellular differentiation. Conversely, blocking the TRKB-BDNF
signaling pathway with TRK-specific tyrosine kinase inhibitors such as CEP-751 may induce apoptosis by blocking
crucial survival pathways.66,86 This targeted approach has
the attractive potential for increased specificity and lower
toxicity than conventional cytotoxic chemotherapy.
Inhibition of Angiogenesis
------------------------------------------------------------------------------------------------------------------------------------------------
Because tumor growth and spread appear to be dependent on

angiogenesis, inhibition of angiogenesis is a logical anticancer
strategy. This approach is particularly appealing for several
reasons. First, despite the extreme molecular and phenotypic
heterogeneity of human cancer, it is likely that most, if not all,
tumor types, including hematologic malignancies, require
neovascularization to achieve their full malignant phenotype.
Therefore antiangiogenic therapy may have broad applicability for the treatment of cancer. Second, the endothelial cells in
a tumors new blood vessels, although rapidly proliferating,
are inherently normal and mutate slowly. They are therefore
unlikely to evolve a phenotype that is insensitive to an angiogenesis inhibitor, unlike the rapidly proliferating tumor
cells, which undergo spontaneous mutation at a high rate
and can readily generate drug-resistant clones. Finally, because the new blood vessels induced by a tumor are sufficiently distinct from established vessels to permit highly
specific targeting,88,89 angiogenesis inhibitors should have a
CHAPTER 28
high therapeutic index and minimal toxicity. The combination

of conventional chemotherapeutic agents with angiogenesis
inhibitors appears to be particularly effective.
The first clinical demonstration that an angiogenesis inhibitor could cause regression of a tumor came with the use of interferon alpha in a patient treated for life-threatening
pulmonary hemangioma.90 An increasing number of natural
and synthetic inhibitors of angiogenesis, which inhibit different effectors of angiogenesis, have since been identified, and
many of these agents have been tested in clinical trials. Examples include drugs that directly inhibit endothelial cells, such as
thalidomide and combretastatin; drugs that block activators of
angiogenesis, such as bevacizumab (Avastin), a recombinant
humanized anti-VEGF antibody, or VEGF trap; drugs that inhibit endothelium-specific survival signaling, such as Vitaxin,
an anti-integrin antibody; and drugs with nonspecific mechanisms of action, such as celecoxib and interleukin-12 (IL-12).
411
as granulocyte-macrophage colony-stimulating factor92 and

interleukin-2 (IL-2),93 current antineuroblastoma antibody
trials are evaluating the use of a humanized, chimeric antiGD2 antibody (ch14.18) with these cytokines and a fusion
protein (hu14.18:IL2) that consists of the humanized 14.18
antibody linked genetically to human recombinant IL-2.
A recently completed randomized phase III trial using
ch14.18 alternating with cycles of granulocyte-macrophage
colony-stimulating factor (GM-CSF) or interleukin-2 added
to maintenance therapy of cis-retinoic acid demonstrated a
significant improvement in 2-year event-free survival for those
who received immunotherapy in addition to retinoic acid.94
General Principles of Radiation

Therapy
------------------------------------------------------------------------------------------------------------------------------------------------
Immunotherapy
------------------------------------------------------------------------------------------------------------------------------------------------
The immune system has evolved as a powerful means to detect

and eliminate molecules or pathogens that are recognized as
foreign. However, because tumors arise from host cells, they
are generally relatively weakly immunogenic. In addition, malignant cells have evolved several mechanisms that allow them
to elude the immune system. These mechanisms include the
ability to down-regulate the cell surface major histocompatibility complex molecules required for activation of many of
the immune effector cells, to produce immunosuppressive factors, and to variably express different proteins that might otherwise serve as targets for the immune system in a process
known as antigenic drift. Nevertheless, because of the large
number of mutations and chromosomal aberrations occurring
in cancer cells, which results in the expression of abnormal,
new, or otherwise silenced proteins, it is likely that most, if
not all, cancers contain unique tumor-associated antigens that
can be recognized by the immune system. Examples include
the fusion proteins commonly found in pediatric sarcomas
and the embryonic neuroectodermal antigens that continue
to be produced by neuroblastomas.
Recruiting the immune system to help eradicate tumor cells
is an attractive approach for several reasons. First, circulating
cells of the immune system have ready access to even occult
sites of tumor cells. Second, the immune system has powerful
effector cells capable of effectively and efficiently destroying
and eradicating targets, including neoplastic cells. Initial
efforts to recruit the immune system to recognize and destroy
tumor cells by using cytotoxic effector mechanisms that are
T-cell dependent or independent focused on recombinant
cytokines. Cytokines act by directly stimulating the immune
system66 or by rendering the target tumor cells more
immunogenic.
Neuroblastoma has been a popular target for immunotherapy in the pediatric population. Although a particular neuroblastoma antigen has not been defined, murine monoclonal
antibodies have been raised against the ganglioside GD2, a
predominant antigen on the surface of neuroblastoma cells.
These antibodies elicited therapeutic responses,37,91 but with
substantial toxicity, particularly neuropathic pain.92 Because
the induction of antibody-dependent cell-mediated cytotoxicity with anti-GD2 antibodies is enhanced by cytokines, such
Radiation therapy is one of the three primary modalities used

to manage pediatric cancers in the modern era. Radiation
therapy is delivered to an estimated 2000 or more children
per year for the primary treatment of tumor types as diverse
as leukemia, brain tumors, sarcomas, Hodgkin disease, neuroblastoma, and Wilms tumor.95 Delivery of radiation therapy
in the pediatric setting differs from that in the adult setting
because of the balance between curative therapy and an anticipated long life span during which long-term morbidity may
result from the therapy.
CLINICAL CONSIDERATIONS
Radiation therapy for the management of pediatric cancer is
most frequently combined with surgery and chemotherapy
as part of a multidisciplinary treatment plan. The sensitive
nature of pediatric tumors requires the use of a combined
therapy approach to maximize tumor control while minimizing the long-term side effects of treatment. Radiation may be
delivered preoperatively, postoperatively (relative to a definitive surgical resection), or definitively without surgical
management. Systemic therapy may also be integrated into
this management approach.
Definitive Irradiation
Definitive radiation therapy is an alternative local approach to
surgical resection of primary solid tumors. It is often the only
local therapeutic approach for children and adolescents with
leukemia or lymphoma.96,97 Definitive radiation therapy for
rhabdomyosarcoma has been used as an alternative to surgical
resection, which has potentially greater morbidity; it has
achieved high rates of local tumor control while allowing preservation of function.38 The Ewing sarcoma family of tumors
may also be considered candidates for definitive radiation
therapy as an alternative to surgery. With careful patient selection, excellent local tumor control rates can be maintained
while reducing or avoiding the morbidity associated with
difficult surgical resections.98,99
Preoperative Irradiation
Targeting of a localized tumor is straightforward in the preoperative setting; the tumor has clearly defined margins undisturbed by a surgical procedure. The volume of normal,
healthy tissues receiving high doses of radiation may be
412
PART III
reduced, because the areas at risk for disease involvement can

be better defined. Preoperative radiation therapy has been
used rarely in the management of Wilms tumor to decrease
the chance of tumor rupture100 and in the management of
nonrhabdomyosarcoma soft tissue sarcoma and Ewing sarcoma to facilitate surgical resection.101,102 One of the limitations may be the slightly higher incidence of postoperative
wound complications noted in the sarcoma population.102
Postoperative Irradiation
Postoperative radiation therapy combined with surgical
resection is the most common application of adjuvant radiation treatment in the United States. Despite some degree of
difficulty in targeting, a postoperative approach allows a
review of tumor histology from the complete tumor specimen,
including identification of the tumor margins and the response to any previous therapy. Wound healing complications
appear to be reduced with this approach, and the radiation
dose can be more accurately tailored to the pathologic findings
after primary resection.
Interactions of Chemotherapy and Radiation
Most childrens cancers are managed with systemic chemotherapy. In children receiving radiation therapy as well as systemic
chemotherapy, issues of enhanced local efficacy and enhanced
local or regional toxicity need to be considered. Solid tumors
that are frequently treated with combined chemotherapy and
radiation therapy include Wilms tumor, neuroblastoma, and
sarcomas. These tumors are subdivided into those in which
chemotherapy is given concomitantly with radiation therapy103,104 and those in which it is given sequentially, before
or after radiation therapy.83,100,105 When delivering radiation
therapy concurrently with or temporally close to a course of
chemotherapy, several issues must be considered.
Chemotherapeutic Enhancement of Local
Irradiation
Several systemic chemotherapeutic agents used against
pediatric tumors may enhance the efficacy of radiation therapy
when delivered concomitantly. Cisplatin, 5-fluorouracil,
mitomycin C, and gemcitabine, for example, are well-known
radiation sensitizers.106108 Concomitant delivery of any of
these drugs with radiation therapy may require that they be
administered at a dose and schedule different from those
typically used when the drugs are delivered alone. Despite
the potential of increased toxicity, significant improvements
in local tumor control have been shown in randomized studies
of concomitant drug and radiation therapy.106,107
is often outweighed by the benefit of continuously delivered

systemic therapy, particularly in tumors associated with a
high incidence of micrometastatic disease.
Agents That Increase Radiation Toxicity
Several agents significantly increase the local toxicity of
radiation. For this reason, these agents are not given concomitantly with irradiation and are often withheld for a period after the completion of radiation therapy. The two most notable
agents are doxorubicin and actinomycin, both of which can
induce significant skin and mucosal toxicity when delivered
concurrently with radiation therapy.38,109 The camptothecins
(including irinotecan and topotecan) also potentiate mucosal
toxicity when delivered concurrently with radiation therapy.110,111 Although this increase in toxicity suggests a possible increase in local efficacy, this benefit has not been noted
with current treatment approaches and chemotherapeutic
dosing guidelines. For this reason, these agents are avoided
during the delivery of radiation therapy and are withheld
for 2 to 6 weeks after the completion of treatment.
The current era of systemic therapy continues to broaden
with the availability of many new agents that target molecular
pathways. It is important to consider the possibility of new
toxicities when combining novel agents with a known therapy
such as radiation.
FRACTIONATION OF RADIATION THERAPY

Conventional, external beam irradiation is delivered in a
fractionated form. Fractionation implies daily doses of radiation delivered 5 days per week and amounting to the prescribed dose for a particular tumor type. Radiation
delivered once daily at a fraction size between 1.5 and 2.0
Gy on 5 days per week is considered conventionally fractionated. This daily dose is well tolerated by normal tissues
adjacent to the tumor and appears to effect local tumor control in many tumor systems. Though adult malignancies may
be treated with increased doses per fraction to overcome the
radioresistance of many carcinomas (termed hypofractionation), nearly all the literature describing radiation therapy,
its efficacy, and its toxicity in children is based on conventional fractionation.
RADIATION THERAPY TREATMENT

TECHNIQUES
Traditional Radiation Therapy
Irradiation Combined with Agents Having

Limited or No Sensitizing Effect
In the management of pediatric malignancies, radiation is
often combined with systemic therapy not to increase its local
efficacy but to allow continued delivery of systemic therapy to
control micrometastatic or metastatic disease. Agents combined with radiation therapy in this setting are common in
the management of pediatric sarcomas and include ifosfamide
and etoposide, which are delivered concurrently with
radiation therapy for Ewing sarcoma, and vincristine and
cyclophosphamide, which are delivered concurrently with
radiation therapy for rhabdomyosarcoma.103,104 Although
local toxicity may be increased by such an approach, this risk
The planning and delivery of traditional, or conventional,

radiation therapy are based on nonvolumetric imaging studies
(i.e., conventional radiographs). Patients are positioned in a
manner that allows the orientation of radiation beams from
the conventional directions: anterior, posterior, and lateral.
Limitations of this approach are related to the ability of
conventional radiographs to accurately convey the location
of tumor-bearing tissue. Although treatment beams are oriented around the tumor, adjacent normal tissues also receive
high doses of radiation. Depending on the accuracy of the
delineation of adjacent normal tissues on radiographs, the
dose to those tissues may not be known. Radiation is delivered
by a photon beam generated by a linear accelerator.
CHAPTER 28
Focal Radiation Therapy

Focal radiation therapy comprises a group of techniques that
deliver radiation to a defined volume, usually delineated by
computed tomography (CT) or magnetic resonance imaging
(MRI). Relatively low doses may be incidentally delivered
to surrounding normal tissues. Radiation therapy may be described as image guided when four criteria are met: (1) threedimensional imaging data (CT or MRI) are acquired with the
patient in the treatment position; (2) imaging data are used to
delineate and reconstruct the tumor volume and normal
tissues in three dimensions; (3) radiation beams can be freely
oriented in three dimensions in the planning and delivery
processes, and structures traversed by the beam can be
visualized with the eye of the beam; and (4) the distribution
of doses received by the tumor volume and any normal
tissue is computable on a point-by-point basis in threedimensional space. Several different methods of delivering
image-guided photon radiation are currently in use and are
discussed here.
Conformal Radiation Therapy
The delivery of three-dimensional conformal radiation
therapy allows specific targeting of tumor volumes on the
basis of imaging studies performed with the patient in the
treatment position. This method of delivery uses multiple
fields or portals, with each beam aperture shaped to the tumor
volume, and it is performed daily. Beam modifiers, such as
wedges, are used to conform the radiation beam to the tumor
and to ensure that the tumor volume receives a homogeneous
dose. Conformal radiation therapy has been intensively studied in adults with head and neck cancer, lung cancer, and
prostate cancer and has been shown to excel when the target
volume is convex and crucial structures do not invaginate the
target volume. Available data demonstrate that it has low toxicity despite high doses of radiation to the target volume.112
Intensity-Modulated Radiation Therapy
Intensity-modulated radiation therapy is another method of
delivering external beam radiation that requires imaging of
the patient in the treatment position and delineation of target
volumes and normal tissues. Radiation is delivered to the target as multiple small fields that do not encompass the entire
target volume but collectively deliver the prescribed daily
dose. Intensity-modulated radiation therapy differs from
conformal radiation therapy in that it (1) increases the complexity and time required for the planning and delivery of
treatment, (2) increases the amount of quality-assurance work
required before treatment is delivered, (3) increases dose
heterogeneity within the target volume such that some intralesional areas receive a relatively high dose, and (4) can be used
to treat concave targets while sparing crucial structures that
invaginate the target volume. The last point holds promise
for better protecting normal tissue and reducing late toxic
effects. Preliminary data from adult patients given intensitymodulated radiation therapy demonstrate its potential for
reducing treatment toxicity when applied to pediatric brain
tumors and other adult tumors.113
Proton Beam Radiation Therapy
Proton radiation therapy and other approaches using heavy
charged particles have been investigated at a limited number
of centers. The primary benefit of therapy with proton or other
413
heavy charged particle beams is the capacity to end the radiation beam at a specific and controllable depth. This may
allow the protection of healthy, normal tissues directly adjacent to tumor-bearing tissues.114 However, the use of proton
therapy has been limited because of the expense of constructing a suitable treatment facility. Several new facilities have
opened in the United States, and pediatric malignancies are
always noted as one of the tumors systems on which the
centers will focus their research efforts. With appropriately
designed studies and comparisons with current state-ofthe-art focal radiation therapy delivered with photon beams,
a determination of the potential benefits of this treatment
modality may be made.
Brachytherapy
Brachytherapy is a method of delivering radiation to a
tumor or tumor bed by placing radioactive sources within
or adjacent to the target volume, usually at the time of surgical
resection and under direct vision. Planning of the dose to be
delivered to the target volume is accomplished after resection
and may use CTor MRI studies; the appropriate strength of the
radioactive source is determined prospectively. Sources
commonly used in children include iridium 192 and iodine 125.
Brachytherapy may consist of either low dose-rate treatments
(approximately 40 to 80 cGy per hour) or high dose-rate treatments (approximately 60 to 100 cGy per minute). Low dose-rate
treatments are delivered during a period of days, often while the
patient remains hospitalized, whereas high dose-rate treatments
are divided into fractions and delivered on several days during 1
to 2 weeks. The primary advantage of brachytherapy is that a radiation source can be placed into or adjacent to the tumor, often
at the time of resection. Preoperative planning and cooperation
between the surgical and radiation oncology teams are necessary
to ensure the appropriate and accurate implementation of
brachytherapy. Nonrhabdomyosarcoma soft tissue sarcomas
and some rhabdomyosarcomas are the pediatric tumors most
commonly treated with brachytherapy.115,116 Most other pediatric solid tumors are not amenable to brachytherapy, however,
because of the tumors behavior (e.g., radioresistance) or its
anatomic location (e.g., retroperitoneal).
Intraoperative radiation therapy has been used intermittently after resection in the management of localized tumors.117 Although of limited availability in the United
States, intraoperative radiation therapy has the distinct advantage of allowing the operative tumor bed to be visible in the
operating theater while radiation is delivered, thereby enhancing the accuracy of delivery and providing the opportunity
to displace or temporarily move mobile crucial structures
(e.g., bowel, bladder) from the field of delivery. The primary
limitation of intraoperative radiation therapy is that it can
deliver only a single fraction of radiation, usually in the 10
to 20 Gy range. Radiation tolerances of normal tissues that
cannot be removed from the treatment field must be respected
and may limit the ability to deliver an effective treatment dose.
PALLIATIVE RADIATION THERAPY

Despite substantial success in the management of pediatric
cancer, some children experience disease recurrence and ultimately die from their malignancy. Palliative radiation therapy
is often a valid intervention for these patients.118 The ultimate
goal of a palliative approach is to maintain quality of life for
414
PART III
radiation-related effects. Essentially all such effects originate from

within the confines of the treatment beams, usually the high-dose
regions of treatment. The most common early and late treatmentrelated effects arising from radiation are listed in Table 28-6.
Despite the arbitrary nature of the division into early and late
effects, this classification distinguishes effects from which the
patient is likely to recover completely from those that are likely
to be permanent. Early treatment-related effects, if managed
appropriately, will resolve as normal, healthy tissues adjacent
to the tumor-bearing tissues, gradually recovering from the effects
of radiation. The period of recovery can range from days to
months, but the patient is often left with minimal sequelae.
Treatment-related effects that are observed later, after the completion of radiation therapy, are more likely to be chronic or permanent. They appear to be related to the normal healing response
of healthy irradiated tissue, resulting in the formation of an
unwanted effect such as fibrosis. Many late treatment effects
can be managed but are not reversible. For children receiving
curative therapy, long-term effects are a primary concern and
are best managed with a preventive approach. Some of the
long-term effects of treatment in children should be ameliorated
by limiting the volume of normal tissue irradiated at high doses
and by implementing approaches that minimize the radiation
dose to adjacent healthy tissues.
patients who will not survive their disease by palliating their

symptoms while minimizing the number of disruptive interventions they must undergo. Painful sites of disease, particularly those with bony involvement, and symptoms resulting
from compression of vital structures, including spinal cord,
peripheral nerves, and respiratory tract are often palliated
with radiation. A palliative course of therapy is highly individualized, and its success or failure depends on the histologic
diagnosis, previous therapy, duration of symptoms, and
symptom(s) being treated.
ACUTE AND LATE TOXICITIES OF RADIATION

THERAPY
The treatment-related effects of radiation therapy, both acute and
chronic, are well described for pediatric and adult patients, but
unfortunately, their incidence and relation to the dose and
volume of treatment are poorly characterized.119 Historically,
treatment-related effects have been classified as acute or late;
an arbitrary time point of 90 days after the completion of treatment defines the division between the two classifications. Current
guidelines for assessing adverse events related to treatment no
longer recognize this arbitrary distinction, but the use of early
and late time points is instructive in the discussion of
TABLE 28-6
Radiation-Related Adverse Events in Children and the Associated Radiation Doses
Organ/Site
Acute
Chronic
Dose Relation
Reference
Skin
Atrophy
Hyperpigmentation
Fibrosis
Ulceration
Necrosis
Myelitis
Decline in cognition
Eye
Conjunctivitis
Thyroid
Heart

Lung
Pneumonitis
Cataract
Retinopathy
Dry eye
Hypothyroidism
Pericarditis
Myocarditis
Valvular disease
Pulmonary fibrosis
Doses more than 40 Gy increase incidence of moist

desquamation

2.5% incidence of brainstem necrosis with doses
of 59.4 Gy
Reduction in intelligence quotient with younger age
and doses of radiation to the supratentorial brain more
than 30 Gy
43% incidence of cataract with doses of total body
irradiation of 12 Gy
121
Subcutaneous tissue
Mucosa
Central nervous
system
Erythema
Desquamation
Edema
Mucositis
Headache
Edema
Bowel
Nausea
Diarrhea
Kidney
Bladder
Muscle
Dysuria
Urgency
Frequency
Edema
Bone
Necrosis
Nephritis
Renal insufficiency
Hemorrhagic cystitis
Fibrosis
Hypoplasia
Hypoplasia
Fracture
Premature physis
closure
122, 123
124
20% incidence at 21 Gy; 61% incidence when > 21 Gy

2.5% incidence of pericarditis at doses of 30 Gy to the
heart
125
126, 127
Increasing risk of pneumonitis with volume of lung

receiving 24 Gy and bleomycin chemotherapy
128

Acute edema in adjacent muscle receiving doses above

40 Gy
Volume of jaw muscles > 40 Gy increases chronic
fibrosis
Increasing reduction in growth above 35 Gy, but effects
seen even at 23.4 Gy
Weight-bearing bones in patient radiated for sarcomas
have a 29% incidence of fracture
129
130132
CHAPTER 28
General Principles of Stem

Cell Transplantation
------------------------------------------------------------------------------------------------------------------------------------------------
Infusion or transplantation of hematopoietic cells capable of

reconstituting the hematopoietic system is used in two broad
instances. First, hematopoietic stem cell transplantation
(HSCT) can be used to replace missing or abnormal components of a defective hematopoietic system. Second, HSCT
can be used to reconstitute elements of the hematopoietic
system destroyed by intensive chemotherapy or radiation
therapy for solid tumors or disorders of the hematopoietic
system itself. The transplanted cells can be the patients own
(i.e., autologous), in which case the cells are obtained before
the administration of myelosuppressive therapy, or they
may come from a donor (i.e., allogeneic) who is generally
a histocompatibility leukocyte antigen (HLA)identical sibling, a mismatched family member, or a partially matched
unrelated donor. The latter two circumstances require immunosuppressive and graft engineering strategies to permit
successful engraftment and avoid graft-versus-host disease.
Hematopoietic progenitor cells are usually obtained from
the bone marrow or peripheral blood. They are the crucial
component of the transplant, because they are capable of
self-renewal and therefore long-term production of cells of
the various hematopoietic lineages. Occasionally, when available, banked umbilical cord blood may be used as the source
of hematopoietic stem cells (HSCs). In general, although
autologous cells are the safest to use for HSCT, they may be
contaminated with tumor cells. Graft-versus-host disease,
which may occur with allogeneic HSCT, can be life threatening, but a modest graft-versus-host reaction may be beneficial
if directed against the hosts tumor cells.
Bone marrow is normally harvested from the posterior
iliac crest to a total volume of 10 to 20 mL/kg body weight
of the recipient. Peripheral blood stem cells are harvested
after their mobilization with recombinant granulocyte
colony-stimulating factor, given daily for up to a week before
harvest. The exact nature of the crucial cellular component
responsible for the reconstitution of the hematopoietic
system is unknown, but the number of cells having the
surface marker CD34 has been shown to be related to the
rate of engraftment.120 Before HSCT, the recipient receives
a preparative (or conditioning) chemotherapeutic regimen.
This treatment serves several purposes, including killing
residual tumor cells, providing immunosuppression for
allogeneic HSCT, and providing space in the marrow into
which transplanted HSCs can engraft. Before reinfusion,
the HSC product may be manipulated ex vivo to enrich
it for putative progenitor cells (e.g., CD34 or CD133
cells), using positive or negative selection methods to facilitate hematopoietic reconstitution; to remove donor T lymphocytes, thereby decreasing the risk of graft-versus-host
disease in allogeneic HSCT; or to purge contaminating tumor
cells from the product used in autologous HSCT.
Complications of HSCT can be significant. The most common early complication is infection, which results from the
transient but profound immunosuppression of the patient,
combined with the breakdown of mucosal barriers. Another
common complication is veno-occlusive disease, which is
characterized clinically by painful enlargement of the liver,
jaundice, and fluid retention. Ultrasound examination shows
415
reversal of flow in the portal vein. Liver biopsy samples show a

classic histologic appearance of obliterated hepatic venules
and necrosis of centrilobular hepatocytes. There is no specific
treatment for this condition; only supportive care can be
given, and mild or moderate veno-occlusive disease is self
limited. Other acute complications of HSCT include graftversus-host disease, a process mediated by donor T cells
targeting host cells with antigenic disparities, and graft
failure. Late complications include chronic graft-versus-host
disease, endocrine insufficiency, secondary malignancies,
growth failure, and other sequelae related to the use of total-body irradiation as part of some preparatory regimens.
Nevertheless, despite the toxicity, HSCT is now an integral
part of successful therapy for many high-risk malignancies
in children.
Clinical Trials
------------------------------------------------------------------------------------------------------------------------------------------------
As previously stated, the past 40 years have seen a significant

increase in overall survival rates for children with cancer. This
increase has been achieved through the development of new
drugs and treatment approaches, improved supportive care,
and better diagnostic modalities to permit earlier cancer detection. The benefits of these advances have been confirmed by
carefully designed and analyzed clinical trials. Because childhood cancer is relatively rare, excellent organization and planning of these trials are essential. In the United States and other
participating countries, clinical trials are largely conducted by
the Childrens Oncology Group, with smaller pilot studies being run by large individual institutions or small consortia.
Clinical trials are generally divided into three phases. Phase
I studies are designed to evaluate the potential toxicity of a
new diagnostic or therapeutic agent. Small numbers of patients are usually required for a phase I study, which typically
uses a dose-escalating design in which cohorts of patients are
observed for signs of toxicity before they advance to higher
doses. The end point of this type of study is generally a determination of the safety of the agent or the maximum tolerated
dose (or both). However, the increasing number of biologic
reagents being introduced and tested may require a shift to
the assessment of the optimal biologic dose. Enrollment in a
phase I toxicity study is often restricted to patients whose disease has not responded to conventional, or standard-of-care,
therapy. Phase II trials are conducted to determine whether a
new agent or treatment approach is sufficiently efficacious to
warrant further study. Phase II agents are often given to newly
diagnosed patients before they begin or just after they complete standard therapy. The testing of new agents in an
upfront window (i.e., before standard therapy) has been
shown not to have an adverse effect on the efficacy of delayed
standard therapy. Finally, phase III studies are designed to
compare the efficacy of an experimental therapy with that
of standard therapy. They are best done as prospective, randomized trials, but often, because of small patient numbers,
a phase III study is done by comparing the efficacy of an
experimental therapy with that of standard therapy given to
historical control subjects. It is through such systematic
assessment of the risks and benefits of new therapies that approaches are rejected or accepted as the new standard of care
and the field of pediatric oncology is advanced.
416
PART III
Conclusion
------------------------------------------------------------------------------------------------------------------------------------------------
Advances in molecular genetic research in the past 3 decades

have led to an increased understanding of the genetic events
in the pathogenesis and progression of human malignancies,
including those of childhood. A number of pediatric malignancies serve as models for the molecular genetic approach to cancer. The pediatric experience highlights the utility of molecular
analysis for a variety of purposes. Demonstration of tumorspecific translocations by cytogenetics, FISH, and RT-PCR confirms histopathologic diagnoses. Detection of chromosomal
abnormalities, gene overexpression, and gene amplification is
used in risk stratification and treatment planning. Elucidation
of pathways involving tumor suppressor genes has increased
our understanding of syndromes associated with cancer and

has led the way for genetic screening and counseling and prophylactic surgical intervention. And in the near future, translation of the molecular profile of a given tumor will form the basis
of a new therapeutic approach. Treatment will be tailored such
that patients with biologically high-risk tumors receive intensified regimens to achieve a cure, whereas patients with biologically low-risk tumors may experience a cure and benefit
from the lower toxicity of nonintensive therapy. Elucidation
of the complex molecular pathways involved in tumorigenesis
will also encourage the production of targeted anticancer agents
with high specificity, efficacy, and therapeutic index.
The complete reference list is available online at www.
expertconsult.com.
CHAPTER 29
Biopsy Techniques
for Children
with Cancer
James D. Geiger and Douglas C. Barnhart
The importance of biopsy techniques in the management of

children with cancer has increased as the use of preoperative chemotherapy has become commonplace for many
childhood cancers. Historically, definitive diagnosis was
made at the time of surgical resection of the primary tumor.
Currently, many children will undergo percutaneous, minimal access surgical, or open incisional biopsy rather than
initial resection. Moreover, with increasing understanding
of the molecular changes associated with these malignancies, definitive diagnosis can be accomplished with smaller
specimens. This should allow a decrease in the morbidity
associated with establishing the diagnosis of solid malignancies in children.
Ironically, this progression to less invasive biopsy has
complicated rather than simplified the selection of technique
in individual cases, as multiple factors must be considered. Percutaneous needle biopsy,1,2 minimal access surgical biopsy,3 and
open biopsy have all been demonstrated to be effective in safely
establishing initial diagnosis as well as verification of recurrent or
metastatic disease. However, the success of these techniques is
obviously dependent upon local institutional experience, which
must be considered in the selection. In addition, it is critical to

realize that many of the advances in risk stratification and
improved therapy of pediatric malignancies has been facilitated by the development of large accessible tumor banks
and the associated biology studies. Without large biopsy
specimens, these tumor banks and the development of research cell lines would not have been possible. For a number
of tumors, including neuroblastoma and Wilms tumor, collection of such specimens remains important to further our
understanding of the disease.
Biopsies may be required in a variety of settings including
primary diagnosis, determination of metastatic disease, and
assessment for viable tumor in residual masses after therapy.
Therefore a biopsy must be considered as a component of
the overall plan of care and not simply as a surgical procedure.
It is, therefore, essential that the surgeon have a thorough understanding of the therapeutic plan prior to performing a biopsy. This is well-illustrated in the current management of a
child with a Wilms tumor and a solitary pulmonary nodule.
Standard therapy for a pulmonary metastasis is lung irradiation. It could, therefore, be important to histologically confirm
this metastasis by excisional biopsy prior to proceeding. This
approach is complicated, however, by a current Childrens
Oncology Group research question of whether children in
whom the pulmonary lesions respond completely after
6 weeks of chemotherapy can be spared lung irradiation
and more intensive chemotherapy. In this research protocol
setting, resection of this solitary lesion would be contraindicated, because it would commit the child to lung irradiation
and more intensive medical treatment.
Current pediatric oncology protocols use risk-stratified
treatment regimens.4 Information needed from biopsy specimens is disease specific. The surgeon must be knowledgeable
about the stratification schema that will be used for multimodality therapy to select the biopsy method that will be least morbid and yet yield all essential information. This concept is
demonstrated by considering two patients with abdominal
masses suggestive of neuroblastoma with apparent bone marrow involvement as they would be treated under the current
Childrens Oncology Group schema. The first patient is less
than 1 year of age. This patients treatment group could be
low, intermediate, or high risk. Risk group assignment will require MYC-N amplification status, Shimada histology status,
and DNA ploidy to determine stratification. This will require
sampling of the primary lesion with an adequate sample of
the tumor to allow Shimada staging. In contrast, an older child
with similar presentation would be classified as high risk, regardless of any of the previous factors. Therefore one could confirm the diagnosis and assign a risk group with bone marrow
biopsy alone.5 Clearly, knowledge of the multimodality therapy
decision making is essential in selection of biopsy technique.
Handling of Specimens
------------------------------------------------------------------------------------------------------------------------------------------------
Historically most diagnoses were made based on hematoxylin

and eosin histology performed on permanent sections. This
was supplemented by immunohistochemistry, which could
similarly be performed on formalin fixed specimens. There
has been extensive progress made in the molecular diagnosis
of childhood malignancies, including recognition of genetic
417
418
PART III
aberrations, which has both diagnostic and prognostic significance.69 Techniques used to detect these changes include reverse transcriptasepolymerase chain reaction (rt-PCR),10
fluorescence in situ hybridization (FISH), microarray analysis,
and flow cytometry. Inappropriate specimen handling can
preclude these analyses. For example, phenotypic classification of lymphoma cannot be performed using flow cytometry
on formalin-fixed lymph nodes. Given the rapidly evolving
field of molecular diagnosis, it is essential that the surgeon
consult with the pathologist regarding specimen handling
prior to performing the biopsy. Additionally, if the patient is
eligible for a research protocol, care must be taken to assure
the specimen is handled in accordance with the protocol requirements. This requires a coordinated effort by the surgeon,
medical oncologist, and pathologist.
Percutaneous Needle Biopsy

------------------------------------------------------------------------------------------------------------------------------------------------
Fine-needle aspiration was first introduced as a technique to obtain specimens for cytopathology by Grieg and Gray in 1904.
Jereb and colleagues reported success with the use of needle biopsy for the diagnosis of pediatric solid tumors in 1978.11 Subsequently, extensive experience from multiple institutions has
confirmed the accuracy and safety of both needle aspiration
and core needle biopsy techniques. The appeal of these techniques is that they both may provide diagnosis without requiring a significant delay in therapy and can be performed as
outpatient procedures. Needle biopsies are often performed under either general anesthesia or sedation. In selected older children, some sites may be biopsied under local anesthesia alone.12
Percutaneous needle biopsies may be performed by palpation in the extremities and other superficial locations such as
lymph nodes. However, deeper biopsies require guidance with
either ultrasonography or CTscan. Ultrasonography that can be
supplemented with Doppler mode allows clear identification of
large vessels and other structures and provides real-time visualization as the needle is advanced.13 Some core needle devices
also deposit a small air bubble that allows verification of the site
that was biopsied. CT scan, on the other hand, allows clear visualization of aerated lung and is not obscured by bowel gas.14
It also allows measurement and planning of depth of biopsy.1
Decision making regarding image guidance is made in conjunction with the radiologist, and ideally, biopsies should be performed with both modalities available if required.
Fine-Needle Aspiration Biopsy

------------------------------------------------------------------------------------------------------------------------------------------------
Fine-needle aspiration biopsy (FNAB) holds the obvious appeal

of being the least invasive of all biopsy techniques. It is typically
performed using a 22- to 25-gauge needle with multiple passes
into the lesion if necessary. Successful diagnosis using FNAB requires coordination with an experienced cytopathologist. To
improve the diagnostic yield, the specimens should be examined immediately by the cytopathologist. Additional aspirations
may be taken if initial samples are inadequate.15 Large series
with fine-needle aspirates in both children and adults have confirmed the safety of the technique.16,17
Historically, diagnosis using fine-needle aspiration was
based primary on cytologic appearance with conventional
stains and light microscopy. Successful diagnosis using FNAB
is dependent upon the availability of an experienced cytopathologist. In adult patients with the higher prevalence of carcinomas, FNA is a popular method for confirming the
presence of malignancy in suspicious lesions. Often, in these
adult cases, a diagnosis of carcinoma and primary site are sufficient to make initial treatment decisions. However, given the
fact that multimodality therapy is histiotype-specific in pediatric patients, FNAB has been used less frequently in children.
Recent application of molecular techniques and electron microscopy to supplement light microscopy has increased the
histiotype specificity of FNAB and may lead to increased application in pediatric solid malignancies.18,19 FNAB has been
used in several pediatric settings with sufficient data reported
for consideration.
The use of FNAB in the evaluation of thyroid nodules in
adults is well-established. Although thyroid nodules are less
common in children, the techniques and interpretation of
FNAB are similar to those used in adults.20 Given the good degree of specificity, FNAB may be considered a standard component of evaluation of thyroid nodules in children.21
Another relatively straight forward application and interpretation of fine-needle aspirate biopsy is in the verification
of metastatic or recurrent disease in the setting of a previously
characterized primary tumor.22 In this context, the verification
of the presence of malignant cells may be sufficient to guide
further clinical decisions. This least invasive biopsy method
is particularly appealing in these patients who may already
be immunologically or physiologically compromised.
There is a limited body of literature on the use of FNAB in
the diagnosis of sarcomas. Osteosarcoma has been diagnosed
by the use of fine-needle aspirates, with definitive diagnosis
being obtained in 65% to 92% of patients. The technique is
as accurate in children as it is in adults.23 The use of FNAB
in soft tissue tumors has been facilitated by the recognition
of cytogenetic abnormalities and fusion proteins that are specific to these tumor types.17,19,24 However, caution should be
exercised in the use of FNAB in this setting, because the reported
series come from a limited number of institutions with extensive
experience in cytologic interpretation. The use of FNAB in diagnosing sarcoma has not gained widespread use.
Fine-needle aspiration has not been widely used for the diagnosis of small, round, blue cell tumors of childhood. However, with the increasing availability of ancillary studies, such
as electron microscopy, immunocytochemistry, DNA ploidy,
cytogenetics, and fluorescent in situ hybridization, its use
may become more common.25 Use of FNAB for the evaluation
of head and neck masses in children has been reported to have
good sensitivity and specificity.26,27 The results of these series,
however, should be interpreted with caution, because the majority of aspirates diagnosed as reactive lymphadenopathy and
the number of new malignant diagnoses was small. In addition, false-negative FNAB diagnosis occurred frequently in patients ultimately diagnosed with lymphoma in other series
(not specifically isolated to the head and neck).15
Core Needle Biopsy

------------------------------------------------------------------------------------------------------------------------------------------------
The advantage of core needle biopsy versus fine-needle aspiration is that it provides a sample sufficient in size to allow histologic examination rather than only cytologic examination.
In addition, it can provide sufficient tissue for molecular
CHAPTER 29
evaluation. Despite the widespread use of this technique in

adults, its application in children has not been as common.
Various core needle devices may be used. These typically
range in size from 14 to 18 gauge. These needles are designed
so that a cutting sheath advances over the core of the needle to
obtain a biopsy that is protected within the sheath as the needle is withdrawn. This cutting sheath may be advanced either
manually (e.g., Tru-Cut, Baxter Travenol, Deerfield, Ill.) or by
a spring-loaded firing system (e.g., Biopty, Bard Urological,
Murray Hill, NJ) (Fig. 29-1). There are no data directly comparing the quality of specimens obtained with these two systems in pediatric malignancies. The faster deployment of the
spring-loaded systems may result in less crush artifact, which
FIGURE 29-1 Two commonly used core needle biopsy devices. The upper device is a 14-gauge Tru-Cut needle (Allegiance, Cardinal Health). It is
advanced into the region of interest, and then the inner needle is advanced. The outer sheath is manually advanced over the inner needle to
obtain a core. The lower device is a 16-gauge Monopty biopsy device
(Bard). It is spring-loaded and is activated after the tip is advanced into
the region of interest. The spring-loaded mechanism automatically
sequentially advances the obturator and the cannula.
BIOPSY TECHNIQUES FOR CHILDREN WITH CANCER
419
has been demonstrated in pediatric kidney biopsies. Regardless of the system used, visual inspection of the core biopsy is
necessary to verify adequate sampling. The number of passes
required with a core needle is dependent upon the purpose of
the biopsy and the consistency of the tissue being biopsied.
For primary tumor diagnosis, multiple cores are typically required to obtain sufficient tissue for biological studies. Alternatively, in pulmonary lesions evaluated for metastatic disease,
a single pass was usually sufficient in most series.
Several large series have demonstrated the utility of core
needle biopsies in children. The larger, more recent series
are summarized on Table 29-1. The three most common
scenarios for which percutaneous biopsies in children with
malignancies are used are diagnosis of primary tumors,
evaluation for possible recurrent disease, and evaluation of
pulmonary lesions.
Success with core needle biopsy has been demonstrated in
a wide variety of anatomic locations. These include neck, mediastinum, lung, peritoneal cavity, liver, retroperitoneum,
kidney, adrenal, pelvis, and extremities.1,2,12,13,28 Core needle
biopsies have been demonstrated to be effective in obtaining
adequate tissue for both primary diagnoses and confirmation
of recurrence in these series. In the largest series of pediatric
oncologic core needle biopsies, multiple passes were typically
performed (median 6 and maximum 17). With this repetitive sampling, adequate diagnostic tissue was obtained for
histologic and biological studies, obviating the need for operative biopsy in a wide variety of pediatric cancers. No patients
in these series suffered procedure-related deaths or required
operative therapy for procedural complications.28
The other common use of percutaneous core biopsies in
pediatric oncology patients is in the evaluation of pulmonary
nodules. Pulmonary nodules can be biopsied under either CT
scan29 or ultrasound guidance, often determined by the size
and location of the lesions.30 These procedures may be performed under either sedation or general anesthesia with
TABLE 29-1
Series of Percutaneous Biopsies in Children
Number of
Children
Number of
Biopsies (Total/
Malignancy)
Skoldenberg
(2002)13
110
147/84
HayesJordan
(2003)52
32
35/23
Cahill
(2004)29
64
75/24
Fontalvo
(2005)30
33
38/32
Author (Year)
Garrett
(2005)28
202/202
Method
Diagnostic
Yield
US-guided core biopsies of wide

range of tumors for initial diagnosis
and evaluation for recurrence
US- or CT-guided core needle
biopsies of pulmonary lesions under
general anesthesia
89%
CT-guided core or FNAB of

pulmonary lesions with sedation or
general anesthesia
US-guided core needle of peripheral
pulmonary lesions with general
anesthesia and controlled ventilation
85%
US-/CT- or fluoroscopic-guided core

needle biopsies of wide range of
tumors for initial diagnosis and
evaluation for recurrence
93% overall
98% initial
diagnosis 88%
suspected
recurrence
CT, computed tomography; FNAB, fine-needle aspiration biopsy; US, ultrasonography.
80%
84%
Comments
Patients with nondiagnostic biopsy

underwent repeat core needle or
thoracoscopic biopsy; 10% small
pneumothorax or hemothoraxnone
required drainage
One false negative for Ewing sarcoma;
one tension pneumothorax required
drainage
Included small lesions (24% < 5 mm);
10% pneumothoraxnone required
drainage (series partially overlaps with
Hayes-Jordan, 2003)
Multiple passes typically taken
(median 6); accomplished diagnosis,
including biological studies without
operative biopsies
420
PART III
controlled ventilation. Typically, general anesthesia would be

used in younger children or in children with smaller or deeper
pulmonary lesions. Current series report success in more than
80% of lesions, including lesions less than 1 cm in size. Surprisingly, pneumothoraces are relatively uncommon, occurring in only 10% of children. The majority of these are
managed without placement of a thoracostomy tube.
Needle tract recurrence represents an oncologic complication specific to this biopsy technique. Estimates of this complication in adults vary widely, ranging from 3.4% in
hepatocellular carcinoma31 to 1:8500 in thoracic tumors.32
Obviously, the incidence of this complication is influenced
by several factors. Immunologic, chemotherapeutic, and radiotherapeutic effects will decrease the likelihood of needle
tract recurrence. The larger needles used for core needle biopsies are associated with a greater risk than the fine needles
used for aspiration.33 The cases series cited previously report
no needle track recurrences in children.
Minimal Access Surgery

------------------------------------------------------------------------------------------------------------------------------------------------
Laparoscopy and thoracoscopy have become commonplace in

general pediatric surgery, and both techniques are now used in
cancer diagnosis and therapy. Gans and Berci first reported experience with multiple endoscopic techniques in children in
1971.34 Interestingly, one of the chief applications for laparoscopy, which they advocated, was for guidance of biopsy of
metastatic implants. Subsequently, the application of both laparoscopy and thoracoscopy has grown in the initial diagnostic
technique for childhood malignancies and for the assessment
of refractory or metastatic disease.
LAPAROSCOPY
Laparoscopy affords several advantages for the evaluation of
the abdominal cavity in children with childhood cancer. First,
it provides the opportunity to completely examine the peritoneal cavity. A systematic examination of all peritoneal surfaces
can be performed. The entire length of the bowel may be examined along with mesenteric lymph nodes. Multiple biopsies
can easily be obtained. The second chief advantage of laparoscopy is decreased physiologic stress in children who may
already be critically ill. Finally, as in all minimally invasive
procedures, postoperative pain is reduced and recovery is
hastened.35 The main disadvantages of laparoscopy are the
limited ability to assess retroperitoneal structures and the loss
of tactile evaluation of deep lesions.
Diagnostic laparoscopy with biopsy has been used in several settings in the management of children with solid malignancies.3,35 Biopsies obtained using laparoscopic techniques
have a high rate of success in yielding diagnostic tissue.3,36
Laparoscopy allows the surgeon to obtain larger tissue samples than may be obtained with core needle biopsy. This is particularly relevant if larger samples are required for biological
studies. In the initial diagnosis, laparoscopy aids in the identification of site of origin of large abdominal masses. Laparoscopy has been shown to be superior to computerized
tomography in assessing intraperitoneal neoplasms and for
the evaluation of ascites. For example, laparoscopy allows direct assessment of whether a pelvic mass arises from the ovary
or bladder neck, which may be difficult to distinguish by
radiographic studies. Direct visualization with laparoscopy

has been used to assess the resectability of hepatoblastoma.
During the course of treatment, laparoscopy may be used to
assess new metastatic disease or to assess initial tumor response as a second-look procedure.3
One area of concern with the use of laparoscopy in oncology has been the issue of port-site recurrence. There are relatively limited data on this issue in children. The Childrens
Cancer Group retrospective study of 85 children noted
no port-site recurrences.3 A survey of Japanese pediatric laparoscopic surgeons reported 85 laparoscopic and 44 thoracoscopic procedures with no port-site recurrences.37 It should
be noted, however, that 104 of these tumors were neuroblastomas, with many being detected by mass screening. The
general applicability of this data may, therefore, be limited.
A port-site metastasis has been reported in a child with Burkitt
lymphoma.38 Given the difference in tumor biology between
adult adenocarcinomas and pediatric neoplasms, which often
have a marked response to neoadjuvant therapy, it is difficult
to draw conclusions from the adult literature. Certainly additional surveillance for this issue in pediatric tumors is merited.
Laparoscopy in children is typically performed under general anesthesia to facilitate tolerance of pneumoperitoneum.
The only absolute contraindication to laparoscopic evaluation
is cardiopulmonary instability, which would preclude safe insufflation of the peritoneal cavity. The supine position is used
most commonly and affords a complete view of the peritoneal
cavity. To facilitate visualization, a 30-degree laparoscope is
used along with at least two additional ports for manipulation
and retraction. Ascites should be collected for cytologic analysis and all peritoneal surfaces inspected. Incisional biopsies
can be performed using laparoscopic scissors. Hemostasis is
achieved with a combination of electrocautery and hemostatic
agents (as discussed later in the section on open incisional biopsy) or by tissue approximation via laparoscopic suturing.
Biopsy specimens are typically retrieved using a specimen
bag. This reduces the chance of specimen destruction during
retrieval and may decrease the risk of port-site recurrence.
Cup biopsy forceps can be used to obtain specimens as well.
Core needle biopsies can be directed by laparoscopy and be
used to sample retroperitoneal, intraperitoneal, or hepatic
masses. For deep-seated lesions, such as intrahepatic lesions,
laparoscopic ultrasonography can be used to guide biopsy
procedures and to compensate for the inability to palpate
tissues.39,40
Complications associated with laparoscopic diagnosis and
treatment of solid tumors in children are infrequent. The need
for conversion to unplanned open operation has similarly
been low.3,35,41,42
Thoracoscopy
------------------------------------------------------------------------------------------------------------------------------------------------
The initial experience with thoracoscopy in children was

reported by Rodgers in 1976 and included two oncology patients (Ewing sarcoma and recurrent Hodgkin lymphoma).43
Since this initial report, thoracoscopy has become widely used
for the evaluation of thoracic lesions in children for several
reasons. Postoperative pain associated with thoracoscopic biopsy or resection is markedly decreased compared with that
seen with thoracotomy. Moreover, thoracoscopy allows nearcomplete visualization of all parietal and visceral pleural
CHAPTER 29
BIOPSY TECHNIQUES FOR CHILDREN WITH CANCER
421
FIGURE 29-2 Computed tomography (CT) scans obtained at the time of diagnosis of a new abdominal mass in a 5-year-old boy. A, Abdominal and pelvic
CT scan shows a large left-sided renal mass. B, Chest CT scan demonstrates a single 8-mm pulmonary nodule in the left upper lobe. No other pulmonary
lesions were identified. At the time of nephrectomy, a thoracoscopic excisional biopsy of the lung lesion was performed. Final pathology of the kidney
demonstrated a stage II-favorable-histology Wilms tumor, and the lung pathology showed a hyalinized granuloma.
surfaces, which cannot be accomplished with a thoracotomy.

Additionally, in most children the mediastinum does not contain a significant amount of adipose tissue and, therefore, can
be inspected thoracoscopically.
Although primary neoplasms of the lung are rare in children, pulmonary lesions are often a confounding issue in
the treatment of children with cancer.44 The most common tumor to have early pulmonary metastases is Wilms tumor. Pulmonary metastases are also common with bone and soft tissue
sarcomas, hepatic tumors, teratocarcinomas, and melanomas.
Thoracoscopy is frequently used to evaluate for the presence
of metastases either at the time of initial diagnosis or after
follow-up imaging. Difficulty in distinguishing an opportunistic infection versus new metastatic disease is a common
clinical scenario during the course of therapy. In areas with
endemic granulomatous disease, thoracoscopy can also be
helpful at the time of diagnosis (Fig. 29-2; case with histoplasmosis granuloma with new diagnosis of a Wilms tumor).
The diagnostic accuracy for thoracoscopic biopsies in this
setting is very high.41,44,45
Mediastinal lesions may also be biopsied or resected using
thoracoscopy.46,47 Thoracoscopy provides clear visualization
of both the anterior and posterior mediastinum, even in small
children; therefore we prefer it rather than mediastinoscopy
for evaluation of mediastinal lesions in children.
The only absolute contraindications to thoracoscopy are
complete obliteration of the pleural space and the inability
to tolerate single-lung ventilation when complete collapse of
the lung is required.
Thoracoscopy in children is typically performed under
general anesthesia with mechanical ventilation. Visualization
is facilitated by single-lung ventilation if possible and supplemented with insufflation. In older children, this may be accomplished with a double-lumen endotracheal tube and, in
smaller children, by mainstem intubation of the contralateral
side. If selective ventilation is difficult to achieve or poorly tolerated by the patient, low-pressure insufflation (5 to 8 cm of
water pressure) with carbon dioxide assists with visualization.
The anesthesiologist must monitor for any adverse effects from

this controlled tension pneumothorax. It can be evacuated
rapidly if need be, but it is typically well tolerated. The child
is positioned in the lateral thoracotomy position. Hyperextension of the chest increases the intercostal spaces and will facilitate movement of the thoracoscopy ports. This positioning
should be adjusted for mediastinal lesions. For anterior lesions, a more supine position is used, and for posterior lesions
the patient is positioned more prone. The initial port is placed
in the midaxillary line using blunt dissection. Additional ports
are placed under thoracoscopic guidance at sites based upon
the location of the lesion of interest. A 30-degree thoracoscope
is helpful to achieve complete visualization of all pleural surfaces. Complete inspection is also facilitated by the use of multiple port sites.
Careful correlation with cross-sectional imaging is essential
to successful thoracoscopic sampling, particularly of smaller
lesions. Pleural-based or subpleural pulmonary lesions are often apparent when the lung is deflated. These can be resected
using endoscopic stapling devices and retrieved using specimen bags. Identification of deeper lesions is more challenging.
Complete collapse of the lung allows identification of larger
lesions. Biopsy of smaller lesions can be based on anatomic
location if the location by CT scan is specific, such as apical,
lingular, or basilar. CT-guided localization may be performed
immediately before surgery to assure correct identification of
the area of concern. The lesion may be marked by injection
with methylene blue or, preferably, stained autologous blood,
which is less prone to diffuse.48,49 Lesions can be concomitantly marked with placement of a fine wire50 or microcoils,51
which can facilitate identification under intraoperative fluoroscopy. These localization techniques have been very effective
in obtaining accurate biopsies in children.48,49,52 Intrathoracic ultrasonography may be helpful in localizing deeper parenchymal lesions.53 However, this technique is not widely
used, and assessment of its efficacy in children is limited.
After sampling of tissues of interest is completed, the
pneumothorax may be evacuated with a small catheter to
422
PART III
form a water seal. Unless extensive pulmonary biopsies are

performed or the lung is otherwise diseased, a thoracostomy
tube is often not required. Most children may be discharged
the next day, and chemotherapy may be started promptly.49
Thoracoscopic techniques are highly effective in achieving
diagnosis. Most pediatric series report a rate of success in
obtaining accurate diagnostic tissue in almost all cases.3,41,44,52
Complications during diagnostic thoracoscopy are rare. Pneumothorax or persistent air leakage may occur in children with
underlying parenchymal lung disease or those requiring highpressure ventilatory support.52
Open Incisional Biopsy

------------------------------------------------------------------------------------------------------------------------------------------------
Incisional biopsy remains the gold standard regarding the

quality of tissue sampling if complete excision is not to be performed. Laparotomy or thoracotomy allows large samples to
be obtained under direct vision, which can provide improved
diagnosis compared with needle biopsies. For example, in the
National Wilms Tumor Study Group-4, open biopsy was more
successful than core needle biopsies at identifying anaplasia in
children with bilateral Wilms tumor. Correlation with preoperative imaging allows multiple samples to be obtained if there
is inhomogeneity within the tumor, which would raise concerns about sampling error.
The ability to obtain larger specimens is beneficial not only
in providing tissue for molecular diagnosis and prognosis, but
in providing samples for tissue banking and creation of cell
lines. Samples obtained from these biopsies have provided
the clinical material that allowed the development of the molecular diagnosis and prognosis techniques referred to earlier
in this chapter. Further stratification of risk to allow more precise risk-based therapy remains a major focus for pediatric oncology trials. Finally, specimens that are tissue-banked from
these larger specimens may be used for investigational therapies, such as tumor vaccines.
Several important factors should be considered in performing an open biopsy. The initial biopsy should consider the ultimate operative treatment of the tumor. For example, the
incision for biopsy of an extremity mass should be oriented
parallel to the axis of the limb, and care should be taken to
avoid undermining subcutaneous or fascial planes. This allows subsequent wide local excision to be performed with
minimal additional resection of tissue because of the biopsy.
Likewise, testicular masses should only be biopsied through
an inguinal approach, because a scrotal biopsy incision could
require the addition of a hemiscrotectomy to the subsequent

orchiectomy. Laparotomy for biopsy should be planned to allow subsequent resection through extension of the same
incision.
Significant distortion of anatomic relations can occur with
large retroperitoneal tumors and attention must be paid to
avoid injury to structures, such as ureters or the bile duct,
which may be distended over the mass. The most common
intra-abdominal tumors in children tend to be vascular, and
bleeding from the biopsy site is the most common serious
complication. Strategies to reduce perioperative hemorrhage
include normalization of coagulation parameters preoperatively and adequate operative exposure. Cauterization of the
tumor capsule may help control bleeding, but we have found
direct pressure after packing the biopsy site with oxidized cellulose combined with procoagulants, as described later, to be
more efficient than generous cautery applied to the base of the
biopsy site. If possible, closure of the tumor capsule can help
with hemostasis.
Supplements to achieving hemostasis include topical
agents and fibrin sealants. Commercially available topical
products include gelatin foam pads, microfibrillar collagen,
and oxidized cellulose, which is available as fabric and cottonoid. Fibrin sealants are composed of fibrinogen, thrombin,
and calcium, which are mixed as they are delivered to the tissue to rapidly form a fibrin clot.
Conclusion
------------------------------------------------------------------------------------------------------------------------------------------------
Prior to performing a biopsy of a potential malignancy, the surgeon should consider the likely possible diagnoses. The biopsy should then be planned so that adequate tissue is
obtained and preserved to determine not only diagnosis but
also risk stratification. Percutaneous, minimal access surgical,
and open surgical techniques each have an appropriate place
in the evaluation of potential pediatric malignancies. The use
of these techniques in a systematic, stepwise fashion is appropriate in some patients. The selection of the appropriate biopsy technique should be driven by both the specific
question to be answered by the biopsy and individual institutional experience and resources. Planning an operative biopsy
must account for the anticipated operative approach for definitive resection.
expertconsult.com.
reviews the most frequent renal tumors in children, including

their biologic properties, multidisciplinary therapies, and future challenges.
Wilms Tumor
------------------------------------------------------------------------------------------------------------------------------------------------
CHAPTER 30
Wilms Tumor
Peter F. Ehrlich and Robert C. Shamberger
WT is the most common primary malignant renal tumor

of childhood and comprises 6% of all pediatric tumors.5,6
Outcomes for children with WT improved dramatically over
the last 50 years, with long-term survival in both North
American and European trials approaching 85% (Fig. 30-1).
Survival rates for many of the low-stage tumors are 95% to
99%.7,8 Current treatment protocols for children with WT
were developed through a series of multidisciplinary cooperative group trials in both North America and Europe by the
Childrens Oncology Group (COG), formerly the National
Wilms Tumor Study Group (NWTSG), and the Societe Internationale dOncologie Pediatrique (SIOP). Their series of
well-designed prospective randomized studies provide a large
body of evidence-based knowledge to establish the optimal
surgical, radiotherapy, and chemotherapy treatments for tumors based on the early studies on stage and histology and,
more recently, also on cytogenetic and response-based factors.
There are differences between the approaches of these two
groups that affect staging and risk classification that are critical
to understand when considering outcomes that will be discussed later in the chapter (Table 30-1).
History
------------------------------------------------------------------------------------------------------------------------------------------------
Renal tumors account for 6.3% of cancer diagnoses for children younger than 15 years of age, with a reported incidence
of 7.9 per million. Including adolescents younger than
20 years of age, this drops slightly to 4.4% of cancer diagnoses, with an incidence of 6.2 per million.1 Renal tumors
include Wilms tumor (WT) (also referred to as nephroblastoma or renal embryoma), renal cell carcinoma (RCC), clear
cell sarcoma of the kidney (CCSK), rhabdoid tumor of the
kidney (RTK), congenital mesoblastic nephroma, cystic renal
tumor, and angiomyolipoma.2,3 WT is by far the most common, accounting for approximately 91% of all renal tumors in
childhood. CCSK and RTK were originally considered subtypes of WT, but are now recognized as separate tumors.
RCC comprises 5.9% of renal malignancies in children and
adolescents.1,4
The treatment strategy for children with renal tumors
evolved in conjunction with the definition of these pathologic
subtypes. Treatment is based on traditional risk factors, stage
and histology, and, more recently, on genetic markers. The
goal of risk-based management is to maintain excellent outcomes but at the same time spare children with low-risk tumors intensive chemotherapy and radiation, with their
long-term side-effects, and to intensify therapy for children
with high-risk tumors in an effort to increase their survival.
Despite these advances, children with rhabdoid, renal cell carcinoma, and anaplastic tumors still do poorly. This chapter
WT is named after Carl Max Wilhelm Wilms, a German pathologist and surgeon. He was one of the first to propose that
tumor cells originate during the development of the embryo.
He published his findings in 1897 and 1899 in an influential
monograph titled Die Mischgeschwulste der Niere, which
described seven children with nephroblastoma as part of a
monograph on mixed tumors.9,10 Although reports of successful excision of renal tumors in children appeared in the
end of the 19th century, his name has been indelibly applied
to them. Dr. Thomas Jessop (1837 to 1903), probably performed the first successful nephrectomy at the General Infirmary in Leeds, England, on June 7, 1877, on a 2-year-old child
with hematuria and a tumor of the kidney.11,12
At the beginning of the 20th century, survival for a child
with WT was 5%. Surgery was the first effective treatment
for nephroblastoma and continues to be a critical component
of successful multimodality therapy. Although surgery at that
time was the only option for cure, it carried a significant operative mortality. In 1916, radiation therapy was added by
Friedlander.13 In the late 1930s, Ladd described removing
renal tumors in selected children. His technique included a
large transverse transabdominal approach with early ligation
of the renal vessels and removal of the surrounding Gerota
fat and fascia. This modification improved the outcome in
children with nonmetastatic nephroblastoma to a 32.2% survival at 3 years, with an operative mortality reduced from 23%
to 7%. The basic tenets of this operative procedure described
by Ladd are used today, with the exception of early ligation of
the renal vessels.1215
423
424
PART III
Percent
5-YEAR WT SURVIVAL
100
90
80
70
60
50
40
30
20
10
0
87
92
92
92
73
60
47
WT
20
5
1899
1938
1954
1960
1975
Year
1981
1987
1995
2003
FIGURE 30-1 This graph shows the improved survival of children with
Wilms tumor (WT) over time.
Epidemiology
------------------------------------------------------------------------------------------------------------------------------------------------
In the United States, there are 500 to 550 cases of WT per year.
It is the second most common malignant abdominal tumor in
childhood after neuroblastoma. The risk of developing WT in
the general population is 1:10,000.16 The incidence is slightly
elevated for American and African blacks compared with
whites and is significantly lower in Asians. The mean age at
diagnosis is 36 months, with most children presenting between the ages of 12 and 48 months. Tumors tend to occur
about 6 months later in girls than in boys. WT is rare at greater
than 10 years and at less than 6 months of age. Tumors can be
unilateral or bilateral (Figs. 30-2 and 30-3). Bilateral Wilms
tumors (BWT) occur in 4% to 13% of patients.5,1719 Children
with congenital syndromes associated with WT, such as Beckwith-Wiedemann, have a higher risk of developing BWT.
Congenital anomalies, either isolated or as part of a congenital syndrome, occur in about 10% of children with
WT.20 WAGR syndrome (WT, aniridia, genitourinary malformation, mental retardation) is a rare genetic syndrome associated with a chromosomal defect in 11p13. Children with
WAGR syndrome are at a 30% higher risk of developing
WT than a normal child. Because of the presence of aniridia,
most children with WAGR syndrome are diagnosed at birth.
Children with WAGR account for about 0.75% of all children
with WT.21
Beckwith-Wiedemann syndrome (BWS) is a congenital disorder of growth regulation, affecting 1 in 14,000 children.
Children with BWS have visceromegaly, macroglossia,
omphalocele, and hyperinsulinemic hypoglycemia at birth.
They also have an increased risk of tumor development.
The risk is greatest in the first decade of life and thereafter approaches that of the general population. Three large studies
of children with BWS reported tumor frequencies of 7.1%
(13/183), 7.5% (29/388), and 14% (22/159).2225 The most
frequently observed tumors in BWS are WT and hepatoblastoma, which comprise 43% and 12% of reported cancers, respectively.22,26 Denys-Drash syndrome (DDS) (nephropathy,
renal failure, male pseudohermaphroditism, and WT) is also
associated with an increased risk of WT. Some investigators
have recommended prophylactic nephrectomy in children
with this syndrome once they develop renal failure.27,28 Other
TABLE 30-1
Childrens Oncology Group (COG) and Societe Internationale dOncologie Pediatrique (SIOP) Staging Systems
COG Wilms Tumor Staging
Stage
I
II
III
IV
V
Criteria
The tumor is limited to the kidney and has been completely resected.
The tumor was not ruptured or biopsied prior to removal.
There is no penetration of the renal capsule or involvement of renal sinus vessels.
The tumor extends beyond the capsule of the kidney but was completely resected with no evidence of tumor at or beyond the margins
of resection.
There is penetration of the renal capsule or invasion of the renal sinus vessels.
Gross or microscopic residual tumor remains postoperatively, including inoperable tumor, positive surgical margins, tumor spillage surfaces,
regional lymph node metastases, positive peritoneal cytology, or transected tumor thrombus.
The tumor was ruptured or biopsied prior to removal.
Hematogenous metastases or lymph node metastases outside the abdomen (e.g., lung, liver, bone, brain).
Bilateral renal involvement is present at diagnosis, and each side may be considered to have a stage.
SIOP Staging
Stage
I
II
III
IV
V
Criteria
The tumor is limited to the kidney or surrounded with a fibrous pseudocapsule, if outside the normal contours of the kidney. The renal capsule
or pseudocapsule may be infiltrated with the tumor, but it does not reach the outer surface, and it is completely resected. The tumor may be
protruding (bulging) into the pelvic system and dipping into the ureter, but it is not infiltrating the walls. The vessels of the renal sinus are not
involved. Intrarenal vessels may be involved.
The tumor extends beyond the kidney or penetrates through the renal capsule and/or fibrous pseudocapsule into the perirenal fat, but it
is completely resected. The tumor infiltrates the renal sinus and/or invades blood and lymphatic vessels outside the renal parenchyma, but
it is completely resected. The tumor infiltrates adjacent organs or vena cava, but it is completely resected. The tumor has been surgically
biopsied (wedge biopsy) prior to preoperative chemotherapy or surgery.
There is incomplete excision of the tumor, which extends beyond resection margins (gross or microscopic tumor remains postoperatively).
Any positive lymph nodes are involved. Tumor ruptures before or during surgery (irrespective of other criteria for staging). The tumor has
penetrated the peritoneal surface. Tumor implants are found on the peritoneal surface. The tumor thrombi present at resection, margins of
vessels or ureter are transected or removed piecemeal by surgeon.
Hematogenous metastases (lung, liver, bone, brain, etc.) or lymph node metastases are outside the abdominopelvic region.
Bilateral renal tumors present at diagnosis. Each side has to be substaged according to above classifications.
CHAPTER 30
WILMS TUMOR
425
are several candidate genes that are been investigated and evaluated or are being evaluated in the clinical setting. These are
described later.
LOSS OF HETEROZYGOSITY
AND DNA PLOIDY
FIGURE 30-2 A computed tomography (CT) scan of a unilateral Wilms

tumor.
syndromes associated with WT include hemihypertrophy

and Perlman syndrome. Urologic abnormalities, such as hypospadias, cryptorchidism, and nephromegaly, are also associated with WT.
Molecular Biology and Genetics

------------------------------------------------------------------------------------------------------------------------------------------------
A number of important advances in WT development have occurred since the early 1990s. A detailed description is beyond
the scope of this chapter. Table 30-2 summarizes some of the
key genes and more detailed references are cited.2956 There
Loss of heterozygosity (LOH) refers to loss of genetic material

and allelic uniqueness. LOH was found initially in children
with WT on chromosomes 11p (33% of tumors), 16q
(20%), and 1p (11%). A major aim of the fifth National Wilms
Tumor Study (NWTS-5) was to determine if tumor-specific
LOH for chromosomes 11p, 1p, or 16q was associated with
an adverse prognosis for children with favorable-histology
(FH) WT, a finding suggested in earlier retrospective studies.34 Chromosomes 11p, 16q, and 1p were prospectively
evaluated. Results demonstrated that outcomes for patients
with LOH at 1p and 16q were at least 10% worse than those
without LOH (Figs. 30-4 and 30-5). These findings are used
as determinants of therapy on the current renal tumor studies
of the COG.
A similar but smaller study was reported from the United
Kingdom (United Kingdom Children Cancer Study Group
Wilms Tumor trials 1 to 3) in which a comparable incidence
of LOH for 16q (14%) and 1p (10%) was found, but in this
study there was no association between poor outcomes and
LOH at 1p.42 The reasons for the different results are unclear;
possible explanations include a smaller sample size of the
British study or that the larger doses of doxorubicin used in
the U.K. studies served to eliminate part of the adverse impact
on prognosis.
Analyses from patients with WT have also identified recurrent deletions and translocations involving the short arm of
chromosome 7.43,48,55 Studies suggest a locus of interest
between 7p13 and 7p21, perhaps the POU6F2 gene at
7p14.5759 Clinical correlates of 7p LOH have not been published, and so the exact prognostic role of this possible Wilms
locus, if any, has yet to be determined.
Another aim of NWTS-5 was to determine whether DNA
ploidy status is associated with worse outcome in children
FIGURE 30-3 Two computed tomography (CT) scans of bilateral Wilms tumor at presentation and after 6 weeks of chemotherapy.
426
PART III
TABLE 30-2
Summary of Current Genes Being Investigated in Wilms Tumor
Gene(s)
Location
Function
Clinical Relevance
WT1
11.13
Tumor Supressor Functions in normal kidney development
WT2
11p15.5
Cadherin-associated
protein b1 gene4
3p21
WTX
Xq11.1
Familial Wilms genes
17q and 19q13.3-q13.4
Several gene loci IGF-2

Cell growth and encodes an embryonal growth factor that is
highly expressed in fetal kidney and WT
Genomic imprinting
Cellular adhesion protein that also associates with members
of the T-cell factor (TCF) family of transcription factors to
promote expression of growth-related genes such as c-MYC
and CYCLIN D1
WTX inhibits the Wnt signal transduction to promote
post-translational modification and degradation
Unknown
WAGR syndrome Deletions

Denys-Drash point mutation
BWS syndrome Genomic
imprinting
Highly correlated with WT1

genes
Unknown
Unknown
BWS, Beckwith-Wiedemann syndrome; IGF, insulin growth factor; WAGR, Wilms tumor, aniridia, genitourinary malformation, mental retardation.
1.0
Proportion relapse free
Proportion relapse free
1.0
0.9
0.8
No LOH
LOH 1p only
LOH 16q only
LOH both loci
0.7
0.6
0.9
0.8
No LOH
LOH 1p only
LOH 16q only
LOH both loci
0.7
0.6
0
2
3
Time since diagnosis (years)
Time since diagnosis (years)
FIGURE 30-4 Relapse-free survival by joint loss of heterozygosity (LOH)

at chromosomes 1p and 16q for stage I/II favorable-histology Wilms tumor
patients. (From Grundy PE, Breslow N, Li S, et al: Loss of heterozygosity
for chromosomes 1p and 16q is an adverse prognostic factor in favorablehistology Wilms tumor: A report from the National Wilms Tumor Study
Group. J Clin Oncol 2005;23:7312-7321.)
FIGURE 30-5 Relapse-free survival by joint loss of heterozygosity (LOH)

at chromosomes 1p and 16q for stage III/IV favorable-histology Wilms
tumor patients. (From Grundy PE, Breslow N, Li S, et al: Loss of heterozygosity for chromosomes 1p and 16q is an adverse prognostic factor in
favorable-histology Wilms tumor: A report from the National Wilms Tumor
Study Group. J Clin Oncol 2005;23:7312-7321.)
with favorable-histology DNA index as a prognostic marker:

DNA index greater than 1.5 was strongly associated with anaplastic histology and predictive of poor outcome. However,
DNA content was not predictive of outcome when stratified
by stage and histology.60
anaplastic WT have TP53 mutations. In the current COG

study, one of the aims of the high-risk protocol is to study
the incidence and association of TP53 mutations.
Clinical Presentation
TP53 GENE
The TP53 gene is located on chromosome 17. The Tp53 protein is a negative regulator of cell proliferation and a positive
regulator of apoptosis in response to DNA damaging agents.
TP53 is the most common mutated gene associated with
human cancer. Li-Fraumeni syndrome is a multicancer predisposition syndrome that has constitutional TP53 mutations.61
However, WT rarely develops in Li-Fraumeni syndrome,
and the majority of WT develop in the presence of wild-type
TP53.62 TP53 mutations in WT are almost exclusively found in
tumors with anaplastic histology. Seventy-five percent of
------------------------------------------------------------------------------------------------------------------------------------------------
Most children with WT present with an asymptomatic abdominal mass, often discovered by either a parent or pediatrician.
Nonpalpable tumors are typically discovered by ultrasonography during evaluation for abdominal pain. Gross hematuria
has been reported in 18.2% of patients and microscopic
hematuria in 24.4%. Ten percent of children with WT have
coagulopathy, and 20% to 25% present with hypertension
because of activation of the renin-angiotensin system.63 Fever,
anorexia, and weight loss occur in 10%. Extension of tumor
thrombus into the renal vein can obstruct the spermatic vein
and result in a left varicocele and, in rare cases, tumor
CHAPTER 30
extension into the atrium may produce cardiac malfunction.

Tumor rupture and hemorrhage are also infrequent events that
can present as an acute abdomen.
The differential diagnosis for an abdominal mass includes
neuroblastoma, hepatoblastoma, rhabdomyosarcoma, and
lymphoma. Neuroblastoma is the most common solid abdominal tumor in children. One clinical observation to help distinguish between WT and neuroblastoma is that children with
neuroblastoma are often ill because of extensive metastatic
disease at presentation. In contrast, children with WT are
generally healthy toddlers with a palpable abdominal mass.
Diagnosis
------------------------------------------------------------------------------------------------------------------------------------------------
After an abdominal mass is identified, radiographic imaging is

performed to determine the anatomic location and extent of
the mass. Ultrasonography (US) is a good screening examination of a mass to determine its site of origin and to assess for
possible intravascular or ureteral extension. About 4% of WT
present with inferior vena cava (IVC) or atrial involvement and
11% with renal vein involvement.5,6 Embolization of a caval
thrombus to the pulmonary artery can be lethal, and the presence of a thrombus must be identified preoperatively to prevent this occurrence. US is a sensitive technique to identify
vascular extension.64,65 A computed tomography (CT) scan
of the abdomen will confirm the renal origin of the mass
and determine whether there are bilateral tumors. Early generations of CT scans missed 7% to 10% of bilateral lesions.
Hence, contralateral exploration of the kidney was recommended in NWTSG protocols to assess for bilateral lesions.66
A recent review of children with bilateral WT, however, demonstrated that only 0.25% of bilateral tumors were missed
with modern helical CT scans, all of which were small.67 Based
on these results, bilateral exploration is not recommended in
current protocols from the COG. Although magnetic resonance imaging (MRI) avoids radiation exposure, it has not
been shown to be superior to CT scanning in standard assessments. MRI is currently being evaluated as a method to help
distinguish nephrogenic rests from WT and may be the preferred method to follow children with bilateral WT after
resection.
The common sites of metastatic spread are the lungs and
the liver. Therefore, in addition to abdominal imaging, pulmonary imaging must be performed. In NWTS-4 and NWTS-5,
13% of patients (575 of 4,006) with unilateral favorablehistology tumors presented with pulmonary disease. Initially
this was routinely evaluated based upon a chest radiograph.
In current protocols, it is based upon CT scans.
18
F-fluorodeoxyglucose positron emission tomography
(FDG PET) has not been fully delineated in pediatric cancers.65 It is recognized that FDG PET has an established
role in Hodgkin lymphoma and increasingly in sarcomas in
children, but its role in WT is unclear.68,69
Screening
------------------------------------------------------------------------------------------------------------------------------------------------
Screening is reserved for children at risk for developing WT.

This includes children with genetic syndromes such as BWS,
idiopathic hemihypertrophy (IHH), WAGR, DDS, and Perlman syndrome. Renal ultrasound examination is the preferred
WILMS TUMOR
427
modality to screen for WT. It is widely available, noninvasive,

does not involve radiation exposure, and generally does not
require use of sedation. It is recommended that children be
scanned every 3 to 4 months. Debaun and colleagues assessed
the cost effectiveness of screening for WT and hepatoblastoma
in children with Beckwith-Wiedemann syndrome (BWS).70 In
this analysis, screening a child with BWS from birth until
4 years of age resulted in a cost per life-year saved of
$9,642, while continuing until 7 years of age resulted in a cost
per life-year saved of $14,740, although it is not truly established that the rate of cure or event-free survival (EFS) is
higher based on this early monitoring protocol. Three retrospective studies have evaluated screening in children at risk
for WT. One study from the United Kingdom of 41 children
with WT and aniridia, BWS, or IHH showed no difference
in outcome or stage distribution between screened and
unscreened populations.71 In a second study of BWS/IHH,
Choyke and colleagues demonstrated that evaluation by US
every 3 months until age 8 years in 12 children with BWS
lowered the proportion of patients with late-stage tumors to
0%, which was significantly reduced compared with the
42% incidence of late-stage tumors in 59 unscreened patients
with BWS/IHH.72 A third study analyzed the impact of surveillance in children with aniridia, BWS, and IHH who had
developed WT.73 There was a higher proportion of stage I
tumors identified in children who underwent routine screening
than in those who did not. Although ultrasonography is easy,
false-positive results have been reported and have led to unnecessary investigations and surgery in patients who had benign
lesions, such as cysts, nephrogenic rests, or foci of renal dysplasia, supporting the use of either MRI or CT to further define
the lesions before surgical intervention.7274 The U.K. Wilms
Tumor Surveillance Working Group suggests that surveillance
should be offered to children who are at a greater than 5%
risk of WT.75
Children with Perlman syndrome are at a significantly
increased risk of WT; therefore surveillance specifically for
WT is warranted. Based on a review by Tan and colleagues,
there is currently insufficient evidence to justify tumor surveillance in Sotos, Weaver, Proteus, and Bannayan-Riley
Ruvalcaba syndromes or the syndrome of macrocephaly-cutis
marmorata telangiectatica congenita. Of interest, children
with Klippel-Trenaunay syndrome (KTS) had been considered
to be at increased risk for developing WT. In a 2004 study
by Fishman and colleagues, the risk of developing WT in
children was assessed using the NWTSG database.76 The risk
of WT in children with KTS was no different than in the
general population, and thus routine ultrasonography surveillance is not recommended.
Pathology
------------------------------------------------------------------------------------------------------------------------------------------------
Tumor histology is a major determinant of therapeutic stratification for children with WT. The diagnostic classification of
pediatric renal tumors has benefited from central review of tumors from patients treated in the cooperative group trials.77
This success has enabled the introduction of disease-specific
and risk-based therapy. For example, clear cell sarcoma of
the kidney (CCSK) and malignant rhabdoid tumor (MRT)
were initially considered to be variants of WT and were
428
PART III
managed with chemotherapeutic agents for WT, but they are

now considered distinct entities with separate therapies.
WT are embryonal tumors containing components seen in
normal developing kidneys. The classic WT consists of three
elements: blastemal, stromal, and epithelial tubules. Tumors
contain various proportions of each of these elements. Triphasic patterns containing blastemal, stromal, and epithelial cell
types are the most characteristic, but biphasic and monophasic lesions occur.78 Less frequently, abnormal mucinous or
squamous epithelium, skeletal muscle, cartilage, osteoid, or
fat are found in WT.79
When the tumors are monophasic, they can be very invasive and difficult to distinguish from other childhood tumors,
such as primitive neuroectodermal tumor, neuroblastoma,
and lymphoma. Monophasic undifferentiated stromal WT
look like sarcomas, such as clear cell sarcoma of the kidney,
congenital mesoblastic nephroma, or synovial sarcoma. Other
WT may have differing amounts of skeletal-muscle differentiation, from well-differentiated (rhabdomyomatous) to poorly
differentiated (rhabdomyoblastic) skeletal muscle. A WT that
is entirely tubular and papillary can be difficult to distinguish
from papillary renal cell carcinoma.79
WT are divided into two groups: those with favorable
histology and those with unfavorable histology. Favorablehistology tumors comprise 90% of the unilateral and bilateral
tumors.
Anaplastic histology is considered unfavorable histology
along with the CCSK and rhabdoid tumors. Unfavorable histology is found in about 10% of childhood renal tumors. It is
rare in the first 2 years of life (2%), then increases in patients
older than 5 years to 13%. It is also more frequent in nonwhite
(African-American and Latino populations) than in white
patients.80 In a report by Bonadio and colleagues, 30.1% of
anaplastic tumors occurred in the nonwhite population. In
a multivariate analysis, older age, being nonwhite, and lymph
node positivity were the significant predictors of anaplastic
WT histology. Finally, anaplasia has been strongly associated
with the presence of TP53 mutations.81
Different treatment protocols for children with anaplastic
versus favorable-histology tumors were first used in NWTS-3.
Anaplasia is defined by multipolar polyploid mitotic figures,
marked nuclear enlargement (giant nuclei with diameters at
least 3 times those of adjacent cells), and hyperchromasia.82
Focal anaplasia is defined as the presence of one or a few
sharply localized regions of anaplasia within a primary
tumor, the majority of which contain no nuclear atypia. The
cells must not be present in any sites outside of the kidney. Tumors with diffuse anaplasia must have at least one of the following four criteria. Anaplastic cells outside of the kidney, presence
of anaplasia in a random kidney biopsy, anaplasia in more
than one region of the kidney, and anaplasia in one region,
with extreme nuclear pleomorphism in another site. The
difference between focal and diffuse anaplasia has been
demonstrated to have prognostic significance.83 Anaplasia is
a marker of resistance to therapy, not of tumor aggressiveness.78,82,84 Although associations between histologic features
and prognosis or responsiveness to therapy have been suggested, with the exception of anaplasia (unfavorable histology),
none of these features have reached statistical significance
and therefore have not been used to determine therapy.78,84
The classic WT is triphasic, but some tumors can have
dominant blastemal, stromal, and epithelial elements. Stromal
dominant tumors are associated with intralobar nephrogenic

rests, and epithelial dominant tumors have been associated
with perilobar nephrogenic rests.
PRETREATED TUMORS AND PATHOLOGY

Tumors that have been treated with chemotherapy before resection differ in their histopathologic findings from tumors
resected primarily. In the SIOP-9 study, the most common
subtype of tumors resected without neoadjuvant chemotherapy was triphasic mixed histology (45.1%), followed by blastemal (39.4%) and epithelial dominant (15.5%), whereas in
tumors that received preoperative chemotherapy, the most
common histology was regressive (37.6%), followed by mixed
(29.4%), stromal (14%), blastemal (9.3%), and epithelial
predominant (3.1%); 6.6% of tumors were completely necrotic.85,86 The SIOP risk classification uses these histologic
findings as prognostic indicators to determine further therapies (Table 30-3). In addition, chemotherapy may produce
tumor differentiation.82,86,87 Anderson evaluated the histologic changes in tumors from 15 BWT patients that did not
decrease in size radiographically following chemotherapy.88
One had complete necrosis, 4 had rhabdomyomatous differentiation, and 10 had mature stromal differentiation.
Despite their absence of regression in size, these patients
had favorable outcomes, especially if there was rhabdomyomatous differentiation.
In SIOP-9, 10% of patients had postchemotherapy tumors
that were completely necrotic. These patients had excellent outcomes. The SIOP-9 study also demonstrated that preoperative
chemotherapy extensively ablates the blastemal component of
WT.87,89,90 The frequency of tumors with dominant blastemal
components was markedly reduced (to 7.7%) by preoperative
treatment compared with the no-treatment group (36%). Furthermore, this response is clearly an important prognostic
factor. If predominant blastemal elements persist after initial
therapy, the tumors were found to be highly aggressive. In
SIOP-9, 5 of 16 (31%) of the postchemotherapy blastemal predominant tumors recurred, compared with none of the tumors
that were predominantly epithelial or stromal after chemotherapy. Prior SIOP studies have also shown the prognosis for the
purely blastemal group (after preoperative chemotherapy) to
be inferior to that for the epithelial and stromal dominant
tumors.
TABLE 30-3
Revised International Societe Internationale dOncologie
Pediatrique Working Classification of Renal Tumors
of Childhood (2001)
Stage
Risk
Histology
Low
II
Intermediate
III
High
Mesoblastic nephroma
Cystic partially differentiated
nephroblastoma
Nephroblastoma epithelial type
Nephroblastoma stromal type
Nephroblastoma mixed type
Nephroblastoma regressive type
Nephroblastoma focal anaplasia type
Nephroblastoma blastemal type
Nephroblastoma diffuse anaplasia type
Clear cell sarcoma of the kidney
Rhabdoid tumor of the kidney
CHAPTER 30
In the SIOP studies, postchemotherapy risk stratification

and stage are used to determine additional therapy after resection. This categorization is different than the risk stratification
used for tumors resected primarily in North America. Low-risk
tumors are those that are completely necrotic following preoperative chemotherapy. Intermediate-risk tumors include all histologies other than completely necrotic, rhabdoid, anaplastic,
or blastemal (less than 66%) dominant. High-risk tumors are
those with diffuse anaplasia, rhabdoid, and blastemal dominance (greater than 66%) after chemotherapy (see Table 30-3).
NEPHROGENIC RESTS AND

NEPHROBLASTOMATOSIS
Nephrogenesis in the normal kidney is usually complete by 34
to 36 weeks gestation. Nephrogenic rests (NR) are areas of
metanephric (embryonal tissue) persisting after the 36th week
of life. The presence of multiple or diffuse nephrogenic rests
is termed nephroblastomatosis.91 Diffuse hyperplastic perilobar nephrogenic rests (DHPLNR) represent a unique category
of nephroblastomatosis in which the rests form a thick rind
around the kidney. The rests that cause the greatest diagnostic
challenge are those that are actively proliferating or hyperplastic, and can be mistaken for WT. Hyperplastic NR can produce
masses as large as conventional WT. Complicating things further is the fact that neoplastic induction of NR can occur. The
diagnosis of DHPLNR is often made based on radiographs
(Fig. 30-6). Histologically, a rest consists of predominantly
small clusters of blastemal cells, but tubules and stromal components can be present. NRs are classified by their growth
phase and location: perilobar or intralobar. Perilobar nephrogenic rests are limited to the periphery (subcapsular) of the
lobes, while intralobar rests occur within the renal lobes
and have an irregular margin. The growth phase of a rest is
divided into (1) incipient or dormant nephrogenic rests that
show few well-formed tubular structures but no evidence of
proliferation and no mitoses, (2) hyperplastic nephrogenic
rests that are composed of epithelial elements with nodular expansive growth, and (3) sclerosing rests that consist of stromal
and epithelial elements with few blastemal nephrogenic elements (Fig. 30-7).
WILMS TUMOR
429
NRs are considered precursor lesions to WT; however, only

a small number develop clonal transformation into a WT. A
child with a WT and NRs in the resected specimen is at increased risk of developing a metachronous tumor in the other
kidney.92 For a child less than 1 year of age, this risk is very significant, and these children need to be followed very carefully
with sequential US examinations. A patient who has a unilateral
tumor and a presumed nephrogenic rest is thought to be at increased risk of developing a metachronous tumor, but data to
support that assumption does not exist. The prevalence of NRs
in unilateral WT has been reported to be 28% to 41% in
unilateral WT and close to 100% in bilateral WT.93 Pathologic
distinction between NR and WT can be very difficult. To make
the diagnosis, it is critical to examine the juncture between the
lesion and the surrounding renal parenchyma to distinguish
between the two entities. Most hyperplastic NRs lack a pseudocapsule at the periphery, while most WTwill have this feature.
An incisional biopsy is of limited value, because it is uncommon for it to contain the interface between the lesion and the
adjacent kidney. This is particularly true for patients with
DHPLNR. In a study by Perlman and colleagues, pathology
alone was insufficient to establish the diagnosis of DHPLNR
in 21 of 33 cases that underwent biopsy at the time of initial
diagnosis.94 In addition, because rests are found within and
adjacent to WT, a biopsy may result in the inadvertent pathologic diagnosis of WT. Alternatively, a small WT may be present
within a large field of nephroblastomatosis, obscuring it for biopsy. Taken together, in these situations where a renal mass
could be a tumor or a rest where a biopsy is performed, Perlman
and colleagues suggest using the term nephrogenic process,
consistent with a WT or a nephrogenic rest.
Staging
------------------------------------------------------------------------------------------------------------------------------------------------
The COG/NWTS and SIOP staging systems are fundamentally

different. In COG/NWTS protocols, initial surgical resection is
recommended in most cases. Thus for unilateral tumors, the
pathology of the tumor is established prior to administration
of chemotherapy or radiotherapy. In contrast, SIOP protocols
generally recommend chemotherapy followed by nephrectomy, and surgicopathologic staging is assessed at that time.
The COG/NWTS staging system has evolved as features associated with prognosis have been defined. A very important
concept for this staging system is that there is a local stage and
a disease stage. Local staging refers to the abdominal disease
Hyperplastic
Wilms
tumor
Incipient
or dormant
Regressing
Obsolete
FIGURE 30-6 Computed tomography (CT) scans showing diffuse hyperplastic perilobar nephrogenic rests.
FIGURE 30-7 Cartoon of growth phases and classification of nephrogenic

rests.
430
PART III
only, whereas disease stage considers both the local and distant hematogenous metastatic disease. Both factors determine
therapy; the use of local radiation therapy to the tumor bed is
based on the local stage, and the use of additional chemotherapy is based on both stage III local disease or distant metastasis.95 The current COG and SIOP staging systems are shown in
Table 30-1.
Treatment
------------------------------------------------------------------------------------------------------------------------------------------------
The successful treatment for children with WT has been the

direct result of prior multidisciplinary studies from cooperative group trials, including the NWTSG, SIOP, and the United
Kingdom Childrens Cancer Study Group (UKCCSG), that
have defined the key components to therapy. These trials have
identified several prognostic factors used for risk stratification
in current protocols, including biologic markers. This section
will review these prognostic factors, operative therapy, chemotherapy, and radiotherapy (with a focus on COG studies).
PROGNOSTIC FACTORS
The current prognostic factors used in COG trials are histology, stage, age, tumor weight, response to therapy, and loss of
heterozygosity at 1p and 16q. The two most important continue to be the histology and the stage of the tumor.7,8,96
Histology: The details and prognostic significance of tumor
pathology have been previously discussed in the Pathology section.
Stage: The tumor stage is determined by the results of the
imaging studies and both the surgical and pathologic
findings at nephrectomy (see Table 30-1).
Rapid response: This is a prognostic category being evaluated in patients who have stage IV disease that is based
on lung metastasis alone. The goal in these patients is to
avoid lung radiation. Response to therapy is also being
assessed in bilateral disease.
Loss of heterozygosity: LOH (described previously) at both
1p and 16q are now used as determinants of therapy
on the current COG renal tumor studies.96
OPERATIVE THERAPY
Surgical therapy is a primary component in the multidisciplinary treatment of WTor other neoplastic renal lesions. Irrespective of whether surgery is performed as a primary therapy or in
a delayed fashion after chemotherapy, there are a number of
fundamental tasks that are required of the surgeon. These are
(1) safe resection of the tumor, (2) accurate staging of the tumor,
(3) avoidance of complications that will upstage the tumor
(rupture or unnecessary biopsy), and (4) accurate documentation of operative findings and details of the procedure in the
operative note. Intraoperative events that negatively affect
patient survival include tumor spill, failure to biopsy lymph
nodes, incomplete tumor removal, failure to assess for extrarenal tumor extension and surgical complications.9799
Technical Concerns: Unilateral Tumors
Ladd and Gross established the basic principles for resection
of a presumed malignant tumor of the kidney, including wide
abdominal exposure, resection of the surrounding Gerota fat
and fascia to remove potential sites of lymphatic spread and

early control of the renal vessels.12,100 Lymph node sampling
is now established as crucial for accurate staging.101 Understaging the extent of the tumor can increase a childs risk of
relapse, and overstaging will result in increasing the intensity
of chemotherapy or radiation. A transverse transabdominal or
thoracoabdominal incision provide the best exposure
and are associated with fewer complications than a flank
incision.98,102104 The thoracoabdominal incision is best for
large tumors, to optimize visualization of the plane between
the tumor and the diaphragm to avoid rupture from excessive
traction on the tumor. Intraoperative events that negatively
affect patient survival include tumor spill and inadequate
staging.9799
Early examination for involvement of the liver, renal vein,
or IVC or peritoneal surfaces is important, as is identification
of preoperative rupture of the tumor. Routine exploration of
the contralateral kidney for bilateral disease was mandated
in NWTS-1 to NWTS-5. In 1995, Ritchey and colleagues
reviewed the accuracy of imaging in assessing bilateral disease
from NWTS-4 (1986 to 1994). He found that bilateral tumors
were missed in 7% of children by using the preoperative imaging studies. Thus, for NWTS-5, routine contralateral exploration was mandated. In 2005, Ritchey and colleagues did a
follow-up study to look at what happened in those patients
whose lesions were missed by imaging on NWTS-4. The size
of the missed lesions was less than 1 cm in six patients and 1 to
2 cm in three patients. Management of missed lesions included enucleation in two cases, biopsy in six, and no surgery
in one. No patient underwent irradiation. The postoperative
chemotherapy regimen consisted of doxorubicin, dactinomycin, and vincristine in six children, and dactinomycin and vincristine in three. Median follow-up was 9 years. There were no
recurrences in any kidney with a missed lesion. All nine patients were alive and disease free at last follow-up. The results
of this study in conjunction with the advances in imaging
quality means that routine contralateral exploration in the
presence of a negative CT is not mandated.66,67 If a clear contralateral lesion is present, then the child should be treated on
the bilateral protocol. If studies suggest a possible contralateral lesion on the kidney, the contralateral kidney should be
formally explored prior to nephrectomy.
Ladd and Gross stressed the need for early vascular ligation
prior to the development of chemotherapy. This is no longer
practiced because of the risk of injury to the vessels, particularly to the superior mesenteric artery in large left-sided tumors. The tumor should be mobilized by opening the
lateral peritoneal reflection and reflecting the colon and its
mesentery off the anterior surface of the kidney. For rightsided tumors, a Kocher procedure is also helpful. When ligating the renal pedicle, it is best to ligate the renal artery first if
it can be safely identified, to avoid increasing the venous
pressure within the tumor, which can result in rupture of
the capsule. Vascular control in most cases is best completed
after the tumor is fully mobilized.99,105,106 The renal vein
should be palpated prior to ligation to be certain there is
no venous extension of the tumor. The adrenal gland may
be left in place if it is not abutting the tumor; but, if the
mass arises in the upper pole of the kidney, the adrenal
gland should be removed with the neoplasm. The ureter is
ligated and divided as low as possible.107 The tumor and
kidney should be handled gently throughout the operation
CHAPTER 30
to avoid rupture, which will increase the intensity of

therapy and risk for local recurrence.99,105,106
Pathologic assessment of hilar and regional lymph nodes is
critical to accurately stage a child with a renal tumor.97,99 Routine lymph node sampling from the renal hilum, the pericaval, or
para-aortic areas must be performed. Simply looking at the
lymph nodes to determine whether they are positive is highly
inaccurate.108 Unfortunately, failure to sample lymph nodes
(whether dealing with a unilateral or bilateral tumor) is the
major technical error noted in WT surgery.97 Furthermore,
studies have demonstrated a higher risk of recurrence in children who did not have their lymph node status documented
at the time of nephrectomy.12,99,109
WTs tend to displace rather than invade the surrounding
vessels. This feature of WT has two implications. First, the surgeon must be certain of the identity of the vessels to ligate.102
Second, most organs can be dissected away from the tumor,
because actual invasion is rare. When actual invasion is identified, radical en bloc resection (e.g., partial hepatectomy
or colectomy) is not warranted as primary therapy.98,99 WTs
are very chemosensitive, and, in these situations, prior
adjuvant therapy will result in a lower rate of complications
than a multiorgan resection.98 A small section of diaphragm,
psoas muscle, or tip of the pancreas, however, is acceptable.
Recent reports have suggested that hepatic metastasis
should be resected at presentation.110,111 To address this
question, the COG renal tumor study group reviewed outcomes for patients with different sites of metastasis and found
no significant difference in outcome for patients with liver
versus lung metastasis. Primary resection of liver metastases
prior to adjuvant therapy is not currently recommended.112
Spill
Spill refers to a break in the tumor capsule during operative
removal, whether accidental, unavoidable, or by design. Studies have shown a higher risk of recurrence in patients who
had tumor spill or rupture, irrespective of the cause or extent
of the soiling.9799 Spill is also considered to have occurred
if the renal vein or ureter are transected where they contain
tumor. In COG protocols, spill is also considered to have occurred if a preoperative or intraoperative needle/open biopsy
was performed. This is not the case for those patients treated
following Societe Internationale dOncologie Pediatrique protocols: Fine-needle or Tru-Cut needle biopsy is allowed in
this study; however, incisional biopsies are considered as
ruptures, automatically stage III, and are contraindicated.
Rupture refers to either the spontaneous or post-traumatic
rupture of the tumor preoperatively, with the result that tumor
cells are disseminated throughout the peritoneal or retroperitoneal space.101 Bloody peritoneal fluid may be a sign of
rupture, and a thorough examination of the tumor surface
is mandated. Rupture is also considered to have occurred if
the tumor penetrates the kidney capsule, with open neoplastic
tissue surface being in free communication with the peritoneal
cavity. If found, all of these situations make the child stage III
and must be carefully documented in the operative note.
Unresectable Tumors
There are clinical situations where it is agreed that primary nephrectomy is contraindicated. These are when (1) there is extension of tumor thrombus above the level of the hepatic
veins; (2) the tumor involves contiguous structures, whereby
WILMS TUMOR
431
the only means of removing the kidney tumor requires removal of the other structures (e.g., spleen, pancreas, and colon
but excluding the adrenal gland); (3) there are bilateral
tumors; (4) the tumor is in a solitary kidney; or (5) there is
pulmonary compromise resulting from extensive pulmonary
metastases. Studies conducted by the cooperative groups have
shown that pretreatment with chemotherapy almost always
reduces the bulk of the tumor.113116 This makes tumor
removal easier and may reduce the incidence of surgical complications.117 Preoperative chemotherapy does not result in
improved survival rates, and it may result in the loss of staging
information and changes the histology of the tumor as noted
previously.118,119
SPECIAL CONSIDERATIONS
Management of Tumor Extension in the Renal
Vein, Inferior Vena Cava, and Atrium
WT patients may present with tumor extension through the
renal vein to the IVC and even up to the right atrium. This
is found in 4% to 11% of children. Surgical treatment is dependent on the extent of vascular invasion. Extension is usually asymptomatic, and many are detected preoperatively by
US, CT, and/or MRI scans. However, those that extend just
into the renal vein may only be detected at operation because
of compression and distortion of the veins by the tumor, reinforcing the need to palpate the renal vein and IVC at the start
of nephrectomy before any mobilization of the kidney that
might dislodge the thrombus.106,120,121 As noted previously,
a primary resection when tumor thrombus extends into the
inferior vena cava at the level of the liver or higher is discouraged. COG protocols recommend that these patients be managed initially with preoperative chemotherapy. This approach
will often achieve significant shrinkage and regression of the
intravascular thrombus, facilitating subsequent surgical removal.106,122 The severity and number of operative complications are reduced with preoperative chemotherapy for those
with vascular extension above the hepatic veins. Alternatively,
if the tumor extends only into the renal vein or renal vein and
IVC below the level of the liver, the tumor and thrombus can,
in most cases, be removed en bloc with the kidney.
Control of renal veins and cava above and below the tumor
with vessel loops is necessary, using standard vascular surgery
techniques. The tumor should not be transected, if possible,
because this will result in spill and upstaging of the patient.
In some cases, the tumor may be adherent to the vessel wall.
A similar technique used for removing plaque for a carotid
endarterectomy is helpful to lift the tumor off the vein wall.
It must be stated in the operative report if the intravascular tumor extension was removed en bloc or if tumor was transected, as well as if the tumor thrombus is removed
completely and if there is evidence of either adherence to or
invasion of the vein wall. If, after preoperative chemotherapy,
the tumor still extends above the hepatic veins, cardiopulmonary bypass is generally needed to remove the vascular extension of the tumor.
Management of Tumor Extension in the Ureter
Extension of WT into the ureter is a rare event.107 In NWTS-5,
the incidence of ureteral extension was 2%. Preoperative imaging detected ureteral extension in only 30% of these
432
PART III
patients; the rest were discovered at operation. Clinical presentations included gross hematuria, passage of tissue per urethra, hydronephrosis, and a urethral mass. The diagnosis
should be suspected in these patients, and cystoscopy with
retrograde ureterogram may aid in preoperative diagnosis. If
extension of tumor into the ureter is detected or suspected,
the ureter should be resected with clear margins.
Horseshoe Kidney, Single Kidney,
and Nonfunctioning Kidney
A WT in a horseshoe kidney presents unique challenges.
Children with a tumor in a horseshoe kidney are treated as
unilateral tumors, NOT as bilateral tumors. Children with
horseshoe kidneys and WT must be carefully imaged prior to
any surgery.123 The blood supply to horseshoe kidneys is
quite variable and must be carefully imaged prior to surgery.123
At the time of operation, the blood supply to the kidney as well
as the location of the ureters must be identified and isolated. Exposure and mobilization of the kidney on the side of the tumor
is carried out as in unilateral resection. The side of the kidney
containing the tumor, the isthmus, and the ipsilateral ureter
are resected. As with other unilateral procedures, the lymph
nodes are sampled for staging purposes. Children with a single
kidney, or a situation where a tumor occurs in one kidney but
the second kidney is nonfunctioning, should be managed using
a renal-sparing approach, with preoperative chemotherapy to
facilitate surgery and preserve more renal tissue.
Patients with Wilms Tumor Treated Only
with Surgery
NWTS-5 evaluated a subset of very-low-risk patients with
favorable-histology tumors who might be treated without
chemotherapy. The criteria for this arm of the study was stage
I FH in patients who had lymph nodes biopsied, had a specimen weight of less than 550 g, and who were less than 2 years
of age. Seventy-five patients were enrolled before closure of
the study, and 8 developed recurrent disease (lung involvement in 5 and the operative bed in 3). Three other patients
developed metachronous contralateral WT. Stringent stopping rules for the study were designed to ensure closure of this
arm of the study if the 2-year EFS was 90% or less based on the
expectation that approximately 50% of the surgery-only children would be salvaged after recurrence, thus attaining the
95% predicted survival of these children treated with vincristine and dactinomycin (EE-4A). This limit was exceeded on
June 14th, 1998, and this arm of the study was closed when
the 2-year disease-free survival estimate reached 86.5%.124
Subsequent patients were treated with EE-4A. A recent
long-term follow-up study of the surgery-only cohort and
the EE-4A group, with a median follow-up of 8.2 years,
reported the estimated 5-year EFS for surgery only was 84%
(95% confidence interval [CI]: 73% to 91%); for the EE-4A
patients it was 97% (95% CI: 92% to 99%, P 0.002).
One death was observed in each treatment group. The estimated 5-year overall survival (OS) was 98% (95% CI: 87%
to 99%) for surgery only and 99% (95% CI: 94% to 99%)
for EE-4A (P 0.70).125 The surgery-only EFS was less than
for EE-4A, consistent with the earlier report. The salvage rate
for the surgery-only cohort, however, exceeded that seen with
children who had received two-drug chemotherapy, which
had been predicted to be 50%. Thus 85% of the infants
avoided any chemotherapy, while those who did receive it
for relapse were treated with three agents (DD-4A). A current

study in the COG is assessing this cohort again and is evaluating biologic markers for this very-low-risk group.126
Neonatal Tumors
Neoplastic renal lesions in the neonate are rare and include
benign and malignant tumors.127,128 Acute and long-term toxicity from therapy is a considerable concern in infants. The
distribution of tumors is age dependent. In the perinatal period, congenital mesoblastic nephroma (CMN) is the leader,
accounting for greater than 50% of the renal tumors, followed
in rank by WT, RTK, and CCSK.127131 WT, CMN, and rhabdoid tumor of the kidney (RTK) are the principal neoplasms of
the kidney occurring after 3 months, when CMN accounts for
less than 10%. An international retrospective study of 750
neonatal renal tumors in children less than 7 months of age
found that 63.4% were WT.127 Eighty-two percent of these
were stage I/II. In contrast, RTK presented with advanced disease (53% stage III/IV). RTK accounted for nine of eleven tumors presenting with metastases. Outcomes paralleled older
children, with excellent results for neonates with WT (5-year
OS of 93.4%) and poor for RTK (5-year OS of 16.4%).127
Acquired von Willebrand Disease in Children
with Wilms Tumor
von Willebrand disease (vWD) is an inherited coagulation
disorder characterized by mucocutaneous bleeding, a prolonged bleeding time (BT), and a reduced level of functional
von Willebrand factor (vWF). Secondary laboratory abnormalities include a decreased level of procoagulant factor VIII (FVIII)
and activity of ristocetin cofactor (FVIII:RCoF) activity.132
Acquired vWD has been reported in patients with WTand other
malignancies and has important implications for the surgeon.133,134 A single prospective study of 50 WT patients found
the incidence of acquired vWD was 8%.134 However, the true
incidence and prevalence in WT is unknown, because a full
bleeding history and factor levels are rarely obtained. Until
recently, the literature has suggested that, when identified, the
bleeding has been clinically insignificant, characterized by
epistaxis, hematuria, gingival bleeding, and easy bruising.135
Recent reports of profuse intraoperative bleeding that only
stopped after ligation of the renal vessels have contradicted
this assumption.136,137 Despite normalization of FVIII and
vWF activity and antigen levels prior to surgery, during surgery
profuse intraoperative bleeding occurred, requiring multiple
transfusions with FVIII, FFP, cryoprecipitate, platelets, and
packed red blood cells.136 Immediately after ligation of the renal
vessels, all abnormal bleeding stopped, with normalization of
FVIII and vWF antigen activity.
The mechanism of acquired vWD in WT is unknown.
Tumor adsorption of vWF has been reported in other malignancies; however, this was not seen in the WT cases where
intraoperative bleeding was significant. vWF inhibitors, rapid
abnormal clearance of vWF, and coagulopathy related to elevated levels of hyaluronic acid and consequent blood hyperviscosity have also been proposed.138,139 Why some cases
had intraoperative bleeding and others do not is also not
known. Baxter136 suggests that these tumors may be more
hypervascular, but this is not proven. The risk of intraoperative
bleeding highlights the importance of recognizing acquired
vWD in children with WT. In all cases, the initial sign was a
prolonged prothrombin time (PT) and partial thromboplastin
CHAPTER 30
time (PTT). When found, this should mandate acquiring a

further history for bleeding and factor analysis. Although correction of factor levels prior to surgery appears to help in most
cases, it does not guarantee that significant intraoperative
bleeding will not occur. In the case reports of profound
intraoperative bleeding, it was observed that, once the renal
vessels were ligated, the bleeding ceased. Thus preoperative
embolization should be considered as a management strategy.
Alternatively, preoperative chemotherapy may also be a safe
option.
BILATERAL WILMS TUMOR

Bilateral Wilms tumors BWT occur in 4% to 13% of patients
(see Fig. 30-3).5,1719 Unfortunately, outcomes for children
with bilateral tumors have not been as good as those of children with unilateral tumors. In NWTS-5, the 4-year OS was
80.8% for a child with favorable histology and 43.8% for a
child with anaplastic histology.84 In 1998, the United
Kingdom Childrens Cancer Study Group published their
experience with BWT patients treated between 1980 and
1995.140 In 57 patients, conservative surgical treatment with
initial biopsy was followed by chemotherapy and delayed
tumor resection, while 13 underwent initial surgical resection
followed by chemotherapy. Overall survival was 69%, with
similar survival in the patients with initial surgery versus
neoadjuvant chemotherapy. BWT with an unfavorable histology was associated with a poor prognosis, with only one of
seven patients surviving. Renal failure was seen in 6% of
the survivors who were conservatively treated and in 20%
of the survivors who underwent initial resection. In 2004,
Weirich reported BWT outcomes from SIOP-9. Twenty-eight
patients were evaluated. Although therapy was individualized,
all 28 patients with BWT were treated with preoperative
therapy. Overall survival at 5 years was 85.1% (95% CI:
71.6% to 98.6%; four deaths), and relapse-free survival was
80.5% (95% CI: 65.2% to 95.8%; five relapses).141
Renal failure is another concern of children with BWT
(Figs. 30-8 and 30-9). The etiology of renal failure in WT patients is multifactorial.142144 Factors that contribute to renal
failure include intrinsic progressive renal disease related to
a genetic predisposition, inadequate renal parenchyma after
one or more tumor resections, the nephrotoxic effects of
WILMS TUMOR
433
chemotherapy and radiation, and the potential for hyperfiltration injury to the remaining renal parenchyma. Ritchey
defined the incidence and etiology of renal failure in patients
treated on NWTS-1 to NWTS-4. BWT was the greatest risk
factor for renal failure (16.4% for NWTS-1 and NWTS-2,
9.9% for NWTS-3, and 3.8% for NWTS-4). Other risk factors
identified were Denys-Drash syndrome, metachronous tumor, progressive disease in patients with bilateral tumors requiring bilateral nephrectomies and radiation nephritis.144
Breslow reported the 20-year end-stage renal disease (ESRD)
outcomes in children treated for WT (see Figs. 30-7 and
30-8).142 The major risk factors he identified for renal failure
were BWT and congenital syndromesDenys-Drash, WAGR,
and genital urinary anomalies (hypospadias or cryptorchidism). Thus preservation of renal tissue without sacrificing
long-term survival is of particular importance for those
with BWT.
Despite 40 years of clinical trials for WT, it was not until
2009 that a formal BWT trial was opened by COG. Several
prior reports contributed to the development of this protocol.
Shamberger and colleagues examined 38 of 188 patients with
BWT with progressive or nonresponsive disease (PNRD).145
The mean duration of chemotherapy was 7 months; 36 patients were treated with two regimens of chemotherapy, and
21 patients received three. Patients with PNRD fell into two
categories: first, patients with anaplasia whose tumors were
not sensitive to the therapy administered (4 patients); second,
patients who had tumors with very mature rhabdomyomatous
or differentiated stromal elements (14 patients) and 1 with
complete necrosis. A second study from Anderson looked
at the histologic changes in BWT patients who did not respond to chemotherapy and the relationship between these
changes and prognosis.146 Their results mirrored those of
the NWTS study. Fifteen patients whose tumors did not respond were evaluated. One had complete necrosis, 4 had
rhabdomyomatous differentiation, and 10 had mature stromal
differentiation. Despite not radiographically responding
to chemotherapy, these patients had favorable outcomes.
Patients in these studies fell into two categories. First, there
were patients with anaplasia whose tumors were not sensitive
to the therapy administered. Anaplastic tumors respond poorly
to chemotherapy and, once the diagnosis of anaplasia is
made, a complete resection is needed.84,140,147,148 Second,
DDS (12/17)
60
WAGR (11/37)
40
20
GU (5/125)
Other (28/5, 347)
0
0
10
15
20
Time since diagnosis of
unilateral Wilms tumor (years)
25
FIGURE 30-8 Kaplan-Meier plot of renal failure rates at 20 years of age in

children with a unilateral Wilms tumor (WT). DDS, Denys-Drash syndrome;
ESRD, end-stage renal disease; GU, genitourinary; WAGR (syndrome),
Wilms tumor, aniridia, genitourinary malformation, mental retardation.
100
Cumulative incidence of
ESRD (%)
Cumulative incidence of
ESRD (%)
80
WAGR (5/10)
80
GU (8/25)
60
DDS (3/6)
40
20
Other (44/409)
0
0
5
10
15
20
25
Time since diagnosis of bilateral Wilms tumor (years)
FIGURE 30-9 Kaplan-Meier plot of renal failure rates at 20 years of age in
children with bilateral Wilms tumor (BWT). DDS, Denys-Drash syndrome;
ESRD, end-stage renal disease; GU, genitourinary; WAGR (syndrome),
Wilms tumor, aniridia, genitourinary malformation, mental retardation.
434
PART III
there were patients who had tumors with very mature rhabdomyomatous or differentiated stromal elements and complete
necrosis, all of whom had an excellent outcome. Again these
patients are best served with resection.146 Therefore if the bilateral lesions do not respond radiographically to therapy, it is critical to establish whether this is due to anaplasia or mature
histology.
Hamilton and colleagues have demonstrated the difficulty
in identifying anaplasia in patients with BWT.148 Twentyseven patients with anaplasia were reviewed from NWTS-4.
Discordant pathology between the kidneys was seen in 20 patients, highlighting the importance of obtaining tissue from
both kidneys. Seven children who were eventually found to
have diffuse anaplasia had core needle biopsies, which failed
to establish the diagnosis in all of these cases. Anaplasia was
identified in only three of nine patients who had an open
wedge biopsy and in seven of nine patients by partial or complete nephrectomy. Thus percutaneous biopsies rarely
establish the diagnosis, and open biopsies were successful
in only a third of the cases.
An important question is to determine how long to treat a
child who has BWT with chemotherapy before intervening
surgically. In SIOP-9, patients with unilateral tumors were
randomized to receive either 4 or 8 weeks of dactinomycin
and vincristine preoperatively. There was an average 48% reduction in tumor volume after 4 weeks that increased to 62%
after 8 weeks of chemotherapy.116,149 A review by the German
Pediatric Hematology Group (GPOH) of their patients with
BWT reported that maximum tumor shrinkage occurred in
the first 12 weeks of chemotherapy.150
The two principal aims of the COG BWT study are to
improve 4-year event-free survival and to prevent complete
removal of at least one kidney in 50% of patients with BWT
by using preoperative chemotherapy. This is a response-based
protocol starting with chemotherapy, followed by evaluation
at 6 and 12 weeks with definitive surgical therapy in all patients by 12 weeks (see Fig. 30-3). This protocol does not
mandate an initial tissue diagnosis because bilateral renal tumors in children are invariably WT; biopsy does not change
the therapy in most cases; anaplasia is hard to diagnose,
and the biopsy will effectively increase the stage of the tumor
and its risk for local recurrence.148 In the current COG protocol, local spill of the tumor is designated as stage III. This classification was changed because of the finding of an increased
incidence of abdominal recurrences in NWTS-4 patients with
tumor spill.99 First, for patients with BWT, the initial regimen
will consist of regimen vincristine, actinomycin D, doxorubicin
(VAD) (vincristine [VCR], dactinomycin [DACT], doxorubicin
[DOX]), a more intensive combination of drugs based on regimens used with good results and minimal toxicities by both
SIOP and the UKCCSG WT groups, which enables patients
to receive two doses of DOX, in addition to six of VCR and
two of DACT, during the first 6 weeks of therapy.151 It differs
from the standard three-drug regimen, DD-4A, in which
the DOX and DACT are administered in separate cycles.152
The three-drug chemotherapy regimen of VAD was chosen to
give an enhanced therapy for possible stage III disease, because
patients rarely have a lymph node biopsy before initiation of
therapy. Second, it was elected to enhance the chemotherapy
rather than administer radiotherapy, which might increase
the occurrence of radiation nephritis in the remaining kidney.
Third, a more intensive therapy was selected for treatment
to avoid the use of a sequential regimen of increasing intensity,

which was seen in the review of the prior cohort of NWTS-4
BWT patients.
CHEMOTHERAPY
In 1963, Farber first reported that dactinomycin had activity
against WT.153 Today, dactinomycin continues to be part of
the backbone of therapy for children with WT. Other active
chemotherapeutic agents have been identified subsequently,
including vincristine, doxorubicin, and cyclophosphamide.
Clinical trials conducted by NWTSG and SIOP have evaluated, stage by stage, different chemotherapeutic protocols to
assess the efficacy of various combinations and duration of
therapy.105,154159 In NWTS-4, 4-year event-free survival
and overall survival averaged 90% for patients with favorable
histology.154,159 Therefore NWTS-5 focused on evaluating
biologic markers of prognosis, such as LOH, developing more
effective therapy for recurrent disease, and reducing therapy
in children with low-risk tumors.
Treatment on the current COG protocols for favorablehistology WT is determined by stage, histology, and LOH.
For children with favorable-histology stage I and II tumors
without LOH, 18 weeks of vincristine and dactinomycin (regimen EE-4A) is recommended. Results from NWTS-5 showed
these children had an overall survival of 98.4% and 98.7%,
respectively. For children with FH stage III and IV tumors
without LOH, 24 weeks of vincristine, dactinomycin, and
doxorubicin is recommended (regimen DD-4A). For those
patients who have positive LOH at both loci (1p and 17q),
treatment will be intensified. If they are stage I or II and
LOH positive, they will receive DD-4A, and if they are stage
III and IV LOH positive, they will receive vincristine, dactinomycin, and doxorubicin with alternating cycles of cyclophosphamide versus etoposide (regimen M). Dosing modifications
are made for children less than 12 months of age.
Anaplastic tumors have been less successfully treated.
NWTS-3 and NWTS-4 were the first studies to prospectively
evaluate the benefit of additional/different chemotherapy
therapy for these tumors. One randomized arm compared
15 months of vincristine, dactinomycin, and doxorubicin,
with or without cyclophosphamide. For patients with stage
II to IV diffuse anaplastic histology, the addition of cyclophosphamide resulted in a 4-year relapse-free survival estimate of
54.8% when treated with cyclophosphamide compared with
27.2% when treated without it (P 0.02).160 In NWTS-5,
patients with focal anaplasia or diffuse stage I were treated
with EE-4A. This was based on prior historical data, with a
goal of reducing therapy. Unfortunately, the 4-year event-free
and overall survival estimates for stage I (focal or diffuse)
anaplastic WT were lower than previous studies (EFS 69.5%
and OS 82.6%). Thus therapy with EE-4A is inadequate.
Patients with focal anaplasia stage II to IV were treated with
DD-4A. Children with stage II to IV diffuse anaplastic WTwere
treated with vincristine, doxorubicin, and cyclophosphamide
(VDC) alternating with cyclophosphamide and etoposide
(CyE) (regimen I). The 4-year event-free survival estimates
for stage II to IV diffuse anaplastic WT on NWTS-5 were
82.6%, 64.7%, and 33.3%, respectively, with similar overall
survival.84 The current protocols and chemotherapy agents
for unilateral tumors are shown in Table 30-4.
CHAPTER 30
TABLE 30-4
Current Childrens Oncology Group Chemotherapy Regimens
for Unilateral Wilms Tumor
Regimen
Agents
EE-4A
DD-4A
Vincristine and dactinomycin

Vincristine, dactinomycin, doxorubicin,
and radiation therapy (XRT)
cyclophosphamide (CPM1), and etoposide
(ETOP), as well as radiation therapy (XRT)
cyclophosphamide, and etoposide;
radiation therapy also to be administered as
part of this regimen
cyclophosphamide, carboplatin, etoposide,
and radiation
cyclophosphamide, carboplatin, etoposide,
irinotecan, and radiation therapy (XRT)
Vincristine and irinotecan in conjunction
with revised UH-1 or revised UH-2,
depending on response
Regimen I
Regimen M
Revised UH-1
Revised UH-2
Vincristine/irinotecan
window therapy
Recurrent Tumor
Treatment of recurrent disease in children with WT is challenging. Recurrence occurs in 15% of patients with favorable
histology tumors and in 50% with anaplastic histology. Recurrence is most frequent within 2 years of the initial diagnosis and
most common in the lungs, tumor bed, and liver.161 Less common sites are bone, brain, and distant lymph nodes.
Recurrent disease is treated by chemotherapy, surgery, and
radiotherapy. NWTS-5 evaluated two protocols for recurrent
disease, avoiding use of agents included in the primary protocols. Stratum B was for patients with stage I and II disease initially treated with EE-4A. The chemotherapy for this relapse
protocol was regimen I (alternating courses of vincristine/
doxorubicin with cyclophosphamide), in addition to surgical
resection and radiation therapy. Event-free survival at 4 years
was 71.1%, and 4-year overall survival was 81.8% for all patients and was 67.8% and 81.0%, respectively, for those who
relapsed only to their lungs.162 Stratum C was for patients initially treated with DD-4A.163 The chemotherapy protocol for
this group was alternating cycles of cyclophosphamide versus
etoposide and carboplatin versus etoposide. Four-year eventfree survival and overall survival were 42.3% and 48.0%,
respectively, for all patients and were 48.9% and 52.8% for
those who relapsed in the lungs only. Bone marrow transplantations have been performed for patients with recurrent disease,
with reported event-free or disease-free survival rates of 36% to
60% in these small series.164166 At present, there is no open
relapsed study in SIOP or COG, because the groups are awaiting new and more effective agents for treatment of this disease.
RADIOTHERAPY
Analogous to surgery and chemotherapy, the cooperative
group trials have refined the indications for radiotherapy. In
addition, technologic advances have helped to deliver irradiation with increased efficacy and less toxicity to surrounding
tissues. The three principle fields for radiotherapy for renal tumors are whole abdominal, flank, and lung (metastatic lung
WILMS TUMOR
435
disease). All five NWTSG studies and the current COG studies
use radiotherapy as part of the multimodality treatment for
advanced-stage tumors.
In 1950, Gross and colleagues demonstrated the efficacy
of radiotherapy as an adjuvant therapy prior to the advent
of chemotherapy. In this series, nephrectomy with postoperative radiation improved survival to 47%.167 Favorable histology tumors are generally very radiosensitive. NWTS-1 to
NWTS-3 helped define the indications, timing, and dose of
radiotherapy. NWTS-1 established that irradiation provided
no advantage in children younger than 24 months with stage
I FH tumors who also received 15 months of dactinomycin.168
That study also demonstrated that in stage III tumors with local tumor spill or previous biopsy, there was no need for irradiation of the whole abdomen, thus sparing patients the
associated toxicity.169 NWTS-2 showed that radiotherapy
could be avoided in all children with stage I WT if they received vincristine and dactinomycin.170 NWTS-3 established
that radiotherapy could be avoided in children with stage II
tumors given vincristine and dactinomycin and also demonstrated that children with stage III favorable-histology tumors
who received 10.8 Gy radiotherapy and vincristine, dactinomycin, and doxorubicin had similar tumor control to those
who received 20 Gy with vincristine and dactinomycin. This
was an important finding, because it eliminated the need for
an age-adjusted dose schedule and significantly reduced the
recommended dose of radiation.157
Timing of radiation following nephrectomy was assessed
on NWTS-2, where a delay of 10 days or more before initiation of radiotherapy was associated with a higher rate of
abdominal relapse, particularly among patients with unfavorable-histology tumors and a small radiation field.157,168,169 A
recent review of this issue from NWTS-3 and NWTS-4 data
confirmed this observation.171 Thus, in the COG protocols, it
is recommended that abdominal irradiation be delivered as
soon as practical after nephrectomy and not later than 14
days after surgery. The current recommendation for radiation
therapy for COG protocols is shown in Table 30-5.
In contrast to FH tumors, the ideal dose for patients with
anaplastic tumors is unknown. Anaplastic tumors are more
resistant to chemotherapy and seem to be more resistant to
radiotherapy as well. Anaplastic tumors have not demonstrated a radiation dose response between 10 Gy and 40 Gy.160
The radiotherapy strategy for patients with anaplastic histology (AH) on NWTS-5 included no irradiation for stage I AH
tumors and 10-Gy radiotherapy for AH stage II and III in conjunction with nephrectomy and regimen I. The outcomes for
both of these treatment strategies were suboptimal. Stage 1 patients had a 4-year EFS and overall survival of only 69.5% and
82.6%, respectively. Stage II, III, and IV patients had a 4-year
OS after immediate nephrectomy, irradiation, and regimen I
chemotherapy of 82.6%, 64.7%, and 33.3%, respectively.84
EFS was similar to OS in all groups. Fifty percent of stage
III recurrences were local, suggesting that the dose of 10 Gy
was not adequate. These results form the basis for the current
COG study that recommends the addition of irradiation for
patients with stage I anaplasia and augmentation of irradiation
for patients with stage III anaplasia.
For liver metastases, only those that are unresectable at diagnosis are irradiated. The treatment portal includes that portion of
the liver known to be involved as identified by CTor MRI studies.
The whole liver is treated in children with diffuse metastases.
436
PART III
TABLE 30-5
Radiotherapy for Favorable-Histology Wilms Tumor
Treatment Site
Clinical Presentation and Dose (Gy)
Flank irradiation
All instances of soilage
will be classified as Stage
III and require abdominal
radiation. Flank radiation
is given to all Stage III
patients with three
exceptions (the patients
meeting any of these
exceptions requiring
whole abdominal
radiation).
Whole abdomen irradiation
(WAI)
Stage III favorable histology

Recurrent Wilms tumor
10.8
10.8
Abdominal stage III

Preoperative tumor rupture
Peritoneal metastases are
found at initial surgery
A large intraoperative tumor
spill affecting areas outside
the tumor bed as
determined by the surgeon/
treating institution.
Abdominal Stage III
Diffuse unresectable
peritoneal implants
Focal metastases
Diffuse metastases
10.5
Liver irradiation
Patients with residual
tumor will receive
supplemental irradiation
with 5.4 to 10.8 Gy.
21
19.8
19.8
(WLI), and the 4-year survival rate improved to 75%.181 A

COG study of patients with pulmonary lesions detected by
CT only (as opposed to CT and chest radiograph) and treated
with only two chemotherapeutic agents showed an inferior
outcome compared with those treated with three drugs irrespective of whether or not they received pulmonary radiation.172 A fourth study examined the value of biopsy prior
to treating patients with lesions detected only by CT.175
Two thirds of the children had tumor on biopsy, suggesting
that histologic evaluation may be valuable in directing therapy.
The current COG study is evaluating the use of radiographic
response to chemotherapy to predict the need for whole lung
irradiation. Those patients with stage IV favorable-histology
WT with pulmonary metastases who have complete CT resolution of the pulmonary lesions after 6 weeks of vincristine/dactinomycin/doxorubicin chemotherapy will continue the same
chemotherapy without whole lung irradiation. Those who do
not have resolution of pulmonary metastases by week 6 will
have the addition of cyclophosphamide and etoposide to the
other three drugs and will receive whole lung irradiation.
LATE EFFECTS
The increasing numbers of survivors of WT have led to a
better understanding of adverse medical conditions related
to treatment of their disease that can develop over time.182
Treatment for WT impacts renal function (discussed earlier),
pregnancy, cardiac and pulmonary function, and second
malignancies may develop.178,183187
Lung Radiotherapy
Pregnancy
Historically, pulmonary metastases were diagnosed based on

lesions found on routine chest radiographs and were treated
with whole lung radiation. For COG studies, it is delivered
in eight treatments of 12 Gy. From NWTS-5, the 5-year EFS
(95% CI) for stage IV category was lung only 76% (72% to
80%) (513 patients) and liver and lung 70% (57% to 80%)
(62 patients).172 Advances in imaging have changed the assessment of lung disease from plain radiograph to widespread
use of chest computed tomography. Lesions are detected on
CT scan that are not found on standard radiographs.173175
Thus more lesions are being identified. Complicating the
use of radiation therapy is the fact that it is a major cause of
long-term morbidity, particularly to the lung and heart, producing congestive heart failure, pulmonary fibrosis, and second malignancy.176178 Recent studies suggest that the
management for pulmonary nodules should be reexamined.
In SIOP-9, by 70 days of therapy, resolution of pulmonary
nodules on CT scan in children with FH tumors was a favorable prognostic indicator.179 In SIOP-9, many of these patients
were spared whole lung irradiation, if complete resolution of
pulmonary metastases occurred after 6 weeks of prenephrectomy chemotherapy with vincristine, dactinomycin, and doxorubicin with or without surgical excision of residual
metastases. The 5-year relapse-free survival (RFS) for stage
IV patients receiving preoperative chemotherapy was
62.5%.179 The results of this study have been controversial.
The United Kingdom Childrens Cancer Study Group
(UKCCSG) Wilms Tumor Study 1 followed a similar protocol;
yet, their 6-year EFS was only 50%.180 In their second study,
UKCCSG-Wilms Tumor Study (UKWT2), the majority of children with lung metastases received whole lung irradiation
Treatment for WT impacts reproductive capacity and increases

the risk of complications during pregnancy. The National
Wilms Tumor Long-Term Follow-Up Study evaluated 700
maternal/offspring pairs.188 If a woman had received flank
radiation for unilateral WT, the dose of radiation correlated
with increased risk of hypertension, fetal malposition, and
premature labor. The children were also more at risk
for low birth weight and prematurity (birth before 37 weeks).
Premature labor was seen in 10.2% of women who did not
receive flank radiation and 22% of those who received
35Gy (P 0.001). Radiation therapy to the abdomen has
resulted in absent/abnormal function of the ovaries, a small
uterus, and premature menopause.189193 Male infertility is
not at risk unless alkylating agents were used.
Secondary Malignancies
Patients who have been treated for pediatric cancer are known
to have an increased risk of second malignancies. This is in
part due to treatment with known carcinogens, such as alkylating agents and radiotherapy.183,194,195 An international cohort of 13,351 children with WT diagnosed before 15 years of
age, from 1960 to 2004, was established to determine the risk
of second malignant neoplasms (SMN).178 One hundred and
seventy-four solid tumors and 28 leukemias were found in
195 people. Median survival after a secondary malignancy
was diagnosed 5 years or more from WT was 11 years; it
was 10 months for leukemia. Age-specific incidence of secondary solid tumors increased from approximately 1 case
per 1,000 person-years at age 15 years to 5 cases per 1,000
person-years at age 40 years. The cumulative incidence of
solid tumors at age 40 years was 6.7%. In those patients whose
CHAPTER 30
WT was diagnosed after 1980, there was a lower age-specific

incidence rate for second tumors compared with those treated
before 1980. Paradoxically, the incidence of leukemia was
higher in those diagnosed after 1990. This may be due to decreasing use of radiation therapy and increasing intensity of
chemotherapy in modern protocols for treatment of WT.
Congestive Heart Failure
Congestive heart failure has been identified as a significant morbidity in children treated with doxorubicin, and this is exacerbated in patients who receive thoracic radiation. The cumulative
frequency of congestive heart failure in patients treated on
NWTS-1 to NWTS-4 was 4.4% at 20 years for patients treated
initially with doxorubicin, but that percentage is expected to be
lower with current cumulative doses.184,185,196 The relative risk
of congestive heart failure was found to be increased in females
(risk ratio [RR] 4.5; P 0.004), and by cumulative doxorubicin dose (RR 3.2/100 mg/m2; P < 0.001), lung irradiation
(RR 1.6 for every 10 Gy; P 0.037), and left abdominal irradiation (RR 1.8/10 Gy; P 0.013).185 Preliminary results
suggest that cardiotoxicity is lower with current radiation doses,
but patients still have a substantial lifetime risk of developing
cardiac disease.183,196
Thoracic
Radiotherapy (RT) has been implicated as a major contributor
to late complications. Acute lung injury is relatively uncommon, occurring in a minority of children.197 The late effects
of pulmonary RT include pneumonitis and restrictive lung
disease, scoliosis, kyphosis, reduced lung capacity, and secondary tumors. In girls, breast hypoplasia and cancer
have been described.176,177 Paulino and his colleagues
reported on the late complications of pulmonary RT in 55
long-term survivors of WT.176 Two thirds of the patients had
at least one complication. Forty-three percent had scoliosis or
kyphosis, and 10% developed benign chest tumors (osteochondromas). Secondary tumors were noted in three patients within
the lung field (two osteogenic sarcomas of the rib and one breast
cancer), and all succumbed to these tumors. Pulmonary function was examined by Attard-Montalto and colleagues.177 Subjectively, 63% percent of patients had mild to moderate exercise
intolerance, and objective measurement of vital capacity and total lung capacity was decreased compared with age and height
predicted values in all. All of the females had breast hypoplasia.
In another study of long-term survival of females, all developed
breast hypoplasia and one had breast cancer.198
Other Renal Tumors

------------------------------------------------------------------------------------------------------------------------------------------------
CLEAR CELL SARCOMA OF KIDNEY

CCSK accounts for 3% of renal tumors reported to the COG
studies. Each year, approximately 20 new cases of CCSK are diagnosed in the United States. CCSK was recognized as a distinct
clinicopathologic entity by Kidd in 1970.199 CCSK has been described as nests of ovoid, epithelioid, or spindled cells separated
by fibrovascular tissue with a chicken wire pattern of small
blood vessels. Most tumors show evidence of this classical pattern, but other reported histologic patterns seen include myxoid,
sclerosing, cellular, epithelioid, palisading, spindle-cell, storiform, and anaplastic patterns.200 Immunohistochemistry is
WILMS TUMOR
437
used to exclude other renal tumors. CCSK is nonspecifically

vimentin and Bcl-2 positive. Gene-expression profiling studies
demonstrate the expression of neural markers (e.g., nerve
growth factor receptor), expression of member genes of the
Sonic Hedgehog pathway and the phosphoinositide-3-kinase/
Akt cell proliferation pathway.201,202 Recently, a translocation
t(10;17) and deletion 14q have also been described in CCSK,
suggesting that they may play a role in its pathogenesis.203
CCSK is characterized by bone and brain metastases and the
increased tendency for late recurrences. Long-term followup of CCSK patients is needed because 30% of relapses occurred more than 3 years after diagnosis, and some occurred as
late as 10 years after diagnosis.204 The tumor is generally unilateral and unicentric, with solid and, occasionally, cystic areas.
On NWTS-1 to NWTS-3, treatment for CCSK was the same as
for WT, and the outcomes were poor. In NWTS-4, patients
were treated with vincristine, dactinomycin, doxorubicin,
and RFS, and overall survival was improved versus NWTS-3
(RFS 71.6% versus 60.2% at 8 years, P 0.11; OS 83% versus
66.9% at 8 years, P < 0.01).204 To further improve survival,
patients on NWTS-5 with CCSK were treated using regimen
I (see Table 30-2), because etoposide and cyclophosphamide
were active against CCSK in preclinical models.205 Four-year
OS for stage I patients was 100%. Stage II, III, and IV had 4-year
OS of 88.9%, 94.8%, and 41.7%, respectively. LOH was
not found in most cases of children with CCSK and is not predictive of outcomes. In the current COG study, patients
with CCSK are treated according to the high-risk study.
Patients with stage I disease will continue to be treated with
regimen I but will not receive radiation therapy. The need to
minimize unnecessary therapy in patients with stage I CCSK
is highlighted by the fact that treatment-related deaths in the
Argani series outnumbered tumor-related deaths, two versus
one.200 In addition, none of the stage I patients from NWTS5 have relapsed, with a median follow-up of more than 4 years.
To improve survival for children with higher-stage disease,
they will be treated with revised UH-1 (see Table 30-4).
RHABDOID TUMOR OF THE KIDNEY

RTK was initially described in 1978 as a rhabdomyosarcomatoid variant of WT.206 Haas used the term rhabdoid tumor
in 1981, because of the absence of muscle differentiation.207
RTKs have been reported to occur throughout the body, including the brain, liver, soft tissues, lung, skin, and heart.
RTK accounts for 2% of all renal tumors, and it is the most
aggressive and lethal of all pediatric renal tumors. Clinical features that help distinguish an RTK from WT clinically include
the presence of hypercalcemia and diffuse lymphatic and hematogenous spread in a young infant. Tomlinson and her colleagues reviewed 142 patients with RTK from NWTS-1 to
NWTS-5.208 Age at diagnosis was found to be a highly significant prognostic factor for survival of children with RTK. Infants have a dismal prognosis, whereas older children have a
slightly more favorable outcome. Higher tumor stage and
presence of a central nervous system (CNS) lesion were also
predictive of a poor rate of survival. Unfortunately, these tumors tend to present at an advanced stage and are resistant
to chemotherapy.209 RTK is associated with second primary
tumors in the brain, including cerebellar medulloblastomas,
pineoblastomas, neuroblastomas, and subependymal giant
cell astrocytomas.210
438
PART III
Grossly, the tumors are solid, unencapsulated, and often

have extensive hemorrhage and necrosis. The tumors are very
invasive. Microscopically, they consist of sheets of cells showing
nuclear pleomorphism and characteristic morphologic features
of open vesicular nuclei, prominent nucleoli, and scattered
hyaline eosinophilic cytoplasmic inclusions composed of intermediate filaments in a whorled pattern. At present, no single
immunohistochemical stain or profile is considered to represent
a diagnostic criterion. Recently, genetic abnormalities of the
hSNF5/INI1 tumor suppressor gene on chromosome 22 have
been shown to be characteristic for both renal and extrarenal
rhabdoid tumors; the gene is important for chromatin remodeling. For all other renal tumors, except RTK, immunohistochemical staining for the wild-type integrase interactor 1 (INI-1)
protein shows nuclear positivity. In renal and extrarenal rhabdoid tumors, this is absent.211 This antibody is being evaluated
for its diagnostic utility in the current COG renal tumor study.
Both SIOP and COG/NWTSG have reported poor outcomes for children with RTK.208,212 The outcomes by stage
from NWTS-5 are stage I 50.5%, stage II and III
33.3%, stage IV 21.4%, stage V 0%. Children with
RTK, on the current COG study, will be treated using revised
UH-1 if they are stage I to IV and have no measurable disease
after surgery. If they have measureable disease (stage III, IV),
they will receive a vincristine/irinotecan window, followed
by revised UH-2 if they have a partial or complete response
(see Table 30-2). The rationale for this treatment strategy
was based on reviewing the outcomes from the intergroup
rhabdomyosarcoma (IRS) studies and several case reports that
documented the successful treatment of advanced or metastatic rhabdoid tumor of the kidney.213215
RENAL CELL CARCINOMA

RCC in childhood accounts for 5% to 8% of all pediatric and
adolescent renal malignancies. They are more common than
clear cell sarcoma of the kidney and malignant rhabdoid.1
The median age at presentation in children is 9 years. By
age 15, RCC becomes as common as WT (Fig. 30-10). In
the pediatric population, there have been limited therapeutic
studies with no randomized controlled trials. Similar to WT,
children with RCC generally present with an asymptomatic
abdominal mass, although hematuria is a frequent finding.216
Imaging studies cannot differentiate RCC from other solid
RENAL CANCER AGE-SPECIFIC INCIDENCE RATES BY TUMOR
SEER 1975-1995
20
Average annual rate
per million
0.1
Wilms tumora
Renal cell carcinoma
15
18.3
10
CONGENITAL MESOBLASTIC NEPHROMA

0.1
5.6
renal tumors. RCC in children can be divided into two broad

pathologic groups.217 The first is the classical clear cell histology. This includes the adult-type RCC with 3p25 (VHL locus)
genetic abnormalities and tumors in patients with tuberous
sclerosis. In addition, there is a unique genetic subtype of clear
cell that presents in adolescents and young adults, accounting
for nearly one third of all cases. These tumors are characterized by the chromosomal translocations involving the TFE3
gene on Xp11.2217219 or the TFEB gene on 6p21.220,221
The abnormal gene fusions produce protein dysregulation
and result in overexpression of either TFE3 or TFEB transcription factors, which contribute to tumor pathogenesis. Immunohistochemistry can detect aberrant expression for TFE3 or
TFEB and can thus be useful in establishing the diagnosis.221,222 In addition, these translocation-positive RCCs have
been described as second malignancies following previous
chemotherapy.223,224
The second subgroup of pediatric RCCs are the papillary
RCCs.225227 Papillary renal cell carcinoma appears more frequently than classical clear cell. Other RCC cell types include
chromophobe or collecting duct types.228 Renal medullary
carcinomas are rare, but highly aggressive, malignancies that
are associated with sickle cell hemoglobinopathy.229,230 Approximately 25% of pediatric RCCs are not able to be classified
because of atypical histologic features.217
Complete tumor resection is the most important determinant of outcome in RCC.228 Younger age at diagnosis is
also a favorable prognostic factor. It has been suggested that
regional lymph node involvement does not portend the same
grave prognosis as it does in adult renal cell carcinoma; however, because this impression was reached based on only
13 patients, further evaluation is required.231 Data collected
from RCC patients enrolled on NWTS-5 showed 5-year OS
survival rates by stage: stage I 92.5%, stage II 73%, stage III
55%, and stage IV 9%. Similar to adult RCC, prognosis
worsens with increasing stage, although direct comparisons
of adult and pediatric data are confounded by the finding that
most reviews of pediatric RCC used the modified Robson
staging system rather than the tumor-node-metastasis
(TNM) system. Neither chemotherapy nor radiation therapy
have demonstrated activity in adult or pediatric patients with
metastatic RCC. To address this lack of knowledge and experience, for the first time these tumors will be addressed in a
COG protocol. To enable comparison with adult tumors,
the staging system proposed by the World Health Organization will be used. The relatively good survival rate for children
with localized RCC combined with the relative inefficacy of
the known adjuvant therapies support treating children without adjuvant therapy. However, the provision of adjuvant chemotherapy is at the discretion of the local physicians. A major
future thrust will be to identify novel agents with activity
against RCC.
0.4
0.8
<5
0.7
0.4
5 to 9
10 to 14
15 to 19
Age at diagnosis (years)
FIGURE 30-10 Incidence of renal cell carcinoma and Wilms tumor by

age. SEER, Surveillance, Epidemiology, and End Results (Program).
Congenital mesoblastic nephroma (CMN) is the most frequent

renal neoplasm of newborns and young infants, accounting
for 5% of all renal tumors.129,232234 The median age at diagnosis
is 2 months. In 1967, Bolande and colleagues were the first to
describe the tumor as a separate entity from WT.234 CMN are
firm on gross examination, and the cut surface has the yellowish gray trabeculated appearance of a leiomyoma. To date,
CHAPTER 30
three histologic subtypes have been described. The classical

type, first identified by Bolande (24% of cases), cellular type
(66% of cases), and mixed type (10% of cases) showing both
classical and cellular patterns.235 The classical variant is characterized by leiomyomatous histology, with spindle cells in
bundles, rare mitoses, and the absence of necrosis. It is histologically similar to infantile myofibromatosis. The cellular variant consists of solid, cellular, sheetlike growth pattern of oval
or round cells with little cytoplasm and frequent mitoses and
necrosis, which resembles infantile fibrosarcoma. The mixed
type of congenital mesoblastic nephroma features areas resembling both classical and cellular morphologies.235237 The relationship between mixed CMN and the two main histologic
subtypes is not clear.238
The observation that classical CMN is similar to infantile
myofibromatosis and cellular CMN resembles infantile fibrosarcoma suggests that these may be two distinct entities, and
genetic studies provide evidence in support of this hypothesis.
Cellular CMN is characterized by the t(12;15) translocation,
resulting in the ETV6-NTRK3 fusion gene, a genetic change
that has not been identified in classical CMN, but is characteristic of infantile fibrosarcoma.238,239 This led to the hypothesis
that cellular CMN is an intrarenal occurrence of infantile fibrosarcoma, whereas classic CMN reflects intrarenal fibromatosis.
The cloning of the resulting gene fusion has allowed the development of molecular detection assays for this subtype of congenital mesoblastic nephroma. The absence of the fusion
product in classical congenital mesoblastic nephroma correlates with its demonstrated absence in infantile myofibromatosis. The challenge then is to explain the existence of the
mixed lesions.
Clinically, most children with CMN have an excellent prognosis and are cured with a radical nephroureterectomy with
lymph node sampling.236,240 However, CMN tends to grow
into the hilar and perirenal soft tissue, and recurrence or metastases are seen.241,242 In 1973, the first reports of local recurrences in children with CMN appeared in the literature.243245
Since then, metastasis to the lung, liver, brain, and heart have
been reported.245249 Recurrence and metastatic disease has
lead to a debate concerning the need for adjuvant therapy
to prevent these rare events in a subset of patients versus
the risks of this therapy in infants.237,241
Subsequent investigations demonstrated that recurrences
were seen preferentially in either the cellular or the mixed
subtypes. Other suggested risk factors for recurrence included
age (more than 3 months of age), stage (stage III resulting
from incomplete surgical resection), and vascular invasion.250,251 In 2006, the German Pediatric Oncology Group
published their experience with 50 children with CMNs, suggesting that a subgroup of children more than 3 months of
age with stage III cellular CMN tends to develop recurrences
more often supporting the earlier findings.252 Alternatively,
Perlman and colleagues evaluated 396 cases of CMN from
the database of the J.B. Beckwith Developmental Renal Tumor
Collection.253 Thirty CMNs were known to have recurred
(7.6% overall recurrence rate and 9.3% recurrence rate for tumors with a cellular histologic component). Recurrences took
place within 1 year of diagnosis (range, 2 to 11.5 months); 20
were local, 8 were metastatic, and 2 were both local and
metastatic. None of the classical CMNs recurred, including
18 that were known or suspected to have residual disease. Recurrences were confined to tumors with a cellular component
WILMS TUMOR
439
or cellular and mixed, which had the same risk of recurrence.

Stage III disease was the second factor associated with recurrence. Intrarenal and renal sinus vascular invasion correlated
with increased potential for recurrence; however, the correlation did not achieve independent statistical significance. Other
clinical or pathologic features previously suggested as prognostic factors, including age at diagnosis, were not proven to be of
additional prognostic significance. This study concluded that
the most important risk factors for recurrence in CMNs are
the presence of a cellular histologic component and stage III
disease. However, in none of these reports has the efficacy of
adjuvant therapy been established.
SOLITARY MULTILOCULAR CYST AND

CYSTIC PARTIALLY DIFFERENTIATED
NEPHROBLASTOMA
Cystic renal tumors are a diagnostic and therapeutic challenge
(Fig. 30-11). Cystic nephroma (CN), cystic partially differentiated nephroblastoma (CPDN), and cystic WT (CWT) are a
spectrum with CN at the benign end, CWT at the malignant
end (these must have both a solid and cystic component),
and CPDN in the intermediate position. The three types cannot be differentiated using imaging techniques and can be
confused with cystic clear cell sarcoma and cystic mesoblastic
nephroma.254 Multicystic dysplastic kidney can generally be
distinguished radiographically from the other entities, because it lacks any normal renal parenchyma that the other
lesions should contain.255
CYSTIC NEPHROMA
CN is an uncommon benign renal lesion that occurs most
commonly in children younger than 24 months of age, with
a male to female ratio close to 2:1. A second peak incidence
occurs in adults around 30 years of age, with an 8:1 female
to male predominance.255258 Grossly, these masses are
FIGURE 30-11 A magnetic resonance scan of a cystic nephroma.
440
PART III
well-encapsulated multilocular tumors composed of varioussized cysts with thin septations that compress the normal kidney. Microscopically, the identifying feature is that of mature
well-differentiated cell types within the septa of the cyst wall.
There are no blastemal or embryonal elements.254,255 Most
cases are unilateral, but some are bilateral.259 Although CN
is benign, cases have been reported with pleuropulmonary
blastoma as well. The relationship between these two entities
is undefined.260,261
CYSTIC PARTIALLY DIFFERENTIATED

NEPHROBLASTOMA
Cystic partially differentiated nephroblastoma is a multilocular
cystic WT composed entirely of cysts separated by delicate
septa. The majority of these lesions occur in the first 2 years
of life.256,262,263 Cystic partially differentiated nephroblastoma
is usually well circumscribed and sharply demarcated from
the adjacent normal kidney. It can be large (up to 18 cm in diameter) and may produce visible abdominal distention. This
neoplasm is composed entirely of variably sized cysts; unlike
CN, the septal stroma contains small foci of blastema, primitive
or immature epithelium, and/or immature-appearing stromal
cells.264,265 In addition, skeletal muscle fibers are commonly

present in cystic partially differentiated nephroblastoma.
Both COG/NWTSG and SIOP have reported their experiences with CN and CPDN.262,266,267 In the NWTSG study,
21 patients were evaluated.262 Thirteen patients received chemotherapy, and 8 patients did not. In the chemotherapy
group, the distribution by stage was 10 children with stage
I, 2 children with stage II, and 1 child with stage V. The
8 no-chemotherapy patients were all stage I with a 100% survival. The SIOP evaluated 14 patients with diagnoses of cystic
nephroma (7 patients) and cystic partially differentiated
nephroblastoma (7 patients). Two patients received preoperative chemotherapy. Primary nephrectomy was performed in
12 patients. Two patients underwent partial nephrectomy.
In 1 child, postoperative chemotherapy was administered.
None of the patients had progression of disease or recurrence.
Overall survival was also 100%.267 There is some concern
about doing partial nephrectomies because of recurrences after incomplete excision as well as distinguishing this tumor
from other malignant lesions.267,268
expertconsult.com.
CHAPTER 31
Neuroblastoma
Barrie S. Rich and Michael P. La Quaglia
Neuroblastoma is one of the most common solid tumors in

infancy and childhood. This is a neoplasm of neural crest origin, arising in the adrenal medulla and along the sympathetic
ganglion chain from the neck to the pelvis. The clinical course
is quite variable, because this highly malignant tumor demonstrates unusual behavior. Although instances of spontaneous
regression and tumor maturation from a malignant to a benign
histologic form have been observed,17 the disease is progressive in many cases. Survival in children with other malignancies, such as Wilms tumor, rhabdomyosarcoma, acute
lymphocytic leukemia, germ cell tumors, Hodgkin disease,
and non-Hodgkin lymphoma, has been significantly improved
by the intensive use of combined treatment modalities, but
the outlook for many children with advanced neuroblastoma
remains dismal.1,5,812 This neoplasm exhibits great heterogeneity in its behavior and represents a significant challenge to
practitioners.
Primitive neuroblasts can be identified in the fetal adrenal
gland in the 10th to 12th intrauterine week. The nodules increase in number by 20 weeks gestation but gradually diminish in number toward the end of gestation. Neuroblastoma in
situ in the adrenal gland is seen in 1 of every 260 neonates
who die of congenital heart disease and in as many as 1 in
39 infants who die from other causes in the first 3 months
of life. The clinical incidence of the tumor is approximately
1 in 7,500 to 10,000 children.1,10,13,14 Neuroblastoma is responsible for 10% of all childhood tumors and 15% of all
cancer deaths. There are 700 cases diagnosed annually in

the United States. Approximately 40% of cases are diagnosed
by 1 year of age, 75% by 7 years, and 98% by 10 years.1
More than half the patients are younger than 2 years at the
time of diagnosis.15 Neuroblastoma is slightly more common
in boys than in girls, with a male-to-female ratio of 1.2:1.0.1,10
It is the most common intra-abdominal malignancy in
newborns, and the most frequently diagnosed malignancy in
children less than 1 year of age.16
The embryonal nature of neuroblastoma has been well
documented by its identification on prenatal ultrasonography, and the tumor has been known to invade the placenta
during the antenatal period, though this is a rare occurrence.1724 More than 55 cases of antenatally discovered
neuroblastoma have been reported in the literature since the
original description by Fenart and colleagues in 1983.25
The masses are usually identified during ultrasound examinations performed after 32 weeks gestation. The earliest reported instance was observed at 18 weeks gestation.26
Mothers of infants with congenital neuroblastoma occasionally experience flushing and hypertension during pregnancy as a result of catecholamine released from the fetal
tumor in utero.
Neuroblastoma has been described in twins on many occasions, and familial occurrences in both mother and child and
father and son have been reported.23,27,28 Concordance for
neuroblastoma in twins during infancy indicates that hereditary factors may be predominant in this age group, whereas
discordance in older twins suggests that a random mutation
may be more important for this population. The median age
for the occurrence of familial neuroblastoma is 9 months, in
contrast to 18 months in the general population. Maris and
colleagues29 observed that 20% of patients with familial neuroblastoma have bilateral or multifocal tumors and reported
evidence for a hereditary neuroblastoma predisposition locus
on chromosome 16p12-13. Neuroblastoma has been observed in infants with Beckwith-Wiedemann syndrome, neurofibromatosis (von Recklinghausen disease), Hirschsprung
disease, central hypoventilation syndrome (Ondines curse),
and fetal alcohol syndrome, and in offspring of mothers taking
phenytoin (fetal hydantoin syndrome) for seizure disorders.3034 Mutations in the PHOX2B gene, which is often
seen in congenital central hypoventilation disorder, have been
documented in those with familial neuroblastoma, and in
2.3% of those with sporadic neuroblastomas.35 Recently, it
has been determined that genetic mutations in the anaplastic
lymphoma kinase (ALK) gene explain most hereditary neuroblastoma. However, activating mutations of this gene can also
be somatically acquired.36 This discovery has initiated the
development of therapy based on ALK inhibition.37 Although
it is unlikely that environmental factors play an important role
in causing this tumor, neuroblastoma has been noted among
infants of mothers receiving medical therapy for vaginal
infection during pregnancy and with paternal occupational
exposure to electromagnetic fields.1
Neuroblastoma may occur at any site where neural crest tissue is found in the embryo. The neuroblast is derived from
primordial neural crest cells that migrate from the mantle layer of
the developing spinal cord. Tumors may arise in the neck, posterior mediastinum, retroperitoneal (paraspinal) ganglia, adrenal medulla, and pelvic organ of Zuckerkandl.5,10,14,38
In 75% of cases, the tumor is located in the retroperitoneum,
441
442
PART III
Mediastinal 20%
Neck <5%
Paraspinal 25%
Pelvic <5%
Adrenal 50%
FIGURE 31-1 Distribution of cases of neuroblastoma at each of the primary tumor sites. Primary tumors most commonly occur in the adrenal
gland.
in either the adrenal medulla (50%) or the paraspinal ganglia

(25%). In 20% of cases, the primary tumor is in the posterior
mediastinum. Less than 5% of tumors occur in the neck
or pelvis (Fig. 31-1).1,5,10,14 Primary intracranial cerebral
neuroblastoma also occurs.39,40 In addition, a teratoma in an
infant may occasionally contain foci of neuroblastoma. Rare
cases of neuroblastoma arising in the bladder have also been
reported.41
The fate of the neuroblasts can follow 1 of 3 clinical
pathways: (1) spontaneous regression, (2) maturation by differentiation from neuroblastoma to a benign ganglioneuroma,
or most frequently, (3) rapid progression to a highly malignant
tumor that is often resistant to treatment.
Mass Screening
------------------------------------------------------------------------------------------------------------------------------------------------
In an effort to identify early cases of neuroblastoma that were

amenable to cure, mass screening programs were initiated in
Japan in 1985, evaluating urinary vanillylmandelic acid
(VMA) and homovanillic acid (HVA) levels in infants at
6 months of age. These studies identified a large number of
infants with neuroblastoma. The survival in these cases was
exceptionally high compared with the survival in patients
who present with clinical disease diagnosed by conventional
methods. The Japanese screening effort doubled the actual
incidence of neuroblastoma in infants younger than 1 year
of age, but neither decreased the number of cases observed
in older children nor improved the survival of children older
than 1 year of age.1,4244 Sawada and colleagues43,44 reported
a 96% survival rate in 170 cases of neuroblastoma identified by
screening. These observations suggest that neuroblastomas
identified by screening were most likely biologically favorable
tumors that spontaneously regressed.43 However, a small
number of screened patients have had tumors with unfavorable biologic markers and a poor prognosis, and a few screened
patients who tested negative at 6 months of age later
(at 12 to 18 months of age) developed highly aggressive neuroblastomas.45
In general, mass screening has provided important information regarding the natural history of this enigmatic tumor
and has identified a group of tumors that clearly regress
and represent a biologically favorable form of tumor, in contrast to that noted in older children.7 Prospective, populationbased, controlled screening trials in Quebec minimized the
rate of false-positive cases, but had an overall sensitivity of
only 45%. The results were similar to the findings in Japan.46

A German study offered screening to 2.6 million infants between 9 and 18 months of age. This effort identified 149 cases
of neuroblastoma in 1,800 screened infants, demonstrating a
predictive value of 8%.47 The German investigators estimated
that two thirds of the tumors detected by screening would
have regressed spontaneously. The potential risks were
highlighted by the fact that all 3 children who died in the
group detected by screening had localized disease and succumbed from complications of treatment. These studies in
North America, Japan, and Europe suggest that screening
may result in an overdiagnosis of neuroblastoma and the performance of unnecessary therapies.48 However, the results observed in screening studies are valuable and should help
minimize treatment in the substantial subset of infants diagnosed with early-stage neuroblastoma that has an excellent
chance of either maturing or spontaneously regressing.49
Because of compelling medical and psychological reasons, especially among parents in false-positive cases, neuroblastoma
screening was discontinued in many countries.50,51 Following
the cessation of screening elsewhere in the world, the Ministry of Health in Japan discontinued its mass screening
program in April 2004.52
------------------------------------------------------------------------------------------------------------------------------------------------
Neuroblastoma is a tumor with multiple clinical manifestations related to the site of the primary tumor, the presence
of metastases, and the production of certain metabolic tumor
byproducts. In 50% to 75% of reported cases, patients present
with an abdominal mass. The tumor may be hard, nodular,
fixed, and painful on palpation. Generalized symptoms include weight loss, failure to thrive, abdominal pain and distention, fever, and anemia.1,5,10,14 Hypertension is found in 25%
of cases and is related to the production of catecholamines
by the tumor. Instances of hypercalcemia have been observed
in association with neuroblastoma, and hemoperitoneum
caused by sudden spontaneous rupture of the neoplasm
has also been reported.53,54
Neoplasms arising in the upper mediastinum or neck may
involve the stellate ganglion and cause Horner syndrome,
which is characterized by ptosis, miosis, enophthalmos,
anhydrosis, and heterochromia of the iris on the affected
side.5,10,14 Metastases to the bony orbit may produce proptosis or bilateral orbital ecchymosisoften referred to as panda
eyes or raccoon eyes (Fig. 31-2). The latter finding in a child
FIGURE 31-2 Child with bilateral orbital ecchymoses (panda eyes or

raccoon eyes) resulting from orbital metastases from neuroblastoma.
CHAPTER 31
NEUROBLASTOMA
443
FIGURE 31-3 A, Plain chest radiograph shows the presence of a left upper thoracic tumor. B, Computed tomography scan documents a mass in the
posterior mediastinum that contains calcium, suggestive of a neuroblastoma.
without a history of trauma should always raise the index

of suspicion for the presence of a malignancy. Mediastinal
tumors may be associated with respiratory distress because
of the tumors interference with lung expansion and
dysphagia caused by extrinsic pressure on the esophagus
(Fig. 31-3).10,5557 Mediastinal and paraspinal retroperitoneal
lesions may manifest with paraplegia related to tumor extension through an intervertebral foramen, resulting in a dumbbell- or hourglass-shaped lesion that may cause extradural
compression of the spinal cord.14,5861 In a few patients,
cauda equina syndrome has also been observed. Pelvic tumors
may be associated with bladder and bowel dysfunction. They
are usually palpable on rectal examination. They must be
differentiated from presacral teratoma, yolk sac tumor,
nonosseous Ewing tumor, and pelvic rhabdomyosarcoma.5,10
Anemia is often related to bone marrow invasion by the
tumor. Excessive catecholamine production by the tumor
may result in flushing, sweating, and irritability. Acute cerebellar ataxia, characterized by opsomyoclonus and nystagmus (dancing eye syndrome), has been observed.6266
This syndrome is seen more frequently (> 60%) in patients
with primary mediastinal tumors, in patients with stage I or
II disease, and in infants younger than 1 year of age.62,66
In addition, they are often more histologically mature. The involuntary muscular contractions and random eye movements
are unrelated to metastases. The cause is suggested to be an
autoimmune phenomenon related to an antigenantibody
complex involving antibodies that cross-react with Purkinje
cells in the cerebellum.62,64,66,67 Poor school performance
and learning deficits may occur as sequelae.64,65 The survival
rate for patients who present with opsomyoclonus and nystagmus is approximately 90%. Presence of the dancing eye syndrome in patients who present with advanced tumors and
N-myc overexpression, however, is associated with a poor outcome.68 Despite tumor resection and adrenocorticotropic hormone treatment, the neurologic symptoms in survivors
(including learning disabilities and attention deficits) may
persist for many years.6365
Infants with neuroblastoma, ganglioneuroblastoma, and,
occasionally, benign ganglioneuroma may present with
intractable diarrhea characterized by watery, explosive
stools and hypokalemia.6971 The diarrhea is related to the
production of vasoactive intestinal polypeptide (VIP) by the

tumor.10,6971 Tumors associated with this syndrome often
have somatostatin receptors and are differentiated, low-risk
tumors. Serum VIP levels can serve as a tumor marker; the
tumor often does not secrete catecholamines. These observations suggest that somatostatin receptor expression is a favorable prognostic factor.72,73
Children with advanced neuroblastoma frequently show
evidence of protein-calorie malnutrition associated with
immunoincompetence, based on anergy to a variety of skin test
antigens.74,75 Rickard and colleagues74,75 demonstrated that
patients with stage IV neuroblastoma who were malnourished
at diagnosis had more treatment delays and a significantly
worse outcome than adequately nourished counterparts with
similar disease severity. These findings suggest that a nutritional
assessment at diagnosis should be a component of the patients
staging.74 In addition, Van Eys and colleagues76 and Rickard
and colleagues74,75 showed that significant nutritional depletion occurs with multimodal cancer therapy and that total
parenteral nutrition can replete and maintain the patients
nutritional status during intensive tumor therapy. In another
study, Sala and colleagues77 reported that the incidence of
malnutrition in children with advanced neuroblastoma was
50%. They stressed the importance of nutritional status and
its possible influence on the course of the disease and survival.
Of interest is a study from Toronto, Canada, that implies
that mandatory folic acid fortification of flourinitially
intended to reduce the incidence of neural tube defects
was associated with a 60% decrease in the incidence of
neuroblastoma in the province of Ontario.78
Neuroblastoma may spread by direct extension into surrounding structures, lymphatic infiltration, or hematogenous
metastases. Regional and distant lymph nodes, liver, bone
marrow, and bone cortex are frequently involved.5,10,11,7981
Patients with bone cortex metastases have an ominous prognosis. Bone metastases occur in sites containing red marrow
and involve the metaphyseal areas of long bones in addition
to the skull, vertebral column, pelvis, ribs, and sternum.1,5,10,11 Bone lesions may cause extreme pain and may
be first identified when a child refuses to walk because of
leg pain. Hematogenous metastases to the brain, spinal cord,
and heart are unusual. Brain metastases usually manifest in
444
PART III
older children with headaches and seizures.8,82 Lung metastases are found on chest radiographs in only 4% of patients.83
This may be the result of direct extension to the lung from
mediastinal lymph nodes or diffuse hematogenous spread,
presenting with a radiographic pattern that may be confused
with pulmonary edema or interstitial pneumonia.83 Lung
involvement by intralymphatic metastases (not seen on chest
radiographs) may be noted at autopsy. Occasionally, patients
with advanced disease present with a bleeding diathesis
related to thrombocytopenia from extensive involvement of
bone marrow and interference with hepatic production of
clotting factors by liver metastases. Multiple subcutaneous
skin nodules and hepatomegaly may occur in infants with
stage IV-S neuroblastoma.
Diagnosis of neuroblastoma is made through a variety of imaging and isotopic studies, serum and urine determinations,
and histologic and genetic evaluation of tumor tissue. On
the plain abdominal radiograph, approximately 50% of cases
may show finely stippled tumor calcification.10,11,14 Radiographs also may show displacement of bowel gas by a mass.
Paraspinal widening is commonly found with celiac axis
tumors. Chest radiographs may show a posterior mediastinal
tumor, a paraspinal widening above the diaphragm from
extension of an abdominal tumor, or a primary thoracic tumor. The diagnostic workup of patients with retroperitoneal
tumors includes an initial upright radiograph of the abdomen,
an ultrasound examination to distinguish a cystic from a solid
lesion, and an evaluation for potential obstruction or compression of the inferior vena cava. As a rule, obstruction of
the inferior vena cava in patients with neuroblastoma suggests
the presence of an initially unresectable lesion.10,84 Computed
tomography (CT) can demonstrate tumor calcification in approximately 80% of cases (Fig. 31-4).14,85 With CT studies
using contrast enhancement, one can often distinguish kidney
and liver from adrenal and paraspinal lesions and evaluate for
intracranial extension of skull metastases.10,85,86 Magnetic

resonance imaging (MRI) is extremely useful in detecting
intraspinal tumor extension and, in some instances, the tumors relationship to major vascular structures. Helical (spiral)
CT with three-dimensional reconstruction is also a useful
method of evaluating this latter relationship. Abdominal CT
is performed with intravenous contrast material, so that an intravenous urogram can be acquired during the same study.84
In most instances, paraspinal or adrenal neuroblastoma causes
lateral or downward displacement of the ipsilateral kidney or
ureter (or both). A separate intravenous urogram is not necessary. Metaiodobenzylguanidine (MIBG) also images both soft
tissue and bony disease. A recent study for the International
Neuroblastoma Risk Group (INRG) task force proclaims
MIBG the most sensitive and specific imaging modality for
staging purposes for neuroblastoma, in addition to recognizing response to treatment, especially when a semiquantitative scoring method is used.87 A long bone survey, isotopic
bone scintigraphy (using the bone-seeking isotopes technetium
and 131I-MIBG), and multiple bone marrow aspirates are
also obtained.10,11,84,85 Isotopic bone scans are also used to
identify bone metastases; they show a close correlation with
the radiographic skeletal survey and are occasionally more sensitive.85 False-positive bone scans can occur in cases of recent
bone trauma or inflammation. The bone-seeking isotopes are
picked up by metastatic foci in the bone and by the punctate
calcifications in the primary tumor (Fig. 31-5).10,14 Demonstration of the bone-seeking isotope in a retroperitoneal or posterior mediastinal mass suggests that the lesion is a
neuroblastoma. Although angiography was once performed to
evaluate many childhood tumors, this test is rarely used
today because vascular structures can be readily identified
with less potential morbidity by other imaging studies such as
helical CT or magnetic resonance angiography (Fig. 31-6).
Because neuroblastoma is a tumor derived from neural
crest cells, it may secrete hormonal products and is likely a
member of the amine precursor uptake and decarboxylation
(APUD) family of tumors. More than 90% of children with
neuroblastoma have tumors that produce high levels of
catecholamines or their byproducts. Quantification of
FIGURE 31-4 Abdominal computed tomography shows a retroperitoneal mass with stippled calcification, consistent with neuroblastoma.
FIGURE 31-5 123I-MIBG scintiscan shows the presence of bone metastases and uptake of the isotope in a primary tumor in the adrenal gland.
Diagnosis
------------------------------------------------------------------------------------------------------------------------------------------------
CHAPTER 31
NEUROBLASTOMA
445
FIGURE 31-6 Helical computed tomography scan with three-dimensional reconstruction of a neuroblastoma arising near the celiac axis. A, Anterior view
indicates that the tumor does not involve the branches of the celiac axis. B, Lateral view demonstrates that the superior mesenteric artery passes through
the tumor.
catecholamine byproduct secretion is best done by 24-hour

urine collection.86 Adrenaline, noradrenaline, dopamine,
metanephrine, HVA, VMA, and vanillylglycolic acid levels
are determined. Children with immature, more undifferentiated tumors tend to excrete higher levels of certain byproducts (e.g., HVA).14 Patients with more mature tumors excrete
more VMA. In rare instances, however, the tumor does not secrete excessive catecholamines. Prasad and colleagues88 suggested that these are parasympathetic neuroblastomas that
secrete increased levels of acetylcholine and fail to metabolize
tyrosine to dopamine. Patients with advanced malignancy
have elevated urine concentrations of cystathionine and
homoserine; increased serum levels of neuron-specific enolase, ferritin, and lactic dehydrogenase; and, in 25% of cases,
sera positive for carcinoembryonic antigen.8993 Hann and
colleagues89 reported that 63% of patients with stage IV disease had high serum ferritin levels, which was predictive of
a poor prognosis, especially in girls older than 2 years. A number of studies showed that neuroblastic tumors produce
increased serum levels of neuron-specific enolase.90,92
Zeltzer and colleagues93 documented that neuron-specific
enolase levels are elevated in 96% of patients with metastatic
disease and that high serum levels are associated with a poor
prognosis, particularly in infants. Elevated serum lactic dehydrogenase levels are also associated with a poor prognosis
in localized neuroblastoma.91 Although these observations
are of historical interest, none of these serum levels are independent prognostic factors, nor are they currently used to
determine treatment. Although histologic examination of
tissue is the key to the conclusive diagnosis of neuroblastoma,
in advanced disease, rosettes of tumor cells in bone marrow
aspirate and increased urinary excretion of VMA or other
catecholamine byproducts are often considered indicative of
the diagnosis. Immunologic analysis of bone marrow aspirate
may be more sensitive than conventional analysis in detecting
tumor cells.94 Serial immunocytologic analysis of peripheral
blood samples has also identified circulating neuroblasts,

documenting tumor dissemination.
Staging
------------------------------------------------------------------------------------------------------------------------------------------------
Various staging schemes for neuroblastoma were used in the

past. In 1988, an international staging system was devised,
establishing a common set of criteria that could be used worldwide and would permit the accrual of large numbers of cases
and allow valid comparisons of data (Table 31-1).95 This system takes into account tumor size and location relative to the
midline, in addition to the presence and degree of metastatic
disease. It depends on the extent of surgical resection of the
primary tumor in patients with nonmetastatic disease. Recently, the INRG developed a new staging system that takes
into account pretreatment imaging of the tumor and bone
marrow morphology, instead of surgical resection, which is
dependent upon the approach of the surgeons and thus
varies from institution to institution; this system appears in
Table 31-2.96 The aim of this system is to better evaluate
pretreatment risk based on image-defined risk factors, and
was developed to be used in tandem with the international
system. In contrast to the International Neuroblastoma
Staging System (INSS), infiltration across the midline is not
included in this classification system.96 Prospective analyses
to validate this new system are ongoing.
Pathology and Histology

------------------------------------------------------------------------------------------------------------------------------------------------
The pathologic classification of neuroblastoma has been revised, and histologic features of the tumor that have important prognostic value have been established.9799 Previously,
the Shimada classification system was used. The Shimada
446
PART III
TABLE 31-1
International Neuroblastoma Staging System
Stage
Description
Localized tumor confined to area of origin; complete excision,

with or without microscopic residual disease; ipsilateral and
contralateral lymph nodes negative (nodes attached to
primary tumor and removed en bloc with it may be positive)
Unilateral tumor with incomplete gross excision; ipsilateral
and contralateral lymph nodes negative
Unilateral tumor with complete or incomplete excision;
positive ipsilateral, nonadherent regional lymph nodes;
contralateral lymph nodes negative
Tumor infiltrating across the midline with or without lymph
node involvement; or unilateral tumor with contralateral
lymph node involvement; or midline tumor with bilateral
lymph node involvement or bilateral infiltration
(unresectable)
Dissemination of tumor to distant lymph nodes, bone, bone
marrow, liver, or other organs
Localized primary tumor as defined for stage I or II with
dissemination limited to liver, skin, or bone marrow (limited
to infants younger than 1 yr)
IIA
IIB
III
IV
IV-S
From Brodeur GM, Pritchard J, Berthold F, et al: Revision of the international

criteria for neuroblastoma diagnosis, staging and response to treatment.
J Clin Oncol 1993;11:1466-1477; Brodeur GM, Seeger RC, Barrett A, et al:
International criteria for diagnosis, staging, and response to treatment in
patients with neuroblastoma. J Clin Oncol 1988;6:1874-1881.
TABLE 31-2
The New International Neuroblastoma Risk Group
Staging System
Stage
Description
L1
Localized tumor not involving vital structures as defined by

the list of image-defined risk factors and confined to one
body compartment
Locoregional tumor with presence of one or more imagedefined risk factors
Distant metastatic disease (except stage MS)
Metastatic disease in children younger than 18 months with
metastases confined to skin, liver, and/or bone marrow
L2
M
MS
Note: Patients with multifocal primary tumors should be staged according to

the greatest extent of disease as defined in the table.
Reprinted from Monclair T, Brodeur GM, Ambros PF, et al: The International
Neuroblastoma Risk Group (INRG) Staging System: An INRG Task Force
Report. J Clin Oncol, 2009 10;27:298-303, with permission. # American
Society of Clinical Oncology.
system divided neuroblastic tumors into age-related favorable

and unfavorable histologic categories, based on whether the
tumor exhibited a stroma-rich or stroma-poor appearance
(Table 31-3).98 Stroma-rich tumors are characterized by extensive Schwannian stroma and signs of neuroblastic differentiation
(i.e., developed nuclear and cytoplasmic features of ganglion
cells). Stroma-poor tumors contain immature, undifferentiated
neural crest cells and have a high mitotic karyorrhexis index
(MKI). The MKI refers to nuclear fragmentation and is determined by the sum of the number of necrotic tumor cells; the
number of cells with mitosis; and the number of cells with
malformed, lobulated, or pyknotic nuclei per 5,000 cells
examined. The MKI varies with age; a high MKI value in infants
younger than 18 months is greater than 200/5,000 cells, and
for those older than 18 months it is greater than 100/5,000
cells. All patients older than 5 years have unfavorable
TABLE 31-3
Modified Shimada Pathologic Classification of Neuroblastic
Tumors
Age
Stroma-rich
appearance
All
Stroma-poor
Age < 18
appearance (i.e., months
neuroblastoma)
Age 18-60
months
Age > 5
years
Favorable
Histology
Unfavorable
Histology
Well differentiated Ganglioneuro(ganglioneuroma) blastoma, nodular

Ganglioneuroblastoma,
intermixed
MKI < 4%
MKI > 4% or
undifferentiated
MKI < 2% and
differentiating
None
MKI > 2%, or

undifferentiated
or poorly
differentiated
All
From Shimada H, Chatten J, Newton WA Jr, et al: Histopathologic prognostic

factors in neuroblastoma: Definition of subtypes of ganglioneuroblastoma
and an age-linked classification of neuroblastoma. J Natl Cancer Inst
1984;73:405-416; Shimada H, Stram DO, Chatten J, et al: Identification of
subsets of neuroblastomas by combined histopathologic and N-myc
analysis. J Natl Cancer Inst 1995;87:1470-1476.
MKI, mitotic karyorrhexis index.
histology. Stroma-poor tumors often have MYCN amplification, a high MKI, and a dismal outcome. A report by
Shimada and colleagues99 documented that both histology
and MYCN amplification provided prognostic information that was independent of staging. Neuroblastomas with
MYCN amplification have a characteristic histopathologic
phenotype and a rapidly progressive clinical course.
The International Neuroblastoma Pathology Classification
(INPC) adopted the Shimada classification with some minor
modifications.90,100102 This age-linked classification is both
prognostically significant and biologically relevant. The current system subdivides the undifferentiated subtype into
undifferentiated and poorly differentiated tumors; changes the
name of stroma-rich, well-differentiated tumors to ganglioneuroma intermixed; and adds a descriptive Schwannian,
stroma-dominant character to ganglioneuroma.103 There is
also a ganglioneuroblastoma nodular (GNBn) group that is
both Schwannian stroma rich/stroma dominant and stroma
poor. Age remains a critical prognostic factor, and the grade
of differentiation and MKI have different prognostic effects,
depending on the patients age at diagnosis. Favorable tumors
are those that are poorly differentiated in children younger
than 1.5 years of age, differentiating in children younger than
5 years of age, ganglioneuroblastoma intermixed, and ganglioneuroma. MKI is low (in those less than 5 years of age) or intermediate (in those less than 1.5 years of age) in this group as
well. Unfavorable tumors are those that are undifferentiated or
poorly differentiated in children older than 1.5 years, or any
subtype of neuroblastoma in children older than 5 years. Patients with high MKI, or patients older than 1.5 years with an
intermediate MKI, also have an unfavorable prognosis.101 Although the presence of calcification was thought to favorably
influence survival, further studies demonstrated that calcification does not have an independent prognostic impact.97,103
Favorable Shimada histology was associated with an 85% survival rate, compared with 41% for unfavorable histologic
CHAPTER 31
NEUROBLASTOMA
AbsentGanglioneuroma
maturing subtype
FH
PresentGanglioneuroblastoma
intermixed
FH
447
AbsentMicroscopic
neuroblastic foci
Macroscopically
visible
50%
nodules
Present
Schwannian
development
0 or 50%
Undifferentiated
Neuroblastoma
Poorly
differentiated
Ganglioneuroblastoma
nodular classic*
UH/FH
GNBn variant (with or

without macroscopic
visible nodule(s)*
UH/FH
%MKC**
Age
Any MKI
Any age
UH
4%
Any age
UH
Any MKI
1.5 yr
UH
4%
1.5 yr
FH
5 yr
UH
4%
1.5 yr
FH
4%
1.5 yr
UH
2%
1.5-5.0 yr
FH
2%
1.5-5.0 yr
UH
Any MKI
Differentiating
Histology
FIGURE 31-7 International Neuroblastoma Pathology Classification. FH, favorable histology; GNBn, ganglioneuroblastoma nodular; MKI, mitotic
karyorrhexis index; %MKC, mitotic and karyorrhectic cells; UH, unfavorable histology; *classic GNBn (single, macroscopically visible, usually hemorrhagic
nodule in stroma-rich, stroma-dominant tissue background; **MKC 2%, 100 of 5,000 cells; MKC 4%, 200 of 5,000 cells. (From Peuchmaur M, dAmore ES,
Joshi VV, et al: Revision of the International Neuroblastoma Pathology Classification: Confirmation of favorable and unfavorable prognostic subsets in ganglioneuroblastoma, nodular. Cancer 2003;98:2274-2281.)
types. All GNBn cases were initially classified as unfavorable

tumors. Umehara and colleagues104 were the first to define
subsets of these specific neoplasms that exhibit different behavior. Peuchmaur and colleagues105 recently revised the
INPC by dividing GNBn cases into two prognostic subsetsfavorable and unfavorable. The favorable type was associated with an 86% event-free survival, whereas the
unfavorable type (two thirds of cases) had only a 32%
event-free survival. Children with the favorable subset of
GNBn have an overall survival of greater than 90%, compared
with 33.2% for those with the unfavorable GNBn subset
(Fig. 31-7).106 Large cell neuroblastoma has been identified
as a distinct phenotype with aggressive clinical behavior.107
These tumors have unfavorable histologic features, including
monomorphous undifferentiated neuroblasts, a low incidence
of calcification, and a high MKI. Immunohistochemical studies showed that large cell neuroblastoma cells stained positive
for neuron-specific enolase, prodrug gene products, and tyrosine hydroxylase, and were negative for CD99.107
On gross examination, neuroblastoma usually appears as a
highly vascular purple-gray mass that is often solid but occasionally cystic. The tumor has an easily ruptured, friable pseudocapsule that may lead to significant hemorrhage during
operative manipulation. The tumor is often necrotic, especially the undifferentiated form. Mature tumors (ganglioneuromas) have a more solid consistency and frequently have a
fleshy white color. The histologic pattern may be quite variable. Primitive stroma-poor neuroblastomas may be indistinguishable from other small, blue round cell tumors, such as
Ewing tumor, rhabdomyosarcoma, or primitive neuroectodermal tumors. The neuroblast is a small round cell consisting
predominantly of the nucleus without much cytoplasm. Immature, undifferentiated tumors are characterized by closely
packed small spheroid cells without any special arrangement
or differentiation.108 Nuclei may appear cone shaped and are
hyperchromic. Rosette formation may be observed and is considered a sign of early tumor differentiation (Fig. 31-8). The
center of each rosette is formed by a tangle of fine nerve fibers.
More mature-appearing, stroma-rich tumors may contain cells
that resemble normal ganglion cells, with an admixture of histologic components characterized by abundant nerve filaments, neuroblastic rosettes, and ganglion cells all seen in a
single microscopic field.28,109 On electron microscopy, neurofibrils and electron-dense, membrane-bound neurosecretory
granules may be observed. The neurosecretory granules may
be the site of conversion of dopamine to norepinephrine.
448
PART III
cell apoptosis and may limit tumor growth by restricting

angiogenesis.1,114
Biologic and Genetic Alterations

------------------------------------------------------------------------------------------------------------------------------------------------
FIGURE 31-8 Histologic appearance of rosettes of neuroblastoma cells

from a bone marrow aspirate, an early sign of tumor differentiation.
These ultrastructural findings and genetic identification of

the tumor tissue can usually separate neuroblastoma from
other small cell tumors. Segregation of neuroblastoma from
other tumors can also be achieved by immunohistochemical
staining that is positive for neurofilament proteins (S-100),
synaptophysin, neuron-specific enolase, ganglioside GD2,
chromogranin A, and tyrosine hydroxylase staining for these
markers is negative in other small round cell tumors.110
Instances of spontaneous maturation from a highly malignant, undifferentiated neuroblastoma to a ganglioneuroblastoma, and subsequently to a benign ganglioneuroma, have
been observed. Ambros and colleagues111 reported that maturing neuroblastomas consist of both Schwann cells and
neuronal cells, including ganglion cells. Schwann cells have
normal numbers of chromosomes and triploid flow cytometry, in contrast to other neuronal cells, including ganglion
cells.102 These observations suggest that Schwann cells may
be a reactive population of normal cells that invade a neuroblastoma, recruited or attracted by trophic factors, and may
be responsible for tumor maturation and serve as an
antineuroblastoma agent.112,113 Schwann cells also
produce angiogenesis inhibitors that induce endothelial
Unique oncogenes are observed in tumors, such as MYCN and

RAS oncogenes.1,8 Amplification of MYCN (> 10 copies) is associated with advanced disease, tumor progression, and a poor
outcome, especially in children older than 1 year.1,8,95,98,115,116
The MYCN proto-oncogene is located on the short arm of chromosome 2p24. Double minutes and long, nonbanding staining
regions have been observed at this site and may represent amplified cellular genes. Studies have determined that the MycN
protein binds DNA and leads to an increase in the level of
endogenous Mdm2 mRNA and protein expression, with consequent p53 inhibition. This modification of Mdm2 levels by
N-myc may partially explain its role in the aggressiveness of
neuroblastoma.117,118 Approximately 30% of patients with
neuroblastoma have tumors with MYCN amplification. More
than 90% of patients with MYCN amplification have rapidly
progressive disease and are resistant to therapy.
Cellular DNA content is a predictor of response to chemotherapy in infants with unresectable neuroblastoma. DNA
flow cytometry studies evaluating tumor ploidy indicate that
children with diploid tumors have a worse outcome than
those with aneuploid (hyperdiploidy or triploidy) tumors.1,10
Similar to MYCN status, DNA ploidy is of prognostic value
independent of stage and age, and the two factors (MYCN status,
and ploidy) together provide important complementary prognostic information for infants.1,111 DNA ploidy flow cytometry
correlates well with response to chemotherapy and outcome.
MYCN amplification is commonly associated with chromosome
1p deletion and diploidy.119,120 Diploid tumors are commonly
associated with an unbalanced gain of chromosome 17q, even
in the absence of MYCN.1,6,116,120 The most common cytogenetic abnormalities in neuroblastoma are 1p deletion
and 17q gain.119 Both abnormalities are poor prognostic
factors and are associated with worse outcomes.1,6,121123 Allelic
loss of 11q and 14q and gains of 4q, 6q, 11q, and 18q have also
been observed (Table 31-4).1
TABLE 31-4
Genetic Alterations in Neuroblastoma
Genetic Feature
Associated Factor
Risk Group
MYCN amplification
Diploidy or tetraploidy, allelic loss of 1p, high Trk-B,

advanced stage (III, IV)
More aggressive tumor associated with MYCN amplification
Occurs concurrently with MYCN amplification
Often associated with MYCN amplification
Few associated with MCYN amplification; correlates
with LOH 14q
Correlates with LOH 11q, inverse relationship with allelic
loss 1p and MYCN amplification
Familial neuroblastoma, multifocal and bilateral
neuroblastoma
Hirschsprung disease
High
Variable
Beckwith-Weidemann syndrome
Low
Allelic gain 17q

Gain at 4q, 6p, 7q, 11q, 18q
Allelic loss 1p36
Allelic loss 11q
Allelic loss 14q
Predisposition of 16p12-13
Association with chromosome 10
(RET-oncogene)
Association with 11p15.5
High
Risk related to MYCN status
High
Intermediate decreased survival in patients
without MYCN amplification
Intermediate
Low
Note: This table does not include changes in the genetic expression of TRK-A, TRK-B, and TRK-C; the multidrug-resistant protein gene; telomerase; or others that are
covered elsewhere in this chapter.
LOH, loss of heterozygosity.
CHAPTER 31
High expression of the neurotropin Trk-A (a high-affinity

nerve growth factor receptor) is associated with a good
prognosis and is inversely related to N-myc.116,124 Trk-A is
observed in young infants and in those with stage I and
stage IV-S tumors, and indicates a very favorable outcome.116,124 Trk-A is associated with neural cell differentiation and tumor regression and may play a role in
angiogenic inhibition. Trk-A downregulates angiogenic factor
expression and decreases the number of microvessels in
neuroblastoma tumor cell lines. Multivariate analysis, however, suggests that N-myc expression is a more important independent prognostic factor. The low-affinity nerve growth
factor receptor gene is another proto-oncogene that has a
prognostic effect similar to Trk-A and probably influences cellular maturation.1,8,125 In contrast, high expression of Trk-B
with its ligand BDNF may provide an autocrine survival pathway in unfavorable tumors, particularly those with MYCN
amplification, possibly by providing a tumor cell survival
or growth advantage.1,126,127 The Trk-BBDNF pathway also
contributes to enhanced angiogenesis, tumorigenicity, cell survival, and drug resistance.1,126 These patients have more advanced disease, are usually older than 1 year, and have a
dismal outcome.1,126,127 Trk-C expression has also been
identified in neuroblastoma and is usually observed in
lower-stage tumors that do not express N-myc.1,128 A recent
report identified targets of TRK gene expression, and recognized upregulation of proapoptotic factors and angiogenesis
inhibitors. Conversely, Trk-B expression was associated
with upregulation of genes related to invasion and therapy
resistance. Its activation is associated with increased proliferation, migration, angiogenesis, and chemotherapy resistance of neuroblastoma cells.126,129
Another gene has been cloned, the multidrug resistance
(MDR)-associated protein gene, that is associated with chemotherapy resistance, overexpression of N-myc, and a poor
outcome.130 The prognostic role of the MDR gene (MDR-1)
in neuroblastoma is controversial.130,131 High levels of the
MDR-associated protein gene (located on chromosome 16),
however, are associated with a poor outcome. This effect is
independent of stage, N-myc expression, and Trk-A status.130
Similarly, elevated P-glycoprotein levels are associated with
progressive disease and a poor outcome.132,133 Telomerase
is increased in tumor cells and maintains cell viability by preserving the telomeres that protect the end of chromosomes.1,134 There is an inverse relationship between
telomerase levels and outcome in neuroblastoma and a direct
correlation between telomerase levels and MYCN amplification.1 CD44 is a glycoprotein found on the cell surface of a
number of tumors, including neuroblastoma. High expression
of CD44 is associated with a favorable outcome and is usually
found in well-differentiated tumors. In contrast, Nm23 overexpression is observed in instances of advanced and aggressive
neuroblastoma.135 The ganglioside GD2 is found on human
neuroblastoma cell membranes, and increased levels are associated with active disease and tumor progression. Gangliosides inhibit the tumor-specific immune response, and GD2
has become a target for immunotherapy.136
Evaluation of the relationship between tumor angiogenesis
and outcome in infants with neuroblastoma demonstrates
that increased tumor vascularity characterized by microvessel
density correlates with advanced disseminated disease and the
likelihood of metastases.137140 Angiogenesis is associated
with MYCN amplification, unfavorable histology, and poor
NEUROBLASTOMA
449
outcome. Neuroblastoma produces angiogenic factors that

induce blood vessel growth, including vascular endothelial
growth factor (VEGF), platelet-derived growth factor
(PDGF-A), stem cell factor, and their respective receptors
Flk-1, PDGFR, and c-Kit.141 Komuro and colleagues142
demonstrated that high VEGF-A expression correlated with
stage IV disease and suggested that it could be a target for
antiangiogenic therapy. Kaicker and colleagues143 noted that
vascular endothelial growth factor VEGF antagonists inhibit
angiogenesis and tumor growth in experimental neuroblastoma in athymic mice with xenograft neuroblastoma cell line
NGP. They also found that thalidomide suppressed angiogenesis and reduced microvessel density but not tumor growth.
Kim and colleagues144 and Rowe and colleagues145 also demonstrated inhibition of tumor growth in experimental neuroblastoma models using antiangiogenic strategies. Imatinib
mesylate, a compound used to treat patients with gastrointestinal stromal tumors, has been shown to decrease the growth
of neuroblastoma in vivo and in vitro, decrease cell viability,
and increase apoptosis (by ligand-stimulated phosphorylation
of c-Kit and PDGFR) in a severe combined immunodeficiency
(SCID) mouse model.141 Davidoff and colleagues138 demonstrated that gene therapy using in situ tumor cell transduction
with retroviral vectors can deliver angiogenesis inhibitors for
the Flk-1 receptor and restrict tumor-induced angiogenesis
and tumor growth.
The Bcl-2 family of proteins is responsible for relaying apoptotic signals that influence tumor cell regression and is
expressed in most neuroblastomas. The BCL-2 gene produces
a protein that prevents apoptosis. The level of Bcl-2 expression is high in advanced cases associated with a poor outcome and low in cases demonstrating tumor apoptosis
(regression) and differentiation. High Bcl-2 expression may
also play a role in acquired resistance to chemotherapy.146
Subgroups of the Bcl family include Bcl-xL, which inhibits
apoptosis, and Bcl-xS, which induces natural cell death.
VEGF upregulates Bcl-2 expression and promotes neuroblastoma cell survival by altering apoptosis and its regulation
proteins.147 Elevated caspase levels (enzymes responsible
for apoptotic signaling) are associated with an improved outcome in neuroblastomas that demonstrate favorable biologic
features.1 It has been shown that CpGisland hypermethylation inactivates caspase-8, TRAIL apoptosis receptors, the
caspase-8 inhibitor, in addition to other proapoptotic factors.148,149 In view of this finding that gene hypermethylation
leads to resistance patterns, demethylating agents, including
decitabine, are currently being investigated in preclinical
studies.150
Neuroblastoma in Infancy
------------------------------------------------------------------------------------------------------------------------------------------------
For many years, the age of the patient and the stage of disease
at the time of diagnosis were the two key independent variables determining prognosis in children with neuroblastoma.
Evans and colleagues3 and others found that infants younger
than 1 year and those with stage I, II, or IV-S disease had
a significantly better outcome.5,11,34,151,152 Historically,
patients older than 1 year and those with advanced disease
(stages III and IV) did poorly. The worst survival data
were observed in patients older than 1 year with stage
IV disease and metastases to cortical bone.1,5,11,14,153
450
PART III
Percent survival
Age
Over 1 yr
Under 1 yr
10
20
30
40
50
60
70
80
Percent
FIGURE 31-9 Bar graph demonstrates the improved survival in infants with neuroblastoma who are younger than 1 year.
However, recent reviews have confirmed that 18 months

serves as a better cutoff to predict outcome.154157
Infants younger than 18 months at diagnosis have a significantly improved outcome. At the Riley Hospital for Children
(Indianapolis, IN), the survival rate was 76% for infants younger
than 1 year and only 32% for older patients (Fig. 31-9).4 This
favorable outlook for patients younger than 1 year extends
across all stages, including infants with stage IV metastatic disease. The incidence of stage IV lesions in infants younger than
1 year is 30% compared with 60% to 70% in older patients.4
Infants with stage IV disease respond better to chemotherapy than do older children; 50% of infants have a complete
response to treatment compared with 22% of older children.158 This observation suggests that resolution of metastases may have a greater impact on length of survival than does
the surgical excision. Further, this implies that surgical resection is beneficial in some infants and should be attempted
when disseminated disease is controlled by chemotherapy.
However, more intensive chemotherapy regimens and bone
marrow transplantation (BMT) may be necessary to achieve
a cure, especially in highly selected infants presenting with
adverse biologic markers.
Stage IV-S
------------------------------------------------------------------------------------------------------------------------------------------------
The most unusual group of patients with neuroblastoma is

those infants younger than 18 months with stage IV-S disease.
This stage is characterized by hepatomegaly produced by
extensive metastatic disease, subcutaneous metastases, and
positive bone marrow with a primary tumor that would
otherwise be classified as stage I or II. Stage IV-S cases account
for approximately 30% of patients with neuroblastoma
recognized in the first year of life.4
Some infants succumb from complications of their stage IVS disease rather than progression of the tumor. Complications
of severe hepatomegaly include respiratory insufficiency,
caused by significant elevation of the diaphragm by the

large, tumor-filled liver; coagulopathy; and renal compromise
resulting from abdominal compartment syndrome produced
by the mass (Fig. 31-10).4,151,159161 Vomiting may occur because of a change in the gastroesophageal angle related to the
diaphragmatic elevation, resulting in gastroesophageal reflux,
protein-calorie malnutrition, and aspiration pneumonia. Total
parenteral nutrition may be a useful therapeutic adjunct.7476
Most fatalities in stage IV-S cases occur in infants younger than
2 months with severe symptoms related to hepatomegaly, who
do not tolerate therapy as well as do older infants.4,162 Symptomatic hepatomegaly caused by tumor infiltration may benefit
from low-dose radiation to the liver in the range of 600 to
1,200 Gy, administered in doses of 100 to 150 Gy/day.4,5,159
Although some early reduction in the size of the liver is seen,
and peripheral edema may resolve in a few weeks, complete
resolution may take 6 to 15 months.4 Resolution of the liver
mass is probably related more to the natural course of stage
IV-S disease than to radiotherapy. Administration of low-dose
cyclophosphamide 5 mg/kg per day is a reasonable treatment
alternative. Although some investigators advocate the insertion
of a Dacron-reinforced Silastic sheet to create a temporary
ventral abdominal wall hernia to accommodate the enlarged liver and reduce intra-abdominal pressure, mortality
resulting from septic complications has been observed.4,159,163
To reduce the risk of infection, Lee and Applebaum164 recommend the use of an internal polytetrafluorethylene patch
to create a temporary ventral hernia. The graft can be removed
in stages as the bulk of the hepatic mass regresses over time.
Survival of infants with remote metastases is greater than
80%, often without specific treatment. Most patients with
stage IV-S disease (> 90%) have favorable genetic and biologic
factors, including high Trk-A expression, no MYCN amplification, favorable histology, and no evidence of allelic loss of
chromosome 1p. This suggests that the majority of stage
IV-S tumors undergo spontaneous regression. Although most
patients with stage IV-S disease do well, Wilson and
CHAPTER 31
NEUROBLASTOMA
451
FIGURE 31-10 A, Six-week-old infant presented with abdominal distention and hepatomegaly. B, Appearance of the liver at laparotomy. There were
multiple metastatic nodules, and the biopsy confirmed the diagnosis of stage IV-S neuroblastoma.
colleagues161 reported 18 cases with a heterogeneous tumor

presentation and a survival rate of only 50%, including
3 patients with MYCN amplification. The presence of adverse
genetic and biologic prognostic factors suggests that this subset of patients (< 10%) requires more aggressive therapy.
Of interest is that infants with multiple subcutaneous nodules
seem to have the most favorable outlook. This may be because
of increased immunologic activity as a result of tumor being
present in multiple sites.4 Increased uptake of major histocompatibility complex (MHC) class I antigen by neuroblastoma cells in vitro and in vivo may influence the
outcome favorably.165 Infants with stage IV-S disease have
normal levels of MHC class I surface antigen expression,
whereas those with stages I to IV have low levels.165 Sugio
and colleagues166 reported that down-modulation of MHC
class I antigen expression is associated with increased amplification of the dMYCN oncogene in patients with advanced
disease.
In 2000, Nickerson and colleagues162 described 80 infants
with stage IV-S disease from the Childrens Cancer Group
(CCG). Fifty-eight cases were managed without specific therapy. All 44 asymptomatic patients survived without treatment.
Symptomatic patients were treated with cyclophosphamide
5 mg/kg per day for 5 days and hepatic radiation at a dose
of 4.5 Gy over 3 days. Five of six deaths occurred in symptomatic infants younger than 2 months. Event-free 5-year survival
was 86%, and overall survival was 92%. Early intervention is
imperative for stage IV-S patients with life-threatening complications (e.g., hepatosplenomegaly, coagulopathy, renal
failure).4,162 Surgical resection did not alter outcome.
More aggressive chemotherapy is also required in those cases
in which the tumor demonstrates more than 10 copies of
MYCN, chromosome 1p deletion, or other adverse biologic
markers.4,162,167 Amplification of MYCN may be observed

in 1 of 12 patients with stage IV-S tumors who develop progressive disease and die, despite having a favorable prognostic
stage. In 2003, Schleiermacher and colleagues167 reported on
94 infants with stage IV-S neuroblastoma in France; they observed an 88% overall survival and recommended a more intensive regimen using cisplatin and etoposide for those who
require therapy. Some infants with stage IV-S disease have
survived without resection of the primary tumor (in some,
the primary tumor may not be identified).
Cystic Neuroblastoma
------------------------------------------------------------------------------------------------------------------------------------------------
Cystic neuroblastomas are relatively rare and are often identified on prenatal ultrasound examinations.168 They characteristically occur in the adrenal gland, and almost all are
diagnosed in early infancy (Fig. 31-11). Few are calcified,
and only 10% are associated with elevation of urinary VMA
and HVA levels.169 They display a benign behavior and a
favorable outcome. Some evidence suggests that they often
regress and undergo spontaneous involution.26 Some investigators have recommended observation alone, with close serial
sonographic monitoring during the first few months of life.
Operative resection should be reserved for tumors that fail
to regress or that increase in size. Adjuvant chemotherapy is
rarely required after resection. The Childrens Oncology
Group (COG) has performed a prospective study of observation alone for cases of perinatal neuroblastoma, with strict criteria for enrollment, including tumor volume (< 16 mL, if
solid, or < 65 mL, if cystic). Results from the study are not
yet available.
452
PART III
Risk Stratification and Risk-Based

Management
------------------------------------------------------------------------------------------------------------------------------------------------
FIGURE 31-11 Photograph of a cystic neuroblastoma of the adrenal

gland in a 5-month-old baby who required complete excision. The patient
was managed by surgery alone and is a long-term survivor.
Multifocal and Bilateral

Neuroblastoma
------------------------------------------------------------------------------------------------------------------------------------------------
Bilateral neuroblastoma is relatively uncommon, occurring

primarily in familial cases and young infants with alterations
at the predisposition locus on chromosome 16p12-13.29
Therapy has included observation alone; unilateral resection, with observation of the second (smaller) lesion or enucleation; and bilateral adrenalectomy, with postoperative
hormonal replacement. Some bilateral tumors resolve spontaneously, while others persist and enlarge, requiring surgical intervention. The prognosis is generally good for these
tumors, and most of the children survive. Occasionally these
infants have other sites of multifocal disease. Tumor enucleation has been performed in cases with favorable biologic
markers to preserve adrenal function. Hiyama and colleagues170 described multifocal neuroblastoma in 8 of 106
cases (7.5%). Seven of eight cases had favorable histology,
and all expressed Trk-A1 mRNA and the Ha-ras p21 protein.
None of the tumors had MYCN amplification or elevated
telomerase levels. Four had near-triploid DNA on flow cytometry, and all 8 had a proliferative index (percentage of cells in the S phase) of less than 25%. Four patients were treated
with multistage resections. Five had bilateral neuroblastoma and were treated with tumor enucleation. All survived
and are free of recurrence, and none require steroid replacement. The authors reviewed 53 additional cases of multifocal disease and noted that 18 had a family history of
neuroblastoma and 25 were detected incidentally. Because
of the excellent prognosis in patients with favorable biologic
features, Hiyamas group recommended conservative surgical excision (enucleation) using minimally invasive surgical
techniques.170
During the past 2 decades, a number of biologic and genetic

factors have been identified that are important prognostic
indicators and currently define therapy in North America.
Based on the INSS, the use of the INPC, and the identification of numerous biologic and genetic characteristics as
risk factors and predictors of outcome, a risk-based management system has been developed to determine
treatment.1,10,100103 Newer treatment protocols individualize
treatment using risk factors as predictors of outcome in an effort to maximize survival, minimize long-term morbidity, and
improve the quality of life. Current protocols now categorize
patients as low, intermediate, and high risk based on their
prognostic factors (Table 31-5). Good outcomes are associated
with stage I, II, and IV-S patients who are younger than
18 months and have hyperdiploid DNA flow cytometry, favorable histology, less than 1 copy of MYCN, high Trk-A expression, and absence of chromosome 1p abnormalities.
In contrast, a poor prognosis is predicted in children older
than 18 months with advanced tumors (stages III and IV),
more than 10 copies of MYCN, low Trk-A expression, diploid
DNA ploidy, allelic loss of 1p36, and unfavorable histology.
The site of the primary tumor was also considered predictive of survival by some investigators. Patients with tumors in
cervical, pelvic, and mediastinal locations had an improved
outlook compared with children with retroperitoneal (paraspinal or adrenal) tumors. Breslow and McCann153 and Koop
and Schnaufer,152 however, suggested that the improved outlook in these cases can be explained by the patients age and
stage of disease. Despite these conflicting views, Filler and colleagues,56 Young,57 and Adams and colleagues171 reported
that site is a beneficial prognostic indicator for mediastinal
lesions, and Haase and colleagues5 noted the same for pelvic
tumors, regardless of other factors.
Some early reports concerning neuroblastoma suggested
that the more mature and differentiated the tumor, the better
the prognosis.108 Others noted that a more mature histology
may be associated with the same dismal outcome as in patients
with undifferentiated neuroblasts.152 In patients with metastatic disease, the presence of more mature elements seemed
to improve the outlook and was associated with increased survival.9,11 Shimada and colleagues101 subsequently classified
the histopathology of neuroblastoma into favorable and unfavorable types, characterized by a stroma-rich appearance
for the former and a stroma-poor appearance for the latter.
The Shimada classification was also age related. The impact
of Shimada histology class on prognosis proved to be important, especially when associated with other prognostic biologic
variables, particularly amplification of the MYCN oncogene and
allelic loss on the short arm of chromosome 1p (1p36).99 The
current INPC (which embraced and modified the Shimada
classification) further divided cases into subsets of favorable
and unfavorable histologic types and is a highly significant independent predictor of prognosis.98,102,103,105,172 MYCN
amplification is seen in approximately 30% of neuroblastoma
cases and has an important role in modulating the malignant phenotype in neuroblastoma.1,124,173 The prognostic value of MYCN
status is independent of tumor stage and patient age. MYCN
amplification is associated with a poor response to treatment,
CHAPTER 31
NEUROBLASTOMA
453
TABLE 31-5
Neuroblastoma Risk Groups*
Risk group
Low
Low
High
Intermediate
Intermediate
High
High
High
Intermediate
High
High
High
Intermediate
High
Low
Intermediate
Intermediate
High
INSS Stage
1
2a/2b
2a/2b
3
3
3
3
4
4
4
4
4
4
4
4s
4s
4s
4s
Age
MYCN Amplification Status{
Any
Any
Any
< 547 days
547 days
Any
547 days
< 365 days
< 365 days
365 to <547 days
365 to <547 days
365 to <547 days
365 to <547 days
547 days
< 365 days
< 365 days
< 365 days
< 365 days
Any
Not amplified
Amplified
Not amplified
Not amplified
Amplified
Not amplified
Amplified
Not amplified
Amplified
Any
Any
Not amplified
Any
Not amplified
Not amplified
Not amplified
Amplified
DNA Ploidy{
INPC (Modified Shimada) Histology
Any
Any
Any
Any
Any
Any
Any
Any
Any
Any
DI 1
Any
DI > 1
Any
DI > 1
DI 1
Any
Any
Any
Any
Any
Any
FH
Any
UH
Any
Any
Any
Any
UH
FH
Any
FH
Any
UH
Any
*Courtesy Childrens Oncology GroupTable 109.4 Childrens Oncology Group Neuroblastoma Risk Stratification.
{
MYCN nonamplified 1 copy, amplified greater than 1 copy.
{
DNA index > 1 (aneuploid) or 1 (diploid).
DI, DNA index; FH, favorable histology; INPC, International Neuroblastoma Pathology Classification; INSS, International Neuroblastoma Staging System;
UH, unfavorable histology.
rapidly progressive disease, and a dismal outcome. Although attempts to stimulate tumor maturation with nerve growth factor,
adrenergic agonists, papaverine, prostaglandins, exogenous cyclic adenosine monophosphate, and hyperthermia were successful in the laboratory setting, there was minimal clinical
evidence of their usefulness.14,109,174177 The use of retinoids
as a promoter of differentiation, however, has rekindled interest
in this concept and has been successful in prolonging survival in
advanced cases during clinical trials.174 The addition of cis-retinoic acid to the treatment protocol for high-risk cases of neuroblastoma following BMTor peripheral stem cell transplantation is
now standard practice.1
A new pretreatment classification system based on
13 prognostic factors has recently been developed by the
INRG, aiming to create international consensus for risk stratification. These factors include age, INRG stage, histology,
DNA index, MYCN amplification status, and presence of
11q abnormality; they classify patients into 1 of 16 groups.
Each group is categorized as very low, low, intermediate,
and high risk, based on event-free survival rates of more than
85%, more than 75% to less than or equal to 85%, greater than
or equal to 50% to less than or equal to 75%, or less than 50%,
respectively.178
For low-risk patients, surgical excision of the tumor is
usually curative and avoids the risks associated with chemotherapy. Intermediate-risk patients are usually treated with
surgery and chemotherapy. Studies aimed at minimizing
treatment regimens for this group of patients are ongoing.
The poor prognosis in high-risk patients justifies a much
more intense treatment regimen, including combination
chemotherapy followed by complete surgical excision
(if possible), radiotherapy to achieve local control, myeloablative treatments with bone marrow rescue, and biologic
therapy.
Operative Management
------------------------------------------------------------------------------------------------------------------------------------------------
Initial surgery for extensive stage III and IV tumors should be

limited to biopsy of tumor tissue, staging, and placement of a
vascular access device. There is an increased rate of surgical
complications when complete resection is attempted during
initial surgery, with no improvement of survival.179 After 4
or 5 cycles of chemotherapy, second-look surgery is performed. Although opinions vary regarding resection in
high-risk patients, most investigators agree that complete
gross resection, which is associated with excellent local control and improved outcome,179,180 should be the goal of
second-look procedures. Complete surgical removal of the
primary tumor remains an essential component of treatment
in the vast majority of cases.
During resection, adequate intravenous access is important
because these tumors are quite vascular, and blood loss may
be significant. Blood pressure must be carefully monitored
intraoperatively to detect sudden hypertension caused by excessive catecholamine release from the tumor.
The surgical approach depends on the characteristics of the
primary tumor. For upper abdominal lesions, particularly
those involving major midline vessels, thoracoabdominal exposure is advantageous and well tolerated. The goal of resection
is a complete dissection of the vasculature and should include
the primary tumor site, in addition to all regional lymph nodes.
Neuroblastoma often adheres to or surrounds the great vessels,
and special care should be taken to identify and spare the blood
supply to important visceral structures, such as the branches of
the celiac axis and superior mesenteric artery.
In most children with localized disease, all or most of the
tumor can be removed successfully. En-bloc contiguous resection of normal surrounding structures, such as the spleen,
454
PART III
stomach, pancreas, and colon, almost always can be avoided.

In some cases, it is impossible to separate an adrenal or paraspinal neuroblastoma from the ipsilateral kidney, so nephrectomy may be necessary. It is important to excise any suspicious
para-aortic and perirenal lymph nodes for staging purposes.
A routine retroperitoneal lymph node dissection is usually
not performed. The margins of the tumor resection are
marked with titanium clips to guide the port if radiation is required and will reduce the scatter effect noted with other types
of metal clips on follow-up CT scanning.
Because neuroblastoma may have a friable pseudocapsule,
careful handling of the tumor during dissection is important to
avoid tumor spill and hemorrhage. Primary adrenal tumors
may be fed by a number of small arteries. The major venous
drainage is usually constant, directly to the inferior vena
cava on the right side, and into the left renal vein and
subdiaphragmatic vessels on the left. Inferiorly located paraspinal and primary pelvic tumors often require careful dissection to separate the lesion from the bifurcation of the aorta and
inferior vena cava. The tumor frequently extends into the
intervertebral foramina (Fig. 31-12).
Minimally invasive surgical techniques have also been
employed for selected cases of neuroblastoma.180 Adrenal
tumors initially detected by mass screening have been
excised laparoscopically by a number of investigators.181,182
Yamamoto and colleagues49 described three cases of
adrenal neuroblastoma in which the lesions were less than
20 mm in diameter. They used a 5-trocar technique and kept
the intra-abdominal pressure for the pneumoperitoneum
less than 4 mm Hg. The well-encapsulated tumors were
completely excised; they were placed in a plastic bag and
removed through the 10-mm trocar site. All had favorable
Tumor
Adrenal
blood supply
Tumor
Duodenum
Ureter
Le
ft i
Right iliac
vein
lia
ca
rte
ry
ry
rte
ht
Rig
a
iac
il
C
FIGURE 31-12 A, Lower retroperitoneal paraspinal neuroblastoma and its relationship to the bifurcation of the aorta and ureter. B, Tumor may extend
into the vertebral foramina. C, Photograph of the operative field after resection of a right-sided pelvic neuroblastoma. Note the vascular loops placed
around the iliac arteries, right iliac vein, and ureter to facilitate a safe dissection.
CHAPTER 31
NEUROBLASTOMA
455
Rib space
Esophagus
Azygos vein
Latissimus
dorsi m.
Sympathetic
chain
Intercostal vessels
Tumor
extensions
C
FIGURE 31-13 A, Right posterolateral thoracotomy incision used for the excision of a posterior mediastinal neuroblastoma. B, Relationship of the tumor
to surrounding tissues. C, The tumor is mobilized and retracted anteriorly, exposing numerous intervertebral extensions. The tumor extensions are divided
at the vertebral foramina, leaving small remnants of residual tumor behind. This does not adversely influence the outcome.
histology, and none had MYCN amplification. No recurrences

were observed. This and other reports suggest that, in selected
cases, laparoscopic biopsy and tumor excision are both safe
and effective.181183
Mediastinal tumors are usually approached through a standard posterolateral thoracotomy incision. Excision of the
pleura and the endothoracic fascia around the tumor usually
allows entry into an appropriate plane of dissection. Mobilization of the tumor from the rib edges is accomplished with both
sharp and blunt dissection. It is important to identify and
either ligate or clip intercostal blood vessels feeding and draining the tumor. The tumor may be attached to a number of
sympathetic ganglia and intercostal nerves and often extends,
in one or more areas, into the intervertebral foramina
(Fig. 31-13).56,57,171,184 It may be impossible to remove every
bit of tumor at the foraminal sites. Small primary tumors have
been successfully removed by thoracoscopic techniques.
Thoracoscopy is also useful in obtaining tissue for biopsy.
In patients with neurologic symptoms (including paraplegia) associated with dumbbell tumors, prompt MRI and an urgent laminotomy to excise extradural tumor and relieve cord
compression are recommended before attempting intrathoracic resection of the tumor. The mediastinal resection can
be delayed a short time to allow the patients neurologic symptoms to improve. If extradural tumor is present on imaging
studies but the patient is asymptomatic, chemotherapy is initiated and may shrink the tumor and avoid the need for laminotomy or laminectomy at the time of resection of the thoracic
tumor. The choice of therapy for intraspinal tumor extension
is still somewhat controversial. Plantaz and colleagues59
reviewed 42 patients in France and recommended initial chemotherapy followed by surgical removal of residual disease.
Yiin and colleagues61 described 13 cases of neuroblastoma
with symptomatic spinal cord compression and neurologic

deficits. All the patients were treated initially with chemotherapy: 3 recovered, 4 improved, and 6 worsened and became
paraplegic. Two of the six recovered after laminectomy. The
authors recommended spinal cord decompression for patients
who have neurologic deterioration on chemotherapy.
Sandberg and colleagues60 described the treatment of 46
patients with epidural or neural foraminal tumor involvement.
Nine were low-risk patients with normal neurologic examinations who remained neurologically intact following operation
or chemotherapy. Four low-risk patients with high-grade
spinal cord compression improved or remained stable after
surgical intervention, but 2 patients who were treated
with chemotherapy had worsening deficits. Eleven of twelve
high-risk patients with normal neurologic examinations and
without radiographic high-grade spinal cord compression
were treated with chemotherapy and had no neurologic deterioration. Of 16 high-risk patients with high-grade spinal cord
compression, 7 of 10 were treated initially with chemotherapy,
and all 6 who underwent initial surgery improved or remained
stable. Spinal deformities occurred in 12.5% (2 of 16) treated
nonsurgically and in 30% (9 of 30) who underwent surgery.
The authors concluded that patients with high-risk tumors
and spinal involvement but normal neurologic examinations
should be offered chemotherapy, with the understanding
that a small percentage may require operations for progressive
neurologic deficits. Chemotherapy may be avoided in
patients with low-risk tumors who can be offered a potentially
curative procedure. Patients and their families should be
made aware that operative intervention may be associated with
subsequent spinal deformity in as many as 30% of cases.60
Cervical neuroblastoma is often localized and has a favorable outcome.185 In a study of 43 cervical neuroblastomas,
456
PART III
Haddad and colleagues186 identified four risk factors that

were associated with increased operative morbidity: adherence to vascular structures, tumor size, friability, and dumbbell tumors. Imaging studies may show a solid mass with
vascular displacement and narrowing.187 Tumors arising in
the neck or upper mediastinum often involve the stellate
ganglion. If not present preoperatively, resection may
result in postoperative Horner syndrome.14,188 This is a
relatively minor consequence outweighed by complete
tumor excision and survival, but the patients family
should be made aware of this possible complication. Special
attention should be given to protecting the brachial plexus
and the phrenic, vagus, and recurrent laryngeal nerves.
Aggressive surgical management is occasionally associated
with late complications in survivors, including ipsilateral atrophy of the kidney following adrenal resection and ejaculatory
problems following pelvic tumor excision.188 Of interest is the
very favorable outlook noted in patients with stage III and IV tumors arising in the pelvis following complete tumor resection.10
chemotherapy. Timed sequential administration of cell

cyclespecific and nonspecific drugs (cyclophosphamide
and doxorubicin, or cisplatin with teniposide or doxorubicin)
was subsequently tested and resulted in improved response
rates.71,192 This improvement led to more aggressive, more
intensive treatment protocols using multiple agents.
Currently, this patient population receives aggressive multimodality therapy, including induction chemotherapy to attain remission, followed by surgery and radiotherapy to
further achieve local control. This treatment is followed by
consolidation of remission with myeloablative therapy, autologous stem cell transplant, 13-cis-retinoic acid, and, possibly,
the addition of immunotherapy.
The purpose of induction chemotherapy is to reduce tumor
burden throughout the entire body, both at the primary site
and at sites of systemic disease. Multiple agents are often used,
including cyclophosphamide, ifosfamide, doxorubicin, cisplatin, carboplatin, etoposide, topotecan, and vincristine.
These agents are now given as dose-intensive regimens.
Chemotherapy
------------------------------------------------------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------------------------------------------------------
Although multiagent chemotherapy has significantly improved

the survival rate of patients with many different types of tumors
(e.g., Wilms tumor), chemotherapy has no such effect in
infants and children with resectable localized neuroblastoma
with favorable biologic and genetic characteristics. For
locoregional disease that does not have MYCN amplification,
surgical resection alone is all that is necessary.5,10,189
Patients with locoregional disease with poor prognostic biologic and genetic factors, however, are at higher risk and
should be treated more aggressively with multiagent doseintensive chemotherapy, in addition to a variety of therapies
thereafter. Patients with stage IV disease who are younger
than 18 months at diagnosis receive low-dose chemotherapy,
in addition to surgery.190
Patients with stage IV disease who are MYCN amplified or
older than 18 months at diagnosis remain the most difficult
population to treat. Historically, these patients received cyclophosphamide, vincristine, and dacarbazine.9,190,191 Patients
in whom this treatment regimen failed received doxorubicin
and teniposide (VM-26). Although these chemotherapy regimens did not effectively increase the cure rate of patients with
stage IV disease, such treatment reduced the size of the primary tumor, often cleared the bone marrow of tumor cells,
and was occasionally associated with histologic maturation
from malignant neuroblastoma to benign ganglioneuroma.9,11
Unfortunately, only 40% of patients with stage IV disease
demonstrated a complete response to chemotherapy; 30%
had a partial response; and 30% were unresponsive.9 When
the clinical estimation of response was subjected to confirmation by laparotomy, many patients thought to be
responders to chemotherapy actually had persistent
tumor not identified by preoperative testing.15,76,172
Numerous studies confirmed the limited effectiveness of
chemotherapy regimens in patients with metastatic disease.9,73 Using cell kinetic data, Hayes and colleagues192 demonstrated that the proliferating fraction of the tumor cell
population in neuroblastoma is exceedingly small. A large
pool of nonproliferating resting cells is resistant to
Radiotherapy
Neuroblastoma is a radiosensitive tumor, so radiotherapy remains an important part of the treatment regimen for patients
with neuroblastoma. In general, in the management of neuroblastoma, radiotherapy is administered after both induction
therapy and surgery, when minimal disease remains. It has
also been used for bulky metastatic disease after a response
is seen with chemotherapy, and for palliation in patients
with refractory end-stage disease or painful metastases.193195
At our institution, a regimen that includes dose-intensive chemotherapy, surgery, and a dose of 2100 cGy of hyperfractionated
radiotherapy to the primary site in patients with stage IV neuroblastoma resulted in a local control rate of greater than 90%.46
Although it is useful, external-beam radiotherapy is associated with considerable toxicity in growing children, resulting
in growth disturbance, bony deformity, endocrine deficiency,
hypoplastic soft tissue changes, skin atrophy, and, of greater
concern, secondary malignancies in the radiation portal. Techniques used to decrease radiation-induced toxicity include
hyperfractionation of the radiation dose, which usually does
not reduce the desired antitumor effect, and avoiding the simultaneous administration of chemotherapy agents that
may enhance the radiation effect. Brachytherapy and intraoperative radiotherapy can better confine the radiation
effect to the target tissue and spare surrounding normal tissues.196,197 Although early local control can be achieved, only
38% of patients with stage IV disease given intraoperative radiotherapy survived after 3 years. Some patients still require
supplemental external-beam radiotherapy; postoperative
ureteral stricture, renal artery stenosis, and neuropathies have
been described.5 Haas-Kogan and colleagues196 described an
experience using intraoperative radiotherapy in 23 cases of
high-risk neuroblastoma and noted that this technique was
effective only in patients who had gross total resection of
the primary tumor.196 All patients with partial tumor resection
had recurrence, despite radiotherapy, and subsequently died.
There are few data to support the efficacy of this therapy.
Intraoperative radiotherapy is sometimes cumbersome to perform, especially in institutions that do not have an operative
suite in the radiotherapy department or radiotherapy
CHAPTER 31
equipment (including linear accelerators) in the operating

room. Under these circumstances, after attempted tumor resection, the patient has to be transported under general
anesthesia for the radiation treatment.
Children with refractory advanced neuroblastoma with
widespread involvement often suffer severe pain due to metastases. Kang and colleagues194 employed targeted radiotherapy using submyeloablative doses of 131I-MIBG to achieve
disease palliation. The treatment stabilized disease, relieved
pain, or improved performance status, with 31% of patients
showing an objective response to treatment. They concluded
that this modality is useful for treating end-stage neuroblastoma. Deutsch and Tersak193 described the use of radiotherapy (300 to 1000 cGy) for palliative treatment of
symptomatic metastases to bone. The most common treatment sites were the skull, spine, hip, and femur. In their study,
29% of patients survived 1 year or longer (range, 1 to
52 months); only 8% survived more than 3 years.
Targeted radiotherapy with MIBG has also been used in
combination with myeloablative chemotherapy and proton
therapy for recurrent and refractory neuroblastoma. Proton
therapy has the advantage of delivering radiation more precisely
than conventional radiotherapy. As it becomes more widely
available, it may play a greater role in the management of
children with neuroblastoma requiring radiation treatment.1
Myeloablative Therapy
------------------------------------------------------------------------------------------------------------------------------------------------
Myeloablative therapy, using near-lethal doses of phenylalanine mustard (melphalan) with autologous bone marrow
rescue, has been shown to improve the tumor response rate.
A combination of melphalan, doxorubicin, teniposide, and
low-dose total-body irradiation followed by autologous
BMTresulted in a relapse-free rate of 40% and, if deaths resulting from toxicity were excluded, there was a 2-year survival
rate of 34% in patients with stage IV disease.198 An alternative
treatment was developed that relies on myeloablative chemotherapy using escalating doses of drugs given by constant
infusion, followed by autologous (purged) bone marrow
infusion but without total-body irradiation.199
Currently, stem cells are harvested during the induction
phase of treatment, and stored for later use during the consolidation phase of treatment.200 Chemotherapeutic agents used
for this phase of treatment include carboplatin, etoposide, and
melphalan. Peripheral blood stem cell infusion is used to reconstitute the marrow after myeloablative treatment. Immunoglobulin G levels are monitored and replaced with
gamma globulin. Sulfamethoxazole and fluconazole are given
prophylactically to avoid opportunistic Pneumocystis carinii
and fungal infection.
The use of purged, peripheral blood hematopoietic stem cells
has shown to have survival benefits over the use of allogeneic
cells, mainly because of a higher toxic death rate and an increased
incidence of graft-versus-host disease in the latter group.94,201
Patients with high-risk stage IV disease have a better outcome
with BMT, especially if they have amplification of the
MYCN oncogene.37,94,124 Preliminary reports are demonstrating improved event-free survival with rapid sequential
tandem transplant consolidation therapy. This has initiated
the ongoing COG phase III trial testing single versus tandem
transplant as consolidation therapy.201 However, to date, there
NEUROBLASTOMA
457
are no prospective studies examining the use of myeloablative therapy in addition to dose-intensive induction
therapy.190 At our institution, the addition of myeloablative
chemotherapy to dose-intensive chemotherapy did not
decrease the rate of systemic or central nervous system
recurrence.202
The addition of 13-cis-retinoic acid (isotretinoin), a biologic response modifier that causes tumor differentiation
and decreases bone marrow tumor involvement, was shown
to be useful.199 Retinoids serve as multistep modulators
of the MHC class I presentation pathway and sensitize
neuroblastomas to cytotoxic lymphocytes.202204 In a phase
III randomized trial in high-risk patients, Reynolds and
colleagues203,204 showed that high-dose pulse therapy with
13-cis-retinoic acid given after completion of intensive chemoradiation (with or without autologous BMT) significantly
improved event-free survival. A CCG phase III randomized
trial showed the use of 13-cis-retinoic acid after myeloablative
chemotherapy had superior event-free survival in patients
with remission, and this protocol is now commonly used
for these particular patients.199 A newly developed retinoid,
fenretinide, has completed a COG phase I trial, and different
oral formulations are being tested in the hopes of improving
its bioavailability.189,205
Immunotherapy
------------------------------------------------------------------------------------------------------------------------------------------------
In the 1970s and 1980s, it was suggested that tumor regression

in neuroblastoma involved an immunologic mechanism, resulting from an unusual tumorhost relationship.206208 Lymphocytes from children with neuroblastoma were observed to
inhibit colonies of neuroblasts in culture but not cells from
other tumors.207 Sera from patients with progressive disease
contain a blocking antibody that prevents a lymphocytemediated cytotoxic response and inhibits lymphocyte blastogenesis to phytohemagglutinins.209,210 Lymphocytes from
patients with neuroblastoma also have a decreased systemic
and in situ natural killer activity.211 In experimental studies,
operative electrocoagulation and hyperthermia resulting from
high-intensity focused ultrasonography induced immunity in
mice with neuroblastoma.206,207 A major problem is that
advanced neuroblastoma cells are MHC class I deficient and
evade immunorecognition.
The mainstay of current immunotherapy for neuroblastoma involves GD2. GD2 is a surface glycolipid antigen that
is copiously found on all neuroblastoma cells. Raffaghello
and colleagues212 employed an anti-GD2 antibody in nude
mice with neuroblastoma and noted increased long-term survival and decreased metastatic spread in a dose-dependent
manner in treated mice compared with controls. Cheung
and colleagues213 suggested that immunotherapy using
ganglioside GD2 monoclonal antibody should be directed at
minimal disease and must be used in conjunction with
dose-intensive chemotherapy to be effective. Kushner and colleagues214 described the treatment of seven patients who relapsed with widespread disease after initial treatment with
surgery alone for locoregional neuroblastoma. They received
dose-intensive chemotherapy; anti-GD2 3F8 antibody;
and targeted radiotherapy using 131I-labeled 3F8, if they
had assessable disease, or 3F8, granulocyte-macrophage
colony-stimulating factor, and 13-cis-retinoic acid, if they
458
PART III
were in remission. Five of the seven patients remained in remission between 4 and 8 years later.214 The same group
reported that high-dose cyclophosphamide, irinotecan, and
topotecan were effective in achieving remission and inducing
an immunologic state conducive to antibody-based passive
immunotherapy (using 3F8 antibody) in resistant neuroblastoma.215 Further studies have been completed that continue
to show improved outcomes with the use of this antibody.
In contrast to 3F8, which is a mouse-derived monoclonal antibody, ch14.18 is a chimeric human/murine anti-GD2 antibody,
and it has shown positive results in both phase I and II clinical
trials (German NB90 and NB97 studies). Preliminary results
from the COG randomized phase III trial using ch14.18 plus cytokines after autologous stem cell transplant versus control
showed improved survival in the treatment group.216
Dendritic cells are potential targets for immunotherapy.
They can enhance growth and differentiation of CD40activated B lymphocytes, directly affect natural killer cell
function, and act as antigen presenters.217,218 Redlinger and
colleagues218 noted that advanced neuroblastoma impairs
dendritic cell differentiation and function in adoptive immunotherapy. It has been shown that neuroblastoma-derived
gangliosides inhibit dendritic cell function. Interleukin-12
(IL-12) is a potent proinflammatory cytokine that enhances
the cytotoxic activity of T lymphocytes and resting natural
killer cells.217 In a murine model of neuroblastoma, Shimizu
and colleagues219 demonstrated that IL-12transduced dendritic cell vaccine (with an adenoviral vector expressing
IL-12) led to a complete and sustained antitumor response.
Tumor regression was associated with a high infiltration of
dendritic cells and viable T cells.
Additional Therapies
------------------------------------------------------------------------------------------------------------------------------------------------
131I-MIBG
infusion provides a way to specifically deliver radiotherapy to neuroblastoma cells, because it is taken up by more
than 90% of neuroblastomas. The use of this treatment alone
results in a response in 18% to 37% of patients with refractory
or relapsed disease.220 Rapamycin (mTOR) plays a significant
role in cell growth and persistence of neuroblastomas, and
phase II trials investigating rapamycin (mTOR) inhibitors are
ongoing. Early results reveal a benefit in patients with recurrent
neuroblastoma.221 Additionally, trials are examining targeted
therapies for the AKT pathway, as activation of AKT correlates
with worse event-free and overall survival.222224 Biologic
agents, such as histone deacetylase inhibitors, tyrosine kinase
inhibitors, IGF-1 receptor inhibitors, N-myc inhibitors, ALK
inhibitors, and various antiangiogenic agents, are also being investigated in ongoing clinical trials. The use of bisphosphonates
is being explored as a possible treatment for bone metastases.
Tumor vaccines and techniques of adoptive immunotherapy
are also being evaluated in clinical trials.
Summary and Future Directions

------------------------------------------------------------------------------------------------------------------------------------------------
This common pediatric malignancy remains an enigma

because of the high variability in tumor behavior. The primitive neuroblastic tumor may follow one of three possible pathways: spontaneous regression (apoptosis); differentiation and
maturation to a benign ganglioneuroma; or, more commonly,

tumor proliferation and rapid malignant progression. Fetal ultrasonography and infant screening programs have clearly
demonstrated that some tumors spontaneously regress.
Age-related histopathologic studies have clarified that some
tumors (those with favorable histology) differentiate and
mature, whereas others (those with unfavorable histology)
are undifferentiated neoplasms that respond poorly to treatment and have a rapidly progressive course and fatal outcome.
Recognition of important biologic (and genetic) characteristics can help to categorize these tumors into risk groups (low,
intermediate, and high) that determine future treatment protocols. Risk-based management permits individualized care
for each patient based on age, INSS stage, INPC histology,
and biologic and genetic characteristics that affect the behavior of each tumor.1,10,102 This avoids unnecessary and
potentially harmful treatment in patients categorized as having
low-risk tumors and who may do well with surgery alone (and
occasionally observation alone in highly selected cases).
It allows the physician to reserve the most intensive treatment
protocols for children with the highest-risk tumors and the
most guarded prognosis. At present, the outlook is best in
low-risk patients: infants younger than 18 months; patients
with localized tumors that can be completely excised (stages
I and II) with favorable INPC histology and low-risk biologic
and genetic factors; and infants with stage IV-S disease.
Infants with cystic or small solid neuroblastomas detected
on prenatal sonograms also have a very favorable outcome.
In patients with stage IV-S disease and those with cystic,
multifocal, or bilateral tumors and favorable biologic characteristics, observation alone may be feasible. Close sonographic monitoring of these cases in the first year of life is
important to ensure that the tumor shrinks and undergoes
regression. Increase in tumor size is an indication for
operative intervention.
Despite some improvements in outcome using highintensity treatments, the outlook for patients with advanced
neuroblastoma remains dismal, and less than half survive.
A better understanding of factors influencing tumor regression and differentiation and tumorhost immune interactions is required. In children with high-risk tumors,
identifying additional tumor markers and targeting effective
monoclonal antibodies against the tumor, developing improved techniques to clear the bone marrow of tumor cells,
using new and more effective chemotherapy regimens, and
using growth factors (e.g., other biologic tumor modulators)
to promote regression and differentiation may control disease progression and improve the outlook for this highly malignant tumor. Molecular profiling of the genetic changes that
occur in neuroblastoma will likely permit a more precise
classification system to predict outcome and further define
the choice of specific therapy, which may include targeting
the genes, proteins, and signaling pathways responsible for
malignant progression of the tumor.1 Also of concern are
the long-term effects of neuroblastoma treatment, which
include cardiac and renal toxicity, scoliosis, adverse effect
on growth and development, delayed sexual maturation,
learning disabilities, and occurrence of second neoplasms
including renal tumors.1,225,226
expertconsult.com.
malignant. Acute abdominal pain may be caused by bleeding

into the mass or into the peritoneum,2 particularly in hepatocellular adenomas, although this problem is rarely seen in
children. Children may present with congestive heart failure
(CHF) and thrombocytopenia, which is known as KasabachMerritt syndrome when associated with a vascular anomaly
such as a liver hemangioma.3 Cutaneous hemangiomas are
seen in about half the children with a liver hemangioma,4,5
and the rapid enlargement of the liver with a diffuse liver
hemangioma can cause abdominal compartment syndrome
and respiratory distress.4 CHF without significant thrombocytopenia can also be seen with liver arteriovenous
malformation (AVM)6 or mesenchymal hamartoma.7 Fetal
hydrops has been identified by prenatal ultrasonography
in some fetuses with liver hemangiomas8 or mesenchymal
hamartoma.9
Diagnosis
------------------------------------------------------------------------------------------------------------------------------------------------
LABORATORY TESTS
CHAPTER 32
Nonmalignant
Tumors of the Liver
Wolfgang Stehr and Philip C. Guzzetta, Jr.
Primary liver tumors constitute less than 3% of tumors seen in

the pediatric population, and only one third of those tumors
are benign.1 Benign tumors may be epithelial (focal nodular
hyperplasia, hepatocellular adenoma), mesenchymal (hepatic
hemangioma, mesenchymal hamartoma), or other (teratoma,
inflammatory pseudotumor). Nonparasitic cysts, although not
technically neoplasms, are also discussed in this chapter. One
of the more interesting aspects of benign liver tumors in
children is their predilection to occur in patients with other
conditions, and this phenomenon will be discussed with each
tumor type.
------------------------------------------------------------------------------------------------------------------------------------------------
Most children with benign liver tumors present with a painless

right upper quadrant abdominal mass or hepatomegaly.
Symptoms of gastrointestinal compression, such as constipation, anorexia, or vomiting, may also be present. If the mass is
painful, the pain is usually dull and aching and is caused by
expansion of the liver capsule or compression of the normal
surrounding structures. Jaundice and weight loss are uncommon except in infants with symptomatic hemangiomas, and
those signs should raise the suspicion that the lesion is
Serum alpha fetoprotein (AFP) is present in very high concentrations at birth (48,000 35,000 ng/mL) and rapidly
declines to adult levels of less than 10 ng/mL by 8 months of
age (Table 32-1).10 Thus in infants younger than 8 months,
AFP levels must be interpreted in the context of this dramatic
change. Markedly elevated AFP levels in a child with a liver
mass almost certainly means that the mass is malignant,
although milder elevation may be encountered with some
benign lesions, such as mesenchymal hamartoma11 or teratoma.12 As mentioned previously, significant thrombocytopenia
associated with a liver mass is usually part of the KasabachMerritt syndrome resulting from a liver hemangioma. Hypothyroidism may also occur in multiple or diffuse forms of liver
hemangioma13; thyroid function tests should be done routinely
in these children, because hypothyroidism significantly impacts
their management.14
IMAGING TECHNIQUES
The initial imaging study in a child presenting with an
abdominal mass should be a supine radiograph of the abdomen, looking for calcifications within the mass. The next imaging study should be an abdominal ultrasonogram with
Doppler spectral analysis, followed by computed tomography (CT) with intravenous contrast (Fig. 32-1).15 Magnetic
resonance imaging (MRI) may be indicated, depending on
the sonogram and CT scan results, especially when surgical
resection is planned and more detailed information about the
vascular anatomy relative to the tumor is desired or in infants
with hemangiomas in whom another diagnosis is being considered because the MRI appearance may be diagnostic.
Arteriography is reserved for children with a liver hemangioma, an AVM, or, rarely, a mesenchymal hamartoma with
CHF, when embolization of the blood supply to the tumor
is needed for treatment. The use of percutaneous biopsy
under sonogram or CT guidance in children with benign tumors is generally discouraged, unless excision of the tumor
would pose a major risk to the child, because establishing a
diagnosis on the basis of a small sample may be problematic
for the pathologist and because resection is the proper
459
460
PART III
TABLE 32-1
Normal Serum Alpha Fetoprotein (AFP) Levels of Infants by Age
Age
Premature
Newborn
Newborn to 2 weeks
2 weeks to 1 month
1 month
2 months
3 months
4 months
5 months
6 months
7 months
8 months
No. of Patients
AFP Level SD (ng/mL)
11
55
16
43
12
40
5
31
6
9
5
3
138,734 41,444
48,406 34,718
33,113 32,503
9452 12,610
2645 3080
323 278
88 87
74 56
46.5 19
12.5 9.8
9.7 ?7.1
8.5 5.5
From Wu JT, Book L, Sudar K: Serum alpha fetoprotein (AFP) levels in normal
infants. Pediatr Res 1981;5:50.
FOCAL LIVER HEMANGIOMA

Focal lesions vary in size but can be as large as 8 cm in diameter.5 They are usually asymptomatic. Some of the children
will have cutaneous hemangiomas as well. On MRI, there is
a solitary liver lesion that is hypodense on T1-weighted
sequences and hyperintense on T2-weighted sequences
compared with normal liver. CT scan similarly shows contrast
enhancing in the periphery of the mass with little contrast
in the center (see Fig. 32-1). These lesions seldom need
treatment, may be a hepatic form of hemangioma similar to
the cutaneous rapidly involuting congenital hemangioma
(RICH), and have generally regressed spontaneously by 1 year
of age.4,5
MULTIFOCAL LIVER HEMANGIOMA

Multifocal lesions are generally widely dispersed, spherical,
and homogeneously enhancing lesions on MRI. Flow voids
may be present in the lesions and may indicate the presence
of arteriovenous shunts that may lead to congestive heart
failure (CHF).15 Cutaneous hemangiomas are almost always
present in these children.5 Treatment by corticosteroids of
patients with CHF is highly successful,4,5 but if steroids fail
to control CHF, embolization of the shunts may be necessary.6,16 Children with this lesion may have evidence of hypothyroidism, and thyroid function tests (TFT) should be
obtained in all children with multifocal lesions. Prognosis is
excellent with 100% survival in one series.5
DIFFUSE LESIONS
FIGURE 32-1 Contrast-enhanced abdominal computed tomography

scan of a 4-day-old infant with a large focal hemangioma of the left hepatic
lobe. Note the central area of necrosis.
therapy for most of these tumors, with the exception of liver

hemangiomas. The findings on imaging studies are discussed
in the sections on each individual tumor.
Diffuse lesions frequently replace nearly all of the liver with

lesions showing centripetal enhancement on MRI or CT.
The clinical course is more complicated and potentially lethal.
Massive hepatomegaly may lead to abdominal compartment
syndrome, multisystem organ failure, and death. Severe hypothyroidism may develop because of overproduction of type III
iodothyronine deiodinase; therefore TFT must be obtained.13
Despite the large tumor burden, CHF is rare. If corticosteroid
therapy does not result in rapid improvement, then liver transplantation should be considered early,17 because the prognosis is otherwise poor.4 Medical therapy with vincristine has
shown some success,18 but this option is often limited by
the rapid clinical deterioration of children with diffuse lesions.
Interferon therapy has been abandoned because of the risk of
spastic diplegia in infants.19 Survival for all children with the
diffuse form of liver hemangioma is approximately 75%.5,13
ARTERIOVENOUS MALFORMATION
Hepatic Hemangioma
------------------------------------------------------------------------------------------------------------------------------------------------
Hepatic hemangiomas are the most common benign liver tumor in children and are more common than all other benign
liver tumors combined. Most of these lesions are identified in
the newborn period or during prenatal ultrasound screening.8
To facilitate discussion about treatment and prognosis, a
new subtype classification was delineated in 2007.4 This classification designates the hemangioma as either focal, multiple,
or diffuse and eliminates confusing terms such as infantile
hepatic hemangioendothelioma.
An AVM may occur within the liver parenchyma or outside the

liver between the hepatic artery and the portal venous system.
Similar to patients with diffuse liver hemangiomas, patients with
hepatoportal AVMs usually present before 6 months of age, many
in the newborn period, with hepatomegaly, CHF, and a bruit over
the liver.21 In older children and adults, hepatic AVM may occur
as part of hereditary hemorrhagic telangiectasia, also known as
Osler-Weber-Rendu disease.15,22 Angiography is diagnostic,
and embolization is therapeutic in some patients, but it is necessary to eliminate the extensive collaterals for successful closure of
the AVM.15,21 Fatal complications from the embolization of liver
CHAPTER 32
AVMs have been reported in adults.23 Steroids have no place in

the management of these lesions, and AVMs not managed
successfully with embolization may be controlled with ligation
of the hepatic artery.24,25
MESENCHYMAL HAMARTOMA
Mesenchymal hamartoma (MH) usually presents as a painless
right upper quadrant abdominal mass in a child younger than
2 years.20,26,27 Some patients may have evidence of CHF,7 and,
similar to liver hemangiomas, MH has been diagnosed prenatally.20,9 Edmondson28 proposed that MH arises from a mesenchymal rest that becomes isolated from the normal portal
triad architecture and differentiates independently. The tumor
grows along bile ducts and may incorporate normal liver tissue. Because the blood vessels and bile ducts are components
of the mesenchymal rest, the biologic behavior of the tumor
varies with the relative predominance of these tissues within
the loose connective tissue stroma (mesenchyma) that surrounds them. Thus the tumor may present as a predominantly
cystic structure (Fig. 32-2) that enlarges rapidly because of
fluid accumulation,29 or it may be predominantly vascular
and present with CHF.7 Von Schweinitz and colleagues30 suggested that fat-storing (Ito) cells of the immature liver may be
involved in the development of MH. There are reports of chromosomal translocations within mesenchymal hamartomas.31
Serum AFP levels are usually normal in children with MH,
but they may be mildly elevated.20,11,32 The radiographic features of these tumors are consistent and distinguishing; abdominal sonography and CT demonstrate a single, usually
large, fluid-filled mass with fine internal septations and no
calcifications.33
Management must be tempered by the understanding that
MH usually follows a benign course,34 although there have
been reports of malignant transformation.35,36 In general,
complete operative resection is the procedure of choice, if it
can be accomplished safely. Huge lesions or those that involve
NONMALIGNANT TUMORS OF THE LIVER
461
both lobes may be treated by unroofing and marsupializing the

cysts, although the lesion may recur after incomplete resection.
MH is an entity distinct from the liver hamartomas associated with tuberous sclerosis. The latter are smaller, multifocal
lesions that may be associated with angiomyolipomas in other
locations, such as the kidney; they are rarely symptomatic and
usually present in children older than 2 or 3 years. These
hamartomas have little clinical significance, but their presence
may be helpful in diagnosing tuberous sclerosis.37
HEPATOCELLULAR ADENOMA
Although isolated lesions are encountered in childhood, hepatocellular adenoma (HCA) is most commonly observed in
adults in association with the use of anabolic corticosteroids
or estrogen. HCA has been described in children treated
with anabolic steroids and multiple blood transfusions for
chronic anemia,38 and it is expected in children with type I
glycogen storage disease.39 Bianchi40 proposed several mechanisms for the development of HCA in patients with type I
glycogen storage disease, including (1) regional imbalance
in insulin and glucagon metabolism, because these hormones
are important in the regulation of hepatocyte proliferation
and regeneration; (2) response to glycogen overload; and
(3) oncogene activation. A giant hepatocellular adenoma has
also been reported in a child treated with oxcarbazepine for
a seizure disorder.41
Microscopic examination of adenomas reveals hepatocytes
in sheets and cords oriented along sinusoids without a ductal
component. The cells have glycogen-filled cytoplasm and
small nuclei without mitoses. Adjacent liver and vessels are
compressed but not invaded. Children usually do not have
coexisting cirrhosis.38 The histologic pattern is similar to that
of a well-differentiated hepatocellular carcinoma, and development of hepatocellular carcinoma within an unresected
HCA has been reported.42,43
In children, HCA generally presents as an asymptomatic
hepatic mass. The mass is solid on ultrasonography and CT.
Liver enzyme and AFP levels are normal. A feature unique
to this lesion is its propensity for intraperitoneal hemorrhage
from spontaneous rupture. In adults, intraperitoneal bleeding
is almost always seen in patients receiving estrogen therapy,
and tumor regression may occur with the cessation of hormone administration. In patients with glycogen storage disease and HCA, tumor regression may occur with the
correction of metabolic disturbances.40 Because of the known
association between HCA and hepatocellular carcinoma,
resection of HCA is recommended when it occurs in a child
who is not receiving steroids and does not have glycogen storage disease. If resection cannot be accomplished without substantial risk, observation of the lesion while monitoring the
serum AFP level may be appropriate. If the AFP level begins
to increase or the lesion is significantly symptomatic, and if
the risk of resection is unacceptably high, liver transplantation
may be the best alternative.
FOCAL NODULAR HYPERPLASIA
FIGURE 32-2 Cross section of a pathology specimen of a left hepatic

lobectomy for mesenchymal hamartoma in a 10-month-old male infant.
Focal nodular hyperplasia (FNH) in children presents as an

irregularly shaped, nontender liver mass. It is frequently found
incidentally at laparotomy for another cause or on radiographic
studies performed for another indication. The female-to-male
462
PART III
comparison with the levels seen with hepatoblastoma. Resection is the procedure of choice for a teratoma because of the
risk of malignancy in any immature elements of the tumor.
INFLAMMATORY PSEUDOTUMOR
Inflammatory pseudotumor of the liver is rare and generally
seen in children older than 3 years but has been reported in
younger children as well. Because this lesion is predominantly solid, it is difficult to differentiate it from other
benign or malignant tumors by imaging studies. Invariably,
the serum AFP level is normal. Fever, leukocytosis, and high
C-reactive protein level in a child with a solid liver mass and
normal AFP level are suggestive of an inflammatory pseudotumor of the liver thought to be an inflammatory reaction
to some insult, although the instigating cause is usually
unknown. It is difficult to diagnose this lesion without
a large biopsy. Most children undergo resection, which is
curative.52,53
NONPARASITIC CYSTS
FIGURE 32-3 Surgical view of focal nodular hyperplasia within the left
lobe of the liver in a 2-year-old child treated by left lateral lobectomy.
ratio for FNH is approximately 4:1.44 FNH is occasionally seen

with vascular malformations and hemangiomas in the liver,45 as
well as in children with type 1 glycogen storage disease,46 and it
has been postulated that the lesions represent an unusual response to injury or ischemia.28 On abdominal sonography,
the lesions may be isoechoic, hypoechoic, or hyperechoic compared with normal liver parenchyma, and multiple lesions may
occur in 10% to 15% of patients. The classic central scar may
not be seen on ultrasonography. CT typically shows a hypervascular lesion with a dense stellate central scar. Conventional arteriography or magnetic resonance angiography show a
hypervascular mass with feeding arteries entering the periphery
and converging on the central portion of the tumor. Some cases
of fibrolamellar hepatocellular carcinoma are radiographically
indistinguishable from FNH, which is a cause for concern if
the diagnosis is being made without a biopsy.47 There are
reports of adult patients who have FNH and hepatocellular
carcinoma simultaneously.48
On gross examination, the lesions are nonencapsulated,
occasionally pedunculated, and quite firm (Fig. 32-3).
Microscopic examination shows proliferation of hepatocytes
and bile ducts and the pathognomonic central fibrosis. These
lesions rarely become malignant or hemorrhage. Therefore
expectant therapy is appropriate when removal might be associated with significant morbidity, the child is asymptomatic,
and the diagnosis has been made conclusively by radiographic
studies, normal AFP levels, and biopsy.49
TERATOMA
There have been fewer than 25 case reports of hepatic teratoma in children invariably younger than 1 year.12,50 Calcification is usually present within the teratoma, helping to
differentiate it from other tumors. Some have met the criteria
for an intrahepatic fetus in fetu.51 Serum AFP levels may be
elevated with a teratoma, but only mildly elevated in
Nonparasitic cysts of the liver are rare and occur more commonly in adults than in children. Although they may be present and symptomatic at birth, most are asymptomatic and are
identified incidentally at autopsy or laparotomy. Symptoms
are related to abdominal distention or displacement of adjacent structures. Nonparasitic cysts occur with equal frequency
in males and females54 and are generally unilocular lined by
cuboidal or columnar epithelium characteristic of bile ducts.
The cyst fluid is typically clear or brown, and bile is rarely
present. Pathologic studies suggest that nonparasitic cysts
arise from congenital or secondary obstruction of peribiliary
glands. These glands normally arise from the ductal plate at
the hepatic hilum around the 7th week of gestation and continue to proliferate until adolescence.54 Symptomatic cysts can
be effectively treated by simple unroofing, marsupialization,55
or sclerotherapy.56 If biliary communication is suspected,
cholangiography may identify the source and allow the communicating ductule to be oversewn.
Cystic dilatation of the intrahepatic ducts may also present as
a mass, although jaundice and cholangitis are often associated
with this problem. Resection of the affected lobe is the preferred
therapy.57 If mesenchymal hamartoma appears to be completely
cystic on imaging, it may be misdiagnosed as a nonparasitic cyst.
Post-traumatic bile cysts result from ductal disruption and intrahepatic accumulation of bile. These lesions can be treated by
percutaneous drainage or, in some cases, by biliary sphincterotomy to reduce the bile duct pressure and lessen the biliary
leak.58 Resection is rarely necessary for post-traumatic cysts.
Multiple parenchymal cysts associated with hereditary polycystic kidney disease are generally asymptomatic and so small that
they do not require intervention.
Epidermoid cysts differ from other nonparasitic cysts, in
that the lining epithelium is squamous rather than cuboidal.
This histologic characteristic has led to the theory that these
lesions may be foregut bud anomalies trapped in the hepatic
substance. Although they are rare, there has been a report of
malignant degeneration. Thus resection is the appropriate
management.59
expertconsult.com.
CHAPTER 33
Malignant Liver
Tumors
Rebecka L. Meyers, Daniel C. Aronson,
and Arthur Zimmermann
Historical Context
------------------------------------------------------------------------------------------------------------------------------------------------
One hundred and thirteen years ago, the first case report of a
hepatoblastoma (HB) was published in the English literature
in 1898 by Misick in Prague.1 He reports A Case of Teratoma
Hepatis in a 6-week-old boy who died of respiratory problems. Autopsy showed a large tumor that occupied the lower
half of the right liver lobe. Cysts, cartilaginous, and bony deposits were seen, as well as venous tumor infiltration. It was
therefore not surprising that the tumor was described as a teratoma, with tissue representatives of the three embryonic
germ cell layers. More than 60 years later in 1962, Willis
introduced the term hepatoblastoma for this type of tumor
that he defined as an embryonic tumor that contains hepatic
epithelial parenchyma.2 At that time, hepatoblastoma usually
was not distinguished from hepatocellular carcinoma (HCC).
Through the work of Ishak and Glunz in 1967, morphologic
criteria were defined for HB and HCC that were refined in the
decennia that followed.3,4
Modern treatment dates to 1975 when Exelby published
a landmark paper that has been cited in most reviews dealing with liver tumors in children. He reports the results of
a survey of the American Academy of Pediatrics Surgical
Section documenting the 1974 treatment practices and outcomes for liver tumors in children.5 Through questionnaires
sent to the members of the Surgical Section of the American
Academy of Pediatrics (AAP), data on liver tumors in children
operated upon during the previous 10 years were requested.
From 110 replies, 375 liver tumors were reported, of which
252 were malignant (129 HB, 98 HCC), and 123 were benign.
All patients with HB underwent primary surgical exploration,
with biopsy only in 43 children and a subsequent attempt at
definitive resection in 86. Seventy-eight of the 86 children in
whom resection was attempted had complete excision of the
tumor and 45 (60% of those resected) survived. Excessive blood
loss was the most common complication during and immediately after operation, after which cardiac arrest occurred in
9 patients. There were 8 deaths in the operating room and
17 deaths in the immediate postoperative period attributable
to the operation. Fifteen HB patients had irradiation of the liver;
53 patients had chemotherapy using a wide variety of agents.
It was apparent that no cures were obtained from irradiation
and/or chemotherapy in the absence of complete surgical resection. The overall survival for HB was 35%; for HCC it was 13%.
With incomplete surgical excision no patient survived. There
was no evidence that radiation therapy or chemotherapy controlled disease that could not be completely excised surgically.
At this time, before the introduction of cisplatin-based chemotherapy and modern surgical techniques, it seemed that complete operative excision carried a high risk of morbidity, even
mortality, but offered the only chance of cure.
The field has progressed considerably since Exelbys 1975
survey. With the introduction of cisplatin-based chemotherapy
regimens in the 1980s, overall survival for HB increased from
35% to 70%6 and has increased further to nearly 80% in the
most recent trials.7 Although our sophistication with chemotherapy and antiangiogenic regimens for both HB and HCC
continues to evolve, the primary advantage of chemotherapy
has been in a neoadjuvant setting to shrink the tumor and
enable surgical resection. Although pediatric HCC is more
likely to respond to chemotherapy than its adult counterpart,
most HCC remains largely chemoresistant, and treatment
efforts often focus on slowing tumor progression with the
newest antiangiogenic agents. Complete surgical excision
remains the cornerstone for cure in both HB and HCC as
evidenced by the recent improvements in survival achieved
with complete hepatectomy and liver transplantation.810
Diagnosis
------------------------------------------------------------------------------------------------------------------------------------------------
CLINICAL PRESENTATION
Most liver tumors present with an asymptomatic abdominal
mass palpated either by a parent or pediatrician.11 In the
youngest children (infants and toddlers) the most common
malignant tumor is hepatoblastoma, which presents as an
asymptomatic right upper quadrant or epigastric abdominal
mass. Some children may have fatigue, fever, pain, anorexia,
and weight loss. Rarely, HB may present with abdominal pain
and hemorrhage after post-traumatic or spontaneous rupture of a previously occult tumor. Hepatocellular carcinoma
and hepatic sarcomas are more common in older children
and are more likely to present at an advanced stage. Nonspecific symptoms of inanition or respiratory failure may appear
insidiously. As the cancer grows, the pain in the abdomen may
463
464
PART III
progress to shoulder or back pain and becomes more pronounced. The child may develop progressive anorexia and
vomiting and appear thin and sickly. Tumor growth may compress or obstruct the normal hepatic architecture causing
(1) ascites secondary to occlusion of the portal or hepatic
veins, (2) gastrointestinal (GI) bleeding or splenomegaly from
the portal hypertension of portal vein occlusion, or (3) jaundice, scleral icterus, and pruritus from obstruction of the
biliary tree.12 Symptoms of biliary obstruction are most
common with biliary rhabdomyosarcoma.13
DIFFERENTIAL DIAGNOSIS
Differential diagnosis of a pediatric liver mass includes malignant tumors, benign tumors, and a wide assortment of congenital and acquired lesions of the liver, listed as other
masses in Table 33-1. For many of the other masses listed
in Table 33-1, the key to the diagnosis might lie in the underlying medical condition. For example, one might expect to see
a bacterial hepatic abscess in a child with chronic granulomatous disease, a fatty deposit in the liver of a child with hyperlipidemia, or perhaps an inspissated bile lake in a child with
biliary atresia as shown in Figure 33-1. Organizing intrahepatic hematoma should be suspected in any child with a history
of hepatic trauma or in newborns with sepsis and coagulopathy, especially if there is a history of perinatal birth trauma or
hemodynamic collapse requiring cardiopulmonary resuscitation. Congenital liver cysts are rare and represent a spectrum
ranging from large simple cysts, intrahepatic choledochal
cyst, and ciliated hepatic foregut cyst. Acquired cysts might
be due to a bacterial, hydatid, or amoebic abscess. A simple,
asymptomatic congenital liver cyst may be safely observed.14
If infectious or large and symptomatic, cyst drainage, marsupialization, or excision may be needed to relieve pain and
prevent risk of rupture. Recent literature suggests a risk of
squamous cell carcinoma arising later in life in those congenital hepatic cysts with a ciliated epithelial lining (ciliated
hepatic foregut cyst), and therefore these should probably
be excised rather than observed or marsupialized.15,16
Neoplastic liver masses, including benign and malignant
tumors, account for about 1.0% to 1.5% of all pediatric
tumors.17 Age at presentation is often the key to differential
diagnosis (Table 33-2).18 In newborns, the most common
tumor is infantile hepatic hemangioma.19 Infantile hepatic

hemangioma is to be distinguished from the much rarer
kaposiform hemangioendothelioma that may present in the
extremities, chest, or retroperitoneum. Kaposiform hemangioendothelioma of the retroperitoneum may present with
Kasabach-Merritt phenomenon and progress to obstruct the
porta hepatis.20 Hepatoblastoma is most commonly diagnosed
between 4 months and 4 years of age. Benign tumors in
toddlers are mesenchymal hamartoma and focal nodular
hyperplasia. Hepatocellular carcinoma and hepatic adenoma
are seen in older children. The other tumors listed in
Table 33-2 are rare. Although the most common benign
tumors often show classical distinguishing features on computed tomography, imaging is not usually a reliable way to
differentiate benign from malignant tumors.21
LABORATORY EVALUATION
Routine laboratory investigation should include complete blood
count; many children with a malignant liver tumor will exhibit
some degree of anemia and thrombocytosis.22 In HB, the thrombocytosis is thought to be caused by tumor production of
thrombopoietin, interleukin-6, and interleukin-1B.2325 Additional laboratory tests include a liver panel (albumin, transaminases, glutamyl transferase, alkaline phosphatase, total and
conjugated bilirubin), lactate dehydrogenase, tumor markers
(alpha-fetoprotein [AFP], beta-human chorionic gonadotropin
[beta-HCG], ferritin, carcinoembryonic antigen [CEA], catecholamines), and viral titers (hepatitis A, B, and C, Epstein-Barr
virus).18
The most important tumor marker is the serum AFP. AFP
will be elevated in 90% of children with hepatoblastoma and
in 50% of children with HCC.26 Although AFP is elevated in
most children with hepatoblastoma, increased AFP is not
pathognomonic for a malignant liver tumor. European,
German, and American multicenter trials have all concluded
that hepatoblastomas that fail to express AFP at diagnosis
(diagnosis AFP level less than 100) are biologically more
aggressive with a worse prognosis.2731 Rarely, the opposite
has been reporteda case of well-differentiated, fetal-type,
favorable prognosis hepatoblastoma that did not express
AFP.32 AFP levels must be interpreted with caution in neonates,
because AFP is the major protein produced by the fetal liver
TABLE 33-1
Differential Diagnosis of Pediatric Liver Masses
Malignant Tumors
Benign Tumors
Other Masses
Hepatoblastoma
Hepatocellular carcinoma
Sarcoma
Biliary rhabdomyosarcoma
Angiosarcoma
Rhabdoid
Undifferentiated
Metastatic/other
Wilms tumor
Neuroblastoma
Colorectal
Carcinoid tumor
Kaposiform hemangioendothelioma
Hemophagocytic lymphohistiocytosis
Langerhans cell histiocytosis
Megakaryoblastic leukemia
Mesenchymal hamartoma
Biliary cystadenoma
Focal nodular hyperplasia
Infantile hemangioma
Hepatic adenoma
Nodular regenerative hyperplasia
Teratoma
Inflammatory myofibroblastic tumor
Vascular malformations
Arteriovenous malformation
Blue rubber nevus syndrome
Congenital/acquired cysts
Simple
Ciliated foregut cyst
Polycystic liver disease
Choledochal cyst
Inspissated bile lake/biliary atresia
Parasitic cysts
Amoebic
Abscess
Bacterial
Chronic granulomatous disease
Hematoma
Fatty liver
CHAPTER 33
MALIGNANT LIVER TUMORS
465
FIGURE 33-1 Differential diagnosis: examples of non-neoplastic liver masses and cysts. A, Multiple small bacterial abscesses in a child with chronic
granulomatous disease. B, Inspissated bile lake in a child with biliary atresia and cholangitis. C, Organizing hematoma in a newborn with sepsis and
coagulopathy. D, Infarction of right lobe liver and hepatic abscess (with air fluid level) in a premature baby with necrotizing enterocolitis. E, Acquired cyst
is an amoebic abscess in a toddler with fever. F, Congenital cyst is a ciliated foregut cyst in an infant with abdominal distension and feeding difficulties.
TABLE 33-2
Age at Presentation, Most Common Liver Tumors of Childhood
Age Group
Malignant
Benign
Infant/toddler
Hepatoblastoma 43%
Rhabdoid tumor 1%
Malignant germ cell 1%
Hepatocellular (including transitional cell tumors) 23%
Sarcomas 7%
Hemangioma/vascular 14%
Mesenchymal hamartoma 6%
Teratoma 1%
Focal nodular hyperplasia 3%
Hepatic adenoma 1%
School age/adolescent
From Von Schweinitz D: Management of liver tumors in childhood. Semin Pediatr Surg 2006;15:17-24.
and is thus produced in high amounts in the normal newborn.

AFP may be especially high in neonates after hepatic damage
and during regeneration of liver parenchyma. The half-life of
AFP is 5 to 7 days, and levels fall throughout the first several
months of life so that by 1 year of age the AFP should be less
than 10 ng/mL.33 Moreover, there are many reports of benign
tumors, especially infantile hemangioma and mesenchymal
hamartoma, in children presenting with high AFP levels.3436
The other tumor markers useful in differential diagnosis are
beta-HCG elevated in germ cell tumors, ferritin elevated in
HCC and metastatic neuroblastoma; CEA elevated in HCC
and metastatic colorectal, lactate dehydrogenase elevated in
many malignant tumors, catecholamines elevated in metastatic neuroblastoma, hepatitis C in HCC, and Epstein-Barr
viral titers in lymphoproliferative disease or lymphoma.
RADIOLOGY
The radiographic appearance of the most common benign and
malignant liver tumors is shown in Figure 33-2. Mesenchymal
hamartoma is classically multicystic with the complex
cysts separated by thick vascular septae. Focal nodular
hyperplasia is generally well demarcated with a characteristic

central stellate scar. Infantile hemangioma classically will
demonstrate bright peripheral contrast enhancement. Infantile hepatic hemangioma may be focal, multifocal, or diffuse,
as shown in Figure 33-2 in its diffuse form. Hepatoblastoma
appears as a large multinodular expansile mass, usually unifocal, but occasionally multifocal. The tumor is generally
well demarcated from the normal liver but is not encapsulated. HB may invade hepatic veins, disseminate to the
lungs, or penetrate the liver capsule to reach contiguous
tissues. An initial ultrasonogram will identify the liver as
the organ of origin; additional testing, usually a contrastenhanced abdominal computed tomography (CT) scan, is
aimed at determining the extent of involved parenchyma
and the presence or absence of macrovascular compression,
displacement, or invasion. Metastatic liver tumors compared
with primary malignant liver tumors have been reported to
be more hypoechogenic on ultrasonography (US) and
have less vessel invasion and contrast enhancement on
abdominal CT.37
In hepatoblastoma and hepatocellular carcinoma, contrastenhanced abdominal CT or magnetic resonance imaging
466
PART III
FIGURE 33-2 Radiographic appearance of the most common hepatic benign and malignant neoplastic masses of the liver in children. A, Mesenchymal
hamartoma, a complex multicystic mass with solid septae. B, Focal nodular hyperplasia with arrow pointing to classic stellate central scar. C, Diffuse infantile
hepatic hemangioma with multiple nodules showing peripheral contrast enhancement. D, PRETEXT 2 hepatoblastoma. E, PRETEXT 4 P hepatocellular
carcinoma with involvement of main portal vein. F, Metastatic tumor, two nodules of metastatic colorectal carcinoma in right anterior and posterior sections.
(MRI) outlines the anatomic extent of the tumor, clarifies its

relationship to the central venous structures, and evaluates
for multicentricity.38 The radiographic appearance of the tumor at diagnosis is used to assign the tumor Pretreatment
Extent of tumor (PRETEXT) (Fig. 33-3). The radiographic
appearance of the tumor after preoperative (neoadjuvant) chemotherapy has been called Post-treatment Extent of tumor
(POST-TEXT).39 A chest CT scan is an essential part of the initial radiographic evaluation, to rule out metastatic pulmonary
disease. In children with HB, about 20% present with metastatic disease in the lungs. In HCC, the number of children
who present with advanced disease is quite high and pulmonary metastases at diagnosis have been reported as high as
50% in some series.40
Malignant Liver Tumors

------------------------------------------------------------------------------------------------------------------------------------------------
After neuroblastoma and Wilms tumor, primary tumors of the

liver are the third most common intra-abdominal neoplasms
in children.41 HB is the most frequent liver tumor in children
in Western countries, whereas in Asia and Africa, hepatocellular carcinoma (HCC) occurs more frequently than HB, probably as a consequence of the higher prevalence of hepatitis B
infection on those continents.42,43 Other less common malignant pediatric liver tumors are listed in Table 33-1.
HEPATOBLASTOMA
Epidemiology, Biology, and Genetics
Hepatoblastoma accounts for about 80% of the malignant
liver tumors in children.12,44 In the United States, the incidence of HB has increased from 0.6 to 1.2 cases per million
population in the last 2 decades.44 It comprises 1% of all
pediatric malignancies and affects mostly young children

between 6 months and 3 years old, but cases in neonates
and school-age children are also seen.
Researchers at the University of Minnesota are conducting a
large epidemiologic study, termed the HOPE study, aimed at
elucidating possible environmental and genetic risk factors
that might account for the increasing incidence of HB seen
over the past 2 decades.45 The HOPE study (hepatoblastoma
origins and pediatric epidemiology) can be reached at www.
cancer.umn.edu/hopestudy. A leading theory is that the increased incidence is due to the growing prevalence of premature birth and very-low-birth-weight (VLBW) babies. Both
prematurity and very low birth weight have been associated
with an increased risk for HB. The association between HB
and prematurity or VLBW was first shown in Japan and has
since been confirmed in multiple studies.4549 No association
has yet been found between prematurity as a risk factor and
the age at which the tumor diagnosis is eventually made or
the histologic subtype of the tumor. Unproven, but postulated, environmental risk factors include occupational exposure of the father to metals, such as welding and soldering
fumes, petroleum products, and paint.50 The list of possible
iatrogenic exposures of the premature or VLBW baby in the
neonatal care unit includes light, oxygen, irradiation, electromagnetic fields, plasticizers, medications, and total parenteral
nutrition.51
HB is also associated with fetal alcohol syndrome and
hemihyperplasia (formerly termed hemihypertrophy).52 Hemihyperplasia is associated with an increased risk of embryonal
tumors, primarily Wilms tumor and HB. Curiously, although
there is clinical overlap between hemihyperplasia and Beckwith-Weidemann syndrome, the genetic abnormalities seen
in HB patients with Beckwith-Weidemann syndrome are not
CHAPTER 33
POSTTEXT
= Extent of tumor after
neoadjuvant chemotherapy
...
...
...
...
467
PRETEXT (PRETreatment EXTent of Disease
PRETEXT
= Extent of tumor at diagnosis
I
II
III
IV
II
3 contiguous sections tumor free

no contiguous sections tumor free
III
b
a
In addition, any group may have:
V ... ingrowth vena cava, all 3 hepatic veins
P ... ingrowth portal vein, portal bifurcation
E ... extrahepatic
C ... caudate
M ... metastasis
IV
Diagram courtesy of SIOPEL liver tumor study group
FIGURE 33-3 PRETEXT.
seen in those with hemihyperplasia.53 In addition to BeckwithWeidemann syndrome, a number of other genetic syndromes
have been associated with an increased risk of HB, including familial adenomatous polyposis (FAP), Li-Fraumeni syndrome,
trisomy 18, and others as shown in Table 33-3.5468 Familial case
reports of HB with familial adenomatous polyposis are striking
and suggest a role in the pathogenesis of HB for chromosomes 5
and 11.56,69 Additional screening for cases in familial adenomatous polyposis kindred families is recommended by testing for
germline mutations in the adenomatous polyposis coli (APC) tumor suppressor gene.55,70 Germline APC mutations are not
commonly seen in children with sporadic HB.71 The association
between Beckwith-Weidemann syndrome and HB is so strong
that experts recommend that children with Beckwith-Weidemann syndrome be screened with abdominal ultrasonography
and AFP at regular intervals until they reach the age of 7 years.72
The genetic abnormality in HB patients with Beckwith-Weidemann syndrome is mapped to the 11p15.15 locus and suggests
the presence of a tumor suppressor gene at this location.73 Additional biological markers may include trisomy 2, 8, 18, 20, and
translocation of the NOTCH2 gene on chromosome 1q12-21.61
Up-regulation of insulin-like growth factor 2 may be mediated
by overexpression of PLGA1 oncogene, a transcriptional activator on the 8q chromosome.74
One of the most provocative genetic findings has been the
association between HB and mutations of beta-catenin and
activation of the WNT/beta-catenin signaling.7577 Microarray
analysis of WNT/beta-catenin and MYC signaling has defined
two tumor subclasses resembling distinct phases of liver
development and characterized by a discriminating 16-gene
signature. The highly proliferating tumor subclass showed
gains of chromosome 8q and 2p and up-regulated MYC signaling.78,79 Histologic subtypes of hepatoblastoma have also
been characterized by different patterns of WNT and NOTCH

pathway activation in DLK precursors.80 The authors speculate that HB may arise from proliferating bipotential precursors with WNT activation most prevalent in embryonal and
mixed histologic subtypes and NOTCH activation more prevalent in the differentiated pure fetal subtype.80 In addition, deregulation of MAPK signaling pathway and antiapoptotic
signaling is preferentially up-regulated in aggressive epithelial
HB with a small cell undifferentiated component.81 These
gene expression signatures may provide prognostic and diagnostic markers, perhaps even therapeutic targets, in the
future.80,81
Other genetic markers that have been associated with biological behavior include multidrug-resistance genes and the
Hedgehog pathway.8285 Increased expression of multidrugresistance genes is seen in response to chemotherapy in many
childhood tumors, and this seems to be particularly true in
HB.82 Chemotherapy has been shown to induce overexpression of the multidrug-resistance gene MDR1, MDR-associated
protein MRP1, and lung-related protein (LRP).84
Pathology
According to the World Health Organization (WHO) Tumor
Classification, hepatoblastoma is defined as a malignant tumor
with divergent patterns of differentiation, ranging from cells
resembling fetal epithelial hepatocytes, to embryonal cells,
and with differentiated tissues, including osteoid-like material, fibrous connective tissue, and striated muscle fibers. In
fact, the morphology of HB seems to reflect distinctive phases
of hepatogenesis, recapitulating cell lineages derived from endoderm fated to become mature liver cells.86 The neoplastic
offspring of these cell systems is present in HB in a variety
468
PART III
TABLE 33-3
Genetic Syndromes Associated with Pediatric Liver Tumors
Disease
Tumor
Chromosome
Gene
Reference
Familial adenomatous
polyposis (FAP)
Beckwith-Wiedemann
syndrome (BWS)
Li-Fraumeni syndrome
HB, HCC, adenoma
5q21.22
APC
HB, Infantile hemangioma
11p15.5
P57KiP2, WNT, others
17p13
TP53, others
Trisomy 18
HB, undifferentiated
sarcoma
HB
Thomas 200354
Hirschman 200555
Steenman 200056
Fukuzawa 200357
Fraumeni 196958
18
Other trisomies
Glycogen storage disease type I-IV
Hereditary tyrosinemia
HB
HB, HCC, adenoma
HCC
2, 8, 20
Several
15q23-25
Alagille syndrome
Progressive familial intrahepatic
cholestasis (PFIC)
Neurofibromatosis
HCC
HCC
20p12
18q21-22, 2q24

Fumarylaceto-acetate
hydrolase
JAG1
F1C1, BSEP
HCC, schwannoma,
angiosarcoma
HCC
17q11.2
NF-1
Kanai 199566
11q22-23
ATM
Geoffroy-Perez67
HCC, adenoma
1q42, 3p, 20q13
FAA, FAC
Touraine 199368
Ataxia telangiectasia
Hepatocellular carcinoma
Fanconi anemia
Bove 199659
Maruyama 200160
Tomlinson 200661
Siciliano 200062
Demers 200363
Keefe 199364
Alonso 199465
HB, hepatoblastoma; HCC, hepatocellular carcinoma.
of proportions, used as the basis of HB classifications, of

which the current International Society of Pediatric Oncology (epithelial) liver tumor study group (SIOPEL) classification is shown in Table 33-4. Untreated HB presents as a
lobulated mass up to more than 20 cm in diameter, being
solitary tumors in 80% of the patients and located to the
right lobe of the liver in about 60%. The lesions usually
show an expanding growth pattern, but conglomerated
masses with satellite nodules are also observed. The color
of the cut surfaces is variegated in many HB, partly caused
by necrosis and hemorrhage, with the exception of fetal HB,
which has the tan color of normal liver. The gross presentation of HB postchemotherapy is characterized by firm and
well-delineated and sometimes multinodular masses with
whitish fibrotic areas and calcifications.
Histologically, the epithelial components range in their
differentiation from a small cell undifferentiated (previously
termed anaplastic) phenotype, resembling other cellular
blue tumors, to cells that are close to mature hepatocytes (the
fetal phenotype). The current, histology-based classification
TABLE 33-4
Classification of Hepatoblastoma Histologic Subtype*
Hepatoblastoma, Wholly Epithelial Type
Fetal
Embryonal/mixed fetal and embryonal
Macrotrabecular (MT)
Small cell undifferentiated (SCU; formerly anaplastic)
Hepatoblastoma Mixed Epithelial and Mesenchymal Type (HB-MEM)
Without teratoid features
With teratoid features
Hepatoblastoma, Not Otherwise Specified (HB-NOS)
*This is the classification used by the SIOPEL (International Society of Pediatric
Oncology [epithelial] liver tumor study group). Classification systems used
by American, German, and Japanese study groups vary.
is not consistent regarding cellular differentiation, because

one subtype (macrotrabecular) reflects a growth pattern
rather than a distinct differentiation step. The fetal subtype,
occurring in a purely fetal and a so-called crowded fetal
variant, displays the highest level of differentiation. Pure
fetal histology HB is associated with both a diploid DNA
complement and a low proliferative activity. About 20% of
epithelial HB shows a mixture of fetal and less differentiated,
embryonal-type cells, with a more pronounced mitotic
activity. The macrotrabecular subtype (less than 5% of the
tumors) reveals a growth pattern with large cell plates
consisting either of fetal-embryonal or hepatocyte-like cells.
The latter variant, macrotrabecular type 1 (MT-1) is difficult
to distinguish from hepatocellular carcinoma and may have
an unfavorable biology.18 Undifferentiated HB mostly occurs
as a small cell neoplasm not associated with elevated serum
AFP (small cell undifferentiated [SCU] HB), but variants
with larger cells also occur. HB-SCU forms a complex group
of tumors in that at least part of the lesions seem to have
a relation to rhabdoid tumors and are INI1 protein
negative.87,88
A large proportion of HB (about 45% when examined after
chemotherapy) reveal a mixed epithelial and mesenchymal
(MEM) phenotype (HB-MEM; see Table 33-4). Osteoid-like
bone tissue is a common mesenchymal (heterologous) component. The same epithelial components as found in the wholly
epithelial HB subtypes occur in variable expression. The relative proportions of the components in HB-MEM undergo
marked changes subsequent to chemotherapy. After exposure
to chemotherapy, often the osteoid dominates the histologic
pattern. A small proportion of HB-MEM exhibit unusual tissues, such as glianeuronal, enteric, or melanocytic tissues.
These tumors are termed HB-MEM with teratoid features. It
has to be emphasized that this term is descriptive and does
not imply that these neoplasms are germ cell tumors. The
prognostic significance of these histologic types and subtypes
is currently under study in large trials.
CHAPTER 33
So far, a prognostic relevance has been worked out for the

fetal subtype (favorable)89 and for HB-SCU (unfavorable).30,31,90 An unfavorable histology of HB-SCU is also
present in cases where the SCU feature is expressed in a focal
pattern only.91 In addition to the lesions listed in Table 33-4,
an increased number of variants of HB, or lesions thought to
be related to HB, have been described, leading to the concept
of tumor families.92
Right
posterior
section
(VI and VII)
469
Right
anterior
Left medial Left lateral
section
section
section
(V and VIII)
(IV)
(II and II)
VIII
IV
VII
PRETEXT, STAGING, AND RISK GROUP

STRATIFICATION
II
I
III
Risk Group stratification determines treatment for hepatoblastoma (HB) in current multicenter group trials. As shown
in Table 33-5, The Childrens Oncology Group (COG) has
low, intermediate, and high-risk treatment groups, whereas
SIOPEL defines a standard-risk and a high-risk group. COG
continues to use traditional COG (Evans) stage I to IV, and
prognostic factors (pure fetal and small cell undifferentiated
histology and AFP < 100) to assign risk groups. Although
COG does not currently use PRETEXT to assign risk group,
it does use PRETEXT to define surgical guidelines, where it
determines whether or not a tumor should be resected at
diagnosis. The timing of resection will determine the tumor
stage for all nonmetastatic tumors. In contrast, SIOPEL uses
PRETEXT and prognostic factors to define risk groups.
PRETEXT (see Fig. 33-3) was devised by SIOPEL 1.93 Subsequent SIOPEL trials (SIOPEL 2, and SIOPEL 3) have used
PRETEXT as a tool to stratify treatment, define risk categories,
and report outcomes in HB. Although the risk stratification
schema differs somewhat between groups, the three other major multicenter pediatric liver tumor study groups, Childrens
Oncology Group (COG), German Pediatric Oncology Hematology (GPOH), and the Japanese Pediatric Liver Tumor (JPLT)
have all chosen to adopt PRETEXT in their current and future
protocols. Although PRETEXT has been found to have a slight
tendency to overstage patients, it is postulated to show good
interobserver agreement (reproducibility.) PRETEXT may also
be used to monitor the effect of preoperative therapy when it is
applied serially to assess tumor response to neoadjuvant
chemotherapy.28 In North America, COG uses PRETEXT to
define surgical resectability (i.e., surgical resection guidelines)
V
VI
Umbilical fissure/ligamentum teres
Right hepatic vein
Middle hepatic vein
Boundaries of each section defined by the right

and middle hepatic veins, and umbilical fissure
FIGURE 33-4 PRETEXT is distinct from Couinaud 8segments (I to VIII)
anatomic division of the liver. PRETEXT defines four sections. Boundaries
of each section defined by the right and middle hepatic veins, and umbilical fissure.
in its current HB protocol (AHEP 0731). Building upon the

Couinaud 8segment anatomic structure of the liver, the PRETEXT system divides the liver into four parts, called sections
(Fig. 33-4). The left lobe of the liver consists of a lateral
(Couinaud segments II and III) and medial section (segment
IV), whereas the right lobe is divided into an anterior (segments V and VIII) and posterior section (segments VI and
VII). Couinaud segment I is the caudate lobe and when
involved is shown in PRETEXT with the annotation C.
As shown by the examples in Figure 33-5, the tumor is classified into one of the following four PRETEXT groups depending on the number of liver sections that are free of tumor:
PRETEXT I, three adjacent sections free of tumor; PRETEXT
II, two adjacent sections free of tumor (or one section in each
hemiliver); PRETEXT III, one section free of tumor (or two
sections in one hemiliver and one nonadjacent section in
the other hemiliver); and PRETEXT IV, no tumor free sections.
Extrahepatic growth and gross vascular involvement is
TABLE 33-5
Hepatoblastoma Staging and Risk Stratification
Traditional COG (Evans) Staging System
Current COG Risk Stratification
Stage I: complete gross resection at diagnosis with clear margins
Very low risk: pure fetal histology, resected

at diagnosis (stage I/II); see resection
guidelines below*
Low risk: any histology resected at
diagnosis (stage I/II); see resection
guidelines below*
Intermediate risk: stage III tumors (includes
SCU histology)
Stage II: complete gross resection at diagnosis with microscopic

residual disease at the margins of resection
Stage III: biopsy only at diagnosis, or gross total resection with
nodal involvement or tumor spill or incomplete resection with
gross intrahepatic disease
Stage IV: metastatic disease at diagnosis
High risk: stage IV tumors, AFP < 100 at

diagnosis
Current SIOPEL Risk Stratification
Standard risk: PRETEXT I, II, III
High risk: PRETEXT IV, metastasis at

diagnosis, SCU histology, AFP < 100
at diagnosis
*COG surgical guidelines recommend resection of tumors at diagnosis (stage I/II) based upon PRETEXT. PRETEXT I and PRETEXT II resected at diagnosis if there is an
anticipated greater than 1-cm surgical margin based upon preoperative imaging.
AFP, alpha-fetoprotein; COG, Childrens Oncology Group; PRETEXT, pretreatment extent (of tumor) staging system; SCU, small cell undifferentiated; SIOPEL,
International Society of Pediatric Oncology (epithelial) liver tumor study group.
470
PART III
+V
+V
FIGURE 33-5 Examples of PRETEXT for hepatoblastoma risk stratification. A, PRETEXT I, left lateral section. B, PRETEXT II, right anterior and right posterior
sections. C, PRETEXT III, V: left lateral section, left medial section, and right anterior section with invasion of all three hepatic veins (V). D, PRETEXT IV, V,
P: tumor involves all four sections and invades vena cava and portal bifurcation.
indicated by adding one or more of the following: V, vena cava

or all three hepatic veins involved; P, main portal or both portal
branches involved; C, involvement of the caudate lobe; E, extrahepatic contiguous growth (e.g., diaphragm or stomach),
and M, distant metastases (mostly lungs, otherwise specify).39
Treatment Strategy, Chemotherapy, and Surgery
In the treatment of hepatoblastoma, complete surgical resection remains the cornerstone of curative therapy. And yet, as
has become increasingly clear in recent large multicenter trials, surgery alone cannot cure patients who present with
advanced disease. More than half of the patients present with
either an initial unresectable tumor or with distant metastases.
In the early years when these children were treated with surgery alone, there was a 30% relapse rate in those patients
whose tumor could be completely resected. Evidence that HB
is a chemosensitive tumor began to accumulate in the early
1970s when responses were seen to combinations of cyclophosphamide, vincristine, 5-fluorouracil, and actinomycin-D,94 but
not until the introduction of cisplatin and doxorubicin
containing regimens in the 1980s was there a major impact
of chemotherapy on survival.6 Twenty years later, cisplatin remains the backbone of current chemotherapy regimen. In fact,
in the most recent study of standard-risk HB by SIOPEL,
SIOPEL 3, treatment results with cisplatin monotherapy were
comparable to those achieved with cisplatin/doxorubicin
combination chemotherapy (PLADO).7 Chemotherapy may
reduce tumor volume, making the tumor resectable, and
may lead to the complete disappearance of lung metastases.
The tumor response rate to the present cisplatin-containing
chemotherapy regimens varies from 70% to 90%, according
to the different series.7,31,95100 Neoadjuvant (preoperative)
chemotherapy not only makes the tumor smaller and consequently more likely to be completely resected, but also more
solid, less prone to bleeding, and better demarcated from the
remaining healthy liver parenchyma.101,102 Also, when chemotherapy is given as soon as possible after diagnosis, occult
(micro)metastases in the lung have no delay in treatment. No
matter how small the primary tumor, SIOPEL recommends
preoperative chemotherapy in ALL patients as shown in
Figure 33-6. This approach is hypothesized to increase the
number of patients for whom complete surgical resection will
be feasible, to reduce the surgical morbidity of resection, and to
provide more time for making definitive surgical plans, including liver transplantation when indicated.103,104 Because of
the large number of countries participating in the SIOPEL
studies, standardization of both sophisticated surgical approaches and supportive care measures has been difficult;
therefore the use of preoperative chemotherapy in every case
has permitted patients from countries with limited resources
to participate in these studies.
In contrast to the SIOPEL approach, the North American
legacy groups Childrens Cancer Group (CCG) and Pediatric
Oncology Group (POG) (now COG) and the German Study
Group (GPOH) have historically recommended primary surgery, whenever prudently possible, as the initial treatment. The
decision about which tumors are resectable, and which ones
are not, has been subjectively made by the treating surgeon;
hence, the approach has been criticized for being highly
variable. Because traditional Evans staging relies on the surgical resection decision at diagnosis (see Table 33-5), and
because this is a surgeon-initiated subjective decision, the
stage has often depended more on the surgeon than on the
tumor. Figure 33-6 shows the North American strategy in
COG study AHEP0731, which recommends tumor resection
CHAPTER 33
471
CT definitively infantile hemangioma?

Yes
Observe if
asymptomatc
Benign
Treatment based
on tumor type
No,
Biopsy
Neoadjuvant
chemotherapy
+ V,P,E,M
Resection
(Transplant POSTTEXT III+V+PM
or POSTTEXT IVM)
Metastatic
workup
Adjuvant
chemotherapy
+ V,P,E,M
Hepatoblastoma (HB)
Hepatocellular carcinoma (HCC)
PRETEXT
I or II
(COG)
Resection
at diagnosis
POSTTEXT II, III
Resection
PRETEXT
III, IV, or
(COG)
Neoadjuvant
chemotherapy
PRETEXT
I, II, III, IV
(SIOPEL/GPOH)
Neoadjuvant
chemotherapy
POSTTEXT III+V+PM
POSTTEXT IVM
Transplant
POSTTEXT I, II, III
Resection
POSTTEXT III+V+PM
POSTTEXT IVM
Transplant
PRETEXT = Extent of disease at diagnosis

POSTTEXT = Extent of disease after neoadjuvant chemotherapy
FIGURE 33-6 Pediatric malignant liver tumor: simplified treatment algorithm. COG, Childrens Oncology Group; CT, computed tomography; E, extrahepatic contiguous growth; GPOH, German Pediatric Oncology Hematology (study); M, metastasis; P, main portal or both portal branches involved; SIOPEL,
International Society of Pediatric Oncology International Society of Pediatric Oncology (epithelial) liver tumor study group; V, vena cava or all three hepatic
veins involved.
at diagnosis dictated by PRETEXT-defined surgical guidelines

(PRETEXT I and II). Tumors can be resected by straightforward segmentectomy or lobectomy, and are resected at
diagnosis. Following these guidelines, approximately one
third of HB patients can successfully achieve a gross total
resection of tumor at diagnosis, and among them it is possible
to identify some that require minimal or no chemotherapy.89,105 Although it has been debated, postsurgical complications do not appear to be more frequent with this approach
in the modern era.30,96,106,107 The potential to reduce cumulative chemotherapy exposure with upfront resection in PRETEXT I and II tumors is important given the ability of HB to
develop resistance to standard chemotherapy.32,82,84 Recent
data on magnitude of AFP response actually suggest that the
majority of chemotherapy tumor kill probably occurs in the
first two cycles.104
The strategy in the German trials HB 89 and HB 94 was
similar to that used in North America, that is, resection at
diagnosis, when feasible, at the discretion of the operating
surgeon. In a review of these studies, 30% of children with
primary tumor resection had macroscopic or microscopic
residual tumor.31 Despite the larger number of advanced
HB in the neoadjuvant chemotherapy group, an incomplete
tumor resection was performed in only 18%. Based upon this
statistically significant difference, GPOH adopted neoadjuvant
chemotherapy for all patients in their latest trial, HB 99, and
recommends against any surgical consideration of atypical,
nonanatomic, or wedge resection.18,98
Surgical guidelines in the current COG trial, AHEP-0731
do not leave the decision about surgical resection at diagnosis
up to the subjective discretion of the individual surgeon; objective resection guidelines are part of the protocol. PRETEXT
is used to define which tumors should be resected at diagnosis
(see Fig. 33-6). Resection at diagnosis is recommended for
stage I/II only when segmentectomy or a facile, nonextended
lobectomy will predictably yield a complete resection, that is,

PRETEXT I or II tumors with at least 1 cm of clear margin
anticipated upon review of diagnostic radiographic imaging.
A POSTTEXT IIId tumor is a central tumor that may be best
resected by mesohepatectomy in the hands of an experienced
liver surgeon (Fig. 33-7). Transplantation is preferred in any
tumor that invades the major vascular inflow (POSTTEXT
P) or outflow (POSTTEXT V). If the PRETEXT (and POSTTEXT) suggests the need for major vascular reconstruction,
which is sometimes called extreme resection (a resection
performed in a patient who would otherwise meet criteria
for liver transplantation) or liver transplantation, a referral
for transplantation evaluation is advisable. Accuracy of
PRETEXT is moderate (because of difficulty in differentiating
tumor vessel compression from vessel ingrowth) with a slight
tendency to overstage.28 Nevertheless, good interobserver
agreement has been reported, and comparing PRETEXT
with POST-TEXT allows for an objective analysis of tumor
response to chemotherapy.39 The predictive value for survival
using PRETEXT is excellent, and combining PRETEXT
with traditional COG staging yields additional predictive
value.28,30
Postoperative (adjuvant) chemotherapy is currently recommended by all study groups, for all patients with one small
exception. Stage I pure fetal histology (PFH) in INT-0098
and COG P9645 received reduced or no chemotherapy89,96,97
and has a 5-year event-free survival (EFS) and overall survival
(OS) of 100% and 100%, respectively. Thus no chemotherapy is recommended for pure fetal histology patients resected
at diagnosis in COG AHEP-0731. Cisplatin remains thebackbone of the chemotherapy regimen, but the drug combinations differ somewhat between study groups. COG
currently uses cisplatin/5-FU/vincristine (C5V) for low-risk
tumors, C5V doxorubicin for intermediate risk, and will
investigate new agents (irinotecan) with upfront window
472
PART III
FIGURE 33-7 Central PRETEXT IIId hepatoblastoma: resection with mesohepatectomy versus complete hepatectomy and transplantation, depends upon
extent of macroscopic vessel involvement. A, Central hepatoblastoma involving left medial and right anterior sections (PRETEXT IIId). B, Resection with right
or left trisegmentectomy would leave very little residual normal liver. C, Much of the normal liver can be saved by mesohepatectomy if the portal vessels
are not encased and a good margin can be obtained. If tumor encases or invades the portal vessels, complete hepatectomy and transplantation is recommended. D, Frozen section shows negative margins.
therapy in high-risk tumors.42 SIOPEL 3 compared cisplatin

monotherapy with PLADO for standard risk95 and
used SUPERPLADO for high risk.7 The current SIOPEL 4
high-risk study uses an intensified platinum regimen.99
The recent GPOH trial HB-94 used IPA (ifosfamide/cisplatin/doxorubicin),31 and the ongoing GPOH trial HB99
uses IPA for standard risk and carboplatin-VP-16 for high
risk.98 The recent Japanese trial JPLT-2 has used CITA
(cisplatin/THP-Adriamycin [doxorubicin]) for standard risk
and ITEC (ifosfamide/carboplatin/doxorubicin/etoposide)
HACE (hepatic artery chemoembolization) for high-risk
patients.108 Irinotecan, with or without doxorubicin, has
been used in both North America and Europe for patients
with relapse.109,110
In terms of overall survival rates, the results of the different
study groups are generally comparable, projecting 3-year
overall survival rates, regardless of the first therapeutic
modality used, of 62% to 70% (Table 33-6).7,31,95100 The improved results in the high-risk group achieved in SIOPEL 3
highlight some important lessons learned over the past 2
decades. (1) With standard treatment, about 25% of patients
who present with metastatic disease are ultimately cured, and
alternative chemotherapy and surgical resection of pulmonary
metastatic disease should be considered in patients who
do not show an excellent early response to chemotherapy.
(2) The presence of a positive microscopic margin may not
portend a poor prognosis in patients who have had an excellent
response to chemotherapy. (3) Liver transplantation or extreme
resection (i.e., mesohepatectomy and major venous resection
and reconstruction) should be considered in every child
with unresectable HB (about 15% of cases).9,10,102,111113
The current COG trial, AHEP-0731, is a risk-stratified
study that seeks to diminish toxicity in low-risk patients,
increase survival in intermediate-risk patients, and identify
new agents(s) in high-risk patients.107 Very-low-risk patients
with pure fetal histology (PFH) hepatoblastoma resected at

diagnosis receive no chemotherapy. Low-risk patients who
have non-PFH histology resected at diagnosis receive two
adjuvant cycles of cisplatin, 5-flouorouracil, and vincristine
(C5V), a reduction from the standard four cycles of chemotherapy used in previous COG trials. For intermediate-risk
patients with stage I SCU, stage II SCU, or any stage III hepatoblastoma the chemotherapy regimen will add doxorubicin
to the C5V therapy (C5VD). High-risk patients with metastatic
tumor or initial AFP less than 100 ng/mL will be treated with
an upfront window of a novel agent (irinotecan) preceding the
backbone therapy with C5VD.
Liver Transplantation for Hepatoblastoma
In 1968, Starzl reported the first long-term survivor of liver
transplantation, a child with hepatoma. From that time until
the cluster of papers published by Al-Qabandi, Reyes, Pimpalwar, Molmenti, and Srivastin from 1999 to 2002,114118 most descriptions of the use of transplantation in hepatoblastoma were
anecdotal case reports. Largely because of early negative experience with liver transplantation in the treatment of adult hepatocellular carcinoma, liver transplantation for the treatment of
hepatic malignancy developed a reputation as a dreaded, last
resort, heroic, and even potentially ethically inappropriate intervention. The biology of pediatric hepatoblastoma has proven to
be very different from that of adult hepatocellular carcinoma,
with cisplatin-based chemotherapy proven to be of significant
value in a number of randomized trials. This availability of effective chemotherapy led credence to the bold statement by Reyes
in his landmark paper in 2000 115 that in these children with
unresectable tumors, the historical barrier of unresectability
can be redefined with the concept of total liver resection and
salvage orthotopic liver transplantation (OLT). Thus beginning
about 2000, liver transplantation began to be offered to more and
CHAPTER 33
473
TABLE 33-6
Summary Results Recent Hepatoblastoma Cooperative Trials
Study
Chemotherapy
No. of Patients
Outcomes
INT0098 (CCSG,
POG)96
C5V vs. CDDP/DOXO
Stage I/II: 50
Stage III: 83
Stage IV: 40
P9645 (COG)97
C5V vs. CDDP/CARBO
Stage I/II: pending publication

Stage III: 38
Stage IV: 50
HB94 (GPOH)31
Stage I/II: IFOS/CDDP/DOXO

Stage III/IV: IFOS/CDDP/DOXO
VP/CARBO
SR: IPA
HR: CARBO/VP16
Stage I: 27; II: 3; III: 25; IV: 14
SIOPEL 295
SR: PLADO
HR: CDDP/CARBO/DOXO
PRETEXT: I: 6; II: 36; III: 25; IV: 21; Mets: 25
SIOPEL 37,99
SR: CDDP vs. PLADO

HR: SUPERPLADO
JPLT-1100
Stage I/II: CDDP (30)/THPA-DOXO

Stage III/IV: CDDP (60)/THPADOXO
SR: PRETEXT I: 18; II: 133; III: 104

HR: PRETEXT IV: 74; VPE: 70; mets: 70;
AFP < 100: 12
Stage I: 9; II: 32; IIIa: 48; IIIb: 25; IV: 20
4-Year EFS/OS
Stage I/II: 88%/100% vs. 96%/96%
Stage III: 60%/68% vs. 68%/71%
Stage IV: 14%/33% vs. 37%/42%
1-Year EFS*
Stage III/IV C5V: 51%; CDDP/CARBO: 37%
*Study closed early due to inferior results
CDDP/CARBO arm
4-Year EFS/OS
Stage I: 89%/96%; II: 100%/100%; III: 68%/76%;
IV: 21%/36%
3-Year EFS/OS
SR: 90%/88%
HR: 52%/55%
3-Year EFS/OS
SR: 73%/91%
HR: IV: 48%/61%
HR mets: 36%/44%
3-Year EFS/OS
SR: CDDP 83%/95%; PLADO 85%/93%
HR: overall 65%/69%; mets 57%/63%
5-Year EFS/OS
Stage I: ?/100%; II: ?/76%; IIa: ?/50%; IIIb:
?/64%; IV: ?/77%
HB99 (GPOH)98
SR: 58
HR: 42
C5V, cisplatin; CARBO, carboplatin; fluorouracil and vincristine; CCSG, Childrens Cancer Study Group; CDDP, cisplatin; COG, Childrens Oncology Group; DOXO,
doxorubicin; EFS, event-free survival; GPOH, German Pediatric Oncology Hematology; JPLT, Japanese Pediatric Liver Tumor (study); IFOS, ifosfamide; HR, high
risk; IPA, ifosfamide, cisplatin, Adriamycin; mets, metastatic disease; OS, overall survival; POG, Pediatric Oncology Group; PRETEXT, pretreatment extent (of tumor)
staging system; SIOPEL, International Society of Pediatric Oncology (epithelial) liver tumor study group; SR, standard risk; SUPERPLADO, CDDP/CARBO/DOXO;
THPA, THP-adriamycin; VP, etoposide; VPE mets, Vena Cava, Portal vein, Extrahepatic metastatic disease.
more children as part of a planned treatment algorithm. With

increased experience defining the optimal timing of transplantation, the outcomes with liver transplantation for hepatoblastoma
have blossomed.
Transplantation Outcomes for Hepatoblastoma In the past
decade, more than a score of reports have appeared in the
literature championing the potential role of liver transplantation
in the treatment of unresectable pediatric hepatoblastoma
(Table 33-7).113130 Transplantation, although potentially lifesaving, carries attendant consequences, including perioperative
morbidity and mortality and the subsequent need for lifetime
immunosuppression. The experience from Birmingham, United
Kingdom illustrates contemporary experience, with 5-year
disease-free survival of 100% when primary transplantation
was performed in patients with a good response to chemotherapy, 60% after primary transplantation in patients with a poor
response to chemotherapy, only 50% in patients with transplantation as a second option or rescue transplantation, and 0% in
patients not undergoing surgery.116 In SIOPEL 1, overall survival at 10 years was 85% with a primary transplantation but
only 40% for the children who underwent a rescue transplantation.120 In a collaborative report of the world experience of
liver transplantation for hepatoblastoma,120 the overall survival
rate at 6 years was 82% for 106 patients who received a primary
transplantation, but only 30% for 41 patients who underwent
a rescue transplantation.
Indications and Contraindications for Transplantation in
Hepatoblastoma The following criteria are currently used
by COG and SIOPEL to select potential candidates for
transplantation: (1) multifocal PRETEXT IV, multifocal tumor

in all four liver sections at diagnosis; (2) unifocal PRETEXT
IV, with neoadjuvant chemotherapy often these tumors will
downstage to a POST-TEXT III and become amenable to
conventional resection by trisegmentectomy; (3) POSTTEXT
IIIV, proximity of the tumor to the vena cava or all three
major hepatic veins makes adequate tumor clearance without
impaired venous outflow doubtful; (4) POSTTEXT IIIP,
proximity of the tumor to the portal venous bifurcation
or both major branches of the portal vein makes adequate
tumor clearance without impaired portal venous inflow
doubtful; (5) Intrahepatic relapse or residual tumor after
previous attempt at resection, known as rescue transplantation. Although these guidelines are very useful, some uncertainty and controversy remains regarding the management
of multifocal tumors, patients with venous involvement
who might be candidates for extreme resection, patients
who present with pulmonary metastasis, and patients who
are referred with relapse or residual tumor and require rescue
transplantation.
Transplantation for Multifocal Hepatoblastoma Both
COG and SIOPEL currently recommend that all patients
with multifocal PRETEXT IV tumors should undergo liver
transplantation, even if one of the liver sections is apparently
clear of tumor nodules after preoperative chemotherapy
(Fig. 33-8). Microscopic foci of viable tumor are seen in
explant livers despite the apparent radiographic disappearance of tumor nodules from these areas after preoperative
chemotherapy.131 In addition, multiple series have shown
excellent results from primary transplantation and poor
474
PART III
TABLE 33-7
Contemporary Outcomes Transplantation for Hepatoblastoma
114
Al-Qabandi et al, J Pediatr Surg, Birmingham, UK

Reyes et al, 2000, J Pediatr, Pittsburgh, Pa115
Pimpalwar et al, 2002, J Pediatr Surg, Birmingham, UK116
Molmlenti et al, 2002, Am J Transplant, Dallas, Tex117
Sinivasan et al, 2002, Transplantation, London, UK118
Chardon et al, 2002, Transplantation, Paris/Brussels113
Cillo et al, 2003, Transplant Proc, Padua, Italy119
Otte et al, 2004, Pediatr Blood Cancer,120
SIOPEL 1 World Experience
Primary transplantation
Rescue transplantation
Tiao et al, 2005, J Pediatr, Cincinnati, Ohio111
Mejia et al, 2005, Clin Transplant, San Antonio, Tex121
Kasahara et al, 2005, Am J Transplant, Kyoto122
Chen et al, 2006, J Pediatr Gastroent Nutr, St Louis123
Avila et al, 2006, Eur J Pediatr Surg, Madrid124
Austin et al, 2006, J Pediatr Surg, UNOS database125
Cassas-Medley et al, 2007, J Pediatr Surg, Dupont, Del126
Beaunoyer et al, 2007, Pediatr Transplant, Stanford, Calif127
Faraj et al, 2008, Liver Transplant, London, UK128
Browne et al, 2008, J Pediatr Surg, Chicago, 129
Kalicinski et al, 2008, Ann Transplant, Warsaw130
No. of Patients
Survival (%)
Follow-up (years)
8
12
12
9
13
4
7
75
83
83
55
85
75
57
0.1-15.4
0.1-9.2
0.5-16
0.1-9
1.1-2
0.2-9
106
41
9
10
14
7
11
135
8
15
25
14
6
82
30
80
70
71
85
82
69
75
86
78
71
66
3.7-18
3.5 ?
0.6-18
1-14
0.6-4.4
3.3 3.5
0.9-14.9
3.8 ?
SIOPEL, International Society of Pediatric Oncology (epithelial) liver tumor study group; UNOS, United Network for Organ Sharing.
FIGURE 33-8 PRETEXT IV multifocal hepatoblastoma. In the presence of extensive multifocal tumors, microscopic satellites should be assumed, and no
distance of surgical margin can ensure complete surgical excision. Extensive multifocal tumors are best treated by complete hepatectomy and liver
transplantation.
results from rescue transplantation.116,120,124,126,129 In a recent series from Padova, predictors of failed conservative therapy included multifocality.132
Major Venous Involvement: Transplantation versus
Extreme Resection Aggressive resections, less than total
hepatectomy, may be successful in select patients with tumor
encroachment on the vena cava or main portal vein. Complex
extensive resection, with vascular reconstruction if necessary,
depends upon surgical expertise and a careful evaluation of
the degree of vascular involvement and realistic ability to
achieve complete, and safe, resection.133,134 Poorly planned
or executed operations risk excessive bleeding, inflow or outflow vascular obstruction, biliary leakage or stricture, cholangitis, and/or hepatic failure. Although most agree that extreme
resection of tumors without liver transplantation will avoid the
need for long-term immunosuppression,102,113,132,133,135137
outcomes with these techniques have not been rigorously
reported. Current recommendations for referral of high-risk
patients with hepatoblastoma to centers that have the ability

to do an extreme resection, with liver transplantation as
an immediately available safety net, should result in an
improved ability to compare the outcomes of these two
approaches.*
Transplantation for Hepatoblastoma with Pulmonary
Metastasis at Diagnosis An absolute contraindication to
liver transplantation is persistent pulmonary metastases
non-responsive to neoadjuvant chemotherapy and not amenable to surgical resection. Stable or progressive disease in the
face of neoadjuvant chemotherapy is a relative contraindication to transplantation.10,133 Lung metastases that disappear
completely with chemotherapy, or with a combination of chemotherapy and surgical resection, do not pose a contraindication, yet the risk of post-transplantation pulmonary relapse is
substantial, and therefore the use of liver transplantation for
children with metastatic disease remains controversial.
* References 9,10,112,131,137,138
CHAPTER 33
475
TABLE 33-8
Liver Transplantation in Children with Hepatoblastoma and Pulmonary Metastasis at Diagnosis, Review of Literature (Meyers and Otte,9 2010)
Pulmonary Metastasis at
Diagnosis
Lung lesions disappeared
with chemotherapy
Pretransplantation
pulmonary metastasectomy
TOTAL
No. of
Patients
Post-transplantation
Pulmonary Relapse
Alive without
Evidence of Tumor
Died of
Other Causes
24
9 (38%)
14 (58%)
1 (4%)
3 (38%)
5 (62%)
32
12 (37%)
19 (60%)
1 (3%)
*Patients listed in table have been separately reported in the following series during the past 10 years: Perilongo, 200495; Casanova, 200999; Schnater, 2002101; AlQabandi et al, 1999114; Reyes et al, 2000115; Avila et al, 2006124; Cassas-Medley, 2007126; Superina et al, 1996139; Nathan 2009140; Otte, 2009.141
Table 33-8 shows the accumulated cases in the literature

and presented at national and international meetings during
the past 10 years.* Overall survival appears to be about
60% with no large difference in outcome when there is lung
metastasis with complete radiographic resolution on chemotherapy versus pulmonary metastasectomy. Some centers do
AFP imaging pretransplantation, PET-CT pretransplantation,
median sternotomy with manual palpation of both lungs pretransplantation, and lobectomy rather than metastasectomy
if the lung has more than four nodules in the same lobe.142
Rescue Transplantation for Local Relapse Hepatoblastoma
Multiple series have shown superior outcome after primary
transplantation (about 80% overall survival) when compared
with rescue transplantation (about 30% to 40% overall survival).116,120,124,126,129 The basis for this is undoubtedly multifactorial, but two important reasons are (1) the likelihood of
chemotherapy resistance in relapsed tumors and (2) the debilitated state of the patients when transplanted in the face of
end-stage disease.
Type of Allograft and Immunosuppression There is a trend
to improved survival of children receiving a live-donor liver
transplantation (LDLT).120,121,128 When a living donor is
available, pretransplantation chemotherapy can be scheduled
optimally, with a rapid decision towards transplantation.121
Whether living donor grafts might require less immunosuppression as suggested by Gras,143 or whether alternative
immunosuppression using rapamycin (sirolimus), a drug
with both antineoplastic and immunosuppressive properties,
will have any impact in children with hepatoblastoma
remains to be seen. Many worry that the toxicity of chemotherapy might be potentiated by immunosuppression, but this
has not been the experience at high-volume centers. With such
a small number of patients in each of the individual
series reported to date, it is not possible to make a clear
recommendation at this time.
Pediatric Liver Unresectable Tumor Observatory
(PLUTO) SIOPEL, together with support from COG,
GPOH, and the Study of Pediatric Liver Transplantation
(SPLIT), has established a worldwide electronic registry for
liver transplantation for childhood liver tumors (hepatoblastoma, hepatocellular carcinoma, and diffuse infantile hemangioma).9,10,141,144 The link to obtain a password to register
patients on this database can be accessed through the PLUTO
Registry website: http://pluto.cineca.org/access.
* References 95, 99, 101, 114, 115, 124, 126, 139141.
New Agents and Treatment Modalities

Hepatic
Arterial
Chemoembolization
(HACE),
Transarterial Chemoembolization (TACE) HACE and
TACE are different acronyms for the same interventional
radiologic procedure, also sometimes referred to as transcatheter arterial chemoembolization. This technique continues to
be quite popular in China where recent experience in infants
and children showed a mean tumor shrinkage of 59%, mean
decrease in AFP of 60%, mean tumor necrosis in the surgical
specimens of 87%.145 Widespread use has been somewhat
limited by toxicity, which includes fever, pain, nausea, vomiting, transient coagulopathy, and, most worrisome, pulmonary
oil (Lipiodol) embolism.145147 Pulmonary oil embolism is
infrequent, and although fatalities have been reported, the
clinical course is usually self-limited oxygen desaturation
for 24 to 48 hours and pulmonary infiltrate for about a
week.148 Chemotherapeutic cocktails have included various
combinations of cisplatin, doxorubicin, doxorubicin-eluting
beads, vincristine, pirarubicin, mitomycin, and Lipiodol,
followed by gelatin foam particles or stainless steel coils,
and radioactive microspheres.131,145148 There are scattered
case reports of cure without the need for surgical resection,149
although it is most often used not as a definitive treatment, but
rather as palliation for large unresectable tumors in the
presence of uncontrolled metastatic disease.145
Ototoxicity Both SIOPEL and COG have put considerable
effort into investigations trying to decrease the significant
ototoxicity induced by the use of cisplatin-based chemotherapy in young patients, especially infants. The risk of cisplatin
causing bilateral moderate to severe high-frequency hearing
loss is significantly increased in children younger than 5 years
of age.150 The COG 9645 trial failed to reduce ototoxicity with
the agent amifostine.151 The recently opened SIOPEL 6 study
will investigate sodium thiosulfate152 as an agent to decrease
the cisplatin-induced ototoxicity. The current COG trial,
AHEP0731, attempts to reduce ototoxicity by limiting the
extended use of cisplatin in the low-risk patients.
Hepatoblastoma Risk Stratification and International
Collaboration Current data suggest that pure fetal histology
and PRETEXT I and II tumors have a favorable prognosis.28,30,89 Risk factors that seem to portend a worse outcome
include metastatic disease at diagnosis (COG Stage IV, PRETEXT M), PRETEXT IV, AFP < 100 at diagnosis, small cell
undifferentiated histology and possibly macrotrabecular and/
or extensive multifocal histology.18,30,90,153 In 2007, SIOPEL,
GPOH, and COG decided to embark on a mutual project that
476
PART III
was called the Childhood Hepatic Tumors International

Collaboration (CHIC). The complete databases of these groups
are in the process of being united to address prognostic questions requiring increased statistical power. To identify these
common data points for prognostication and risk stratification,
data regarding prognostic factors (i.e., histology AFP, stage, multifocality, biological markers) can thus be studied in much larger
patient groups in which the clinical outcome is known. These
developments are the starting point of a new trans-Atlantic
converging cooperation on a large intercontinental scale that
will be of eventual benefit for children with liver tumors.
New Agents, Tumor Relapse The prognosis for a patient
with recurrent or progressive hepatoblastoma depends on many
factors, including the site of recurrence, prior treatment, and
individual patient considerations. It was recently shown that
in patients who initially received only cisplatin/5-FU/
vincristine cure may be possible with a multidrug relapse
regimen including doxorubicin.109 Surgical resection of pulmonary relapse is possible and has been reported to produce
long-term cure, but does not carry as good a prognosis as
resection of pulmonary metastatic lesions present at diagnosis
that simply fail to completely resolve on chemotherapy.154156
If possible, isolated metastases should be resected completely
in patients whose primary tumor is controlled.138,142,157
Success with autologous peripheral blood stem cell transplantation with a double conditioning regimen has been reported
in a child with pulmonary relapse after liver transplantation.158
Irinotecan has been used in chemotherapy relapse regimens
with some success.159 In recurrent refractory disease, phase I
and II clinical trials may be appropriate and should be considered. Multidrug chemotherapy resistance is a key factor for the
poor outcome of relapsed HB. Novel gene-directed treatment
approaches, such as adenovirus-mediated cytosine deaminase/
5-fluorocystine suicide gene therapy, may offer hope for treatment of these chemotherapy-resistant tumors in the future.82,83
Information on current COG trials can be found at www.
childrensoncologygroup.org.
HEPATOCELLULAR CARCINOMA
Epidemiology, Biology, and Genetics
In Western countries, hepatocellular carcinoma occurs approximately half as often as hepatoblastoma (HB) or in 23% of all
primary pediatric liver tumor cases, most often in school-age
children and adolescents. Although described previously, it
was not until 1967 that childhood HCC was identified by Ishak
and Glunz3 as an entity to be distinguished from HB. In 1974,
Exelby and colleagues5 analyzed the clinical course of childhood HCC and found an overall dismal outcome.
HCC occurs predominantly in the setting of underlying liver
disease and cirrhosis. Compared with adults, in children cirrhosis is less commonly part of the antecedent process, while congenital or acquired disorders of the liver, such as metabolic
disease, are common.17 Table 33-9 shows the conditions that
are associated with HCC in children.43,66,160175 Patients with
tyrosinemia seem to be a particularly high risk and should be
vigilantly screened with serial AFP and imaging.174 In East Asia
and Africa, HCC is more common than HB because of the widespread prevalence of hepatitis B and C.43 In Taiwan, where HCC
is most often seen in carriers of the hepatitis B virus, vaccination
TABLE 33-9
Conditions Associated with Hepatocellular Carcinoma
in Children
Alpha-1 antitrypsin deficiency160
Anomalous abdominal venous drainage161
Alagille syndrome162
Biliary atresia163
Congenital hepatic fibrosis164
Familial polyposis/Gardner syndrome165
Focal nodular hyperplasia166
Hemochromatosis167
Hepatic adenoma168
Hepatitis B and C43
Glycogen storage disease (type I and III)169
Methotrexate therapy166
Neurofibromatosis66
Oral contraceptives170
Parenteral nutritionassociated liver disease (PNALD); total parenteral
nutrition (TPN) cholestatic liver failure171
Progressive familial intrahepatic cholestasis (PFIC)172,173
Tyrosinemia174
Wilms tumor/Bloom syndrome175
programs targeted against hepatitis have led to a significant

decrease in the incidence of HCC.42 In contrast to hepatitis B,
the cirrhosis and the subsequent development of HCC in
the hepatitis C population usually takes several decades to
develop.176 The genetic syndromes associated with HCC are
shown in Table 33-3.54,55,6268
Pathology
In the pediatric age group, more than two thirds of HCC occur
in children older than 10 years of age, but only 0.5% to 1% of
all HCC manifest before 20 years of age, and very few HCCs
are diagnosed in children less than 5 years old. About 20%
to 35% of children with HCC have underlying chronic liver
disease. It is still disputed whether classical (adult-type)
HCC in the pediatric age group is the same or a different
disease with respect to HCC in adult patients. It is currently
suggested that HCC forms a tumor family, consisting of
adult-type HCC and its variants, fibrolamellar HCC, and a
novel entity occurring in older children and young adolescents, transitional liver cell tumor (TLCT).92
HCC presents grossly as solitary or multiple (multifocal)
lesions. Solitary tumors display four main growth patterns,
that is, expanding (or pushing) mass lesions, pedunculated
(or hanging) lesions, invading tumors with poor delineation,
and multifocal tumors resembling metastatic disease. These
growth patterns exert a considerable influence on the surgical
resectability of the tumors. The color of the cut surfaces of
HCCs depends, apart from bleeding and necrosis, on differentiation features of the tumor cells, for instance bile synthesis
and accumulation.
The microscopic features of pediatric classical HCC are
similar to or the same as those in adult patients. Many tumors
exhibit a trabecular growth pattern with intervening sinusoidlike vascular channels and a reduced reticulin network.
Regarding grading, Edmondson and Steiner developed a
system comprising a scale of I to IV.177
CHAPTER 33
Fibrolamellar Hepatocellular Carcinoma (FL-HCC) This

tumor usually arises in noncirrhotic livers of adolescents or
young adult patients and is encountered more frequently in
Western countries.178 Overall, FL-HCC accounts for less than
10% of all HCCs. Recent data show that FL-HCC has biological features similar to that of adult-type HCC. FL-HCC
shows vascular invasion in up to 35% of cases, frequently
metastasizes into locoregional lymph nodes (about 50% of
cases), and tends to show unusual spreading patterns, including intraperitoneal spread. FL-HCC is typically a solitary
lesion that has a predilection for the left liver lobe (two thirds;
unusual for hepatic primary tumors). It reveals well-defined
margins and a central scar in 70%. The cut surface often shows
a firm, tan to brown tissue with radiating septa, sometimes
closely resembling focal nodular hyperplasia. The leading cell
is a large and polygonal, hepatocyte-like cell with a granular
cytoplasm of large vesicular nuclei. These cells form strands
embedded in the typical fibrosclerotic stroma that may form
a central stellate scar. A considerable proportion of the tumor
cells contain large, ground glasslike inclusions, the so-called
pale bodies, which are helpful in bioptic diagnosis. Periodic
acidSchiff (PAS)-positive globular inclusions in part contain
alpha-1-antitrypsin and other glycoproteins. Typically, cells of
FL-HCC show marked immunostaining for cytokeratin 7.92
Transitional Liver Cell Tumor (TLCT) Transitional liver
cell tumor is a recently identified liver neoplasm that occurs
in older children and young adolescents. The term transitional
had been proposed to denote a putative intermediate position
of the tumor cells between hepatoblasts and more mature
hepatocyte-like cells. TLCT are highly aggressive lesions that
have a treatment response pattern clearly different from hepatoblastoma.179 The usual presentation is that of a large or very
large solitary hepatic tumor (mostly in the right liver lobe),
commonly associated with very high serum AFP levels.
Grossly, the tumors display an expanding growth pattern
and sometimes exhibit a large central necrosis. Histologically,
the tumor cells vary between HCC-type cells and cells found
in hepatoblastoma, sometimes with formation of multinuclear
giant cells. The lesions markedly express beta-catenin,
typically in a mixed nuclear and cytoplasmic pattern.180
PRETEXT and Staging
Childrens Oncology Group staging for hepatocellular carcinoma does not use risk stratification and simply follows the
traditional COG stage I, II, III, and IV shown in Table 33-5.
Nevertheless, discussions with colleagues describing the
extent of tumor involvement of the liver are based upon
PRETEXT to aid in making key decisions about surgical
resectability.
Treatment Strategies
Hepatocellular carcinoma is relatively chemotherapy resistant
and therefore carries a poor prognosis with a dismal rate of
cure.181,182 Complete surgical resection or hepatectomy and
transplantation for tumor localized to the liver is often the only
hope. Unfortunately, HCC is most often advanced at diagnosis, and cure is rarely possible in the setting of metastatic
disease. Even with aggressive attempts at surgical resection,
tumor relapse is common and tumor-free survival rates of
not more than 25% to 30% can be achieved. These mostly
477
depend on the extent of disease, and the main prognostic

factor for childhood HCC is resectability. The first multicenter
clinical trials on pediatric liver tumors were conducted in
North America by the Childrens Cancer Study Group (CCSG)
and POG, some of which included HCC in addition to HB.181
These studies confirmed the poor response of HCC to chemotherapy and radiation and the dismal rate of cure in the
majority of patients.
The North American cooperative study (INT-0098) as well
as SIOPEL 1181,182 used pre-operative chemotherapy in an
attempt to increase surgical resectability for children and
adolescents with HCC, because this is the foundation for
curative therapy of liver tumors. Of the 46 patients entered
onto INT-0098, only 8 had completely resected tumors (stage
I) at study entry, 25 had unresectable tumors (stage III), and
13 presented with metastatic disease (stage IV). Patients were
randomized to receive cisplatin with either doxorubicin or
5-fluorouracil and vincristine. No differences were seen in
response or survival rates between the two treatment regimens. Seven of the 8 stage I patients (88%) with complete
tumor excision at time of diagnosis, followed by adjuvant
cisplatin-based chemotherapy, survived. This is a significant
improvement when compared with only 12 of 33 patients
(36%) treated before the consistent use of adjuvant chemotherapy. This result suggests that adjuvant chemotherapy
may be of benefit for patients with completely resected
HCC. However, because one third of these initially resected
patients have fared well without any additional chemotherapy,
the question of the necessity for adjuvant chemotherapy will
only be answered in a randomized trial. In contrast, outcome
was uniformly poor for patients with advanced-stage disease.
The 5-year event-free survival for stage III and IV patients was
23% and 10%, respectively (Table 33-10).
Hepatocellular carcinoma patients have been treated in
three consecutive studies of the German Society for Pediatric
Oncology and Hematology (see Table 33-10).26 In the first
study, HB89, neoadjuvant and adjuvant chemotherapy consisted of conventionally dosed ifosfamide, cisplatin, and doxorubicin (IPA), which did not show any substantial benefit.26
Of the registered 12 patients, only 4 with resectable tumor
survived. In the second study (HB94), patients with nonresectable HCC received conventionally dosed carboplatin
and etoposide in addition to IPA, which seemed to produce
at least short-term benefit.26 Of the registered 25 patients,
9 had locally unresectable and 11 metastatic HCC. Three of
the 9 and 1 of the 11 patients survived free of disease in
addition to 4 of 5 patients with resectable tumor (total 8 of
25 32%).
Results of SIOPEL 1, 2, and 3 are shown in Table 33-10.182,183
Only 2 of the 39 patients entered onto the SIOPEL-1 study
underwent complete resection of the tumor at diagnosis,
followed by chemotherapy, while the remaining 37 patients
had preoperative chemotherapy with cisplatin and doxorubicin. Metastases were identified in 31% of the patients, and
extrahepatic tumor extension, vascular invasion, or both in
39%. Although partial tumor response to chemotherapy was
observed in 49% (18 of 37) of the patients, complete tumor
resection was achieved in only 36% (14 of 39) of the
patients. Outcomes of patients on this study were also unsatisfactory, with a 5-year event-free survival of 17%. All
long-term survivors had complete surgical excision of their
tumor. Twenty-one patients were enrolled on the subsequent
478
PART III
TABLE 33-10
Summary Results Hepatocellular Carcinoma Cooperative Trials
Study
Chemotherapy
No. of Patients
Outcomes
INT0098 (CCSG, POG)
CDDP/DOXO
HB89 (GPOH)26
CDDP/DOXO
HB94 (GPOH)26
CDDP/DOXO
HB99 (GPOH)26
CDDP/DOXO
SIOPEL 1182
CDDP/DOXO
Stage I: 8
Stage II: 0
Stage III: 25
Stage IV: 13
Stage I/II/IIIa: 6
Stage IIIb, IV: 6
Stage: I/II/IIIa: 5
Stage IIIb, IV: 20
Stage: I/II/IIIa: 14
Stage IIIb, IV: 27
PRETEXT: I, 1; II, 14; III, 11; IV, 13, VPEM, 8
SIOPEL 2182
CDDP/DOXO
PRETEXT: I, 1; II, 3; III, 1; IV, 7; VPEM, 5
SIOPEL 3183
CDDP/DOXO
PRETEXT: I, 4; II, 22; III, 14; IV, 21; VPEM, ?
5-Year EFS/OS
Stage I/II: 88%/88%;
III: 8%/23%;
IV: 19%/34%
5-Year DFS
Stage I/II/IIIa: 50%; IIIb, IV: 17%
5-Year DFS
5-Year DFS
5-Year EFS/OS
17%/28%
5-Year EFS/OS
23%/23%
3-Year EFS/OS 10%/16%
181
CARBO, carboplatin; CCSG, Childrens Cancer Study Group; CDDP, cisplatin; DFS, disease-free survival; DOXO, doxorubicin; EFS, event-free survival; GPOH, German
Pediatric Oncology Hematology (study); IFOS, ifosfamide; IPA, ifosfamide, cisplatin, Adriamycin; OS, overall survival; POG, Pediatric Oncology Group; PRETEXT,
pretreatment extent (of tumor) staging system; SIOPEL, International Society of Pediatric Oncology (epithelial) liver tumor study group; VP, etoposide; VPEM,
Vena cava, Portal vein, Extrahepatic, Metastatic disease.
study SIOPEL 2. Data were available for 17 of these. One

patient died 17 days after diagnosis from massive GI bleeding and never received treatment. Thirteen of the 16 treated
patients received preoperative chemotherapy with cisplatin,
carboplatin, and doxorubicin. Partial response to preoperative chemotherapy was observed in 6 of 13 cases (46%).
Gross total tumor resection was achieved in 8 patients
(47%), 3 at the time of diagnosis and 1 through liver transplantation. Nine tumors (53%) never became operable. One
patient was lost to follow-up just before planned surgery.
Four of the patients having resection of their tumors were
alive at a median follow-up time of 53 months (range of
35 to 73 months). Twelve patients died because of progressive disease and one from surgical complications. The
three-year overall survival for this study was 22%.
In comparing the results of these studies, the outcome for
patients with HCC has shown no significant improvement,
despite the progress in surgical techniques, chemotherapy
delivery, and patient support. It seems obvious that a new
treatment approach is needed to increase the rate of cure of
childhood HCC.
In adults, fibrolamellar type of hepatocellular carcinoma
has been traditionally associated with a higher resection rate
and better survival when compared with the typical pathologic variant of HCC both in adolescents and young
adults.184,185 The higher resection rate for children and
adolescents with the fibrolamellar variant of HCC was not
supported by the studies reported by either Katzenstein181
or Czauderna.182 Patients with the fibrolamellar variant did
not have a better outcome when compared with those with
typical HCC, the 5-year event-free survival was 30% compared with 14%, respectively (P 0.18), although the median
survival was longer for patients with the fibrolamellar variant.
Given the poor response of HCC to chemotherapy and
radiation, the mainstay of treatment is surgery. This means
that, in contrast to hepatoblastoma, a primary radical tumor
resection has to be attempted whenever possible using all
available techniques in order to achieve this goal.26 Therefore
in school-age children and adolescents with a primary liver
tumor the surgeon has to be prepared to perform highly

sophisticated liver surgery after confirmation of the diagnosis
by pathologic investigation of intraoperative frozen sections.
Patients with the clinical constellation for advanced HCC
should always be treated in consultation with a specialized
center with experience in childhood liver surgery.
Liver Transplantation for Hepatocellular
Carcinoma in Children
Outcomes, Indications, and Contraindications Published
outcomes for liver transplantation in children with HCC are
shown in Table 33-11.* The following guiding principles have
been formulated by centers with particular expertise in pediatric liver transplantation. They are in a greater state of controversy and evolution than are the guidelines for HB. In most
centers, the criteria for transplantation of multifocal and
unifocal HCC are the same as for HB and do not follow adult
limitations on size and number of nodules. Unlike HB, however, any history of pulmonary metastatic disease or extrahepatic disease is considered an absolute contraindication.
Major vascular involvement, of the portal vein for example,
is a relative contraindication depending upon the degree
and severity of involvement.142 It is important that consultation with a transplantation center with special expertise in
pediatric liver surgery be considered early in the treatment
to prevent delays and unwanted extended courses of chemotherapy while awaiting resection and transplantation.
Response to Chemotherapy HCC tumor progression while
on chemotherapy is a relative contraindication to transplantation, because occult extrahepatic micrometastatic disease is
increasingly likely in this situation.
Milan Criteria The Milan criteria, introduced by Mazzaferro
in 1996, restrict transplantation in adults with HCC as follows:
(1) single tumor diameter less than 5 cm; (2) not more than
* References 115, 124, 125, 127, 130, 139, 182, 186194.
CHAPTER 33
479
TABLE 33-11
Literature, Transplantation for Pediatric Hepatocellular Carcinoma
No. of
Patients
{Olthoff et al, 1990, Arch Surg, UCLA
{Penn et al, 1991, Surgery, Transplant Registry187
Tagge et al, 1992, J Pediatr Surg, Pittsburgh, Pa188
Yandza et al, 1993, Transplant Int, Paris189
Broughan et al, 1994, J Pediatr Surg, multicenter190
Otte et al, 1996, Transplant Proc, Brussels191
Achilleos, et al 1996, J Pediatr Surg, Birmingham, UK192
Superina et al, 1996, J Pediatr Surg, Toronto139
Reyes et al, 2000, J Pediatr, Pittsburgh, Pa115
Tatekawa et al, 2001, J Pediatr Surg, Kyoto193
Czaudema et al, 2002, J Clin Oncol, SIOPEL 1182
Avila et al, 2006, Eur J Ped Surg, Madrid124
Austin et al, 2006, J Pediatr Surg, UNOS database125
Beaunoyer et al, 2007, Pediatr Transplant, Stanford,
Calif127
Kalicinski et al, 2008, Ann Transplant, Warsaw130
Ismail et al, 2009, Pediatr Transplant, Warsaw194
186
Survival (%)
Tumor Recurrence
Small Incidental*{
Died Comp OLT{
10
22

44
100
75
60
0
100
63
100

100
41
83
8/16
158/429
3/9
2/5
0/3
6/19
0

63%
1/10

31/429

0
0
3/3
7/19

12/41
4/10
4/16

1/9

0
0
0
2/12
0

2/10
8
11
75
72
1/8
1/11

3/11
1/8
2/11
16
429
9
2
4
5
2
3
19
2
2
1
*Most are patients with tyrosinemia, other metabolic liver disease, familial intrahepatic cholestasis, hepatitis, or biliary atresia.
{
Died as a result of complications of orthoptic liver transplantation.
{
Did not separately analyze pediatric cohort.
Comp, complications; OLT, orthotopic liver transplantation; SIOPEL, International Society of Pediatric Oncology (epithelial) liver tumor study group; UCLA,
University of CaliforniaLos Angeles; UNOS, United Network for Organ Sharing.
three foci of tumor, each one not exceeding 3 cm; (3) no angioinvasion; (4) no extrahepatic involvement. Since the introduction
of these criteria, long-term recurrence-free survival after liver
transplantation in adults with HCC improved from 30% to
75%.195198 The problem with the Milan criteria in children
is that 50% to 70% of children present with large de novo
tumors and a large tumor burden in otherwise healthy livers,
and the Milan criteria were developed in adults with small tumors
and underlying cirrhotic liver disease. In children, the number
of nodules, as stipulated by the Milan criteria, is usually not
considered a contraindication to transplantation as long as
the disease is confined to the liver. Furthermore, de novo pediatric HCC often shows features on a continuum with pediatric
HB, and these transitional liver tumors may have a more favorable biology.92200 In view of the lack of improvement in results
of conventional treatment of pediatric HCC during the past 2
decades, most clinicians treating pediatric HCC do NOT
recommend adherence to Milan criteria in children who present
with large de novo tumors, no cirrhosis, and no evidence of
extrahepatic disease.201
Metastatic Disease Metastatic disease is considered an
absolute contraindication to liver transplantation in HCC,
and a very careful and thorough evaluation to exclude metastatic microdeposits is essential.
Post-transplantation Chemotherapy Guidelines for posttransplantation immunosuppression in HCC are the same as
with transplantation for HB with one possible difference.
Many centers would consider post-transplantation adjuvant
antiangiogenic therapy with sorafenib in HCC. Experience
in the transplantation population of patients is limited, but
in any patient considered to be at high risk for tumor relapse,
options for possible antiangiogenic therapy should be
discussed. Similarly, many centers have begun to experiment

with rapamycin (sirolimus) as a post-transplantation immunosuppressant because of its antineoplastic, antiangiogenic
properties.203205
New Agents and Treatment Modalities
Antiangiogenesis, Sorafenib New treatment modalities
including metronomic chemotherapy,206 and adjuvant antiangiogenic therapy207 are the target of investigation based
upon some early promising results. Most promising has been
the recent adult experience with sorafenib, an antiangiogenic
tyrosine kinase inhibitor, where a survival advantage has
clearly been shown in prospective trials of sorafenib in the
treatment of HCC in adults with unresectable tumors.202 Interestingly, this seems to be also the case in some preliminary
investigation in childhood HCC.208
Chemoembolization and Theraspheres Hepatic arterial
chemoembolization (HACE) and transarterial chemoembolization (TACE) refers to the intra-arterial administration of chemotherapeutic and vascular occlusive agents (generally gelatin
or Lipiodol) along with cytotoxic drugs. The drugs most frequently used for chemoembolization are doxorubicin, mitomycin, and cisplatin. Intra-arterial injection of cytotoxic
agents results in higher local concentration of drugs with reduced systemic side effects, while the intra-arterial embolization causes ischemic necrosis of the tumor. This therapeutic
strategy has been used in a small number of children and adolescents with recurrent HCC while awaiting the availability of
a liver donor, or as adjuvant therapy in an attempt to facilitate
tumor resection.145,209 There are no large trials in children;
however, in a study of adult HCC patients without liver failure
or cirrhosis, although TACE successfully reduced tumor
growth, it frequently caused acute liver failure and did not
480
PART III
improve survival.210 A related approach that combines

radiation therapy with angiographic embolization has been
the intra-arterial injection of yttrium-90 radioactive microspheres, called Theraspheres.211
Portal Venous Embolization Portal venous embolization
has been used in adults with liver disease to induce hypertrophy of the remaining liver remnant212 and reported experimentally in children.213 The portal venous branch on the
side of the tumor is cannulated percutaneously, and polyvinyl
alcohol and coils are inserted to induce portal vein occlusion
under fluoroscopic control. This has a dual effect of alcohol
thrombosis of the embolized tumor and compensatory hypertrophy of the unharmed opposite liver lobe, increasing the
potential hepatic functional reserve in patients with cirrhosis
and underlying liver dysfunction in preparation for hepatic
resection.
Percutaneous Ablative Therapies Ablative percutaneous
methods of local control may be considered, especially in
recurrent tumors. They include percutaneous radiofrequency
ablation (RFA), percutaneous ethanol injection (PEI), and
cryotherapy. Cryotherapy refers to cold injury produced by
cryoprobe delivery of liquid nitrogen, and although once
popular in adults, it has now fallen out of favor because of
superior results achieved with RFA and PEI. In most cases,
these treatment approaches are palliative and are suitable
for smaller tumors only, generally below 3 cm to 4 cm maximum diameter. RFA provides slightly better tumor kill than
PEI (90% vs. 80% complete tumor necrosis) with fewer
sessions (mean of 1.2 vs. 4.8).199 It is also associated with
fewer side effects; thus in many centers, RFA is now preferred
versus PEI; however, RFA is contraindicated in lesions located
adjacent to the major biliary ducts or to bowel loops. Complications of these ablative techniques occur in about 8% to 9%
of cases, mainly in the form of pain, fever, bleeding, tumor
seeding, and gastrointestinal perforation.214 Percutaneous
ablation has not been well studied in children.
Rhabdoid Tumor
Although pediatric rhabdoid tumors are most common in the
kidney and brain, they do occur at other sites, including the
mediastinum and liver. When primary to the liver, rhabdoid
tumor is difficult to distinguish from the small cell undifferentiated (SCU) variant of hepatoblastoma. Given the aggressive
biological behavior and poor prognosis seen with the SCU variant of HB, it has been suggested that some tumors previously
classified as HB-SCU may actually have been hepatic rhabdoid
tumors. The differentiation of an HB-SCU from a rhabdoid
tumor is challenging and is important in terms of research,
but possibly clinically irrelevant at present because both are
biologically aggressive with poor response to chemotherapy.
Malignant rhabdoid tumor of the liver is a rare and aggressive
tumor of toddlers and school-age children that may present
with spontaneous rupture.216,217 These rare tumors are often
chemoresistant and fatal,216 although a recent case report
documents the potential for cure with multimodal therapy,
including ifosfamide, vincristine, and actinomycin D.217 As
with all locally aggressive liver tumors that respond poorly
to chemotherapy, the most important treatment goal is
complete surgical excision.
Undifferentiated Sarcomas
Undifferentiated (embryonal) sarcoma of the liver is a rare
childhood hepatic tumor that has historically been considered
an aggressive neoplasm with an unfavorable prognosis. These
tumors may arise in a solitary liver cyst.218 Survival has
improved in recent multimodal approaches, designed for
patients with soft tissue sarcomas at other sites, including
conservative surgery at diagnosis, multiagent chemotherapy,
and second-look operation in cases of residual disease. Using
these techniques several small series have reported survival in
up to 70% of children.219221
Angiosarcoma
Primary hepatic sarcomas are rare. Outcome depends primarily on tumor histology, sensitivity to chemotherapy and/
or radiotherapy, and the ability to achieve complete tumor
resection.215
Although rare, personal experience and multiple case reports

in the literature support the potential for malignant transformation of an infantile hemangioma to angiosarcoma.222,223
Histologic verification of malignancy may be difficult, and
this rare entity must be suspected if the biological behavior
of an infantile hemangioma shows unusual progression or
recurrence after a period of relative quiescence. Relatively
chemoresistant, prognosis is generally poor.
Biliary Rhabdomyosarcoma
Aggressive Hemangiomatous Tumors
The classic presentation of biliary rhabdomyosarcoma is in

young children (average age 3 years) with jaundice and abdominal pain, and it is often associated with distension,
vomiting, and fever.13 Histology is exclusively embryonal or
botryoid, both histologic subtypes of rhabdomyosarcoma that
are known to have a favorable prognosis. Because the tumor
most often involves the central biliary tree and porta hepatis,
the ability to achieve gross total resection is rare. Fortunately,
the tumor is often sensitive to both chemotherapy and radiation, and long-term survival is seen in 60% to 70% of patients.
Surgical intervention has two goals: to establish an accurate
diagnosis and to determine the local-regional extent of
disease. Although chemotherapy is generally effective at relief
of the associated biliary obstruction, patients remain at risk for
biliary sepsis until the obstruction abates.
Locally Aggressive Infantile Hepatic Hemangioma Infantile

hemangioma is the most common benign tumor of the liver
in infancy19 with striking variability of the three subtypes of
infantile hemangioma: focal, multinodular, and diffuse. Many
focal lesions are often discovered incidentally and are localized and small enough to be of little clinical significance.
Symptoms seen with larger lesions may include abdominal
distention, hepatomegaly, congestive heart failure, vomiting,
anemia, thrombocytopenia and consumptive coagulopathy,
jaundice secondary to biliary obstruction, and associated
cutaneous or visceral hemangiomas.11 Contrast-enhanced
CT scan shows an area of diminished density, and after bolus
injection of intravenous contrast, there is contrast enhancement from the periphery toward the center of the lesion.
Further, after a short delay, there is complete isodense filling
Hepatic Sarcomas
CHAPTER 33
of the lesion and liver. Magnetic resonance angiography

(MRA) has been used in complex cases to identify atypical
radiographic features that may portend a poor prognosis.224
Unfavorable radiographic features include central varix with
arteriovenous shunt, central necrosis or thrombosis, and
diffuse hemangiomatous involvement of the liver with abdominal vascular compression.224 Arterial angiography may be
used in infants with refractory symptoms in whom either
hepatic artery ligation or embolization is considered.
If a definitive diagnosis of simple infantile hepatic hemangioma can be made radiographically, management can be
noninvasive because spontaneous regression occurs in most
cases, especially focal tumors. The terminology is confusing,
however, with different authors often using the terms hepatic
hemangioma, infantile hepatic hemangioma, hepatic hemangioendothelioma, or kaposiform hemangioendothelioma
interchangeably.225 True kaposiform hemangioendothelioma
with Kasabach-Merritt (as opposed to the high-output heart
failure from intrahepatic shunts seen in diffuse infantile
hemangioma), rarely, if ever, presents as a primary hepatic
tumor.226 Hemangioendotheliomas are occasionally primary
to the retroperitoneum, where they can invade the liver and
obstruct the porta hepatis, causing portal hypertension. These
tumors are discussed in more detail in Chapter 125.
Sometimes a large rapidly growing infantile hepatic hemangioma can be life threatening with intractable high-output
cardiac failure from intralesional arteriovenous shunting, intraperitoneal hemorrhage, respiratory distress as a result of
pulmonary congestion, and massive hepatomegaly compressing abdominal vasculature and producing abdominal compartment syndrome (Fig. 33-9). Historically, the initial
medical intervention for symptomatic tumors has been corticosteroids. Many other medical treatment options exist,
although no single treatment has been shown to be universally
effective. Congestive heart failure is treated with supportive
care, digitalis, and diuretics. Anemia and coagulopathy are
treated with corrective blood product replacement therapy.
Both success and complete failure have been reported
variously with many other treatments, including epsilonaminocaproic acid, tranexamic acid, low-molecular-weight
heparin, vincristine, cyclophosphamide, interferon-2-alpha,
AGM-1470, and newer generation antiangiogenic drugs.227231
Recent studies have shown that the large tumors may produce
481
antibodies to thyroid-stimulating hormone (TSH), and screening to rule out secondary hypothyroidism is recommended.232,233 Most recently, propranolol has been shown
to inhibit the growth of infantile hemangioma.234 Although
rare, malignant transformation to angiosarcoma has been
reported, and close follow-up is recommended.223,224,235,236
In infants who fail medical management, symptomatic
solitary tumors may be treated by excision, hepatic arterial
ligation, or selective angiographic embolization.237 Treatment
algorithms may stratify treatment based upon whether or
not the tumor is solitary, multifocal, or diffuse.238,239 About
65% of tumors are solitary or unifocal with a survival of
86% and death usually not caused by the tumor but by a
comorbidity.19 Thirty-five percent of tumors are multifocal
or diffuse, with a survival somewhere between 60% to
100%, with death usually secondary to cardiorespiratory
compromise caused by tumors refractory to medical and
interventional management.19,237,238
Metastatic and Other Liver Tumors
Metastatic Liver Tumors Unlike the large body of literature
concerning liver resection for metastatic colorectal tumors in
adults, there is little published data that addresses the treatment of metastatic tumors in the liver following from abdominal solid tumors in childhood. A recent series from a large
metropolitan childrens cancer center reported only 15 such
patients during a 17-year period, including neuroblastoma
(7), Wilms tumor (3), osteogenic sarcoma (2), gastric epithelial (1), and desmoplastic small round cell tumor (2).240
Eleven of the 15 patients died of progressive disease; 4 had
a local recurrence. These results lead the authors to conclude
that the overall prognosis in these patients remains poor, and
the decision to perform hepatic metastasectomy should be
made with caution. The treatment approach should not, however, be uniformly nihilistic, because not all liver lesions in
children with abdominal solid tumors turn out to be metastatic disease. Both nodular regenerative hyperplasia and focal
nodular hyperplasia have been reported to mimic hepatic metastasis in children241; definitive diagnosis requires biopsy
and/or resection.
Occasionally, a pancreatoblastoma may present with extensive hepatic metastasis (Fig. 33-10). Despite the alarming radiographic appearance at diagnosis, this tumor was, in fact,
FIGURE 33-9 Symptomatic multifocal/diffuse infantile hepatic hemangioma. These tumors are benign but occasionally will be refractory to aggressive
attempts at medical and percutaneous management. This tumor showed progressive growth despite chemotherapy and percutaneous embolization of
largest nodules. The baby developed abdominal compartment syndrome and vena cava obstruction. Treated with temporizing abdominal decompressive
laparotomy and, definitively, with hepatectomy and live-donor liver transplantation.
482
PART III
FIGURE 33-10 Metastatic pancreatoblastoma. A, Infant with extensive metastatic tumor in the liver at diagnosis. B and C, Appearance at laparoscopic
biopsy. Primary tumor is a pancreatoblastoma involving the body of the pancreas. Although the tumor metastases were extensive at diagnosis, they prove
to be exquisitely chemosensitive with cisplatin/doxorubicin chemotherapy.
exquisitely chemosensitive, and the child did well after neoadjuvant chemotherapy, subtotal pancreatectomy, hepatectomy,
and live-donor liver transplantation. Pancreatoblastomas are
treated with multiagent chemotherapy analogous to hepatoblastoma and have a fair prognosis if chemosensitive.
Liver Tumors as Secondary Malignancies
We recently saw a case of multiple lesions of focal nodular
hyperplasia in the liver of a 10-year-old boy 9 years after treatment for stage IV neuroblastoma with double autologous stem
cell transplantations. Given the history, we initially suspected
metastatic neuroblastoma, but diagnostic laparoscopy and
laparoscopic biopsy of multiple lesions showed focal nodular
hyperplasia (FNH). Another case report of FNH in a child with
a history of stage IV neuroblastoma showed foci of small cell
undifferentiated hepatoblastoma in the resection specimen;
so, very close follow-up is necessary if treatment of the
FNH is nonoperative.242 Liver tumors have been recognized
as potential late effects and/or secondary malignancies in
children who have previously undergone chemotherapy and
radiation as toddlers.
Hemophagocytic Lymphohistiocytosis
Hemophagocytic lymphohistiocytosis (HLH) may occasionally present as an abnormal liver mass in a newborn with
coagulopathy. Predisposing factors include familial, herpes
simplex virus, and severe combined immunodeficiency.243
Diagnostic criteria according to HLH-2004 include fever,
splenomegaly, bicytopenia, hypotriglyceridemia, hypofibrinogenemia, hemophagocytosis, low natural killer (NK) cell
activity, hyperferritinemia, and high interleukin-2 (IL-2)
receptor levels.244 Treatment is with combination chemoimmunotherapy, including etoposide, dexamethasone, cyclo-
sporine A, and anticipated mortality of about 40% is increased

if the diagnosis or appropriate therapy is delayed.
Langerhans Cell Histiocytosis
Morphologic changes and clinical findings in Langerhans cell
histiocytosis (LCH) of the liver may resemble primary sclerosing cholangitis or a chronic nonsuppurative destructive cholangitis.245 Therefore LCH is an important differential
diagnosis of chronic destructive cholangitis with cholestatic
liver disease, especially in children and young adults. Other
involved organs include bone, pituitary, thyroid, and lungs.246
The diagnosis can be verified by S-100 and CD-1a (antigen)
immunohistochemistry. There have been rare reports of
pediatric liver transplantation in toddlers with multisystem
LCH, children who developed end-stage liver disease despite
intensive chemotherapy.247,248
Megakaryoblastic Leukemia
Rarely, congenital acute megakaryoblastic leukemia (AMKL)
may present isolated to the liver, with ascites caused by massive infiltration of hepatic sinusoids by leukemic cells.249 The
bone marrow by microscopy and flow cytometry and the
peripheral blood smear may not initially show the presence
of blasts. Because the marrow fibrosis may not manifest until
after the massive hepatic infiltration, it may initially be difficult to diagnose as leukemia. In most children with liver
involvement the spleen, lymph nodes, and marrow will also
be involved at diagnosis. But even in these cases, the diagnosis
may be difficult both clinically and pathologically, and the
hepatic and lymph node involvement is not uncommonly
misinterpreted as solid tumor.250
expertconsult.com.
in adults, in children they are frequently multifocal, with

a third involving the entire esophagus and 70% extending
into the proximal stomach. Children typically present with
esophageal obstruction and dysphagia, food regurgitation,
and chest pain. Barium swallow findings mimic achalasia,
and a biopsy is required for definitive diagnosis. Leiomyomas
in children are occasionally associated with familial syndromes, such as familial leiomyoma and Alport syndrome.
Extensive surgical resection is necessary in the majority
of cases.4,5
Esophageal and Gastric

Adenocarcinoma
------------------------------------------------------------------------------------------------------------------------------------------------
CHAPTER 34
Pediatric
Gastrointestinal
Tumors
Joseph T. Murphy and Robert P. Foglia
Primary gastrointestinal (GI) tumors are uncommon in infants

and children, and GI malignancies account for less than 2% of
all cases of pediatric cancer.1 The presentation and histopathology of pediatric GI tumors differ significantly from those
seen in adults. Although rare, GI malignancy should be considered in any child with signs and symptoms of intestinal
obstruction, intractable pain, alteration in bowel habits, or
GI bleeding that are not attributable to other more common
and established diagnosis. Symptoms often persist for several
weeks and may progress to intestinal obstruction requiring
emergency surgery.1,2 Children with unexplained gastrointestinal symptoms require a detailed diagnostic evaluation.3
Esophageal Smooth Muscle

Tumors
------------------------------------------------------------------------------------------------------------------------------------------------
Esophageal leiomyomas and leiomyosarcomas are rare in

children, with fewer than two dozen patients accounting for
all documented pediatric esophageal smooth muscle tumors.
Although esophageal smooth muscle tumors are often solitary
Esophageal and gastric cancer in children is extremely rare.

Between 1988 and 1996 the Surveillance, Epidemiology,
and End Results (SEER) database documented esophageal
malignancy in only three patients between 10 and 19 years
of age, and none younger than 10 years.6,7
The development of Barrett esophagus secondary to
chronic gastroesophageal reflux disease (GERD) is a primary
risk factor for the development of esophageal adenocarcinoma. Children with severe neurologic deficits, such as cerebral palsy, and those with congenital defects involving the
esophagus, such as esophageal atresia and tracheoesophageal
fistula, are at increased risk for development of Barrett esophagus.8 The incidence of Barrett esophagus has been estimated
to be 0.02% among children with severe GERD and associated
risk factors. Nevertheless, adenocarcinoma of the esophagus
has been documented in adolescents with long-standing
GERD, and surveillance with upper endoscopy and multiple
longitudinal biopsies is appropriate for those children who
have the mucosal changes of Barrett esophagus.9,10
Barrett changes can also be seen in the retained cervical
esophagus following esophageal replacement surgery. Postoperative care of these patients requires control of gastric pH and
long-term surveillance endoscopy with biopsy of the retained
upper esophageal segment. Esophageal replacement surgery
for a patient with esophageal atresia can be performed with
retention of the distal esophageal segment. This remnant
can develop severe chronic esophagitis and Barrett changes
requiring resection. Because Barrett esophagus is a premalignant condition, the distal segment of the esophagus should be
removed at the time of esophageal replacement surgery.11
Esophageal carcinomas may also occur in children after caustic esophageal injuries. Endoscopic evaluation with biopsies
should be considered for patients with chemical injuries to
monitor for the development of premalignant changes.12
Between 1975 and 2007, the SEER database reported a gastric cancer incidence of 9.25 per 100,000 individuals, with an
age-adjusted incidence of 0.1% for patients younger than
24 years.6,13 Despite the rarity of this entity, there are case
reports of adenocarcinoma of the stomach in children as
young as 2.5 years. The tumors can arise from any anatomic
location in the stomach, with nonspecific symptoms including
epigastric pain, weight loss, vomiting, anemia, and symptoms
associated with esophageal achalasia. Surgical resection is
the primary therapeutic modality; however, curative resection
is rare and mortality rates are high for children with this
tumor.1416
483
484
PART III
Gastrointestinal Stromal Tumors

------------------------------------------------------------------------------------------------------------------------------------------------
EPIDEMIOLOGY
Gastrointestinal stromal tumors (GISTs) are rare mesenchymal
tumors whose classification is hindered by anecdotal reports,
failure to distinguish between primary and secondary GISTs,
and the mixing of benign and malignant tumors in the
reports.17 In addition, GISTs arising from various anatomic
sites have been reported together, making prediction of their
clinical behavior difficult.18 The most common site is the
stomach (50% to 70%), followed by the small intestine
(20% to 30%), colon or rectum (10%), and esophagus (5%).19
Patients with GIST tumors present with nonspecific symptoms, often generalized abdominal pain, dyspepsia, and occult
GI bleeding. Iron-deficiency anemia should prompt an investigation to exclude a GI tract malignancy as the cause.1 Less
commonly, patients present with a palpable abdominal mass
or intestinal obstruction.20 Standard imaging studies may
assist in the diagnosis (plain radiographs and computed
tomography [CT]). Endoscopy can identify a tumor mass in
the stomach, duodenum, or colon.21
PATHOLOGY
GISTs are classified as mesenchymal tumors of the GI tract
thought to originate from the intestinal cell of Cajal, an
intestinal pacemaker cell.22 Historically, smooth muscle tumors, such as leiomyomas and leiomyosarcomas, and neural
tumors, such as nerve sheath tumors, have been categorized as
GISTs. GISTs are now defined as cellular spindle cell, epithelioid, or occasional pleomorphic mesenchymal tumors that
express the KIT (CD117, stem cell factor receptor) protein,
as detected by immunohistochemistry. Additional cell type
markers, such as CD34, smooth muscle actin, desmin, and
S-100 protein, are also used to establish a diagnosis of
GIST. These histologic and immunohistochemical features
now distinguish GISTs from leiomyomas, leiomyosarcomas,
neural tumors, and other tumors of smooth muscle origin.23
Prognosis relies on traditional pathologic staging criteria, such
as size, extent of tumor invasion into mucosa or surrounding organs, mitotic index, and nuclear pleomorphism. However, no single feature is consistently reliable in predicting
outcome.24
Determining prognosis of pediatric patients with GIST
tumors can be controversial. The usual criteria for assessing
risk of malignancy (i.e., tumor size, mitotic activity, anatomic
location) are not reliable in pediatric GIST. Children frequently present with multiple gastric nodules, making identification of a dominant mass difficult. Secondly, there exists a
wide variation in proliferation index between patients and
even among multiple tumors within the same patient. Furthermore, some pediatric patients develop GIST metastasis despite
being classified as low risk by adult criteria, and others with
low proliferation indices develop recurrent disease in perigastric nodal basins, the peritoneum, or liver.25 Pediatric GIST
is distinguished as a separate clinical, pathologic, and molecular subset with a predisposition for females, multifocal gastric tumors, and wild-type KIT/PDGRA genotype. This is in
contrast to older-age, adult GIST and even GIST in young

adults. All these factors must be considered when distinguishing benign from malignant pediatric gastric stromal
tumors.18,26
ASSOCIATED CONDITIONS
The Carney triad consists of a gastric leiomyosarcoma, functioning extraadrenal paraganglioma, and pulmonary chondroma. The gastric stromal tumors are usually located along
the lesser curve or antrum and produce few symptoms; however, continued growth leading to mucosal ulceration, GI
bleeding, and serosal involvement is common. Despite the
possible development of additional gastric tumors in the
remaining stomach, if feasible, partial gastrectomy is recommended as the initial operation, to avoid the complications
of more extensive gastric resection, particularly in teenaged
patients. Because the multifocal nature of the tumor can lead
to local recurrence, regular follow-up is mandatory to assess
for new gastric tumors. Adjuvant therapy has been unsuccessful in treating metastatic disease. Evaluation for adrenal tumors in patients with gastric stromal sarcomas and
pulmonary chondromas should be considered, and a family
history should be obtained from patients with the Carney
triad. Recently, an autosomal dominant inheritance of paragangliomas and gastric GIST, called the Carney-Stratakis syndrome, has been identified, representing a separate condition
affecting both males and females. Succinate dehydrogenase
subunit gene mutations, typically associated with familial
paragangliomas, have been implicated in the pathogenesis
of Carney-Stratakis syndrome.2730
An uncommon, histologically distinct subset of GIST,
called a GI autonomic nerve tumor (GANT), has been described in children. Pediatric GANTs have a female prevalence
and symptoms that may include anemia, abdominal pain, fullness, emesis, and a palpable abdominal mass. Although adult
GANTs are found predominantly in the small intestine, pediatric GANT lesions are primarily gastric tumors. The majority
of pediatric patients have localized disease at the time of diagnosis. Younger age, localized disease, gastric location, and
small tumor size at diagnosis are associated with favorable
prognosis. Immunocytochemical and ultrastructural evaluation is required to differentiate these tumors from GIST. Established pathologic criteria for malignancy are not well defined
for the pediatric GANT because of the low incidence of these
tumors. Surgical resection of the tumor is the treatment of
choice, because there appears to be no definite role for chemotherapy or radiation.10
TREATMENT
Complete surgical excision of GISTs, along with the pseudocapsule, is the treatment of choice. Achieving negative pathologic
margins is frequently possible, because GISTs tend to hang from
and do not diffusely infiltrate the structure from which they
arise. Consequently, wedge resection of the stomach or segmental resection of the intestine provides adequate therapy; wide
resection is not necessary.17 In addition, because the status of
microscopic margins does not appear to be important for survival, vital structures should not be sacrificed if gross tumor
clearance has been attained. GIST rarely metastasizes to lymph
nodes; so, lymphadenectomy is seldom warranted.19
CHAPTER 34
The high rate of local and distant recurrence underscores

the need for adjuvant therapy. GIST has traditionally been
resistant to radiotherapy; however, imatinib mesylate, a selective KIT, PDGF-RA, PDGR-RB, and BCR-ABL tyrosine kinase
inhibitor, has been successful as a first-line agent in treating
advanced and metastatic GIST in adult patients. Imatinib
blocks the constitutive activity of KIT receptor in GIST cells.21
Recently, a mutation in the c-KIT gene on exon-11 associated
with increased risk of recurrence and higher mortality was
identified.19 The efficacy of imatinib is related to GIST genotype, with KIT exon-11mutated GISTs being more sensitive
to imatinib than wild-type (WT) tumors. While imatinib
mesylate has been effective adjuvant therapy for adult GISTs,
pediatric GIST lesions are frequently less responsive. The lack
of efficacy may result from pediatric GISTs being predominantly WT genotype and lacking the KIT mutations
more commonly detected in adult GIST tumors.31,32 Secondgeneration kinase inhibitors (i.e., sunitinib, nilotinib, sorafenib, and dasatinib) have demonstrated in vivo and in vitro
efficacy in treatment of malignancy with KIT mutations.3336
Although investigations of adjuvant and neoadjuvant tyrosine kinase inhibitors are ongoing, surgical excision remains
the initial option for pediatric GISTs. Adjuvant chemotherapy with imatinib and other agents may be used in cases
of incomplete resection, tumor spillage, or other high-risk
factors. For recurrent or metastatic GIST, a trial of a kinase
inhibitor, followed by surgical resection, may be effective.
Neoadjuvant tyrosine kinase inhibitor chemotherapy may
similarly reduce unresectable GIST lesions making surgical
resection possible. These therapies may decrease the incidence of postoperative GIST recurrence and spread, and
thereby extend survival.34,35
SURVIVAL
The long-term survival following surgical resection of pediatric GIST is difficult to determine, because most reports contain
small numbers of children or include adults. Moreover, given
recent changes in the recognition and pathologic identification of these tumors, many older series contain tumors that
are actually not GISTs. Factors associated with long-term
survival following surgical resection include small tumor size,
low mitotic index, genotype, and gastric primary location.20
Pediatric GISTs present with a higher incidence of metastasis
than comparable adult gastric tumors. However, the biology of
pediatric lesions appears more indolent than adult disease
with significant long-term survival, despite the presence of
metastatic disease and with or without effective adjuvant
chemotherapy.25
Intestinal Tumors
------------------------------------------------------------------------------------------------------------------------------------------------
MYOFIBROMATOSIS
Infantile myofibromatosis is a mesenchymal tumor that can
arise in the skin, muscle, bone, subcutaneous tissue, or
viscera. It is the most common fibrous tumor of infancy. Myofibromatosis presents with either solitary or generalized
lesions, with or without visceral involvement. Most lesions
spontaneously regress; however, extensive intestinal myofibromatosis is associated with significant morbidity and
mortality.37,38 Various chemotherapeutic interventions have
PEDIATRIC GASTROINTESTINAL TUMORS
485
demonstrated limited efficacy, significant treatment toxicity,

and long-term morbidity. However, the combination of lowdose chemotherapy and long-term total parental nutrition
for life-threatening infantile myofibromatosis can provide
symptomatic relief and inhibit disease progression.39,40
LYMPHOMA
Lymphoma is the most common small bowel malignancy in
children, with high-grade non-Hodgkin lymphoma comprising 74% of these tumors. Burkitt lymphoma constitutes the
most common histologic subtype. The majority of patients
(50% to 93%) present with lymphoma localized to the distal
small bowel, although tumor may occur anywhere from the
stomach to the rectum.41
Patients may present with chronic GI distress, occult blood
per rectum, hematochezia, and/or an abdominal mass. An
acute worsening of symptoms may result in emergency
surgery for treatment of ileocolic intussusception, with lymphoma creating the lead point (46%), acute appendicitis
(22%), perforation (11%), or obstruction (8%). Higher mortality is associated with advanced disease stage, intestinal
perforation, high-grade histology, and T-cell lymphomas.42
Surgical management depends on disease presentation, as
well as extent of disease at presentation. Bulky disease is
usually not completely resectable. Extensive resection of bulky
retroperitoneal or mesenteric disease does not enhance survival; nevertheless, complete surgical resection (including
bowel resection), if possible, significantly enhances the prognosis of patients with intestinal lymphoma, especially when
included in a multimodality treatment approach. Tumor
downstaging by complete resection allows for decreased
duration and intensity of post-operative chemotherapy. When
operating for a complication of intraperitoneal disease, the
extent of the procedure should be limited to resolution of
the complication and resection of sufficient tissue to ensure
an accurate diagnosis. If limited disease is encountered, complete resection and an evaluation of mesenteric, perihepatic,
and periaortic nodes should be undertaken to assess for
regional metastatic spread. Two-year cumulative survival for
intestinal B-cell lymphoma is 94% and 28% for intestinal
T-cell lymphoma. The overall 5- and 10-year survival rates
for all intestinal lymphoma patients treated with multimodality therapy (surgery, radiation, chemotherapy) are 52% and
44%, respectively. The corresponding disease-free survival
rates are 43% and 38%, respectively.4348
Carcinoid Tumors
------------------------------------------------------------------------------------------------------------------------------------------------
EPIDEMIOLOGY
Carcinoid tumors originate from neuroendocrine cells within
the GI tract. These neoplasms derive from GI epithelial and
subepithelial endocrine progenitor cells that function as part
of the amine precursor uptake and decarboxylation (APUD)
system.49 Carcinoids can also be found in the lungs, mediastinum, thymus, liver, pancreas, bronchus, ovaries, prostate,
testes, and kidneys.50 Pediatric carcinoid tumors typically
occur in the GI tractstomach, small intestine, appendix
(most common), and rectum. Carcinoid tumors of the appendix occur with an estimated incidence of 1 case per million
children per year, with a slight female predominance.5153
486
PART III
DIAGNOSIS
Carcinoid tumors are classified according to the location of
origin in the primitive gut (foregut, midgut, and hindgut).
Foregut tumors include carcinoids of the lung, bronchus,
stomach, proximal duodenum, and pancreas. Midgut tumors
arise from the distal duodenum, jejunum, ileum, and right
colon, including the appendix. These account for 60% to
80% of all carcinoids in adults and children.5456 Hindgut
tumors arise in the transverse and distal colon and rectum.
Tumors can also arise from a Meckel diverticulum, enteric
duplications, and the mesentery. Appendiceal carcinoids are
the most common, with more than 70% of these tumors
developing at the appendiceal tip. Pediatric carcinoid tumors
are often discovered incidentally during an operation for presumed appendicitis or another unrelated diagnosis. Although
clinical signs of acute appendicitis or gynecologic pathology
may prompt exploration, true inflammatory changes of acute
appendicitis are not often induced by the carcinoid, possibly
because of the distal location of the tumor and absence of
proximal luminal obstruction.51,53,55
The most serious complication of carcinoid tumors is a carcinoid crisis, which is most often associated with foregut tumors,
larger tumors, and high serum/urine 5-hydroxyindoleacetic
acid (5-HIAA) levels. Although pediatric carcinoids vary in
size, carcinoid syndrome (flushing, diarrhea, abdominal
pain, tachycardia, hypertension, hypotension, altered mental
status, and coma) has not been typically associated with tumors
confined to the appendix.53,54 In contrast, pediatric patients
with extra-appendiceal carcinoid tumors, such as in the
lung or liver, are often symptomatic. Biologically active
amines (serotonin, catecholamines, histamine) and metabolites
(5-HIAA) are characteristically elevated in the plasma and urine
of patients with symptomatic carcinoid tumors.57 Patients with
extra-appendiceal carcinoids frequently present with disseminated disease at the time of diagnosis and have a higher
incidence of recurrent tumor following the initial diagnosis
and resection.58
TREATMENT
Tumor size at presentation dictates surgical decision making
for carcinoid tumors of the appendix. For appendiceal carcinoid
tumors less than 2 cm in diameter, surgical resection of the
appendix and mesoappendix is considered curative. Long-term
follow-up demonstrates minimal disease recurrence and a rare
likelihood of metastatic disease.51,52,58,59 Carcinoid tumors
greater than 2 cm, those with cecal involvement, lymphatic
invasion, lymph node involvement, mesoappendix infiltration,
positive resection margins, goblet cell malignancy, or cellular
pleomorphism with a high mitotic index require a more
extensive resection (i.e., a right hemicolectomy with associated
resection of the mesocolon).55,60,61
SURVIVAL
Complete resection of localized appendiceal carcinoid tumors
can result in cure, with greater than a 90% survival rate.
Diminished disease-free and overall survival is associated with
carcinoids larger than 2 cm, older age, positive lymph nodes,
extra-appendiceal spread, distant metastatic disease, and
tumors with atypical histologic features.62
Colorectal Adenocarcinoma
------------------------------------------------------------------------------------------------------------------------------------------------
Adenocarcinoma of the colon and rectum is the most common

cancer of the GI tract, with approximately 142,570 new cases
and 51,370 deaths in the United Sates in the past year. The
lifetime risk of developing colorectal cancer in the general
population is 1 in 19.6 However, colorectal cancer in children
is rare, with an estimated incidence of 0.3 to 1.5 cases per
million.63,64 Although reported as early as 9 months of age,
the median age at diagnosis for pediatric cases is 15 to 19 years.
Pediatric colorectal cancer accounts for 2% of malignancies in
adolescents.6567
Colorectal cancer differs greatly between adults and children. These differences include the presenting signs and
symptoms, primary site of the tumor, pathologic findings,
stage, and prognosis. Carcinoma of the colon is associated
with several predisposing factors, including ionizing radiation
(e.g., CT scan, therapeutic radiation treatments), polyposis
syndromes, urinary diversion with previous ureterosigmoidostomy, and chronic parasitic infection. Various environmental factors, including herbicide exposure, may also be
associated with tumor formation.68
Polypoid Disease of the

Gastrointestinal Tract
------------------------------------------------------------------------------------------------------------------------------------------------
Polyps are common, occurring in 1% of all children, and

are the most frequent source of rectal bleeding in the young
child (2 to 5 years old). Most polyps are benign lesions and
are either hamartomas or result from lymphoid hyperplasia.
Some hamartomas, however, have the potential for dysplastic,
adenomatous or neoplastic transformation because of germline mutations and somatic inactivation of STK11, SMAD4,
BMPR1A, and PTEN genes.6871
Isolated juvenile polyps (i.e., retention polyps, inflammatory polyps, cystic polyps) are considered hamartomas. They
constitute 80% of polyps in children with 40% to 60% found
in the rectosigmoid colon. If multiple (typically two to five
polyps) they may be found in the proximal colon as well. They
are one of the most common sources of GI bleeding in young
children, but are rarely seen in adolescence. Colonoscopy of
the entire colon is diagnostic and therapeutic if endoscopic
removal is warranted.
Lymphoid polyps (lymphoid nodular hyperplasia) account
for 15% of pediatric polyps and are submucosal lymphoid
aggregates, specifically localized to distal small bowel, colon,
and rectum (Peyer patches). Bleeding results from mucosal
erosion and can usually be managed expectantly. Uncontrolled bleeding or irreducible intussusception requires
surgical intervention.
Juvenile polyposis coli syndrome is transmitted in an autosomal dominant fashion. Affected individuals are at increased
risk for colorectal malignancy with cumulative risk for cancer
of nearly 50% to 70% by age 60 years.63,64,72 A diagnosis of
juvenile polyposis coli requires at least 5 polyps throughout
the GI tract, or 1 polyp and a family history of juvenile polyposis. Most patients typically have 50 to 100 polyps including
gastric and small bowel polyps. Higher numbers of polyps are
associated with more severe symptoms, including chronic
bleeding, anemia, hypoproteinemia, and failure to thrive.
CHAPTER 34
These patients and their families require long-term endoscopic surveillance (semiannual panendoscopy) with subsequent total abdominal colectomy if mucosal dysplasia,
persistent bleeding, or rapid increase in polyp number is
detected. Depending on individual circumstance, there also
appears to be a role for prophylactic total colectomy and rectal
mucosectomy with an endorectal pull-through procedure.
Diffuse juvenile polyposis of infancy is a nearly universally
fatal disease typically diagnosed within the first few months
of life. Patients present with diarrhea, lower GI bleeding, intussusception, prolapse, obstruction, protein-wasting enteropathy, macrocephaly, and hypotonia. Despite involvement
of the entire GI tract, bowel rest and total parenteral nutrition
(TPN) permit selective surgical resection. However, survival
beyond 2 years of age is rare.
Diffuse juvenile polyposis presents with hematochezia, abdominal pain, and prolapse from hamartomatous polyps in
the colon and rectum in infancy to 5 years of age. Although
hamartomas typically do not have premalignant potential,
chronic polyp inflammation is thought to result in reactive
hyperplasia that then progresses to dysplasia or adenomatous
changes.
Several genetic disorders carry significant risk for the
subsequent development of colon carcinoma and are characterized as polyposis syndromes. They include Gardner syndrome (adenomatous polyposis and soft tissue and bone
tumors), Turcot syndrome (familial adenomatous polyps
and central nervous system tumors), and familial polyposis
coli. Both Gardner syndrome and familial polyposis are
autosomal dominant disorders and are associated with adenomatous polyps in the colon and the small intestine. Because
the entire surface of the colon can be carpeted with thousands
of polyps, the ability to carry out effective surveillance and
identify suspicious lesions is low. Recommendations for and
the timing of colon resection are based on the likelihood of
the development of malignancy. There is little question that
colectomy is the appropriate treatment for patients with
familial polyposis coli (familial adenomatous polyposis),
Gardner syndrome, and Turcot syndrome.
Peutz-Jeghers syndrome is defined by polyposis of the intestinal tract and melanotic skin lesions. It is inherited as an
autosomal dominant trait. Germline mutations in LKB1,
STK11, and ENG genes may have a causative role in the pathogenesis of this syndrome.73,74 Despite equal sex distribution,
symptoms appear earlier in males. Brown and black melanotic
spots occur in the rectum, around the mouth, lips, buccal
mucosa, feet, nasal mucosa, and conjunctivae, typically presenting at puberty. Adolescents characteristically complain
of frequent defecation, rectal bleeding, abdominal pain,
vomiting, and may present with anemia or recurrent episodes
of intussusception. Polyps are found in the small intestine
(55%), stomach and duodenum (30%), and the colorectal
bowel (15%). The risk of death because of cancer for those
with Peutz-Jeghers syndrome is 50% by 60 years of age. There
is a 13-fold increased risk of death because of GI cancer and
a 9-fold increased risk for all other malignancies. Rapid
growth, severe dysplasia, villous changes, or larger polyps
(greater than 15 mm) may indicate GI malignancy and necessitate aggressive surgical intervention. However, repeated,
extensive intestinal resections may result in short-bowel syndrome resulting from the multifocal and recurrent nature of
these polyps.
487
Gardner syndrome patients present with adenomatous,

rather than hamartoma polyposis, and extraintestinal lesions,
including bone tumors (80%), sebaceous/inclusion cysts
(35%), and desmoid tumors (18%). Bone lesions include cysts
of the mandible, fibromas, and osteomas of the skull and face.
These patients also may develop hypertrophy of the retinal
pigmented epithelium. The syndrome is inherited in an
autosomal dominant pattern. Various mutations of the adenomatous polyposis coli (APC) gene are associated with Gardner
syndrome (APC polymorphism in exons 13 and 15), implicating this as a phenotypic variant of familial adenomatous
polyposis (FAP). The intestinal polyps that characterize this
syndrome have a 100% likelihood of undergoing malignant
transformation.75,76 Desmoids are fibroblastic tumors of the
abdominal wall and mesentery that present as dysplasia or a
malignant fibrosarcoma. They often become apparent after
diagnosis of GI disease and carry a high mortality. Small desmoid tumors, if amenable to excision, have a 10% local
recurrence. However, many desmoid lesions are unresectable at presentation. Slow-growing tumors can be treated
with sulindac, tamoxifen, vinblastine, and methotrexate, while
symptomatic, aggressive tumors require doxorubicin and
dacarbazine or high-dose tamoxifen and radiation therapy.
Turcot syndrome, also considered a variant of FAP, is characterized by polyposis and brain tumors (e.g. gliomas, ependymomas). Carcinoma of the colon is prevalent in young adults.
Chronic bloody diarrhea, hypoproteinemia, weight loss, anemia, malnutrition, bowel obstruction, and intussusception are
common presenting symptoms. Medulloblastoma development is associated with APC-related mutations, while microsatellite gene instability (typical of hereditary nonpolyposis
colon cancer) is associated with glioblastoma multiforme diagnosis.77 Cronkhite-Canada syndrome is typified by multiple
hamartomatous polyps in the stomach and colon. It is a
variant of juvenile polyposis and is associated with early-onset
skin hyperpigmentation, alopecia, and nail changes. Chronic
diarrhea results in malabsorption, hypovitaminosis, hypoproteinemia, and fluid and electrolyte imbalance.
Osler-Weber-Rendu syndrome is characterized by childhood
(less than 10 years of age) GI bleeding from cutaneous and
hepatic telangiectases and vascular malformations in 50% of
affected individuals. Telangiectases are found on the lips, oral
and nasopharyngeal membranes, tongue, and perilingual
areas. Lesions may also involve the brain, lungs, and liver.
Within the GI tract, they occur commonly in the stomach
and small bowel, causing significant recurrent GI bleeds
throughout childhood. It is inherited in an autosomal dominant manner, and 80% of patients have a family history of
the disease. With a high incidence of colon carcinoma or multiple juvenile colonic polyps, all Osler-Weber-Rendu patients
with new-onset anemia or GI bleeding require lower GI tract
evaluation.78,79
Hereditary Associations
------------------------------------------------------------------------------------------------------------------------------------------------
Although the majority of childhood colorectal carcinomas are

not associated with hereditary factors, approximately 25% of
childhood cases have some associated predisposing condition,
that is, at least two first-degree relatives with colon cancer,
genetic/polyposis syndromes (1%), inflammatory bowel
disease (1%), and hereditary nonpolyposis syndromes (5%
488
PART III
to 6%). The progression toward tumor development may

occur secondary to tumor suppressor gene mutation, loss of
heterozygosity, or a mutational event.80 Phenotypically normal colonic epithelium may develop hyperplasia as a result,
then progress to adenoma formation, dysplasia, and finally,
invasive carcinoma. Mutations associated with development
of colon cancer may result from exposure to environmental
influences or be the result of accumulated DNA transcription
errors. Typical of these genetic changes are APC inactivation,
K-ras activation, and TP53 gene mutations.
Familial adenomatous polyposis (FAP) is inherited as an
autosomal dominant trait that accounts for less than 1% of
all colorectal cancer. A diagnosis of FAP requires greater than
5 colonic polyps, polyps throughout the GI tract, or polyps
associated with a family history of juvenile polyposis. Extensive colonic polyposis (i.e., greater than 100 adenomatous
polyps) is common, with some patients having thousands of
polyps. Symptomatic patients often present with frequent
bloody stools, anemia, and abdominal pain. Long-standing
symptoms may signify the presence of a malignant lesion.
Patients identified through family history should be assessed
in early adolescence prior to the development of symptoms.
All patients require early colonoscopic screening to determine
the extent of polyposis and the possibility of malignancy.
Colorectal carcinoma occurs by age 20 years in 7% of patients
and by age 25 years in 15% of patients. Untreated FAP characteristically progresses to colorectal cancer by age 39 years. In
contrast, gastric polyps seen with FAP are usually benign
hamartomas. FAP patients are also at risk to develop desmoid
tumors, congenital hypertrophy of retinal pigment epithelium, duodenal and periampullary adenocarcinomas, thyroid
malignancy, and hepatoblastoma.81,82
The APC gene, a tumor suppressor gene on the long arm
of chromosome 5, is known to contain a mutation in 80%
to 90% of FAP patients. If a defective APC allele is inherited
from one parent, a mutation acquired during childhood in
the other APC gene results in the loss of function of the
tumor suppressor gene product.83 Inactivation of the APC
alleles results in activation of subsequent signaling pathways leading to uncontrolled cell growth. Malignant progression from adenoma to dysplasia, then to malignancy,
may occur. Site-specific APC gene mutations correlate with
various FAP phenotypes and the development of associated
tumors. Classic FAP is associated with central gene mutations,
while a less aggressive, attenuated FAP presentation correlates with peripheral APC gene mutations. The development
of malignancy is also associated with accumulation of
other oncogene/tumor suppressor gene mutations, such as
K-ras activation and TP53 mutation, in otherwise quiescent
adenomas.
Sulindac, a nonsteroidal anti-inflammatory drug (NSAID),
and celecoxib, a cyclooxygenase-2 (COX-2) inhibitor, have
been used to reduce polyp numbers by induction of epithelial
cell apoptosis. Despite the unique mechanism of action of
these agents, they have not completely eliminated the risk
of colorectal cancer in FAP patients.84,85 FAP patients with
few polyps are still at risk of early colorectal cancer. Resection
is indicated even if extensive polyposis does not develop.
Surgical options include total proctocolectomy with permanent ileostomy, total abdominal colectomy with ileorectal
anastomosis, coloproctectomy with perseveration of the anal
sphincter, coloproctectomy with ileoanal pull-through, and
total colectomy with rectal mucosectomy and endorectal

(J-pouch) pull-through. Each of these procedures has their
proponents. Total colectomy with a rectal mucosectomy and
endorectal pull-through has gained popularity in recent years.
This procedure removes all at-risk colonic mucosa and laparoscopic techniques have been demonstrated to be practical,
effective, and safe. Endorectal pull-through procedures typically incorporate a distal J-pouch ileal reservoir. Although
straight ileal pull-through procedures initially have higher
stool frequency, differences in stool frequency between
straight pull-through and J-pouch patients have been reported
by some authors to be negligible by 24 months. A number of
patients treated with a J-pouch may require later treatment
for intermittent pouchitis.86,87 Total proctocolectomy with
permanent ileostomy carries significant risk of postoperative
urinary bladder atony, impotence, and retrograde ejaculation
because of disruption of nervi erigentes during the pelvic dissection. More commonly used for adult colorectal pathology,
this technique has limited utility for treatment of pediatric
patients because of the psychological and physiologic impact
of a permanent stoma.88 Procedures involving the preservation of the distal rectum can result in the development of colorectal cancer. Forty-four percent of patients undergoing an
ileorectal anastomosis require subsequent treatment for rectal
polyps that develop in the remaining mucosa. The risk of rectal cancer in these patients is 10% at age 50 years and 29% by
age 60 years. Polyps remaining or developing in preserved colorectal segment significantly increases the risk for subsequent
cancer. Those with retained rectal mucosa at risk require annual flexible endoscopic surveillance of the pelvic pouch.8991
The significant long-term risk of rectal cancer in these patients
makes this procedure unacceptable for treatment of FAP in the
adolescent population.
Hereditary nonpolyposis colon cancer (HNPCC) has an autosomal dominant inheritance, is the most common hereditary
colon cancer syndrome, and accounts for 2% to 3% of all
colorectal cancers. It is characterized by early onset, multiple
family members affected, and is 5 times more prevalent than
familial polyposisrelated colon cancer.92 In contrast to familial adenomatous polyposis, HNPCC malignancy may develop
in the absence of adenomatosis of the colon and rectum. Unlike sporadic colorectal tumors, HNPCC colorectal cancer
usually develops in a proximal colon lesion and occurs at a
younger age (approximately 45 years).
Disease may be limited to the colon in the Lynch syndrome
I, where malignancies occur in the cecum and ascending
colon more often than in other colorectal sites (70%).
These tumors are characterized by poorly differentiated and
mucin-producing lesions (i.e., signet cell). Lynch syndrome
II is further defined by the development of synchronous
and metachronous extracolonic cancers, such as carcinomas
of the endometrium, uterus, ovary, stomach, small bowel,
pancreas, hepatobiliary tract, brain, genitourinary system,
and upper uroepithelial tract. They usually manifest in the
second decade of life. HNPCC patients are categorized by
the Amsterdam criteria: colorectal cancer in at least three
relatives spanning two generations. One of these individuals
is a first-degree relative of the other two and one of these individuals must have a diagnosis prior to age 50 years. Patients
with the Lynch syndromes have a 50% to 70% lifetime risk of
developing cancer and a threefold increased incidence of
colorectal cancer compared with the general population.9396
CHAPTER 34
Hereditary nonpolyposis colon cancers, unlike FAP, do not

have inherited defects in the APC gene. HNPCC tumors are
characterized by mutations in genetic loci (MSH2, MLH1,
PMS1, PMS2, and GTBP) resulting in defective DNA nucleotide mismatch recognition and repair. Greater than 90% of
these mutations are in MSH2 and MLH1 genes on chromosome arms 2p and 3p, respectively. These genes are inherited
in a dominant manner with 90% penetrance. Although benign
adenomas appear with the same incidence in HNPCC patients
as in the general population, DNA-repairdeficient HNPCC
adenomas are more likely to grow and progress to invasive
cancer than in the general population. As a result, a benign
tumor may progress to cancer in as few as 3 to 5 years.97
Patients suspected of carrying MSH2 and MLH1 mutations
may be tested for DNA mismatch-repair gene mutations. A
total abdominal colectomy, rather than hemicolectomy or a
segmental resection, is recommended, because the risk of
recurrent colorectal cancer is 45% spanning 10 years. Patients
who have undergone subtotal colectomy must undergo lifelong endoscopic evaluation of their remaining rectal segment.
Subtotal colectomy with a rectal mucosectomy and endorectal
pull-through has not been studied in this population. Patients
who are poorly compliant with colonoscopic surveillance may
be candidates for prophylactic colectomy. Asymptomatic
HNPCC gene carriers may reduce their risk of invasive cancer
through prophylactic colectomy or surveillance colonoscopy
and polypectomy, starting with biannual colonoscopy at age
25 to 30 years and annually after age 40 years. All Lynch
syndrome patients must undergo lifelong screening for
extracolonic malignancies as well.98,99
Other Associations
------------------------------------------------------------------------------------------------------------------------------------------------
There is a strong association between long-term inflammatory

bowel disease and the development of colon carcinoma. After
the first 10 years with ulcerative colitis, the likelihood of
cancer development increases from 1% to 2% per year.100
Those with ulcerative colitisassociated carcinoma typically
present with malignancy at a young age, have multifocal
lesions, and have had a history of colitis involving the entire
colon rather than isolated, left-sided disease. Crohn disease
is an inflammatory bowel disease in which the risk for colon
cancer is significantly greater (more than 20 times) than that in
the general population.101 Crohn-associated colon cancer may
develop in an area of colon that appears grossly normal, making the diagnosis of malignancy more difficult than with
ulcerative colitis. Routine surveillance contrast enema and
colonoscopy are recommended for all patients with either
ulcerative colitis or Crohn disease. Biopsies should be performed on suspicious areas as well as on random areas of
the colon during the colonoscopy.102
Ureterosigmoidostomy performed for urinary diversion
predisposes to the subsequent development of malignancy
in the colonic segment used as a diversion conduit. Five percent of patients with ureterosigmoidostomy develop colon
cancer, often at the site of ureteral implant. Chronic inflammation, possibly resulting from exposure to intermittently
infected urine, has been shown to predispose to the development of colon cancer. Close follow-up and annual sigmoidoscopy is warranted for all patients following this type of urinary
diversion, which is now infrequently used.
489
In the United States, approximately 600,000 abdominal

and head CT studies are performed annually on children
under 15 years of age. The risk of later malignancy related
to diagnostic pediatric CT scan is directly proportional to
the age of the child at the time of the study. It is estimated
500 of these individuals may ultimately die from a malignancy
attributable to the CT irradiation received as a child. Secondary colorectal malignancies may also result from therapeutic
radiation, especially if the abdomen was included in the
field of primary irradiation. Radiation colitis and adenomatous polyps can develop years after radiation exposure and
colonic adenocarcinoma decades later. Immunohistochemical
studies suggest a radiation-induced TP53 mutation may lead
to the eventual development of colorectal cancer in these
individuals.67,103105
Diagnosis
------------------------------------------------------------------------------------------------------------------------------------------------
In children with colorectal tumors, presenting symptoms are

nonspecific and include abdominal pain, nausea, and vomiting, and changes in bowel habits with the development of
constipation, particularly with left-sided lesions. Physical
findings include abdominal distention, tenderness, and a palpable mass. Many will have guaiac-positive stools. Lower GI
(rectal) bleeding may be present in a third of patients and,
as with adults, is more prevalent in those with cancer of the
left colon and rectum. Significant weight loss affects 20% to
30% of patients.
The median length of time from onset of symptoms to
presentation is often months; at least one report cited almost
a year between the onset of symptoms and the diagnosis.106
Diagnosis can be delayed if symptoms are repeatedly attributed to common pediatric conditions such as chronic gastroenteritis.107 Delay in diagnosis may also be related to
adolescents tendency to minimize or hide embarrassing
symptoms and the low index of suspicion of pediatricians
for this rare entity. Rectal bleeding, commonly a sign of benign
pathology, such as polyps or hemorrhoids, should also raise
the suspicion of a colorectal malignancy. Delay in diagnosis
contributes to the advanced stage of disease in many children.
Most colonic lesions in adults are rectosigmoid in location and
are identified by sigmoidoscopy. In contrast, childhood colorectal cancer is relatively evenly distributed throughout the
colon, with one third of the tumors located in the right
colon.63,107 Colonoscopy, therefore, is required to obtain a
biopsy for diagnosis.
Sporadic Colorectal Carcinoma

------------------------------------------------------------------------------------------------------------------------------------------------
Sporadic colon cancer in the young is an aggressive disease

whose morphology and natural history differ from those of
familial adenomatous polyposis, hereditary nonpolyposis
colorectal cancer, and adult colon cancer. The location, stage,
and histologic type of pediatric colorectal tumors differ
markedly than the same disease in adults. Primary colorectal
tumors in children occur frequently in the right and transverse
colon, as opposed to the predominant rectosigmoid distribution found in adults. Approximately 40% of adults with colon
cancer have involvement of regional lymph nodes or have
distant metastases (Dukes stage C or D lesions) at diagnosis.
In children, more than 80% of tumors are Dukes stage C or
490
PART III
D at diagnosis.63,64,106,107 In addition, more than half of

the colorectal tumors in children are mucinous adenocarcinomas. The mucinous subtype has an aggressive course and
is known to metastasize early. Both advanced stage of disease
at presentation and the preponderance of mucinous subtype
contributes to a poorer prognosis in children.66 Accumulation
of DNA base-pair mistakes resulting from defective mismatch
repair processes (microsatellite instability) is an early step in
the process leading to malignant transformation. Patients with
colorectal cancer and high microsatellite instability are more
likely to have multiple synchronous or metachronous colorectal cancers and are diagnosed at a younger age than those
without microsatellite instability.71,108 Microsatellite instability is not, however, associated with a family history of colorectal cancer or of phenotypic features.81
The utility of carcinoembryonic antigen (CEA) has been
well established in adults with colon cancer; however, there
is little evidence of similar utility in pediatric patients. CEA
levels correlate with a change in tumor burden in only 60%
of children with colorectal tumors. A number of children
with Dukes stage C or D lesions have been shown to have
normal antigen levels at the time of diagnosis.107 In addition,
CEA levels do not correlate well with long-term response to
treatment in children and therefore should not be used as a
definitive marker of recurrence.64
Treatment
------------------------------------------------------------------------------------------------------------------------------------------------
The primary treatment for colon cancer in children is surgical

resection consisting of a wide excision of the involved colon,
the mesentery, and the lymphatic drainage area. Unfortunately, resection for cure is possible in only 40% to 70% of
pediatric patients because of advanced stage at diagnosis;
these percentages are much lower than for adults. The ovaries
and the omentum are common sites of metastasis. If resection
for cure is performed, omentectomy, and in female patients,
oophorectomy, is appropriate in patients identified with associated ovarian disease.109

No rigorously tested or widely accepted therapeutic
protocols are available specifically for children with colorectal
carcinoma. The use of adjuvant chemotherapy consisting of
irinotecan, oxaliplatin, and leucovorin has been described
in conjunction with 5-fluorouracil. The use of adjuvant
chemotherapy, combined with second-look surgery in select
cases, may improve survival.3,68,107,110 As in adult rectal
cancer, preoperative radiation therapy may convert unresectable rectal carcinoma to resectable tumors in selected patients.
Although anecdotal case reports indicate these therapies can
be beneficial, no data documents the utility of either chemotherapy or radiation therapy for cure or palliation; hence,
prognosis and survival are most directly related to successful
complete resection. The overall rate and duration of diseasefree survival among children with colon carcinoma are low,
with less than 30% of patients surviving 5 years. In patients
who have resection for cure, predictors of survival include
node involvement and histologic grade.
Summary
------------------------------------------------------------------------------------------------------------------------------------------------
The biology of pediatric colorectal carcinoma is different from

colorectal malignancy in adults. The presentation, histologic
type, stage, and prognosis differ sharply. Most cases of childhood colorectal cancer arise from previous adenomas. Children
with syndrome-associated adenomas are at increased risk for
colorectal carcinoma. This suggests that genetics play a greater
role than previously thought, and fewer cases are truly sporadic.
Understanding the molecular basis of colon carcinoma in
children should facilitate the identification of patients at high
risk and result in prophylactic intervention or earlier diagnosis
and reduce mortality.
expertconsult.com.
Rhabdomyosarcoma Patient
Demographics
------------------------------------------------------------------------------------------------------------------------------------------------
Rhabdomyosarcoma is the most common type of soft tissue

sarcoma diagnosed during the first 2 decades of life, accounting for 4.5% of all cases of childhood cancer.5 It is the third
most common extracranial solid tumor of childhood after
Wilms tumor and neuroblastoma. Age at presentation follows
a bimodal distribution, with peak incidences between 2 and
6 years and again between 10 and 18 years of age.6 This distribution reflects the incidences of the two major histologic
subtypes of RMS. The incidence of embryonal RMS is highest
at birth and extends through childhood before declining,
while alveolar RMS peaks during childhood and adolescence.7
Approximately 65% of all RMS cases occur in children younger than 6 years of age. Slightly more males (58.4%) are
affected than females (41.6%), and whites have a higher
incidence than African Americans (rate ratio 1.2).
CHAPTER 35
Diagnosis and
Treatment of
Rhabdomyosarcoma
Kevin P. Mollen and David A. Rodeberg
Historically, the mainstay of therapy for rhabdomyosarcoma

(RMS) has been aggressive surgical resection, often including
a significant amount of normal tissue along with the tumor.1
As a result, operations were often disfiguring and outcomes
disappointing, with survival rates from 7% to 70% depending
on tumor location. It was not until chemotherapy was added
to the RMS treatment algorithm in 1961 that outcomes began
to improve. The addition of radiotherapy in 1965 to select
patients further improved outcomes and decreased the need
for aggressive radical operations. Recognition of the crucial
contribution of multimodal therapy to the treatment of RMS
led to the establishment of the Intergroup Rhabdomyosarcoma
Study Group (IRSG) in 1972. The goal of the IRSG was to oversee the development of treatment protocols for RMS. Now called
the Soft Tissue Sarcoma Committee of the Childrens Oncology
Group, this collaborative group has completed a number of
cooperative group trials evaluating new drug combinations,
chemotherapy dosing, imaging evaluation of tumors, radiotherapy, and surgical strategies for local tumor control and tumor
biology.2 During this time period, the overall 5-year survival rate
of RMS has increased from 25% to 70% (Fig. 35-1).3,4
Rhabdomyosarcoma Tumor
Biology
------------------------------------------------------------------------------------------------------------------------------------------------
Rhabdomyosarcoma is a malignant tumor of mesenchymal

origin.5 RMS also falls under the greater category of small,
blue, round-cell tumors of childhood that includes neuroblastoma, lymphoma, and primitive neuroectodermal tumors
(PNET). The two major histologic subtypes of RMS are embryonal and alveolar. Embryonal RMS (ERMS) is the most
common type of RMS, affecting two thirds of all patients with
disease. ERMS can be further broken down into spindle-cell
and botryoid subtypes. ERMS is typically composed of spindleshaped cells with a rich stroma. In addition to occurring in
younger patients, ERMS has a favorable survival rate of 60%.
Tumors occur more frequently in the head and neck region
as compared with the extremities. Spindle-cell histology is
common in paratesticular lesions, whereas botryoid lesions
are generally polypoid masses filling the lumen of hollow
viscus, such as the vagina, bladder, and extrahepatic bile ducts.
Alveolar RMS (ARMS) occurs in older children, and tumors
are most commonly located on the trunk or extremities. These
lesions are composed of small, round, densely packed cells
arranged around spaces resembling pulmonary alveoli. However, this histologic classification of RMS may, in the near
future, be supplanted by gene array analysis.8 Prognosis is
worse in ARMS than ERMS, with a 5-year survival rate of
54%. For all histologic types of RMS, outcome is heavily dependent on age at diagnosis, the primary anatomic site, extent of
disease (tumor size, invasion, nodal status, metastatic disease),
and the completeness of surgical excision. The Soft Tissue
Sarcoma Committee is investigating the outcomes of patients
by disease characteristics and tumor biology to refine riskadapted therapy for the treatment of RMS.2
The exact nature of the pathogenesis of RMS is unclear;
however, many hypotheses exist. It is largely thought that
RMS arises as a consequence of regulatory disruption of
skeletal muscle progenitor cell growth and differentiation.9
Pathogenic roles have been suggested for the MET protooncogene, which is involved in migration of myogenic
491
492
PART III
80
IRS-III
5-year survival
70
60
IRS-IV
IRS-II
IRS-I
50
40
Pre30 cooperative
20
locus.29,30 Autopsy findings suggest that one third of children

with RMS also have congenital anomalies, suggesting that
prenatal events may also contribute to tumor development.31
Although no specific carcinogens have been identified, benzenediazonium sulfate has been shown to induce RMS in
mice.32 Maternal marijuana or cocaine use in pregnancy
may be an environmental factor that contributes to the development of RMS.33,34
10
0
FIGURE 35-1 Improvement in survival for RMS during the past 40 years.
precursor cells, and the TP53 proto-oncogene, which is responsible for tumor suppression.10,11 At the chromosomal level,
ERMS is characterized by a loss of heterozygosity at the
11p15 locus, with a loss of maternal information and duplication of paternal genetic information. Within this locus lies the
insulin growth factor II (IGF-II) gene.1214 Both ERMS and
ARMS overproduce IGF-II, which has been shown to stimulate
RMS tumor growth, suggesting that IGF-II plays a role in unregulated growth of these tumors.15 Although the significance is
unclear, ARMS is frequently tetraploid, whereas ERMS lesions
are generally diploid. Translocations of the FKHR transcription
factor gene from chromosome 13 with either the PAX3 (chromosome 2) or PAX7 (chromosome 1) transcription factor genes
occur frequently in ARMS.1618 In these PAX/FKHR fusions,
the DNA binding domain of PAX is combined with the regulatory domain of FKHR. This results in increased PAX activity
leading to the de-differentiation and proliferation of myogenic
cells. Understanding the role of these fusion proteins in tumor
development may provide insight into treatment strategies and
potential biomarkers for the diagnosis of RMS.8 For example, it
has been demonstrated that approximately 25% of ARMS
tumors are translocation negative. By gene array analysis, these
fusion negative ARMS tumors more closely resemble ERMS
overall and have a similar prognosis to ERMS. It has therefore
been proposed that tumors should be divided into PAX/FKHR
fusionpositive and negative tumors rather than the more
ambiguous alveolar and embryonal histologies.
Although most cases of RMS occur sporadically, the disease
is associated with familial syndromes, including Li-Fraumeni
and neurofibromatosis I. Li-Fraumeni is an autosomal dominant disorder and is usually associated with a germline mutation of TP53. Patients with this syndrome present with RMS
at an early age and have a family history of other carcinomas,
especially premenopausal breast carcinoma.1922 Neurofibromatosis is an autosomal dominant genetic disorder characterized by optic gliomas, cafe-au-lait spots, and neurofibromas.23
The association of RMS with Li-Fraumeni and neurofibromatosis appears to involve malignant transformation through
the inactivation of the TP53 tumor suppressor gene and
hyperactivation of the RAS oncogene.24,25 Nevoid basal cell
carcinoma (Gorlin syndrome) is an autosomal dominant
disorder caused by mutations in the PTCH tumor suppressor
gene mapping to chromosome 9q22.3.26 Animals with
mutations in the PTCH gene have elevated levels of the tumor growthpromoting IGF-II and develop spontaneous
RMS.27,28 The association of mutations in the PTCH gene in
human disease with spontaneous development of RMS is supported by the finding that up to 30% of sporadic cases of
ERMS demonstrate molecular abnormalities at the 9q22.3
Presentation of
Rhabdomyosarcoma
------------------------------------------------------------------------------------------------------------------------------------------------
Rhabdomyosarcoma typically presents as an asymptomatic

mass found by the patient or the parents of younger children.5
Specific symptoms vary based on the site of occurrence and
extent of disease. These symptoms are generally related to
mass effect or complications of the tumor. The most common
sites of primary disease are the head and neck region, the
genitourinary tract, and the extremities.
Preoperative Workup
------------------------------------------------------------------------------------------------------------------------------------------------
Patients with suspected RMS require a complete workup

prior to surgical intervention.5 Standard laboratory work, including complete blood counts (CBC), electrolytes, and renal
function tests, liver function tests (LFTs), and urinalysis (UA)
should be performed. In addition, imaging studies of the primary tumor should be performed with computer tomography
(CT) or magnetic resonance imaging (MRI). CT is advantageous for the evaluation of bone erosion and abdominal
adenopathy, whereas MRI provides better definition of the
tumor and surrounding structures. MRI is preferable for limb,
pelvic, and paraspinal lesions. Metastatic workup includes a
bone marrow aspirate and bone scan, CT of the brain, lungs,
and liver, and lumbar puncture for cerebrospinal fluid collection. Tumor imaging defines the proximity of tumors to vital
structures and determines size. Both factors are important
when determining if the tumor can be primarily resected or
if neoadjuvant therapy is required to decrease tumor size
and thereby decrease the morbidity of resection. It has been
demonstrated that the size of the primary mass, as determined
by pretreatment imaging, carries prognostic significance.
Recent evidence would suggest that tumor volume and
patient weight may be superior predictors of failure-free survival than tumor diameter and patient age in patients with
intermediate-risk RMS.35,36 Evaluation of regional and distant
lymph nodes by clinical and radiographic means should
be performed, because this is an important component of
pretreatment staging.
Metabolic imaging using 18F-fluorodeoxyglucose positron
emission tomography (FDG PET) has become widely used in
the adult population to determine the extent of disease in the
setting of many cancers; however, there is limited experience
in the pediatric population. Recent studies have suggested that
FDG PET would be both a sensitive and specific tool in the
clinical determination of the extent of disease in childhood sarcomas.3740 Further, when combined with CT, it may be more
accurate than conventional imaging modalities in staging
patients or re-staging patients at the time of recurrence.41,42
CHAPTER 35
Surgical Principles
------------------------------------------------------------------------------------------------------------------------------------------------
BIOPSY
Open biopsy of a mass suspected to be RMS should be
performed to confirm the diagnosis. Care should be taken to
obtain adequate specimens for pathologic, biological, and treatment protocol studies. For small lesions in areas that will
be treated with chemotherapy and radiation or for metastatic
disease, core needle biopsy may be appropriate for diagnosis.44,45 Although less invasive than open biopsy, core needle
biopsy obtains a smaller tissue sample, which increases sampling error and the number of inconclusive findings. This
smaller volume of tissue may prevent the performance of
adequate molecular biology studies. Image guidance with
ultrasonography may increase the accuracy of sampling while
helping to avoid inadvertent puncture of surrounding structures.46 Clinical and radiographic positive lymph nodes should
be confirmed pathologically. Open biopsy is recommended;
however, fine-needle aspiration or core needle biopsy of lymph
nodes may be performed at the discretion of the surgeons
judgment and pathologists recommendations.44,47 Sentinel
Pretreatment Clinical Staging

------------------------------------------------------------------------------------------------------------------------------------------------
Staging of RMS is determined by the site of the primary tumor,

primary tumor size, degree of tumor invasion, nodal status, and
the presence or absence of metastases, and it is based solely on
the preoperative workup of imaging and physical examination.
This is expressed in a tumor-node-metastasis (TNM) classification system modified for the site of tumor origin (Fig. 35-2).
Adequate pretreatment clinical staging requires a thorough
Sites
Size
Orbit
Head and neck (excluding
parameningeal)
GU nonbladder/
nonprostate
T1 or T2
a or b
N0 or N1 or Nx
Bladder/prostate,
extremity, cranial
parameningeal, other
(includes trunk,
retroperitoneum, etc.)
T1 or T2
Bladder/prostate,
extremity, cranial
parameningeal, other
(includes trunk,
retroperitoneum, etc.)
T1 or T2
a
b
All
T1 or T2
a or b
Definitions:
Tumor T(site)1
T(site)2
Regional nodes
N0
N1
Nx
Metastasis
M0
M1
493
physical examination and preoperative imaging. Several investigators have validated the modified TNM staging system as a
reliable predictor of patient outcome.43
It is unclear what role FDG PET will have in the clinical evaluation of RMS, although there are several settings in which this
imaging modality may improve our pretreatment staging and
thus alter treatment for patients. FDG PET may enhance the
evaluation of regional adenopathy versus traditional modalities.
Similarly, FDG PET may offer improved detection of occult
metastases, helping to differentiate them from normal structures. Finally, this modality may offer a guide to the diagnosis
and treatment of recurrent disease. The diagnosis of a recurrence in a previously operated field is often difficult to obtain
with conventional imaging methods. FDG PET/CT may offer
an enhanced diagnostic tool and, more important, may offer
tumor viability information which will guide further surgical
therapy. One of the goals of ongoing trials will be to investigate
the role of FDG PET in RMS.
Stage
DIAGNOSIS AND TREATMENT OF RHABDOMYOSARCOMA
N0 or NX
N1
N0 or N1 or Nx
N0 or N1
Confirmed to anatomic site of origin

(a) <5 cm in diameter (b) >5 cm in diameter
Extension and/or fixative to surrounding tissue
(a) <5 cm in diameter (b) >5 cm in diameter
Regional nodes not clinically involved

Regional nodes clinically involved by neoplasm
Clinical status of regional nodes unknown
(especially sites that preclude lymph
node evaluation)
No distant metastasis
Metastasis present
FIGURE 35-2 TNM Pretreatment Staging Classification. Staging before treatment requires thorough clinical, laboratory, and imaging examinations.
Biopsy is required to establish histologic diagnosis. Pretreatment tumor size is determined by external measurement or MRI or CT, depending on anatomic
location. For less accessible primary sites, CT also will be used for lymph node assessment. Metastatic sites will require some form of imaging confirmation
(but not histologic confirmation, except for bone marrow examination). CT, computed tomography; GU, genitourinary; MRI, magnetic resonance imaging.
494
PART III
node biopsy may offer a safe and less invasive means of lymph
node evaluation for extremity and truncal lesions, although its
role in RMS is yet to be determined but will soon become the
focus of a clinical trial.4851
RESECTION OF THE MASS

Surgical biopsy of a primary lesion is often performed prior to
a definitive surgical resection. If this is the case, pretreatment
reexcision (PRE) is advisable. PRE is a wide reexcision of the previous operative site with adequate margins of normal tissue prior
to the initiation of adjuvant therapy. PRE is most commonly performed on extremity and trunk lesions but should be considered
the treatment of choice whenever technically feasible.52
The primary goal of surgical intervention is wide and
complete resection of the primary tumor with a surrounding
rim of normal tissue. A circumferential margin of 0.5 cm is
considered adequate; however, there is minimal data to support this recommendation. Such a margin may be unobtainable, however, especially with head and neck tumors.
Because of these limitations, adequate margins of uninvolved
tissue are required unless excision would compromise adjacent organs, result in loss of function or poor cosmesis, or
is not technically feasible. All margins should be marked
and oriented at the operative field to enable precise evaluation
of margins. If a narrow margin occurs, several separate biopsies of normal tissue around the resection margin should be
obtained. These specimens should be marked and submitted
separately for pathologic review. Communication between the
pathologist and surgeon is mandatory to ensure that all
margins are accurately examined. The surgeon should not
bisect or cut the excised tumor into specimens prior to sending it to the pathologist. Any microscopic or gross tumor
should be marked with small titanium clips in the tumor
bed to aid radiotherapy simulation and subsequent reexcision. Published outcomes analyses have shown that a clear
margin and no residual disease (group I) is superior to residual
microscopic margins (group II) or gross residual disease
(group III).2,3,5254 Tumors that are removed piecemeal are
considered group II even if all gross tumor is removed.
LYMPH NODE SAMPLING/DISSECTION

Lymph node status is an important part of pretreatment
staging and therefore directly impacts risk-based treatment
strategies in RMS. Regional lymph node disease (N-1) has
been identified in ARMS as an independent poor prognostic
factor in stage 3 patients.2 Data from IRS-IV would suggest
that N-1 disease in patients with ARMS is associated with
tumor characteristics that carry a poor prognosis, such as
older age, more invasive tumors (T2), large tumor size
(>5 cm), and unfavorable primary sites.55 In addition, N-1
disease was present in 23% of all RMS patients, predominantly
in primary tumor sites, such as perineum, retroperitoneum,
extremity, bladder/prostate, parameningeal, and paratesticular. N-1 disease alters both failure-free survival (FFS) and overall survival (OS) for ARMS but not ERMS.55 For patients with
N-1, ARMS outcomes were more similar to patients with single-site metastatic disease than those with only local disease.
However, for ERMS other prognostic factors, such as patient
age, tumor invasion (T stage), site of primary tumor, and
the presence of metastasis at initial presentation, were more
important prognostic factors than N-1 disease. In addition,

it has previously been shown that in patients with otherwise
localized disease, such as an extremity, N-1 disease may be associated with an inferior outcome.56,57 Clinical and radiographic positive nodes should therefore be biopsied to
confirm tumor involvement, thus ensuring correct assessment
of disease risk and assignment of optimal therapy. Lymph node
removal has no therapeutic benefit, therefore prophylactic
lymph node resection plays no role in therapy.57 Therefore clinical and/or radiographic negative nodes do not require pathologic evaluation except in extremity tumors and for children
older than 10 years of age with paratesticular tumors.58,59 In
both of these sites, the high incidence of nodal disease and
false-negative imaging necessitates pathologic evaluation of regional nodal basins.
The use of sentinel node mapping to determine regional
node status has proven to be beneficial in adult breast cancer
and melanoma. For childhood RMS, sentinel node mapping is
not yet the standard of care but may prove to be effective.48
Sentinel node mapping has proven its utility in determining
nodal status in pediatric skin and soft tissue malignancies
and will likely become the standard of care for identifying
the regional nodes involved with tumor.60
If regional nodes are positive then distant nodes should be
harvested for pathologic evaluation. Tumor identified in these
nodes would be considered metastatic disease and would therefore alter therapy using the current risk-based protocols. For
upper extremity lesions, the distant nodes would be the ipsilateral supraclavicular (scalene) nodes. In the lower extremity, the
distant nodes would include the iliac and/or paraaortic nodes.
For paratesticular RMS, the ipsilateral paraaortic lymph nodes
above the renal vein are considered distant nodes.60
CLINICAL GROUP
The extent of residual disease after resection is one of the most
important prognostic factors in RMS. For this reason, a clinical
grouping system was developed in 1972 to stratify patients
into groups that would more accurately reflect their prognosis
and treatment options. Currently, patients are assigned to a
clinical group based on the completeness of tumor excision
and the evidence of tumor metastasis to the lymph nodes or
distant organs after pathologic examination of surgical specimens (Fig. 35-3). This system differs from TNM staging in that
determination of each patients clinical group is based on the
extent of the surgical resection instead of tumor size and site.
Group
Criteria
Localized disease, completely resected

A. Confined to organ or muscle of origin
B. Infiltrating outside organ or muscle of origin: regional nodes
not involved
II
Compromised or regional resection including:

A. Grossly resected tumors with microscopic residual tumor
B. Regional disease, completely resected, with nodes involved
and/or tumor extension into an adjacent organ
C. Regional disease, with involved nodes, grossly resected, but with
evidence of microscopic residual tumor
III
Incomplete resection or biopsy with gross residual disease remaining
IV
Distant metastases present at outset

FIGURE 35-3 Clinical grouping for RMS patients.
CHAPTER 35
Failure free survival rate (%)
100
Group II
60
Group III
40
Group IV
20
0
0
495
In general, an aggressive surgical approach is used for

recurrent RMS. Data would suggest that resection of recurrent
RMS confers a 5-year survival of 37% compared with 8%
survival in a group of patients without aggressive resection.63
Given these results, SLO and aggressive resection for recurrence
can be important tools for the treatment of RMS.
However, resection of residual masses after completion of
adjuvant therapy may not be warranted. Associated morbidity
of resection and the inability to achieve complete resection
in some cases need to be considered. Further, it is not uncommon to find an absence of viable tumor tissue in resected
samples.63a This brings into question the utility of aggressive
re-resection and suggests that better means of detecting viable
tumor is crucial. As discussed, PET/CT may provide the
crucial information required to make these decisions.
Group I
80
Years
FIGURE 35-4 Rhabdomyosarcoma survival based on completeness of
surgical resection (clinical group).
Chemotherapy
------------------------------------------------------------------------------------------------------------------------------------------------
Data from IRS-III and IRS-IV demonstrate that five-year

failure-free survival rates vary according to clinical grouping
and by histologic type (Fig. 35-4).2 One criticism of clinical
grouping is that variation of surgical techniques make comparisons of clinical grouping between different institutions
problematic.61 Nonetheless, this system offers a tremendous
companion to preoperative staging in determining patient risk
assessment and prognosis (Fig. 35-5).
It was not until the 1960s that chemotherapy was recognized

as an important adjunct to surgery in the treatment of RMS.
Today, all patients with RMS receive some form of chemotherapy. Standard therapeutic regimens consist of a combination of vincristine, actinomycin-D, and cyclophosphamide
(VAC). Although tremendous advances have been made in
improving the outcomes of patients with isolated local and
regional disease, little progress has been made in improving
outcomes for advanced RMS tumors. The limiting factor has
been an inability to improve significantly upon standard chemotherapeutic regimens. Dose intensification of vincristine
and actinomycin-D is not possible because of their neurotoxic
and hepatotoxic side effects. Studies evaluating dose intensification of cyclophosphamide found that although patients
tolerate higher doses, outcomes of intermediate-risk tumors
are not changed.64 These findings have lead to the evaluation
of new drug combinations and the development of risk-based
treatment protocols.65
The combination of ifosfamide and etoposide was tested in
a Phase II therapy window in IRS-IV. When combined with
VAC, ifosfamide, and etoposide therapy resulted in a better
3-year survival rate, with less bone marrow toxicity when
compared with the use of vincristine and melphalan with standard VAC regimens.66 Other chemotherapeutic regimens
being developed to treat advanced rhabdomyosarcoma have
SECOND-LOOK OPERATIONS AND

AGGRESSIVE RESECTION FOR RECURRENCE
After completing adjuvant therapy, patients with RMS are
reimaged with CT or MRI. If residual tumor remains, or if
the outcome of therapy remains in doubt, a second-look
operation (SLO) may be considered. SLO can be performed
to confirm clinical response, to evaluate pathologic response,
and to remove residual tumor in order to improve local
control.62 As with the initial operation, the goal of SLO is
complete resection of disease. Data from IRS-III suggested
that SLO results in the reclassification of 75% of partial
responders to complete responders after excision of residual
tumors. These operations were most effective in extremity
and truncal lesions.
Risk group
Pretreatment
stage*
Low 1
1 or 2
I or II
Favorable or unfavorable
EMB
III
Orbit only
EMB
III
Favorable
EMB
I or II
Unfavorable
EMB
III
Unfavorable
EMB
13
IIII
ALV
IV
EMB
IV
ALV
Low 2
Intermediate 2 or 3
High
Clinical
group#
Site#
Histology
* Pretreatment stage dependent on site of disease

# Favorable sites: Orbit, genitourinary tract, biliary tract nonparameningeal
head and neck
FIGURE 35-5 Risk-based stratification of patients to guide

degree of therapy and prognosis for RMS patients. ALV, alveolar; EMB, embryonal.
496
PART III
incorporated doxorubicin and the topoisomerase inhibitor

irinotecan. Although used as a single agent, irinotecan is of
little value, it may be a useful adjunct to current VAC regimens
for the treatment of advanced RMS.6769 Another topoisomerase inhibitor, topotecan has shown some promise in patients
with stage 4 disease when combined with cyclophosphamide.7072 However, alternating these drugs with standard
VAC therapy has not shown any benefit in intermediate-risk
patients.73 Multiple drugs are currently being evaluated for
the treatment of RMS in Phase I and II trials.
Radiation Therapy
between active tumors and scar. It is possible that FDG PET

may offer useful clinical information in patients treated or
partially treated for RMS.
SPECIFIC ANATOMIC SITES

Rhabdomyosarcomas are unique among solid tumors in that
they may occur in many different areas of the body. Tumors
in different parts of the body may behave differently than
those in other areas. In addition, some areas of the body offer
unique obstacles to surgical resection. As such, some specific
anatomic sites of tumor occurrence will be discussed
separately.
------------------------------------------------------------------------------------------------------------------------------------------------
Radiotherapy is an important adjunct to therapy for many

children diagnosed with RMS, offering improved local control and outcomes. Candidates for radiotherapy primarily
include those with group II (microscopic residual disease)
or group III (gross residual disease) disease. The impact of
therapy is influenced by the location of the primary tumor
and amount of local disease (tumor stage and clinical grouping) at the time radiotherapy is initiated.74,75 Among patients
with group II disease, low-dose radiation (40 Gy at 1.5 to
1.8 Gy/fraction) is associated with local tumor control rates
of at least 90%.76 For patients with group III disease, radiation doses are more commonly 50 Gy.77 A randomized study
within the IRS-IV protocol demonstrated that twice-daily
irradiation at 110 cGY per dose, 6 to 8 hours apart (hyperfractionated schedule) for 5 days per week is feasible and safe.
This schedule, however, is difficult to accomplish in small
children who require twice-daily sedation for treatment.
Unfortunately, the hyperfractionated schedule demonstrated
no improvement in local control over conventional radiation
therapy.78
Radiation therapy in very young children with RMS poses
a unique therapeutic challenge. Concerns over the technical
difficulties associated with external beam radiotherapy in
young children and late side effects of therapy have led to
the evaluation of strategies that reduce the total burden
of therapy without sacrificing local control. Modern techniques, such as intensity modulated radiation therapy
(IMRT) and proton beams, may improve outcome without
compromising long-term function.79,80 Ongoing studies
continue to evaluate the dose of radiation necessary for local
control of the tumor.
Assessment of Response
to Treatment
------------------------------------------------------------------------------------------------------------------------------------------------
Although European RMS trials have incorporated the use of

conventional radiologic modalities to evaluate the response
to induction therapy and help tailor subsequent therapy, this
has not been employed in the United States. IRS-IV data
demonstrated no predictive value of radiographic response
after 8 weeks of induction therapy.81 Further, radiographic
evidence of a complete response to therapy in group III
RMS was not associated with a reduction in disease recurrence
and death.63a Clearly, the significance of persistent radiographic masses in patients treated for RMS is unknown. Conventional imaging modalities offer no information about
the biology of these masses and are unable to differentiate
Head and Neck (Superficial Nonparameningeal)

Approximately 35% of RMS arises in the head and neck
region. Of these tumors, 75% occur in the orbits. Other sites
include the buccal, oropharyngeal, laryngeal, or parotid
areas.3 The histologic variant of RMS correlates to some extent
with the location of the orbital tumor. ERMS and differentiated
types more commonly arise in the superior nasal quadrants,
whereas ARMS generally originate within the inferior orbit.83
For all head and neck RMS, biopsy is required for the confirmation of diagnosis. Resection may be limited by the inability
to obtain an adequate margin, and therefore the success of
resection is heavily dependent on location.8486 Lymph nodes
are rarely involved in childhood head and neck RMS; however, clinically or radiographically positive nodes must be
biopsied.87 Outcomes correlate strongly with tumor location.
Orbital RMS carries the best prognosis and is least likely to
extend to the meninges. These tumors generally present
earlier in the course of disease. Tumors arising in nonorbital
parameningeal locations have a high likelihood of meningeal
extension. If meningeal extension occurs after chemotherapy
and radiation therapy, the outcome is often fatal.88
Parameningeal Sites
Parameningeal RMS includes tumors arising in the middle ear/
mastoid, nasal cavity, parapharyngeal space, paranasal sinuses,
or the pterygopalatine/infratemporal fossa region. These tumors are considered high risk because of their propensity to
cause cranial nerve palsy, bony erosion of the cranial base,
and intracranial extension.89 Wide local excision is recommended but is often not feasible because of the location of
the tumors. Craniofacial resection for tumors of the nasal
areas, paranasal sinuses, temporal fossa, and other deep sites
are reserved for expert surgical teams. The recognition of poor
outcomes associated with meningeal extension has lead to a
propensity for early radiation therapy of primary tumors
and adjuvant chemotherapy.87 For patients with unresected
tumors and/or lymph node-positive disease, the use of
three-drug chemotherapy regimens (including an alkylating
agent) plus local or regional radiation may be beneficial.
The optimal dosing and timing of radiation are not yet
determined.84
Trunk
Accounting for only 4% to 7% of tumors, RMS of the trunk is
associated with a poor prognosis. Symptoms for RMS of the
trunk often occur late in the progression of disease, which
leads to late diagnoses. Complete surgical resection is difficult,
particularly when the pleura and peritoneum are involved. In
CHAPTER 35
addition, resections are frequently morbid and associated with

poor cosmetic outcomes. Resection may necessitate major
chest wall or abdominal wall reconstruction with prosthetic
materials or with flaps.90,91 Indicators of poor prognosis
include advanced stage at presentation, alveolar histology,
recurrence disease, tumor size greater than 5 cm, lymph node
involvement, and the inability to undergo gross total
resection.92,93
Abdominal Wall
Abdominal wall RMS generally presents as a painless, firm
mass. Many abdominal wall primaries can be removed
completely at presentation or following neoadjuvant chemotherapy. However, tumors arising from the interior abdominal
wall may not be noticed until significant tumor progression
has occurred, thus rendering resection much more challenging. Tumor excision should include full-thickness resection
of the abdominal wall, including the skin and peritoneum
with a margin of normal tissue. Reconstruction of the abdominal wall can be performed with mesh or myocutaneous
muscle flaps in an attempt to preserve function and cosmesis
after resection. Data would suggest that localized tumors of
the abdominal wall can be resected with good outcomes
and that younger children with abdominal wall RMS fare
better than adolescents, possibly because of a higher proportion of unfavorable histology in the older group of children.94
If the size or location prevents adequate excision, neoadjuvant
chemotherapy should be initiated to reduce tumor size and
facilitate subsequent resection.
Chest Wall
The differential diagnosis for malignant chest wall masses
includes Ewing sarcoma, primitive neuroectodermal tumors
(PNET), and RMS. Diagnostic biopsies are performed in the
long axis of the tumor, parallel to the ribs. Wide local excision
of chest wall lesions with a 2-cm margin, including the previous biopsy site, involved chest wall muscles and involved ribs,
as well as wedge excision of any involved underlying lung, is
recommended. Thoracoscopy performed at the time of resection may be helpful in determining the extent of pleural
involvement and tumor extension to the underlying lung.
Chest wall reconstruction can be performed using a number
of techniques employing prosthetic mesh, myocutaneous
flaps, and titanium ribs. Chest wall lesions have a worse prognosis than other trunk lesions, with a 1.8-year survival rate of
only 42%.90 Although radiotherapy may be beneficial for local
control of tumor, this option is associated with significant
morbidity, including pulmonary fibrosis, decreased lung
capacity, restrictive defects from altered development of the
thoracic cavity, and scoliosis.95 There is also no proven
survival benefit.
Biliary Tract
Classically, patients with biliary RMS present at a young age
(average age 3.5 years) with jaundice and abdominal pain,
often associated with abdominal distension, vomiting, and
fever. Workup reveals a significant direct hyperbilirubinemia
and a mild elevation of hepatic transaminases. Gross total
resection of biliary tract RMS is rarely possible and is often
unnecessary because of good outcomes with chemotherapy
and radiation. Currently, open biopsy is the only definitive
role of surgery in the treatment of biliary RMS, although this
497
is controversial. The histology of these tumors is often the

botryoid variant of embryonal RMS, which carries a good
overall prognosis.96 Biliary obstruction can be relieved by
stenting, but external biliary drains should be avoided because
of infectious complications. Overall, outcomes are good unless
distant metastases are present at the time of diagnosis.97,98
Paraspinal Sites
Paraspinal RMS is rare (3.3% of all RMS) and carries a poor
prognosis. These tumors tend to spread along anatomic structures, such as neurovascular bundles and fascial sheaths, occasionally causing spinal cord compression. Complete excision
of paraspinal lesions is often difficult to perform because of
large tumor size at presentation and proximity to the vertebral
column and spinal canal.92,99 Recurrence rates for paraspinal
RMS are high (55%) with the majority of these occurring at
distant locations. The lung is the most common site of distant
metastasis followed by the central nervous system.99
Retroperitoneum/Pelvis
Like paraspinal tumors, retroperitoneal/pelvic lesions are
often discovered at an advanced stage and thus generally carry
a poor prognosis. These tumors can envelop vital structures,
making complete surgical resection challenging. Neoadjuvant
chemotherapy may play a role in tumors that cannot be safely
resected at the time of diagnosis. With the exception of group
IV metastatic disease, aggressive resection is recommended
and has been shown to offer improvement in survival.100
Group IV patients with embryonal histology and those who
present at less than 10 years of age may also undergo surgical
debulking.101 It has been demonstrated that excising greater
than half of the tumor before chemotherapy resulted in
improved rates of failure-free survival when compared with
patients who did not undergo debulking.102 This is the only
setting in which surgical debulking of RMS has shown any
benefit.
Perineal/Perianal Sites
Perineal tumors are rare and usually present at an advanced
stage. Characteristics associated with improved survival
include a primary tumor size less than 5 cm, less advanced
clinical group and stage, negative lymph node status, and
age less than 10 years of age. Interestingly, histology does
not affect overall outcome for these tumors. Resection of these
tumors can be challenging because of proximity to the urethra
and anorectum. At resection, particular care should be taken
to preserve continence. If anorectal obstruction exists, a temporary colostomy may be necessary. Patients presenting in
clinical group I had 100% overall survival at 5 years compared
with 25% for group IV patients.103
Extremities
Rhabdomyosarcoma of the extremities accounts for 20% of all
new diagnoses. The majority of these tumors have alveolar histology and thus a poor prognosis. The cure rate for children
with extremity RMS has, however, improved steadily from
47% in IRS-I to 74% in IRS-III.104,105 As with many types
of RMS, complete gross resection at initial surgical intervention is the most important predictor of failure-free survival.
The primary goal of local tumor control in extremity tumors
is limb-sparing complete resection. Amputation is rarely necessary for tumor excision. Positive regional lymph nodes are
498
PART III
found in 20% to 40% of patients and are associated with

decreased overall survival (46% survival rate for node-positive
patients compared with 80% survival for node-negative
patients). Seventeen percent of IRS-IV patients with clinically
negative nodes were found to have microscopic nodal disease
on biopsy. In light of this, surgical evaluation of lymph nodes
is necessary to accurately stage children with extremity RMS,
even in the absence of clinically positive nodes.60 Currently,
axillary sampling is recommended for upper extremity
lesions, and femoral triangle sampling is recommended for
lower-extremity lesions. Sentinel lymph node mapping may
be a useful adjunct in the setting of extremity RMS. If regional
nodes are involved, then x-ray therapy (XRT) fields are
adjusted to incorporate regional lymph node basins. This
approach is associated with decreasing rates of local and
regional recurrence.57 In-transit nodal involvement at the time
of diagnosis, present in 4% of IRS-IV patients, has also been
identified as a factor contributing to regional treatment failure.
This may be evaluated by MRI, or possibly FDG PET, at the
time of diagnosis. Radiation therapy (RT) should be used at
regional lymph node sites in these patients.106
Genitourinary Sites: Bladder/Prostate
Rhabdomyosarcoma of the bladder or prostate typically
presents with urinary obstructive symptoms. These lesions
are typically of embryonal histology (73%). The major goal
of surgery is complete tumor resection with bladder salvage.
This can be achieved in 50% to 60% of patients.107,108 Partial
cystectomy has resulted in similar survival rates and improved bladder function compared with more aggressive
resections.109,110 Bladder dome tumors frequently can be
completely resected, whereas more distal bladder lesions
frequently require ureteral reimplantation or bladder augmentation. Prostatic tumors require prostatectomy, often combined with an attempt at bladder salvage with or without
ureteral reconstruction.53 Continent urinary diversion may
be necessary if tumors are unresectable or have a poor
response to medical therapy. Lymph nodes are involved in
up to 20% of cases. Therefore during biopsy or resection, iliac
and para-aortic nodes should be sampled, as well as any other
clinically involved nodes. An analysis of patients with bladder
or prostate RMS in IRS-IV revealed that 70% of these tumors
arose from the bladder with an overall 6-year survival of
82%.111 Of these patients, 55 retained their bladder without
relapse, but only 36 had normal bladder function. Urodynamic studies have been used to evaluate bladder function
after treatment.112
Genitourinary Sites: Vulva/Vagina/Uterus
Traditionally, females with primary tumors of the genital tract
underwent aggressive resection followed by chemotherapy with
or without radiation.113115 Newer treatment approaches rely
more heavily on neoadjuvant chemotherapy to reduce tumor
size and minimize the extent of resection in an attempt to
preserve organ function. Primary tumors of the vagina are about
5 times more common than cervical tumors. The vast majority
of these tumors are classic embryonal or are of the botryoid
subtype. This may account for the more favorable prognosis that
these tumors display.116 These tumors respond well to chemotherapy, with impressive tumor regression that often precludes
the need for radical operations such as pelvic exenteration. Vaginectomy and hysterectomy are performed only for persistent or
recurrent disease. Primary uterine tumors require resection with

preservation of the distal vagina and ovaries if they do not
respond to chemotherapy. Oophorectomy is only indicated in
the setting of direct tumor involvement. For those patients presenting with nonembryonal RMS of the female genital tract,
more intensive chemotherapeutic regimens are recommended
to reduce the risk of recurrence. Prognosis for this tumor site
with only locoregional disease is excellent, with an estimated
5-year survival of 87%.117
Paratesticular Sites
Paratesticular RMS generally presents as a painless scrotal
mass. Histology is generally favorable, with most tumors
showing the spindle-cell subvariant of embryonal histology.
Survival rates are greater than 90% for patients presenting
with group I or II disease.118,119 Radical orchiectomy via an
inguinal approach with resection of the spermatic cord to
the level of the internal ring is the standard of care. Open
biopsy should be avoided, because the flow of lymphatics
in this region facilitates spread of the disease. If a transscrotal
biopsy/resection has been performed, subsequent resection
of the hemiscrotum is required. If unprotected spillage of
tumor cells occurs during tumor resection, these patients
are considered clinical group IIa regardless of the completeness of resection.120 The incidence of nodal metastatic disease
for paratesticular RMS is 26% to 43%.121,122 Unfortunately,
studies have demonstrated that CT is a poor means of evaluating lymph node positivity in the retroperitoneum.123 In
addition, patients older than 10 years of age or those with
enlarged nodes have a much higher incidence of node positivity.59 Those patients should therefore undergo an ipsilateral
retroperitoneal nodal resection. Suprarenal nodes should be
evaluated, because positive nodes in this area place a patient
in group IV with disseminated metastatic disease.
Metastatic Disease
------------------------------------------------------------------------------------------------------------------------------------------------
Rhabdomyosarcoma metastasizes both through hematogenous and lymphatic routes. Children with metastatic RMS
have very poor survival rates. For the IRS studies I through
III, children with metastatic disease had a 5-year disease-free
survival of 20%, 27%, and 32%, respectively, in each of the
successive studies. Recently studies have employed the use
of upfront window studies to address potential chemotherapeutic regimens that would improve the disease-free survival
period when given to patients with newly diagnosed metastatic RMS. One such study evaluated the combination of
ifosfamide and doxorubicin for the treatment of children
with metastatic disease who are less than 10 years of age,
have embryonal histology, and lack nodal, bone, or bone
marrow involvement. This treatment strategy increased 5-year
failure-free survival to 28% and 5-year overall survival to
34%.68 Despite these improvements, more intensive research
into chemotherapeutic regimens for group IV disease should
be investigated to improve overall outcome.
Prognosis
------------------------------------------------------------------------------------------------------------------------------------------------
The prognosis of patients with RMS is dependent on many

factors. Favorable prognostic factors include embryonal/
botryoid histology, primary tumor sites in the orbit and
CHAPTER 35
nonparameningeal head/neck region and genitourinary

nonbladder/prostate regions, a lack of distant metastases at
diagnosis, complete gross removal of tumor at the time of
diagnosis, tumor size less than or equal to 5 cm, and age less
than 10 years at the time of diagnosis.77 Clinical grouping
was identified as one of the most important predictors of
failed treatment and tumor relapse.2,77 These factors become
important in the designation of treatment groups for riskbased therapy.
For group II patients, Smith and colleagues performed a retrospective review of patients enrolled in IRS-I through IRS-IV to
determine the risk factors for relapse. Those patients in group II
at highest risk for treatment failure had alveolar/undifferentiated
histology, unfavorable primary sites, regional disease with residual tumor after gross resection and node involvement, or were
treated with early therapeutic regimens (IRS-I or IRS-II). Current
therapy for patients with group II tumors results in 85% survival
long term, indicating that risk-based therapeutic strategies have
assisted with failure-free survival.124
Patients with group III disease have incomplete resection or
biopsy only prior to chemotherapy and irradiation. Wharam
and colleagues determined that predictors of failure-free survival in group III include tumor size less than 5 cm, primary
sites of orbit and bladder/prostate, and TNM staging equivalent to T1/N0Nx tumors in stage I or stage II. Since radiotherapy is important for local control of group III disease,
the incidence of local failure was stratified by radiotherapy
dosing (<42.5 vs. 42.5 to 47.5 vs. > 47.5 Gy) and was not
significantly different among these dose ranges.125
499
Approximately 15% of patients with RMS present with

metastases (group IV) at the time of diagnosis.104 Patients
in group IV have poor outcomes despite aggressive multimodality treatments, with only 25% expected to be free of
disease 3 years after diagnosis.104,105 A review of prognostic
factors and outcomes for children and adolescents with metastatic RMS in IRS-IV found that 3-year overall survival and
failure-free survival was improved if there were two or fewer
metastatic sites and the histology of the tumor was embryonal. Compared with patients without metastatic disease,
group IV patients in the IRS-IV study were more likely to
be older (median age 7 years vs. 5 years), had a higher incidence of alveolar histology (46% vs. 22%), had tumors
that were more invasive (T2: 91% vs. 49%) and larger
(>5 cm: 82% vs. 51%), a higher incidence of lymph node
involvement (N1: 57% vs. 16%), and had a greater proportion of extremity and truncal/retroperitoneal primary sites
(48% vs. 25%). This study concluded that not all children
with metastatic RMS have uniformly poor prognoses, suggesting that therapy should be tailored according to these
factors.126
Future clinical trials and a better understanding of the
molecular biology driving RMS tumor behavior may assist
with customized clinical therapies that will improve outcome and failure-free survival in patients diagnosed with
RMS.
expertconsult.com.
CHAPTER 36
Other Soft Tissue

Tumors
Andrea Hayes Jordan
Nonrhabdomyosarcoma Soft
Tissue Sarcoma in Children:
Background and Overview
------------------------------------------------------------------------------------------------------------------------------------------------
Approximately 8% of childhood malignancies are soft tissue

sarcomas. Half of these are nonrhabdomyosarcoma soft tissue
sarcomas (NRSTSs). There are more than 50 histologic types,
and genetic patterns are poorly understood. When surgical resection is feasible, 60% of patients are expected to achieve
long-term survival with or without radiation therapy.1 Patient
outcome is largely based on age, the presence of metastasis at diagnosis, and size and depth of the lesion. Here we focus on the
most common primary histologic types and differences in presentation and surgical treatment of childhood NRSTS and other
common pediatric soft tissue tumors.
The treatment for children and adolescents with NRSTS
has not previously been standardized, nor have there been
any pediatric cooperative group trials as for rhabdomyosarcoma (RMS). Because there are many histologic subtypes of
NRSTS, standardization of treatment is difficult. The first
risk-based prospective trial of NRSTS in children and adolescents will complete enrollment soon, with results anticipated
in 2013. In this trial, patients with NRSTS are treated as low,
intermediate, or high risk based on criteria previously ascertained in a thorough review of 121 patients by Spunt.2,3
In patients with surgically resected NRSTS, univariate analysis
revealed clear risk factors. Positive surgical margins (P
0.004), tumor size greater than or equal to 5 cm (P < 0.001),
invasiveness (P 0.002), high grade (P 0.028), and intraabdominal primary site (P 0.055) had a negative impact on
event-free survival (EFS). Multivariate analysis confirmed all
of these risk factors, except for invasiveness. Local recurrence
was predicted by intra-abdominal primary site (P 0.028),
positive surgical margins (P 0.003), and the omission of
radiation therapy (P 0.043). As expected, the biology of the
tumor, assessed by tumor size greater than 5 cm, invasiveness,
and high grade, predicted distant recurrences. Children and
adolescents with initially unresectable NRSTSs are a subgroup
with pediatric NRSTSs that is particularly high risk. These are
large tumors, greater than 5 cm, which involve critical neurovascular structures of the extremity, trunk, abdomen, or pelvis.
In these patients, the 5-year estimated overall survival and EFS
were 56% and 33%, respectively, and postrelapse survival was
poor, 19% despite multimodality therapy.4
In addition to the tumor being unresectable, age is a prognostic indicator in pediatric NRSTS. Patients less than 1 year of
age have an excellent prognosis, whereas the adolescents and
young adults have the worse prognosis compared with younger patients or older adults.2 A 34-year review of patients
treated at St. Jude Childrens Research Hospital (SJCRH)
revealed the overall 5-year survival estimate for children less
than 1 year of age was 92% compared with 36% in those
15 to 21 years of age. Patients between 1 and 15 years of
age had an intermediate survival of approximately 60%.
Survival after relapse was poor in all age groups less than
18 years, except those less than 1 year of age. The 5-year
estimate of postrelapse survival in patients less than 1 year
of age was 80% compared with the 15- to 25-years cohort
in which survival was 21%. The type of chemotherapy used
in these patients was variable; surgical excision was generally
completed for lesions less than or equal to 5 cm, and for most
patients, incisional biopsy was performed for lesions greater
than 5 cm, followed by chemotherapy, reexcision, and
radiation therapy or amputation.5
INFANTILE FIBROSARCOMA
Patients in the study above who were less than 1 year of age
had infantile fibrosarcoma (IF). This is a very rare form of
NRSTS that occurs primarily during the first year of life, but
can appear up to year 4. IF presents as a rapidly growing mass
in the trunk or extremities. It can erode bone and usually
reaches a large size.
Most cases of IF have a specific translocation t(12;15)
(p13;q25)68 leading to fusion of ETV6 (TEL), a member of
the ETS family of transcription factors, on chromosome
12p13, and NTRK3 (TRKC), which encodes a tyrosine kinase
receptor for neurotropin-39,10 on chromosome 15q25. Other
cytogenetic abnormalities include trisomy 11; random gains of
chromosomes 8, 11, 17, and 2011; deletion of the long arm of
chromosome 1712; and a t(12;13) translocation.13 The helixloop-helix dimerization domain of ETV6 fuses to the protein
tyrosine kinase domain of NTRK3. The fusion protein results
in ligand-independent chimeric protein tyrosine kinase activity with autophosphorylation. This leads to constitutive
501
502
PART III
activation of Ras-MAPK and P13K-AKT pathways through insulin

receptor
substrate-1,
which
is
tyrosinephosphorylated,1416 and through the activation of c-Src.17
The fusion protein also associates with TGF-beta II receptor,
which can be oncogenic by leading to inhibition of TGF-beta
receptor signals that mediate tumor suppression.18
Identical genetic findings have been reported in the cellular
variant of congenital mesoblastic nephroma, a microscopically
similar tumor of the kidney,19,20 and in secretory carcinoma
of the breast21 and acute myeloid leukemia,22 implying
oncogenesis by lineage-independent activation of kinaserelated signaling pathways.
SYNOVIAL SARCOMA
Synovial sarcoma (SS) and malignant peripheral nerve sheath
tumor (MPNST) are the most common pediatric NRSTSs. SS
is characterized by a very specific fusion gene 18[t(X;18)
(p11.2;g11.2)]. Its etiology is unknown.23 In evaluating the
three largest reviews of pediatric SS, common principles are evident. For children 0 to 16 years old and tumors less than 5 cm in
size, overall 5-year survival (OS) is 71% to 88%. In this group,
the addition of chemotherapy did not improve survival.
In patients 17 to 30 years old, the addition of chemotherapy does
improve metastasis-free survival. In patients with SS tumors
greater than 5 cm that are deep and invasive and without metastasis, OS is 50% to 75%, and chemotherapy responsiveness is
50% to 60%.24 It is clear that for SS survival does not depend
on surgical margins but depends on size (>5 cm) and local
invasiveness. Brecht and colleagues found event-free survival
was 92% and 56%, respectively, when SS tumors were less than
or equal to 5 cm or greater than 5 cm.24 Figure 36-1 shows the
leg of a child with synovial sarcoma that was not responsive
to chemotherapy and required resection down to the periosteum
of the tibia. Radiotherapy does have a role in this disease and
is recommended after marginal resection or before anticipated
marginal resection, such as the one pictured.23
MALIGNANT PERIPHERAL NERVE

SHEATH TUMOR
Malignant peripheral nerve sheath tumor (MPNST), also
called schwannoma or neurofibrosarcoma, usually arises in
proximity to nerve sheaths. MPNST develops in a preexisting
neurofibroma in approximately 40% of patients, particularly
those with neurofibromatosis type 1 (NF-1).25 In a review

of 171 patients the 5-year OS and progression-free survival
was 51% and 37%, respectively. Multivariate analysis revealed
absence of NF-1 and tumor invasiveness to be poor prognostic
variables. The overall response of the patients who received
neoadjuvant chemotherapy was 45%. Some partial responses
were seen in patients with initial unresectable disease, because
of neurovascular involvement.25 Neoadjuvant radiotherapy
failed to maintain or achieve local control in 45% of patients
(26 of 58). Neither chemotherapy nor radiotherapy produced
any statistically significant difference in outcome. This article
concluded by stating . . .complete surgical resection is the
mainstay of successful treatment.25 In another much smaller
series, the same patterns in outcome were seen.26
Surgical Approach
and Presentation
of Nonrhabdomyosarcoma
Soft Tissue Sarcoma
------------------------------------------------------------------------------------------------------------------------------------------------
Unlike rhabdomyosarcomas, NRSTSs are relatively chemoinsensitive. In the above pediatric studies and in adult multiinstitutional studies, the impact of chemotherapy on outcome
is minimal. In large American Joint Commission on Cancer
(AJCC) stage 3 tumors, overall survival was no different
whether or not chemotherapy was added to surgery and also
if neoadjuvant or adjuvant radiation therapy was added.27
Complete surgical excision provides the best outcome.
Patients usually present with a painless mass, sometimes
identified after a recent episode of trauma. Pediatric patients
who have an extremity or trunk mass that is greater than
5 cm, should have a magnetic resonance imaging (MRI)
examination, followed by core needle or open biopsy. If
NRSTS is identified and no mutilating limb-sparing surgical
excision is feasible, resection should be completed. If margins
are microscopically positive, postoperative radiotherapy
should be given in high-grade tumors and tumors larger than
5 cm. Low-grade tumors that are less than 5 cm can be reexcised or just watched closely. If surgical excision is not feasible
without amputation or severe morbidity, whether less than or
greater than 5 cm, preoperative chemotherapy and radiotherapy should be administered. If surgical excision is feasible, but
R1 resection is anticipated, the type of radiotherapy, whether
FIGURE 36-1 A-C, Magnetic resonance (MR) image of a child with synovial sarcoma abutting the tibia. Neoadjuvant chemotherapy was not successful in
reducing the size of the tumor. Marginal resection with postoperative radiation or brachytherapy is a preferred alternative to amputation.
CHAPTER 36
preoperative or postoperative brachytherapy, proton beam

therapy, or external beam therapy, should be discussed with
the radiation oncologist, with the goal in pediatric extremity
tumors to avoid the growth plate in younger patients who
are still growing. In tumors less than 5 cm, complete surgical
excision with negative microscopic margins is the goal. In the
case of unexpected malignant pathology, primary reexcision is
recommended. For all NRSTSs, negative microscopic margins
should be achieved; however, there is no consistent reliable
evidence to establish the appropriate width of the margins.
NRSTSs are graded histologically to help predict outcome.
Grade 1 is any NRSTS with low malignant potential, such as
infantile fibrosarcoma, with mitotic activity less than 5 mitoses
per high-powered field (HPF). NRSTSs with tumor necrosis
less than 15% and mitotic activity of 5 to 10 mitoses per
HPF are graded 2, and specific histologic subtypes with
known aggressive behavior and/or any sarcoma with tumor
necrosis of more than 15% or mitotic activity of more than
10 mitoses per HPF are graded 3.28
Cytotoxic chemotherapy (Adriamycin, ifosfamide, vincristine,
dactinomycin, etc.), will be effective, at best, in 45% to 50% of
patients from the evidence we have to date.4 (This does not
include targeted therapy, because there are not yet sufficient
data to analyze at this time.) Very close observation by imaging
is warranted if neoadjuvant chemotherapy is chosen, because
an increase in tumor size may preclude limb-sparing, nonmutilating surgery, and an abdominal or pelvic tumor may become
unresectable.
Sentinel lymph node biopsy, although recommended for
rhabdomyosarcoma to evaluate normal-appearing lymph nodes,
is only recommended in histologic subtypes of NRSTS that have
high risk of lymph node metastasis. These include epithelioid
sarcoma and clear cell sarcoma, which have an approximate
incidence of lymph node metastasis of up to 30%. Synovial
sarcoma metastasizes to the lymph nodes about 15% of the time.
Computed tomography (CT) scan of the chest is a necessary part of the workup to exclude lung metastasis. Lung metastasis occurs in approximately 30% of patients with NRSTS.
Because NRSTSs are relatively chemoinsensitive, surgical
resection of lung metastasis is recommended. Thoracotomy
is the recommended approach in order to palpate the lung
for any tumors that may have been missed on imaging.
Desmoplastic Small Round

Cell Tumor
------------------------------------------------------------------------------------------------------------------------------------------------
Desmoplastic small round cell tumor (DSRCT) is a malignant

neoplasm in the soft tissue sarcoma family that arises from the
peritoneal surface of the abdomen and pelvis. No more than
200 cases have been reported worldwide since the disease
was first described in 1989 by Gerald and Rosai29 and Ordonez.30 The tumor is most prevalent in young white males.2930
Presenting symptoms include abdominal pain, constipation,
and abdominal distension with ascites. Overall survival is
approximately 30% to 55% despite chemotherapy, radiotherapy, and aggressive surgical resection.31,32 Because most
DSRCT patients present with multiple abdominal tumor
implants (Fig. 36-2), microscopic tumor cells can be left behind, despite the complete resection of dozens to hundreds
of tumors. The most widely accepted standard of care for
OTHER SOFT TISSUE TUMORS
503
FIGURE 36-2 Desmoplastic small round cell tumor in the omentum of a

5-year-old boy after six cycles of chemotherapy. Peritoneal disease has similar appearance. This child had 402 nodules removed at this operation.
DSRCTwas multimodality therapy with the P6 regimen: cyclophosphamide, doxorubicin, and vincristine, alternating with
ifosfamide and etoposide for seven total courses,31 followed
by aggressive debulking surgery to remove all visible
disease.32 It is clear that without complete resection of all visible disease survival is poor.32 Hyperthermic intraperitoneal
chemotherapy (HIPEC) is a new therapeutic modality recently
used in children; its results are promising, but studies are
ongoing. Hyperthermia and chemotherapy have synergistic
cytotoxicity that is of value in the treatment of microscopic
disease in adult carcinomas. HIPEC has been applied
successfully in adults with extensive peritoneal disease,
commonly observed with mesothelioma, appendiceal, colon,
and gastric carcinoma.3337 A recent publication shows that
DSRCT can now be treated safely with aggressive cytoreductive surgery followed by (HIPEC) in children.38 The study
included 23 pediatric adolescent and young adult patients
with DSRCT. HIPEC was compared with standard chemotherapy, radiation therapy, and surgical debulking. The patients
were mostly males (96%). The age of the HIPEC patients
ranged from 5 to 25 years of age. Complete resection (CR0)
to less than 1.0-cm tumor size was achieved in all 8 patients
who underwent HIPEC. Operative times ranged from 7 to 16
hours. Figure 36-3 shows the setup used in the operating
room to deliver HIPEC. In the pediatric patients, the estimated
12-month disease-free survival (DFS) rate was 53% for the
HIPEC group, compared with 14% for the non-HIPEC group.
Median 3-year survival in this small group of patients was 29%
with chemotherapy and radiotherapy alone, compared with
71% in the HIPEC with cytoreductive surgery group. The
severe morbidities that occurred were partial bowel obstruction managed nonoperatively, prolonged ileus/gastroparesis,
transient renal insufficiency, and one patient developed
cardiomyopathy secondary to resection of more than 3 kg of
tumor, causing release of tumor necrosis factor. HIPEC is an
option in treating this rare tumor.38
Desmoid Tumors
------------------------------------------------------------------------------------------------------------------------------------------------
Desmoid tumors are very different than DSRCT. These are

intermediate-grade sarcoma-type tumors that are locally very
aggressive and can be fatal, but usually do not metastasize. Desmoid fibromatosis is a mesenchymal neoplasm. It is encountered
504
PART III
Drugs
Thermistor
Cardiotomy
Reservoir
Heat
Exchanger
Monitor
FIGURE 36-3 Setup for hyperthermic intraperitoneal chemotherapy (HIPEC) therapy for children with
sarcomatosis after cytoreductive surgery.
Water Heater
in two settingswithin the context of familial adenomatous

polyposis (FAP) and sporadically.39 Here we focus on the
sporadic group. Desmoid tumors can arise in any body site
and are much more common in women. Surgery has been the
therapeutic mainstay, but radiotherapy plays an important role
in treatment as do systemic therapies, such as the tamoxifen
and sulindac combination and nonsteriodal anti-inflammatory
drugs (NSAIDs).4046 Desmoids have a very unique course
in that they can recur locally and can be more aggressive or
regress spontaneously. However, they have no capacity for metastasis.39 Resecting recurrent tumors can be potentially mutilating. Some large retrospective studies42,47,48 demonstrated
that microscopically positive (or grossly positive) margins were
Roller
Pump
predictive of increased frequency of local recurrence on

retrospective multivariate analysis, although radiation improved
outcome in one study. Other studies40,4951 have failed to
demonstrate an effect of microscopic margin on recurrence.
Some of these differences may result from the mixture of disease
sites, pattern of application of adjuvant radiotherapy, and
selection of patients treated by surgical approach. In the end,
surgical therapy must be tailored to what is achievable in terms
of margins with preservation of functional status for the individual patient.39 Incomplete resection or positive microscopic
margins in desmoid tumors should be treated with adjuvant radiotherapy. Figure 36-4 provides a helpful algorithm to
follow.
Positive
Observation
Margin
assessment
(microscopic)
Resection
Significant
morbidity?
Adequate
response
Yes
Primary/
recurrent
desmoid
Yes
*Radiation
therapy
Resectable?
No
Extraabdominal
Negative
Observation
No
Resectable
without
morbidity?
Resection
Yes
Inadequate
response
*Systemic
therapy
Adequate
response
Location
Intraabdominal
Observation
Systemic
therapy
Inadequate
response
Reassess with progression/recurrence
No
Radiation
therapy
?Radiation
therapy
FIGURE 36-4 General treatment protocol for desmoid tumors at the University of Texas M.D. Anderson Cancer Center. The route of initial observation for
certain cases, to avoid overtreatment advocated by some, is depicted in gray. Given the propensity for progression on treatment and local recurrence, all
treatment pathways ultimately end in observation. *Radiation therapy can be preceded, and even precluded, by systemic therapy in certain cases of initially
unresectable extraabdominal desmoid tumors.
CHAPTER 36
Dermatofibrosarcoma
Protuberans
------------------------------------------------------------------------------------------------------------------------------------------------
Dermatofibrosarcoma protuberans (DFSP) is a relatively common soft tissue tumor. Its peak age is in young adulthood,
but it is frequently present in children and at birth. DFSP occurs
primarily on the trunk and extremities. It can present as a plaque on the skin or in a more diffuse multinodular pattern.52,53
The latter is more common in children. Pigmented dermatofibrosarcoma, giant cell fibroblastoma, and fibrosarcoma can
arise in DFSP.54,55
Dermatofibrosarcoma protuberans has a reciprocal translocation, t(17;22)(q22;q13.1), resulting in fusion of the genes
COL1A (encoding the alpha 1 chain of collagen type 1, a heterotrimer) on 17q21-22 and PDGFB1 (encoding the beta
chain of platelet-derived growth factor, a homodimer) at
22q13.55,56 The same fusion is also seen in supernumerary
ring chromosomes derived from t(17;22),57 which are found
in adult cases of dermatofibrosarcoma. Fusion gene transcripts can be detected by reverse transcriptasepolymerase
chain reaction (RT-PCR).53,58 This is not usually required
for diagnosis but might be useful in guiding therapy, especially
for superficial fibrosarcomas.
Dermatofibrosarcoma protuberans has a high local
recurrence rate, especially if incompletely excised, and can
metastasize in 5% of cases, usually after multiple local
OTHER SOFT TISSUE TUMORS
505
recurrences. Therefore complete excision with negative margins is crucial, and 2- to 3-cm margins are recommended. However, in areas such as the head and neck, lesser margins are
acceptable.
Platelet-derived growth factor receptor (PDGFR) is a receptor
tyrosine kinase, which in dermatofibrosarcoma protuberans is
constitutively activated by autocrine or paracrine mechanisms
as a result of overproduction of its ligand platelet-derived growth
factor-beta (PDGFB),59 leading to cellular proliferation.60 This
has suggested the use of the tyrosine kinase inhibitors imatinib61
and, more recently, sunitinib or sorafenib in locally advanced or
metastatic disease,6263 but fibrosarcomatous variants without
the translocation do not respond64,65 so that genetic analysis
is indicated before targeted therapy. In the only multicenter
Phase 2 study published to date, imatinib was found to be effective preoperative therapy in 36% of patients (n 25) by reducing
tumor size by an average of 20%. Response was measured by
physical exam, ultrasonography, and MRI. Decrease in average
diameter by 1 cm on physical exam, 1 cm by ultrasonography,
and 2 cm by MRI were observed, respectively. In 21 of 25
patients, the fusion gene COL1A1-PDGFB was detected. Therefore when DFSP is located in places where a decrease in size provides a significant advantage in wound closure, neoadjuvant
imatinib is a viable option.
expertconsult.com.
CHAPTER 37
Teratomas and
Other Germ Cell
Tumors
Frederick J. Rescorla
Pediatric germ cell tumors are rare tumors that are unique due
to their varied clinical presentation and locations. Approximately 20% of pediatric germ cell tumors are malignant,
and they represent 1% to 3% of all malignant tumors in childhood and adolescence.1,2 Three features distinguish these
childhood tumors from many other malignancies as well as
their counterparts: In children, the extragonadal tumor site
is more common than the gonadal site, whereas in adults, only
10% are at extragonadal sites; yolk sac tumor is the predominant malignant histology, and a serum marker (alpha fetoprotein, AFP) exists to follow response to therapy and monitor for
recurrent disease; and the introduction of modern chemotherapy with cisplatin and bleomycin significantly increased
survival for affected children and has allowed neoadjuvant
therapy with vital organ preservation in initially unresectable
cases.
Abnormal or arrested migration of primordial germ cells
results in deposition of cells in the sacrococcygeal region,
retroperitoneum, mediastinum, and pineal gland of the brain,
resulting in the potential of extragonadal germ cell tumors
at these sites. Whereas in adults 90% of germ cell tumors are
at gonadal locations, in childhood, the extragonadal site is
more common until puberty, at which time the gonadal sites

are more common. The totipotential nature of these cells
results in a wide variety of histologic patterns, and in addition,
one quarter of pediatric tumors have more than one histologic
component.2 The management of these tumors is dependent
upon complete surgical resection at diagnosis or after neoadjuvant therapy, accurate and thorough histologic examination,
and selective use of chemotherapy. Prior to the late 1970s,
the survival of advanced-stage tumors was dismal; however,
Einhorns introduction of cisplatin, vinblastine, and bleomycin
for disseminated testicular cancer in 1977 changed the treatment of all germ cell tumors with dramatic results.3 Subsequent studies validated the use of chemotherapy in a
neoadjuvant fashion, thus allowing vital organ preservation
in advanced cases with frequent massive tumor shrinkage.
The role of the surgeon in determining resectability and performing a proper staging operation is vital.
Current therapy within the Childrens Oncology Group
(COG) is risk based: with surgery alone for stage 1 testes
and ovary tumors and all immature teratomas, with anticipated survival of 95% to 100%; surgery and chemotherapy
for all remaining gonadal tumors (except stage IV ovary)
and low-stage (I-II) extragonadal, with anticipated survival
of 90% to 100%; and surgery and intensive chemotherapy
for high-risk (stage III-IV) extragonadal and stage IV ovary,
with survival between 75% and 90%, depending on site
and stage.
Embryology and Classification

------------------------------------------------------------------------------------------------------------------------------------------------
Primordial germ cells arise near the allantois of the embryonic

yolk sac endoderm and are evident at the fourth fetal week.
They migrate along the midline dorsal mesentery to the genital
ridge, arriving by the end of the sixth fetal week. The migration of the germ cells appears to be mediated by the c-KIT
receptor and stem cell factor; the latter is expressed in increasing levels from the yolk sac to the genital ridge.4,5 Arrested
migration is presumed to account for the extragonadal locations
in the normal path of the germ cells (retroperitoneum), whereas
aberrant migration results in cells at other extragonadal sites
(pineal, sacrococcygeal).
CLASSIFICATION
Teilum6 proposed the germ cell origin of gonadal tumors, and
the pathway of differentiation is listed in Figure 37-1. Seminoma (or dysgerminoma) is a primitive germ cell tumor that
lacks the ability for further differentiation. It is unusual in
childhood and occurs most frequently in the mediastinum,
pineal gland, and at the gonadal sites during the adolescent
years. Embryonal carcinoma is composed of cells capable of
further differentiation into embryonic or extraembryonic tumors. Teratomas are the most common germ cell tumor and
are composed of elements from one or more of the embryonic
germ layers and contain tissue foreign to the anatomic site of
origin.7,8
Mature and immature teratomas are considered benign
lesions. It is, however, imperative to have a thorough and
accurate pathologic review, because 25% of germ cell tumors
in childhood are mixed tumors with more than one histologic
507
508
PART III
Normal fetal yolk sac

Germ cell production
Migration
Normal
Gonads
Abnormal
Extragonadal germ cell

Neoplastic cell
Suppressed
differentiation
Seminoma/
Dysgerminoma
Differentiation
Embryonal carcinoma
Embryonic
Mature or
immature teratoma
Extraembryonic
Choriocarcinoma
Yolk sac tumor
(endodermal sinus tumor)
FIGURE 37-1 Classification system for development of germ cell tumors.
component.2 Certain sites are more likely to have mixed tumor

histology, with ovary (46%) and mediastinal (61%) the most
common.9,10 Mature teratomas contain well-differentiated tissue, whereas immature teratomas contain neuroectoderm and
are graded between 1 and 3 based on the number of low-power
fields of primitive neuroepithelium.11 There has been debate
about the treatment of immature teratomas. Many adult reports
of ovarian tumors have considered grade 3 lesions malignant,
and these patients have been treated with chemotherapy.
A review of childhood immature teratomas demonstrated an
association between high-grade immaturity and the presence
of microscopic foci of endodermal sinus tumor,12 with malignant foci observed in 83% of grade 3 immature teratomas as
the only risk factor for recurrence.13
Yolk sac tumors (endodermal sinus) and choriocarcinoma
are well-differentiated, highly malignant tumors. Yolk sac is
the more common histology in childhood and occurs primarily in the sacrococcygeal region, ovary, and prepubertal testes.
Genetics and Risk Factors

------------------------------------------------------------------------------------------------------------------------------------------------
Germ cell tumors demonstrate a bimodal age distribution with

peaks at 2 and 20 years of age. Pediatric germ cell tumors differ
in several aspects from their adult counterparts. Pediatric yolk
sac tumors are more likely to have DNA ploidy, whereas adolescent and adult germ cell tumors are usually aneuploid.14 In
children younger than 4 years of age, the primary malignant
germ cell tumor is yolk sac, and these are diploid or tetraploid;
the teratomas are diploid with normal karyotypes and are
benign.1517 Childhood yolk sac tumors have also demonstrated deletion of chromosomes 1p and 6q in 50% of specimens.18 In addition, a smaller percentage demonstrates
amplification of c-MYC. The isochromosome i (12p), which
is identified in most pubertal or postpubertal testes tumors,
is not observed in prepubertal tumors. Gains of 12p have been
noted in malignant ovarian germ cell tumors but not in ovarian immature teratomas.19
The presence of intersex disorders is a known risk factor for
gonadoblastoma, an in-situ germ cell tumor with the ability to
differentiate into dysgerminoma, immature teratoma, yolk
sac tumor, or choriocarcinoma.20 One risk group includes
testosterone deficiency, androgen insensitivity syndromes,
and 5-alpha-reductase deficiency, which are androgendeficient males. The presence of any portion of a Y chromosome is considered a risk factor in these children.21 Risk of
malignancy in androgen insensitivity is 3.6% at age 20 and
22% at age 3022; in view of this, gonadectomy usually in adolescence, is recommended. Gonadal dysgenesis is associated
with a risk of malignancy of 10% at age 20 and 19% at age 30.
Undescended testes have an increased risk of malignancy,
with the rate highest for intraabdominal testes. Approximately
0.4% of all males have undescended testes, however, it is observed in 3.5 to 12% of the testicular cancer population.23
One study noted that although intraabdominal testes only account for 14% of undescended testes, they account for nearly
50% of tumors in the undescended testes group. The effect of
orchiopexy on the risk of testes cancer is not known, and 20%
of the tumors in patients with undescended testis occur in the
descended testis.24 Seminomas occur in a higher percentage of
undescended testes (60%) compared with the descended
testes tumors (30% to 40%),25 and one study observed that
orchiopexy decreases the incidence of seminoma.26 The early
identification of these children is important, because a recent
report noted a 2-year-old boy with a large yolk sac tumor in
an intraabdominal testis with lymph node involvement.27
Surgery and chemotherapy yielded a successful outcome.
Risk-Based Therapy
------------------------------------------------------------------------------------------------------------------------------------------------
The survival of patients with advanced-stage germ cell tumors

was poor prior to the introduction of modern chemotherapy,
with most survivors having had low-stage surgically excised
tumors. Surgery and chemotherapy consisting of vincristine,
actinomycin, cyclophosphamide, and doxorubicin was the
primary therapy in the 1960s and 1970s.28 In 1975, Samuels
and colleagues29 introduced bleomycin with vinblastine for
advanced-stage testicular tumors, and in 1977, Einhorn and
Donohue3 reported success with cisplatin, vinblastine, and
bleomycin in disseminated testicular cancer. This therapy
dramatically transformed the treatment of germ cell tumors.
Even after the introduction of cisplatin-based regimens, the
early results in children were poor. A report from the Childrens Cancer Group (CCG) of children treated between
1978 and 1984, using cisplatin and bleomycin alternating
with other agents (cyclophosphamide, dactinomycin, and
doxorubicin), reported 4-year survival and event-free survival
(EFS) of 54% and 49%, respectively, with ovarian tumors
higher at 67% and 63%, respectively, and extragonadal tumors
at 48% and 42%, respectively.30 The lower survival in the early
study may have been due to the inclusion of less effective chemotherapy that lengthened the intervals between the courses
of the more effective cisplatin and bleomycin.
The subsequent CCG/Pediatric Oncology Group (POG)
intergroup studies conducted between 1990 and 1996 used
only cisplatin, etoposide, and bleomycin (PEB). The overall
6-year survival was 95.7% for stage I and II ovarian and
CHAPTER 37
testes and 88.9% for stage III-IV gonadal and stage I-IV
extragonadal.3133 The higher-risk group (stage III-IV gonadal
and stage I-IV extragonadal) were stratified to either standard
or high-dose cisplatin, and the overall survival was not different between the groups, but the toxicity was higher with the
high-dose cisplatin, and it has therefore not been incorporated
in the current study.
Based on these past studies, the current COG protocol for
malignant germ cell tumors is risk based (Fig. 37-2). The overall goal is to maintain the excellent survival from the past
intergroup study while decreasing the toxicity of the chemotherapy. Mature teratoma is considered to be a benign lesion,
and these tumors are not entered on the current protocol.
Immature teratomas at all sites are treated with surgery and
observation. The 3-year survival for immature teratomas on
the last study was 93% among 73 patients with immature teratoma, and four of the five recurrences were salvaged with
platinum-based chemotherapy.13,34 Stage I ovarian and testes
tumors are treated with surgery and observation, although this
portion of the protocol is currently suspended (see Ovary section). Stage II-III ovary and stage II-IV testes currently receive
three cycles of PEB administered during 3 days compared with
four cycles during 5 days on the prior study, thus resulting in
significantly less total chemotherapy. Higher-risk tumors
(stage IV ovary and stage III-IV extragonadal), are currently
not a part of a protocol but would received PEB.
Testes
------------------------------------------------------------------------------------------------------------------------------------------------
CLINICAL PRESENTATION AND INITIAL

EVALUATION
Testicular germ cell tumors in children are one of the rarer
germ cell tumor types, with an incidence of 0.5 to 2.0 per
100,000.35 The bimodal age distribution of testes tumors,
with a small peak in the first 3 years of life and a much larger
peak in young adults, suggests a difference in the tumors of
these age groups. The malignant germ cell tumors in the younger group are predominantly yolk sac tumors, whereas most
adolescent and adult testes tumors are seminomas and mixed
tumors. Several other factors provide evidence of differences
between pediatric and adult testes tumors. Intratubular germ
cell neoplasia (ITGCN), which is a carcinoma in situ, is commonly identified in adults with malignant germ cell
tumors but does not occur in association with prepubertal
yolk sac tumor. Adult testes tumors usually have a chromosomal gain of the short arm of chromosome 12p (isochromosome 12p), whereas this is not seen in prepubertal
yolk sac tumors.
Testicular tumors are rare in boys prior to puberty, and during this time nongerm cell Sertoli tumors and paratesticular
rhabdomyosarcomas are more common, whereas germ cell tumors predominate in pubertal and adult males. Paratesticular
neuroblastoma has also been reported arising from an embryonic adrenal rest along the spermatic cord.36,37 Although it is
difficult to determine the incidence of malignancy in prepubertal testes tumors, several reports would suggest that it is
less common than in adults. In one large series,38 74% of
all tumors were benign, with teratoma accounting for 48%
and yolk sac tumors only 5%. This has affected the initial surgical evaluation of these children in order to avoid unnecessary radical orchiectomy.
TERATOMAS AND OTHER GERM CELL TUMORS
Low risk
Stage 1 ovary
Stage 1 testes
Immature teratoma
Intermediate risk
Stage IIIII ovary
Stage IIIV testes
509
Surgery alone
COG, AGCT 0132
Stage III extragonadal
Surgery and
Chemo-PEB x 3
COG, AGCT 132
High risk
Stage IIIIV extragonadal
Stage IV ovary
Surgery and
PEB
FIGURE 37-2 Low- and intermediate-riskbased scheme for pediatric

germ cell tumors. Childrens Oncology Group AGCT 0132, opened
November 2003.
Most testicular tumors present as a painless scrotal mass. In

the intergroup CCG/POG study (1990 to 1996)31 of malignant
testes tumors, 76% of the stage 1 boys presented with a testicular mass and 17% with generalized scrotal swelling. The preoperative diagnosis was tumor in 79%, hydrocele in 11%,
hernia in 3%, and acute scrotum or torsion in 3%.
Preoperative workup includes a thorough physical examination, looking for signs of androgenization as well as metastatic disease. Metastatic disease is relatively uncommon in
prepubertal testes cancer, but if present, is usually in the retroperitoneum or chest. Testicular ultrasonography is useful to
identify extratesticular lesions and may be useful to identify
or raise the suspicion of a teratoma. Benign testes tumors tend
to be well circumscribed with sharp borders and decreased
blood flow on Doppler studies.39 Preoperative AFP levels
should be obtained, and this level was elevated in 98% of
the children with malignant tumors in the most recent study.31
If the preoperative diagnosis is a testicular malignancy (elevated AFP), it is reasonable to obtain an abdominal computed
tomography (CT) scan, because the presence of enlarged
nodes after an inguinal exploration can be due to either a
reactive or malignant process.
OPERATIVE MANAGEMENT
The standard approach consists of an inguinal incision, with
initial control of the vessels at the level of the internal inguinal
ring with subsequent mobilization of the testes. A preoperative elevation of AFP indicates the presence of yolk sac tumor
and thus precludes consideration of testes-sparing surgery,
and a radical orchiectomy is performed with ligation of the
cord at the internal ring. If the AFP is normal, there is a much
greater chance that the mass represents a benign lesion, and in
these instances, the field can be draped off and the tunica
opened. Enucleation is often possible, leaving a large amount
of residual normal testes.40 If frozen section analysis reveals a
benign lesion, the tunica is closed, and if malignant, an orchiectomy is completed. Unfortunately, this is not always possible, and in a recent review from the U.K. Childrens Cancer
Group, 48 of 53 boys with mature or immature teratoma
had radical orchiectomy.41 There were no recurrences in the
five treated with enucleation. Bilateral testes-sparing surgery
510
PART III
Stage
Extent of disease
II
III
IV
Limited to testis (testes), completely resected by high

inguinal orchiectomy; no clinical, radiographic or histologic
evidence of disease beyond the testes.
Transscrotal biopsy; microscopic disease in scrotum or high
in spermatic cord (<5 cm from proximal end). Tumor
markers fail to normalize or decrease with an appropriate
half-life.
Retroperitoneal lymph node involvement, but no visceral or
extraabdominal involvement. Lymph nodes > 4 cm by CT;
or > 2 cm and < 4 cm with biopsy proof.
Distant metastases, including liver.
FIGURE 37-3 Current Childrens Oncology Group staging system for

childhood testes cancer.
has been reported for testes teratoma.42 A more recent report

noted no atrophy or recurrence with enucleation in a large
group of benign testes tumors.43
POSTSURGICAL TREATMENT
Testicular teratomas are benign lesions and are treated with
enucleation, if possible, and then postoperative observation.
Testicular immature teratomas are also benign germ cell tumors, and surgery alone (enucleation if possible) is definitive
treatment. Higher-grade immature teratomas are, however, associated with yolk sac tumors. In a (CCG/POG) review, grade
1 and 2 immature teratomas were not associated with yolk sac
tumors, whereas 2 of 3 grade 3 lesions were associated with
yolk sac tumors.13
Yolk sac tumor is the primary malignant prepubertal testes
cancer. The current staging is noted in Fig. 37-3. The role of
surgery alone for stage I testes tumors was reported in the
1980s44 and confirmed in an initial small series.45 The U.K.
Childrens Cancer Study Group46 and the Testicular Tumor
Registry of the Section of Urology of the American Academy
of Pediatrics,47 in larger series (73 and 181 children, respectively), confirmed the safety of surgery alone for stage I malignant testes tumors.
The intergroup trial of testes cancer (CCG/POG; 1990
1996)31 confirmed the excellent outcome with stage 1 testes
tumors treated with surgery alone (Table 37-1). This study
of 63 boys (median age 16 months) reported AFP elevation
in 98%. In patients with the preoperative diagnosis of tumor,
the surgical guidelines were followed in 84% of boys but were
followed in only 27% with a nontumor diagnosis. Although
overall adherence to surgical guidelines did not affect outcome, scrotal violation was associated with a 75% recurrence
rate compared with 15.5% in those without scrotal violation.
All recurrences were successfully treated with surgery and
chemotherapy.
Stage 2 boys on the CCG/POG study included only 17 patients, and 11 were stage II because of a transcrotal procedure.32 Survival was excellent (see Table 37-1) with surgery
and chemotherapy. Higher-stage 3 and 4 boys received surgery
and were then randomized to standard or high-dose cisplatin,
both with etoposide and bleomycin.33 Sixteen were recurrences from stage 1 disease (median age 3.1 years), and the
rest were newly diagnosed and much older (median age
16 years). Despite the advanced disease, outcome was
TABLE 37-1
Survival for Testes Cancer, POG/CCG 9048/8891; 9049/8882,
1990-1996
Stage
Treatment
6-Year
EFS (%)
I
II
III
IV
63 S
17
17
43
78.5
S PEB 4
S HDP/EB vs. PEB
S HDP/EB vs. PEB
100
100
94.1
88.3
6-Year
Survival (%)
100
100
90.6
CCG, Childrens Cancer Group; EB, etoposide and bleomycin chemotherapy;

EFS, event-free survival; HDP, high-dose platinum chemotherapy; PEB,
platinum, etoposide, and bleomycin chemotherapy; POG, Pediatric
Oncology Group; S, surgery.
excellent (see Table 37-1). The toxicity with high-dose cisplatin was significant without added benefit, and it has therefore been eliminated from current protocols.
The current protocol of the Childrens Oncology Group is
designed to reduce the total dose and days of chemotherapy
(Fig. 37-2). As noted in the staging, if the retroperitoneal
nodes are greater than 4 cm in size, it is assumed to be due
to tumor, whereas nodes between 2 and 4 cm require biopsy
to confirm status. There is no role for retroperitoneal lymph
node dissection in prepubertal yolk sac tumors at diagnosis
and simple biopsy is adequate.
Ovary
------------------------------------------------------------------------------------------------------------------------------------------------
CLINICAL PRESENTATION AND EVALUATION

Ovarian tumors are the most common site for germ cell tumors in children and adolescents. Eighty to 90% percent of
all ovarian masses are benign (epithelial cyst, mature teratoma), often with predominant cystic components.10,48 Presenting symptoms often include pain and gradual onset of
lower abdominal fullness. Approximately 10% present with
an acute abdomen secondary to torsion or tumor rupture.10
Of all girls presenting with ovarian torsion, only 1.8% to
3% are malignant tumors; however, 33% are benign tumors,
including teratoma and cystadenoma.48
In nonacute cases, preoperative evaluation should include
assessment of AFP and beta-HCG, as well as ultrasonography
and usually abdominal and pelvic CT scan. Unfortunately, reliable tumor markers are absent in many tumors. Germinoma
is present in one third of malignant tumors, and they have normal markers or mild elevation of beta-HCG, and embryonal
carcinomas have normal markers.2 Benign lesions are primarily cystic, and a 2% risk of malignancy in cystic lesions is frequently quoted based on adult series.4951 This, however, is
also unreliable, because in the recent intergroup study from
the Childrens Oncology Group (COG), 57% of malignant tumors had cystic components.10 A recent study attempting to
identify risk factors noted that markers were elevated in only
54% of malignant tumors. The best predictors were a mass
with solid characteristics and a mass greater than 8 cm in
diameter.52 They also noted as, in other series, that girls
between 1 and 8 years have the greatest incidence of malignancy. In view of these observations, great care should be
taken to perform a proper staging operation with lesions with
solid components.
CHAPTER 37
1. Collect ascites or peritoneal washings for cytology

2. Examine peritoneal surface and liver; excise suspicious lesions
3. Unilateral oophorectomy
4. Examine contralateral ovary and biopsy if suspicious lesion
5. Examine omentum and remove if adherent or involved
6. Inspection of retroperitoneal lymph nodes, biopsy of enlarged
nodes
FIGURE 37-4 Operative procedure for malignant ovarian germ cell
tumor.
Stage I: Limited to ovary (ovaries) peritoneal washings negative;

tumor markers normal after appropriate half-life decline
(AFP 5 days, HCG 16 hours).
511
TABLE 37-2
Event-free Survival (EFS) and Survival in Pediatric Ovarian Germ
Cell Tumors, POG/COG Intergroup Study 1990-1996
Stage
Treatment
I
II
III
IV
41
16
58
16
S
S
S
S
PEB
PEB
HDP/EB vs. PEB
HDP/EB vs. PEB
6-Year
EFS (%)
6-Year
Survival (%)
95
87.5
96.6
86.7
95.1
93.8
97.3
93.3
CCG, Childrens Cancer Group; EB, etoposide and bleomycin chemotherapy;

HDP, high-dose cisplatin chemotherapy; PEB, cisplatin, etoposide, and
bleomycin chemotherapy; POG, Pediatric Oncology Group; S, surgery.
Stage II: Microscopic residual; peritoneal washings negative for

malignant cells, tumor markers positive or negative.
Stage III: Lymph node involvement; gross residual or biopsy only;
contiguous visceral involvement (omentum, intestine,
bladder); peritoneal washings positive for malignant cells;
tumor markers positive or negative.
Stage IV: Distant metastases, including liver.
FIGURE 37-5 Childrens Oncology Group ovarian staging system. AFP,
alpha fetoprotein; HCG, human chorionic gonadotropin.
The staging procedure endorsed by COG is listed in

Figure 37-4 and the current staging system in Figure 37-5.
The importance of an accurate and complete staging procedure and accurate pathologic evaluation cannot be overemphasized. The recent COG intergroup study of 131 girls
reported positive ascites/peritoneal fluid in 23 of 100 girls,
and 5 of these would have otherwise been stage I tumors.10
This is particularly relevant, because the current low- and
intermediate-risk COG study manages stage I girls with surgery alone.
The survival rates of children in the most recent intergroup
study is listed in Table 37-2. The current therapy for ovarian
malignant tumors is noted in Figure 37-2. In the most recent
study,33 the results for stage IV ovarian tumors did not allow
them to be included in the current low- and intermediate-risk
COG study (AGCT 0132) using reduced chemotherapy.
Some tumors are noted with invasion into surrounding
structures, and in these cases, recommendations are for initial
biopsy, neoadjuvant chemotherapy, and delayed resection.
Bilateral ovarian tumors were observed in 8% of girls on the
recent study, and 4 of the 11 contralateral tumors were benign
teratomas. The current recommendation for bilateral tumors
is to attempt ovarian preservation, if possible, on the least involved side, attempting to find a plane of demarcation between the tumor and normal ovarian tissue. The larger
tumor should be removed and sent for frozen section. If the
first side is malignant and the contralateral side is greater than
10 cm, it should also be removed.
The treatment algorithm for malignant ovarian tumors is
surgery and observation for stage I and surgery and chemotherapy for higher-stage tumors (see Fig. 37-2). The surgeryonly arm was based on a German and French series of a total
of 39 girls with stage I tumors treated with surgery alone
who experience a 67% EFS with salvage of 12 of 13 recurrences
with chemotherapy for an overall survival of 97.4%.53,54 The
CCG/POG intergroup study noted excellent results in girls with
stage I immature teratoma, with microscopic yolk sac tumor
treated with surgery alone as well as stage I girls treated with

surgery and PEB.55 The current low-risk arm of the study
has been closed because of a higher than expected recurrence
rate in stage I ovarian tumors. These girls had a less than 70%
three-year EFS, thus leading to suspension of the trial; however,
with salvage chemotherapy, they have an overall survival of
over 95%.56
Laparoscopy has been widely used for ovarian cystic disease, and the application of this for malignant procedures
has been controversial. The primary concern is adequate completion of the staging procedure (potential understaging) and
avoidance of intraperitoneal spill or tumor rupture, which
could upstage a stage I to a stage II tumor. The COG germ cell
committee and others10,57 recommend laparotomy for known
malignancy; however, this is difficult to determine preoperatively, although preoperative elevated markers and a large
solid mass are very suggestive of malignancy. A recent French
study suggested that size greater than 7.5 cm or predominately
solid components predicted malignancy and thus required
laparotomy.57
Most primarily cystic lesions, some of which are large, are
benign, and a laparoscopic approach is appropriate. One option to avoid spill is to either excise the cyst, as a cystectomy or
oophorectomy, and then place it in a retrieval bag, which is
then delivered out of the umbilical opening, allowing decompression of the cyst while in the bag without spill and then
removal of the bag and cyst. A second option is to glue a
bag to the cyst through a small laparotomy, using one of the
adhesives, such as cyanoacrylate, as described by Shozu and
colleagues.58 The cyst is incised by cutting through the center
of the bagcyst interface, allowing removal of the fluid without
spill, and the decompressed cyst is then delivered from the abdominal cavity. The cyst can then be separated from the normal ovary as a cystectomy, or if not possible or if there is
concern for malignancy, an oophorectomy.
Sacrococcygeal Tumors
------------------------------------------------------------------------------------------------------------------------------------------------
CLINICAL PRESENTATION AND INITIAL

EVALUATION
Tumors of the sacrococcygeal region, referred to as sacrococcygeal teratomas (SCTs) in most reports, generally present in
two distinct fashions: neonates with large predominantly external lesions, which are detected in utero or at birth and are
rarely malignant (Fig. 37-6); and older infants and children
512
PART III
who present with primarily hidden pelvic tumors with a much

higher rate of malignancy (Fig. 37-7). Sacrococcygeal teratomas are the most common extragonadal tumor in neonates,
accounting for up to 70% of all teratomas in childhood. A 3
to 4:1 female to male ratio is generally reported.59 Newborns
typically present with a mass protruding from the sacral region, and many are detected with prenatal ultrasonography.
Abdominal delivery should be considered if the external mass
is greater than 5 cm, to avoid dystocia and rupture.60 In-utero
shunting can lead to fetal hydrops, which is associated with
high mortality. Adzick and colleagures61 performed the first
successful fetal resection in a fetus that developed placentomegaly and polyhydramnios and, at 25 weeks, underwent
FIGURE 37-6 A newborn with a large ruptured sacrococcygeal teratoma.
fetal resection of a 400-g immature teratoma. After delivery

at 29 weeks, the child underwent exploration, with no residual tumor identified.
Makin and colleagues62 reported a 77% survival among 41
antenatally diagnosed SCTs but noted survival of 50% in those
undergoing fetal interventions and survival of only 14% if the
intervention was for hydrops. Intervention included nonresection procedures, such as cyst drainage, laser ablation, or
alcohol sclerosis. Another study of prenatally detected lesions
noted the highest survival (100%) in lesions less than 10 cm
with predominantly cystic tumors, whereas survival was only
48% in tumors greater than 10 cm and in those with increased
vascularity, vascular steal syndrome, or rapid growth.63 This is
a difficult group, and the University of California San Francisco experience with fetal resection noted a survival of 20%.64
Older infants and children typically present with symptoms related to compression of the bladder or rectum. If a
mass has been noted at birth and left in place, an increased
rate of malignancy has been noted.65 AFP levels, which can
be normally elevated in newborns, should be obtained and
then followed to ensure that they return to normal by 9 months
of age. An association of the triad of presacral teratoma, anal
stenosis, and sacral defects was first reported by Ashcraft and
Holder, who also confirmed the autosomal dominant nature of
the condition.66 Currarino proposed that adhesions between
the endoderm and ectoderm form, causing a split notochord
that results in this association, and the triad now bears
his name.67
CLASSIFICATION AND ASSOCIATION

WITH MALIGNANCY
Altman and colleagues68 developed the classification system
of SCTs based on a survey of the Surgical Section of the
American Academy of Pediatrics (Fig. 37-8). In this study,
the malignancy rate increased with the more hidden (type
III and IV) lesions. This survey also noted the low rate of malignancy in neonates and young infants (2 months of age,
7% girls and 10% boys have malignant tumors) and the higher
rates in older infants and children (2 months of age, 48%
girls and 67% boys have malignant tumors). Several subsequent studies have confirmed this and noted malignancy rates
as high as 90%.65,69
B
FIGURE 37-7 A, Three-month-old boy with a small external mass noted
since birth. B, Underlying presacral mass noted on magnetic resonance
imaging.
SURGICAL MANAGEMENT
In neonates presenting with large external masses, the degree
of pelvic and abdominal involvement should be assessed
preoperatively with either ultrasonography, CT, or magnetic
resonance imaging (MRI), and these studies may also offer a
clue as to the characteristics of the vascular supply. An open
or laparoscopic abdominal exploration may be required to mobilize the pelvic portion and to divide the middle sacral artery.
The neonatal type I and II lesions can usually be
approached with the child in the prone position (Fig. 37-9).
Removal of the coccyx is an essential step, because Gross
and colleagues70 reported a 37% recurrence rate if it was
not removed. In view of the anterior displacement caused
by the large mass, the rectum is often brought back to a more
posterior location at the time of closure. Fishman and colleagues71 described a buttocks contouring closure bringing
CHAPTER 37
Type I
Type III
513
Type II
Type IV
FIGURE 37-8 Classification of sacrococcygeal teratomas based on Altmans study: Type I (46.7% of reported cases) predominantly external, type II
(34.7%) external with intrapelvic extension, type III (8.8%) visible externally but predominantly pelvic and abdominal, type IV (9.8%) entirely presacral.
(Adapted from Altman RP, Randolph JG, Lilly JR: Sacrococcygeal teratoma: American Academy of Pediatric Surgical Section Survey1973. J Pediatr Surg
1974;9:389-398.)
the ventral portion of the lateral flaps to a more central posterior location, thus resulting in a transverse posterior incision
and two vertical incisions in the midportion of each buttock.
The operative approach in older infants and children is similar; however, due to the presence of malignancy in many of
these cases with invasion of adjacent structures or massive
size, initial resection is not possible, and an initial biopsy
followed by neoadjuvant chemotherapy is the best mode of
management (Fig. 37-10). In the CCG/POG Intergroup study,
there was no survival difference between initial and delayed
rejections, supporting surgical delay in these cases.72
POSTOPERATIVE MANAGEMENT
The staging system for extragonadal tumors is noted in
Figure 37-11. Most neonatal tumors are mature or immature
teratomas that can be managed by surgery and postoperative
observation. Recurrent tumors are noted in 10% to 20% of
initially benign tumors, and 50% of these are malignant
recurrences.65,73 The recurrence may be due to a sampling
error of the original tumor, incomplete resection of a malignant focus, or transformation of a small benign remnant into
a malignant lesion. The large size of the neonatal tumors and
frequent cystic components can often result in rupture during resection. Follow-up of these neonates should include
serial AFP levels to ensure return to normal by 9 months

of age and rectal examination every 3 months until 3 years
of age, because the latest reported recurrence has been at
33 months.65
The management of the older infants with malignant tumors has been influenced by the chemosensitive nature of
these yolk sac tumors. In the intergroup study of 74 infants
and children (median age 21 months; 62 girls, 12 boys),
59% had metastatic disease at diagnosis, and the initial procedure was biopsy in 45 patients and resection in 29 patients.72
All patients received chemotherapy, and postchemotherapy
resection was accomplished in all but three patients. Definitive
resection required a sacral approach in 63% and a combined
abdominal-sacral approach in 35%. The 4-year EFS and survival was 84 6% and 90 4%, respectively, with no significant difference noted between timing of resection or presence
of metastatic disease. In view of these results, it is strongly
recommended to avoid resection of normal structures at initial
exploration.
Long-term follow-up of the newborns and older children
is necessary, because neuropathic bladder or bowel abnormalities have been reported in 35% to 41% of survivors.74,75
A recent report from the U.K. Childrens Cancer Group noted
that 10 of 95 survivors of sacrococcygeal tumors had a neuropathic bladder, and two had leg weakness.41 In a large survey of
514
PART III
D
FIGURE 37-9 A and B, Operative excision of sacrococcygeal teratoma in a neonate with an inverted-V incision. C and D, The tumor along with the coccyx
is excised with careful preservation of the rectum.
79 patients from the Netherlands, 9.2% reported involuntary

bowel movements, 13.2% suffered from soiling, 16% had constipation, and 30% reported difficulty with urinary control,76
with all of these correlating with decline in their quality of life.
Interestingly, the Altman classification of the tumor did not correlate with the occurrence of these long-term complications.
Mediastinal Germ Cell Tumors

------------------------------------------------------------------------------------------------------------------------------------------------
Mediastinal tumors are relatively common in childhood and

adolescence and are more common in boys than girls. Germ
cell tumors compromise approximately 6% to 18% of mediastinal tumors,77 and of these, 86% are benign.78 Mediastinal
germ cell tumors are typically located in the anterior
mediastinum. Younger children present predominantly with

respiratory symptoms. The most common symptoms during
adolescence include chest pain, precocious puberty, or facial
fullness related to superior vena caval obstruction. Klinefelters
syndrome is also observed in the adolescent group as are hematologic malignancies. The histology of the malignant mediastinal germ cell tumors is more heterogeneous than other
sites. In the intergroup study of 38 children, yolk sac was seen
in boys less than 5 years of age and in all girls; the older boys
had mixed malignant tumors in greater than 50%.9 Reflective
of this, the AFP was elevated in 29 cases and beta-HCG in
16 cases.
Anterior mediastinal tumors pose significant anesthetic
risks because of airway compression and may affect the anesthetic from compression as well as the weight of the tumor,
CHAPTER 37
515
A
Diagnosis
B
Postchemotherapy
FIGURE 37-10 A, Appearance of a large unresectable malignant yolk sac
tumor treated with biopsy and neoadjuvant chemotherapy. B, Residual
postchemotherapy tumor.
B
leading to further compression with loss of spontaneous ventilation. An early report suggested increased risk of respiratory
collapse upon induction of anesthesia if the trachea was compressed by one third of the cross-sectional area.79 Shamberger
and colleagues,80 added pulmonary function tests and observed that general anesthesia was well tolerated if both the
tracheal area and the peak expiratory flow rate were greater
than 50% of predicted. Alternatives to general anesthesia for
diagnostic procedures in children in these situations include
aspiration of pleural fluid and needle biopsy or open biopsy
Extragonadal germ cell tumors:

Stage
Extent of disease
Complete resection at any site, coccygectomy for

sacrococcygeal site, negative tumor margins.
II
Microscopic residual: lymph nodes negative.
III
Lymph node involvement with metastatic disease. Gross

residual or biopsy only; retroperitoneal nodes negative or
positive.
IV
Distant metastases, including liver.
FIGURE 37-11 Staging system for extragonadal germ cell tumors.
FIGURE 37-12 A, Appearance of a large mediastinal mass causing

tracheal compression and cardiac displacement. B, Appearance after
neoadjuvant chemotherapy.
under local anesthesia. Open biopsy can be performed using

an anterior thoracotomy (Chamberlin procedure) with excision of a segment of costal cartilage.81
In the intergroup study (N 38) 14 children underwent
initial resection, with 12 survivors.9 Twenty-two patients
underwent biopsy followed by neoadjuvant chemotherapy
and subsequent resection in 18, with 13 survivors. The size
of the mass was reduced by a mean of 57% in 12 of the
patients and was stable or increased in 6 (Fig. 37-12). Four
patients had no further surgery, because of complete radiographic resolution in 1, progressive disease in 1, and death
from toxicity in 2. Eight of 10 image-guided biopsies were
successful. Of 31 resections, 20 were by median sternotomy
and 11 by thoracotomy. Excision was frequently reported
as difficult because of adherence to the major arteries and
veins as well as the phrenic and vagus nerves and the lung
and thymus. The overall survival was 71%, which is higher
than the historical series but lower than survivals reported
for the other extragonadal sites. The outcome was superior
516
PART III
in the patients with yolk sac tumors, and all of the tumor
deaths were noted in adolescent boys with mixed germ cell
tumors.
Genital (Vaginal) Germ

Cell Tumors
------------------------------------------------------------------------------------------------------------------------------------------------
Abdominal and Retroperitoneal

Germ Cell Tumors
------------------------------------------------------------------------------------------------------------------------------------------------
Retroperitoneal and abdominal germ cell tumors account

for approximately 4% of germ cell tumors in children. Most
present in infancy, although several have been indentified
antenatally.82 Eighty percent were less than 5 years of age in
the recent intergroup (CCG/POG) study.83 Mass and pain
are the most common presenting symptoms, but fever, weight
loss, constipation, and acute abdomen are also reported.
An unusual group within this cohort are the infants with
choriocarcinoma, which are thought to be primary placental
tumors with metastases to the fetal liver. The beta-HCG production can lead to precocious puberty, and these infants usually
present with hepatomegaly and anemia in the first 7 months
of life.
Most retroperitoneal germ cell tumors are mature and immature teratomas; reports have noted malignancy rates between 0% and 24%, with the highest percentage occurring
in infants.82,8487 The histologic pattern of the malignant
tumors is most commonly pure yolk sac (63%), but also includes choriocarcinoma and mixed tumors. In the intergroup
study,83 19 of 24 of the malignant tumors had elevated AFP,
indicating yolk sac components were present but also illustrating the difficulty of determining malignancy preoperatively.
Prior to attempting resection, a search for metastatic disease
is appropriate, because nearly 90% of those with malignancy
have stage III or IV disease at presentation.83
Primary resection should be attempted if preoperative imaging suggests lack of contiguous organ involvement or metastatic disease. Unfortunately, the benign tumors can also
encase blood vessels, and the hazardous nature of these operations was demonstrated by two recent reports about several major vascular, biliary, and intestinal injuries.82,86 In the
intergroup study of 25 children, only 5 underwent initial resection, 13 had resection after chemotherapy and biopsy, or
there was partial resection in 7.83 Of note, 4 had no residual
tumor after chemotherapy. The outcome with modern chemotherapy has dramatically improved the outcomes of children with these lesions from a historical survival of less than
20%59 to current 6-year EFS of 82.8 10.9% and overall
survival of 87.6 9.3%.83 There are other rare abdominal
sites that may present later in life, and yolk sac tumors
of the pelvis and uterus have been reported in adult patients.8890
Genital lesions are rare and most commonly involve the vagina
in girls. Although early reports of surgery alone reported survival rates of 50%, survival has improved with the addition
of platinum-based adjuvant chemotherapy.91,92 Vaginal lesions
generally occur in girls less than 3 years of age who usually present with vaginal bleeding. A mass is typically identified within
and often protruding from the vagina and uterus, and the actual
site of origin may be difficult to ascertain. The CCG/POG report
of 13 genital lesions (12 vaginal, 1 penile) confirmed the efficacy of platinum-based chemotherapy administered in a neoadjuvant fashion, with ultimate preservation of the vagina in 10 of
12 girls.93 This is best accomplished by initial biopsy, followed
by chemotherapy, and subsequent excision of the residual tumor, with the goal of partial vaginectomy. Although there is
no role for initial total vaginectomy or hysterectomy, this rarely
may be required in chemoresistant cases.
Cervicofacial Teratomas
------------------------------------------------------------------------------------------------------------------------------------------------
This rare site accounts for 5% to 6% of teratomas, which

generally present in the neonatal period with large tumors.
Most are mature or immature teratomas, but up to 20% are malignant.94 A review of 20 neonates noted that 35% presented
with airway obstruction.94 A more recent report of seven giant
fetal cervical teratomas observed that four developed hydrops
(two died, one aborted), with one undergoing fetal resection.95
Three neonates without hydrops underwent ex utero intrapartum treatment (EXIT) with intubation, tracheostomy, and resection on placental support in one each. If there is no evidence of
hydrops, these can be followed to term. If the fetus is sufficiently mature (28 weeks) and hydrops is present, the fetus
can undergo delivery; however, if the gestational age is less than
28 weeks, fetal resection should be considered.95
Gastric Teratomas
------------------------------------------------------------------------------------------------------------------------------------------------
Tumors at this location generally present within the first few

months of life with abdominal distention, bleeding, or symptoms of gastric outlet obstruction because of the gastric
mass.96 We have seen older children present with pain and obstructive symptoms as a primary cystic component has
enlarged. These tumors occur primarily in males, and there
are no reported malignancies at this site. Resection with
primary closure of the stomach is the treatment of choice.
expertconsult.com.
described the distinctive multinucleated giant cell with the

prominent nucleoli that are characteristic of Hodgkin disease
(HD) (Fig. 38-2). They showed that these cells, now referred
to as Reed-Sternberg cells, are derived from germinal center
B cells.5,6 Radiotherapy was the first reported curative treatment for HL in the 1930s.7 In 1950, Peters published the first
long-term series of survivors (20 years) treated with radiotherapy.8 Single-agent chemotherapy (nitrogen mustard) was used
to treat HL in 1946, and multiagent treatment with MOPP
(Mustargen [mechlorethamine], Oncovin [vincristine], procarbazine, prednisone) was reported in 1967.9,10 In the
1960s, the staging laparotomy was increasingly used to identify sites of involvement and for research purposes.11 In the
1980s, oncologists began to appreciate the long-term morbidity of the chemotherapy and radiotherapy regimens used to
treat patients with HL. Thus multimodality therapy designed
to maintain outcomes while reducing toxicity were initiated.
Currently, biologically based therapies, both immunotherapy
and small molecules, are being investigated for use as primary
and relapse therapy.
CHAPTER 38
Hodgkin
Lymphoma and
Non-Hodgkin
Lymphoma
Peter F. Ehrlich
Hodgkin Lymphoma
------------------------------------------------------------------------------------------------------------------------------------------------
Hodgkin lymphoma (HL) is one of the few cancers that affect

both adults and children with a wide spectrum of histopathologic and clinical presentations. Unlike many other cancers,
the adult and pediatric forms have similar biology and natural
history. Pediatric HL accounts for 12% of all HL cases and represents 6% of all childhood cancers. Cure rates for pediatric
HL are excellent, approaching 90% to 95% (Fig. 38-1).1,2
Despite these excellent rates of cure, treatment can result in
significant short-term and long-term morbidity. The aims of
current therapeutic trials are to maintain or improve on outcomes while reducing short-term and long-term complications of therapy.3
Hodgkin lymphoma is named after Thomas Hodgkin, a
British pathologist, who in 1832 described the disease in a paper titled On Some Morbid Appearances of the Absorbent Glands
and Spleen.4 One hundred and fifty years later, with the advent
of microscopic histology, Sternberg (1898) and Reed (1902)
INCIDENCE AND EPIDEMIOLOGY

Hodgkin lymphoma accounts for 6% of all pediatric malignancies, with an incidence of about 6 cases per 1 million, with
a bimodal distribution with peaks in adolescence (15 to
19 years) and after age 55 years. HL is exceedingly rare in children less than 5 years of age.12 Epidemiologic studies identify
three forms of HL: two that involve the pediatric population
and one in adults. Childhood HL is found in children less than
14 years old and accounts for 10% to 12% of cases; adolescent
young adults (AYA) HL is defined as occurring in those 15 to
35 years of age and accounts for greater than 50% of the cases.
It is the most commonly diagnosed cancer among adolescents
15 to 19 years of age. Older adults HL occurs in those older
than 55 years of age and comprises 35% of the cases.13
Childhood HL is more common in males, and the histology
is more likely to be mixed cellularity or nodular lymphocyte
predominant. Risk factors include increasing family size,
lower socioeconomic status, and exposure to the Epstein-Barr
virus (EBV).1417 The EBV viral infection appears to precede
tumor cell expansion, and EBV may act alone or in conjunction with other carcinogens.
The AYA form has no gender predilection, and the most
common form is nodular sclerosis. Risk factors include higher
socioeconomic status, early birth order, smaller family size,
and EBV. In the AYA forms, it is hypothesized that EBV exposure is delayed (as opposed to the childhood form), suggesting
that delayed exposure to EBV or some other unidentified common infectious agent may be a risk factor for AYA HL.1619
Hodgkin lymphoma is derived from a single transformed B
cell that has undergone monoclonal expansion. Classic cells
include Reed-Sternberg, lymphocytic, and histiocytic cells.
There are also many cytokine-producing and cytokineresponding cells that are responsible for the nonspecific signs
and symptoms seen with this tumor. Immune system dysfunction is hypothesized to be one of the primary causes for
Hodgkin lymphoma. In the childhood form, it is thought to
result from immune immaturity, whereas in the adult form,
it is thought to result from immune dysregulation. Support
for this hypothesis is found in diseases with altered immune
states in which an increased incidence of HL is seen, including
517
518
PART III
Percent
5-YEAR SURVIVAL
100
95
90
85
80
75
70
65
60
55
50
1975
1981
1987
1995
2003
Year
Bone
HD
Wilms
Neuroblastoma
ALL
FIGURE 38-1 Graph shows survival statistics of different pediatric cancers
from 1975 to 2003.
B
FIGURE 38-3 A, Chest radiograph demonstrating a large anterior mediastinal mass. B, Computed tomography scan demonstrating a large anterior mediastinal mass.
FIGURE 38-2 High-power hematoxylin and eosinstained slide of a
patient with nodular sclerosis Hodgkin lymphoma with typical Reed-Sternberg cells.
patients with human immunodeficiency viral infection, other

acquired immunodeficiency states (postsolid organ or hematopoietic stem cell transplantation), and autoimmune disorders or a family history of autoimmune disorders.2024
Hodgkin lymphoma must be considered in any child with
lymphadenopathy. Involved nodes are described as firm, nodular, and painless. Children and adolescents most frequently
present with cervical and or supraclavicular lymphadenopathy (80%). Patients presenting primarily with enlarged axillary
nodes (25% of all cases) or inguinal nodes (5%) are far less
common. Associated mediastinal disease is found in up to
75% of adolescents and 33% of children.13,2527 Mediastinal
involvement must be assessed prior to any operative intervention; involvement may be extensive and produce major complications upon the induction of anesthesia (Fig. 38-3)
Patients may also present with B symptoms, including fever
greater than 38 C, soaking night sweats, and weight loss of
10% or more. These symptoms are not specific to HL and
can occur in non-Hodgkin lymphoma. The presence or absence of B symptoms, which occur in up to a third of children,
has prognostic significance and is reflected in the staging of
HL.13,25 Respiratory symptoms may also result from large mediastinal masses, including dyspnea on exertion or orthopnea.
Itching or pruritus is a frequent finding but is nonspecific.28
DIAGNOSIS
A full history and physical examination focusing on nodal
areas and the abdomen should be performed. At present there
is no specific laboratory test for HL. An excisional biopsy of a
suspicious lymph node should be the initial step to diagnosis
CHAPTER 38
HODGKIN LYMPHOMA AND NON-HODGKIN LYMPHOMA
519
TABLE 38-1
Hodgkin Lymphoma Staging: Ann Arbor Classification
with Cotswolds Modification
Stage 1
Stage II
Stage III III1 III2
Stage IV
Involvement of a single lymph node region or

lymphoid structure (e.g., spleen, thymus, Waldeyer
ring) or involvement of a single extralymphatic site
Involvement of two or more lymph node regions
on the same side of the diaphragm
Indicates that the cancer has spread to both sides
of the diaphragm, including one organ or area near
the lymph nodes or the spleen
With or without involvement of splenic, hilar, celiac,
or portal nodes With involvement of paraaortic,
iliac, and mesenteric nodes
Indicates that the cancer has spread to both sides
of the diaphragm, including one organ or area near
the lymph nodes or the spleen
Modifiers:
A or B: The absence of constitutional (B-type) symptoms is denoted by adding
an A to the stage; the presence is denoted by adding a B to the stage.
E: Used if the disease is extranodal (not in the lymph nodes) or has spread
from lymph nodes to adjacent tissue.
X: Used if the largest deposit is greater than 10 cm large (bulky disease),
or whether the mediastinum is wider than one third of the chest on a
chest x-ray.
S: Used if the disease has spread to the spleen.
The nature of the staging is (occasionally) expressed with:
CS: Clinical stage as obtained by doctors examinations and tests.
PS: Pathologic stage as obtained by exploratory laparotomy (surgery
performed through an abdominal incision) with splenectomy (surgical
removal of the spleen). Note: Exploratory laparotomy has fallen out of favor
for lymphoma staging.
of Hodgkin lymphoma. Prior to surgery, a chest radiograph

must be obtained to assess the presence of mediastinal disease.
If a mediastinal mass is detected, a computed tomography
(CT) scan of the chest is mandated to assess the tracheal area,
and pulmonary function tests further define the extent of
respiratory impairment. In some cases, the procedure may need
to be performed under local anesthesia because of the size of
the mediastinal mass and the resultant respiratory compromise
(see Fig. 38-3). Minimally invasive techniques have been used
to biopsy mediastinal masses, if no suspicious extrathoracic
lymph nodes are available for biopsy. Care must be taken when
using a thoracoscopic or laparoscopic technique to ensure that
adequate specimens are obtained. A report from the Childrens
Oncology Group Hodgkins Lymphoma Committee demonstrated that up to 50% of mediastinal cases required a second
diagnostic biopsy when a thoracoscopic biopsy was performed.29 thoracoscopic biopsy should also be avoided in children with respiratory compromise.
HISTOPATHOLOGY
Reed-Sternberg cells are the pathognomonic cells of HD (see
Fig. 38-2). The classification systems for HL have evolved over
time from the Rye classification to the Ann Arbor Classification and the Cotswolds modification (Table 38-1).3032 The
current World Health Organization classification system separates HL into two broad categories: classical and lymphocyte
predominant. Classical has four subtypes: lymphocyte depleted, nodular sclerosing, mixed cellularity, and classical
lymphocyte rich. Classical HL accounts for 90% of all cases.
For children, nodular sclerosis is the most common subtype,
accounting for 65% of cases. Immunohistochemical studies
FIGURE 38-4 CD30-positive staining for Reed-Sternberg cells in a patient

with Hodgkin lymphoma.
define a common immunophenotype for classical Hodgkin,

characterized by CD15-positive and CD30-positive ReedSternberg cells (Fig. 38-4). Classical HL expresses CD30, a
marker of activated B-lymphoid and T-lymphoid cells, in
almost all cases.25,28,30 About 87% of classical Hodgkin lymphomas express CD15, the carbohydrate X hapten. Classical
Hodgkin lymphoma rarely expresses CD45, also known as
common leukocyte antigen, which is expressed by nearly all
non-Hodgkin lymphomas and can serve as a useful differential
marker between HL and non-Hodgkin lymphoma.
The lymphocyte predominant (LPHD) subtype accounts
for 10% of all cases and is characterized by lymphocytic
and histiocytic (L&H) cells that express markers not typically
seen in the classical subtype (Fig. 38-5). These cells are also
known as popcorn cells and are CD20 positive. Other B-cell
immunomarkers found in LPHD include CD79a, CD75,
epithelial membrane antigen, and CD45. The lymphocyte
predominant subtype historically carries the best prognosis.
However, since the development of highly effective multiagent
and multidisciplinary treatment regimens, all histologic subtypes have become responsive to therapy.
STAGING
Staging has both clinical and pathologic features. The Ann
Arbor staging system and its Cotswolds modification
remain the standard for adult and pediatric HL (see
Table 38-1).30,33,34 The original Ann Arbor staging system developed in 1974 was based principally upon the use of staging
laparotomy and lymphangiogram, both of which have been
abandoned.
Clinical staging requires a complete history and physical
examination. Basic tests should include a complete blood cell
count with differential, lactate dehydrogenase, alkaline phosphatase, erythrocyte sedimentation rate, or C-reactive protein
(CRP), baseline hepatic and renal function tests, and electrolytes. Radiographic studies include a chest radiograph and a
computed tomography (CT) scan of the neck, chest, abdomen, and pelvis. Chest radiographs often reveal the presence
of a mediastinal mass, and the ratio of its maximal diameter to
that of the thoracic cavity on a posteroanterior view is important prognostically. A mass with a ratio greater than 1:3 places
520
PART III
number of involved nodal regions, and extranodal involvement of disease. High-risk patients are those with stage IIIB
and IVA/B disease.3840 LPHD is considered a low-risk disease
but is often separated from the classical HL studies.
Surgery
FIGURE 38-5 High-power hematoxylin and eosinstained slide of a

patient with lymphocyte predominant Hodgkin lymphoma demonstrating
classical popcorn cells as defined by the arrows.
the patient in the subcategory of bulky mediastinal disease associated with a worse prognosis. Bone marrow biopsy is reserved for those patients with B symptoms or stage III-IV
disease. (18F)-2 fluoro-D-2-deoxyglucose positron emission
tomography (FDG PET) is replacing gallium scans, and recent
studies have assessed the ability of PET scans to replace CT
scans and as possible prognostic indicators for response to
therapy.25,28,3537 Magnetic resonance imaging (MRI) provides a more accurate evaluation of disease in the abdomen
compared with CT, with better visualization of fat-encased retroperitoneal nodes, but whether or not this provides clinically
significant information has yet to be established.
TREATMENT
Risk Classification
Children and adolescents with HL are divided into three risk
categorieslow-, intermediate-, and high-risk disease
based on clinical and pathologic staging data, histology, stage
at presentation, presence or absence of B symptoms, number
of involved sites, and/or presence of bulky disease (>10 cm).
The exact definitions of each stage will often change between
studies and clinical trial consortiums, such as the Childrens
Oncology Group (COG).34 In general, low-risk disease is
defined as classical Hodgkin lymphoma patients, with clinical
stage I or II disease showing no B symptoms or bulky nodal
involvement and disease in fewer than three nodal regions.
Intermediate-risk disease includes stage I, II, and sometimes
IIIA disease with criteria that vary from trial to trial.26,30 Some
trials have included B symptoms, bulky disease, a large
The role of surgery in the initial diagnosis and staging for HL

has been reduced. With the wide application of chemotherapy
in all stages of HL, surgical staging has become irrelevant,
because the additional information it provides does not alter
treatment.41,42 The surgeons primary role is to obtain tissue
for diagnosis. Biopsies should be taken from the most easily
accessible site, and adequate tissue must be obtained and sent
fresh to pathology for immunohistochemistry, immunophenotyping, cytogenetics, and flow cytometry. Fine-needle aspiration is generally discouraged, because it is inaccurate and
inadequate tissue is obtained to properly stage and classify
the patient. Thoracoscopic biopsy or a Chamberlain procedure can be used for diagnosis in patients with only mediastinal involvement. Retroperitoneal lymphadenopathy is often
accessible through laparoscopic biopsy. However, thoracoscopic and laparoscopic, as well as core needle biopsies, have
a higher incidence of misdiagnosis and can require multiple
procedures to obtain an adequate sample.29 The second role
for surgery is to provide central venous access for chemotherapy. Bilateral oophoropexies are also performed in girls who
will receive abdominal radiotherapy.
Chemotherapy and Radiation Therapy
Chemotherapy and radiotherapy (RT) are the mainstay treatments of HL. Although the outcomes for children with HL have
improved dramatically, the short-term and long-term toxicity of
therapy has been substanial.43,44 Therefore recent and current
therapeutic protocols for HL have focused on maintaining excellent outcomes but reducing toxicity. Ideal chemotherapy regimens use drugs that are individually effective with different
mechanisms of action and toxicities, to allow for a maximal
dose. The first widespread successful regimen was MOPP (Mustargen, oncovin, procarbazine, and prednisone). In a long-term
study of 188 patients from the National Cancer Institute, who
were treated with MOPP, the complete remission rate was
89%, and 54% of patients remained disease free at 10 years.45
In this study, 95% of patients had stage III or IV disease, and
89% had B symptoms. ABVD (Adriamycin [doxorubicin], bleomycin, vinblastine, and dacarbazine) was the second regimen
used in the treatment of HD.46 It was developed for the treatment of patients failing MOPP therapy and contains individually
effective drugs with nonoverlapping toxicities.
Historically, radiation therapy was based on the concept of
contiguous lymph node basin involvement.47 The whole
nodal region was included as defined by Kaplan and Rosenberg, sometimes additionally covering uninvolved adjacent
lymph node region(s), extended field radiotherapy (EFRT).47
However, radiation therapy is one of the major contributors to
early and late toxicity in children with HL. Similar to chemotherapy, treatment has evolved, however, to reduce the radiation necessary. EFRT has been supplanted by involved field
radiation therapy (IFRT). Over time, improvements in equipment and targeting have reduced the exposure of uninvolved
areas. These practices aim to reduce salivary gland and oral
cavity morbidity and to optimally spare the heart from irradiation. A further reduction of RT volume to cover just the nodal
CHAPTER 38
tissue involved by disease, without any attempt to include

whole nodal region(s), is termed involved node radiation therapy (INRT).48 Relapses in patients treated with chemotherapy
alone occur primarily in the initially involved lymph nodes.49
Using FDG PET analysis of residual disease and advances in
radiation planning, it is possible now to confine the radiation
to the initially involved nodal tissues rather than the whole
nodal chain. The hope is that a reduction in irradiation volume will result in a lower incidence of late complications. This
goal may be particularly important in young females with anterior mediastinal disease, where exclusion of the hilar and
subcarinal nodes from the radiation field would lead to significant reductions in radiation dose to the breasts. This is important because the most common malignancy following
treatment for HL is breast cancer. In addition, children have
been shown to be particularly susceptible to thyroid toxicity
following RT, and the transition to INRT may potentially exclude the thyroid from the treated volume for many patients
with supradiaphragmatic HL. Preliminary data reported from
British Columbia in Canada indicated no increase in relapses
with INRT compared with IFRT or EFRT using a current multiagent chemotherapy regimen.50
Therapy for Low-Risk Disease
Optimal therapy for low-risk Hodgkin disease in children and
adolescents continues to evolve. Protocols using chemotherapy followed by low-dose radiation therapy have achieved
cure rates of greater than 90% for patients with low-risk
Hodgkin disease and represent the standard of care for children and adolescents with Hodgkin disease. Several multiinstitutional trials demonstrate that children and adolescents
with low-risk HL can be effectively treated with two to four
cycles of chemotherapy followed by 15- to 25-Gy IFRT, with
series reporting 90% or better event-free survival (EFS), with
overall survival (OS) greater than 95%.5154 The most recent
COG low-risk HL study used a response-based chemotherapy
regimen of AP-PC (Adriamycin [doxorubicin], vincristine,
prednisone, and cyclophosphamide) with or without IFRT. After three cycles, those with complete response do not receive
IFRT. IFRT consists of 21 GY in 14 fractions of 1.50 Gy per
day for 14 sessions.37 This study closed in the fall of 2010.
Therapy for Intermediate-Risk Disease
Intermediate-risk trials for HL have documented the need for
adjuvant radiotherapy in most patients.5456 In a German trial,
patients who completely responded to induction therapy had
radiation therapy omitted, but their event-free survival was
lower than expected.40 In the CCG 5942 trial, intermediaterisk children with complete response were randomized to receive either IFRT or no further treatment.54 Three-year EFS
was 82% with OS of 93%, but the patients who received IFRT
had three-year EFS of 88%. Both these studies support the
need for IFRTwith most intermediate HL patients. The current
intermediate-risk COG trial is a randomized trial to see if early
complete responders can have a dose reduction of both chemotherapy and radiation therapy without a decrease in their
EFS. Induction chemotherapy consists of ABVE-PC (Adriamycin [doxorubicin], bleomycin, vincristine, etoposide, prednisone, and cyclophosphamide) for two cycles. It is a double
randomized responsebased protocol with both IFRT and
chemotherapy intensifications following induction.36 This
study closed in the fall of 2010.
521
Therapy for High-Risk Disease

Patients with high-risk tumors require both intensification of
chemotherapy and radiation therapy. The German (GPOH)
HD-DAL 90 protocol treated high-risk patients with two or
four cycles of COPP (cyclophosphamide, vincristine, procarbazine, and prednisone) plus 20- to 35-Gy IFRT. Five-year
EFS in high-risk groups was 93% and 86%, respectively.53
The EFS in the high-risk groups was comparable to that seen
in the low-risk group. The Childrens Cancer Group (CCG)
5942 protocol treated those with high-risk disease with two
courses of intensive multiagent chemotherapy with cytarabine/etoposide, COPP/ABV, and cyclophosphamide, vincristine, doxorubicin, and methylprednisolone/prednisone with
granulocyte colonystimulating factor support. Complete responders were randomly assigned to 21-Gy IFRT or no further
therapy. Three-year EFS rates in intermediate- and high-risk
patients receiving IFRT were 88% and 91%, respectively.54
In both the GOPH HD-95 and CCG 5942 trials, the benefit
of IFRT in reducing relapse rates was most pronounced among
high-risk patients. The most recent studies suggest that outcome of patients with high-risk factors can be improved with
intensification of chemotherapy and lowering RT based on
response. Pediatric Oncology Group (POG) 9425 study
reported 2-year EFS for intermediate- and high-risk disease
in a response-based paradigm. In this study, 63% of patients
received 9 weeks of chemotherapy and 21 Gy of IFRT because
of good response, whereas the others received more intensive
therapy.56 The most current COG high-risk study recently
opened. This is a nonrandomized response-based protocol.
Induction therapy is with ABVE-PC, and patients will be
divided into rapid early responders and slow early responders.
Response will be determined by PET scan, and further chemotherapy and radiotherapy targets are based on the PET scan.57
Therapy for Lymphocyte-Predominant
Hodgkin Disease
Lymphocyte-predominant Hodgkin disease (LPHD) is recognized as a distinct clinical-pathologic entity, with a favorable
outcome, but also associated with a higher risk of late relapse
and subsequent development of non-Hodgkin lymphoma
(NHL). LPHD comprises up to 10% of cases in adult and
pediatric series. Most patients with LPHD reported in the literature have been treated similarly to patients with classical
HD, using chemotherapy, RT, or combined modality treatment.3,58 However, treatment by surgical resection alone
has been reported in adult and pediatric patients. The outcomes suggest that patients with low-stage disease may be
effectively treated with surgery alone, particularly considering
the toxicity of treatment.5961 In 2007, European researchers
reported 100% survival in 58 LPHD patients treated initially
with surgery alone; 50 had a complete response (CR) and received no adjuvant therapy.62 In this group, 14 (28%) recurred, but 73% required no other therapy at 43 months
follow-up. A recently completed COG protocol treated patients with stage I single node disease with surgery only. Results have not been published.35
Novel Therapy
Novel therapies are being investigated for children with HD at
diagnosis and relapse. These include rituximab (an anti-CD20
monoclonal antibody) and small molecule agents, such as
522
PART III
bortezomib, a reversible proteosome inhibitor that leads to the

blockage of NF-kappa beta being explored in HD. Other
agents include histone deacetylase inhibitors, such as
MGCD0103; however, none of these agents are being incorporated in standard therapies.3
Treatment Toxicities
Toxicities of treatment include decreased stature, cardiopulmonary dysfunction, thyroid disease, infertility, second malignancies, impaired psychosocial functioning, and decreases in
health-related quality of life.28
Growth Problems Full-dose (35- to 44-Gy) RT produces
bone and soft tissue hypoplasia in prepubertal children. For
patients treated with mantle fields, this manifests as spinal
and clavicular shortening and underdevelopment of the soft
tissues in the neck.
Cardiopulmonary Dysfunction Long-term survivors of HL
treated with full-dose RT have an increased risk of atherosclerotic heart disease, valvular dysfunction, and pericardial disease.6264 A study reported a 45-fold mortality risk from
acute myocardial infarction in children treated before the
age of 20 years with more than 30 Gy of mediastinal radiation.65 Heart disease and valvular disease tends to occur
late8 to 10 years after therapy. Lower doses and cardiac
shielding reduce this risk. Pericarditis can occur, especially
if the tumor involved the pericardium. Anthracyclines, such
as doxorubicin, cause dose-dependent myocardial heart failure and coronary artery disease.66 In children, a cumulative
dose of 300 mg/m2 of doxorubicin increases heart failure rate
by 11-fold at 15 years after therapy.67,68 Contemporary chemotherapy regimens delivering 250 mg/m2 doxorubicin with
low-dose IFRT appear to be associated with minimal early
cardiac toxicity.69 Bleomycin results in both short-term and
long-term lung toxicity with impaired diffusion capacity
and restrictive lung disease.70 RT can also produce breast
hypoplasia and contribute to the pulmonary fibrosis.
Thyroid Hypothyroidism, hyperthyroidism, as well as benign and malignant thyroid nodules have been recognized as
problems occurring in long-term survivors of HL.71,72 In the
Childhood Cancer Survivor Study, 34% of 1,791 5-year survivors of HL treated between 1970 and 1986 reported thyroid abnormalities.72 Thyroid nodules appear late in the follow-up,
often 10 or more years after completion of therapy. The relative
risk (RR) is 18.3 (confidence interval, 11.4 to 27.6) compared
with the general population. Children receiving neck RT also
appear to be at greater risk of hypothyroidism than adults.71
Infertility Sterility/infertility is a significant risk of alkylating
agents, most commonly cyclophosphamide and/or procarbazine.73 Males in the German GPOH studies receive etoposide
in place of procarbazine, because testicular germinal function
is more sensitive to alkylating agents than is ovarian function,
and current COG protocols limit alkylating agents to doses
compatible with preservation of fertility.3537,57 Gonadal failure is also a result of pelvic RT. In boys, doses greater than
3 Gy can produce irreversible azospermia.74 Low-dose IFRT
to iliac or inguinal lymph nodes may impair fertility among
females if the direct or scattered dose to the ovaries exceeds
2 to 3 Gy. Oophoropexy can help limit the adverse effects of
radiation therapy. Also reported in females is a high risk of prematurity and premature menopause.75
Second Cancers (SC) The risk of second cancers is significantly increased in the long-term survivors of HL treated with
full-dose RT.7680 The Late Effects Study Group estimated the
30-year cumulative incidence of SC to be 26.3% among survivors diagnosed before age 16. The two most frequent cancers are breast cancer (20% risk at 45 years of age),
followed by thyroid carcinoma (36-fold increased rate).80 Exposure to alkylating agents, particularly in conjunction with
extended-field RT, is associated with an increased risk of leukemia. Leukemias tend to arise 2 to 10 years after therapy. The
risk of SC after modern treatment is not yet known, because
reduction in exposure to alkylating agents and the use of lowdose IFRT became standard practice within the last 15 to 20
years. The transition from extended-field RT to IFRT significantly reduces the radiation dose to breast and lung tissue.81
It is thought that modern IFRT should lead to lower SC rates
than have been documented in the past.
Non-Hodgkin Lymphoma
------------------------------------------------------------------------------------------------------------------------------------------------
Non-Hodgkin lymphomas (NHLs) comprise a heterogeneous

group of tumors that has a constantly evolving classification
system. The current World Health Organization (WHO) pathologic classification identifies almost 60 unique subtypes
based on morphologic, immunophenotypic, and genetic differences, as well as clinical behavior (Table 38-2). NHL can
be broadly divided based on the cell of origin (B cell or T cell)
or on clinical behavior (indolent, aggressive, or highly aggressive). There are distinct differences between adult and pediatric NHL, with a strong bias toward precursor B-lymphoblastic
and T-lymphoblastic lymphoma, anaplastic large cell lymphoma, and Burkitt lymphoma in childhood.
Indolent lymphomas are slowly progressive but incurable
diseases, with a median survival time of 8 to 10 years. Aggressive lymphomas, such as Burkitt and Burkitt-like lymphomas,
are rapidly progressive at presentation but curable in 70% to
90% of patients, with outcome strongly dependent on clinical
and biological features (as identified by current molecular and
immunologic approaches) at presentation.
INCIDENCE EPIDEMIOLOGY
AND CLASSIFICATION
There are 750 to 800 new cases of non-Hodgkin lymphoma
each year in the United States.82 Non-Hodgkin lymphoma
(NHL) accounts for 7% of cancer in children and adolescents,
with an incidence of 10 per 1 million population annually in
the United States.83 NHL is rare at less than 5 years of age, with
an incidence of 2.8 per million cases but increases dramatically after age 20. NHL is more common in males (1.1 to
1.4:1), with a higher frequency in whites than in blacks or
Asians. Certain NHL types cluster according to race, for example, the natural killer (NK) T-cell lymphomas are most frequently encountered in Asian populations. A family history
of a hematologic malignancy produces an increased risk,
but it is not NHL-disease specific.82
DNA and RNA viruses are thought to play an important role
in the pathogenesis of NHL.17,82 The Epstein-Barr virus (EBV)
CHAPTER 38
523
TABLE 38-2
World Health Organization and Clinical Classification of Selected Subtypes of Non-Hodgkin Lymphoma
Clinical Behavior
WHO Pathologic
Category
Mature B-cell
neoplasms
Mature T-cell
and NK-cell
neoplasms
Indolent
Aggressive
Highly
Aggressive
Follicular lymphoma
Chronic lymphocytic leukemia/small
lymphocytic lymphoma
Hairy cell leukemia
Extranodal marginal zone lymphoma
Lymphoplasmacytic lymphoma/Waldenstrom
macroglobulinemia
Splenic B-cell marginal zone lymphoma
Mycosis fungoides
Sezary syndrome
Diffuse large B-cell lymphoma, NOS

Primary mediastinal large B-cell lymphoma
Burkitt
lymphoma
Mantle cell lymphoma
Hepatosplenic T-cell lymphoma

Peripheral T-cell lymphoma, NOS
Angioimmunoblastic T-cell lymphoma
Anaplastic large cell lymphoma, ALK type
Anaplastic large cell lymphoma, ALK type
Adapted from Jaffe E, Harris NL, Stein H, et al: Introduction and overview of the classification of lymphoid neoplasms. In Swerdlow SH, Campo E, Harris NL,
et al (eds): WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues. Lyon, France, IARC Press, 2008, p 158-166.
ALK, anaplastic lymphoma kinase; NK, natural killer; NOS, not otherwise specified; WHO, World Health Organization.
is the most prominent. EBV was first detected in cultured

African Burkitt lymphoma cells and is known to be present
in greater than 90% of such cases. EBV is important as a trigger
for lymphoproliferations/lymphomas occurring in congenital
immunodeficiencies, iatrogenically immunosuppressed organ
transplant recipients, patients receiving maintenance chemotherapy, and patients receiving combined immunosuppressive
therapy for collagen disorders.84 EBV is also found in HL
(mostly the mixed cellularity type), and patients who have
had infectious mononucleosis are at increased risk of HL.
Other viruses implicated in the pathogenesis of NHL include
the retrovirus human lymphotropic virus type 1 (HTLV-1),
with adult T-cell lymphoma and human herpesvirus
8 (HHV-8) as a cause of primary effusion lymphoma, a rare
type of large cell lymphoma confined to serous-lined body
cavities, which occurs with highest frequency in the HIVpositive population. Bacterial overgrowth can also promote
the occurrence of a lymphoma. In gastric lymphoma of
mucosa-associated lymphoid tissue (MALT) type, Helicobacter
pylori infection has been shown to be necessary for the development and early proliferation of the lymphoma.85
NHLs in children are typically high grade.86 Ninety percent
are from three main groups. These are (1) mature B-cell NHL,
which includes Burkitt lymphoma (BL), Burkitt-like lymphoma (BLL), or diffuse large B-cell lymphoma (DLBCL);
(2) lymphoblastic lymphoma (LL); or (3) anaplastic large T-cell
lymphoma (ALCL). The other 10% are similar to types seen in
adults, such as MALT and mature T-cell natural killer (NK) cell
lymphoma (see Table 38-2). NHL subtypes have different cell
lineages and cell cycle kinetics with different propensities to
invade the bone marrow and central nervous system.87
in the right iliac fossa and can be confused with appendicitis

or an appendiceal abscess. Children may also present with intussusception bleeding, ascites, or a bowel perforation.
As with HL, the presence of mediastinal disease must be
assessed, because these masses can be exceedingly large and
result in significant morbidity or mortality (see Fig. 38-3).
Superior vena cava syndrome and respiratory distress are
more common in patients with NHL. In these cases, immediate treatment with corticosteroids with or without cyclophosphamide or radiation may be required. The concern in these
situations is that the treatment will make it difficult to establish the pathologic diagnosis. However, there is some thought
that treatment for up to 48 hours is beneficial and unlikely to
obscure subsequent pathologic diagnosis, but if there is significant resolution of the mass, the tissue may be necrotic.90
Many children with NHL will present with advanced-stage
disease, including bone marrow involvement and malignant
pleural or pericardial effusions. Pleural fluid and pericardial
fluid often require drainage; cytologic examination of the fluid
can be diagnostic.90 Patients may also present with B symptoms. The presence or absence of B symptoms has prognostic
significance and is reflected in the staging of these lymphomas. B symptoms occur in up to one third of children with
NHL.13,25 In NHL, the proper diagnosis allows for classification to the distinct biological subgroups. Specimens should be
sent fresh to pathology. Regular histopathology, cytology,
immunopathology, cytogenetics using fluorescence in situ
hybridization (FISH, to look for chromosomal translocation),
polymerase chain reaction (PCR), and growth patterns all are
needed to properly classify a case of NHL.91,92
Staging
CLINICAL PRESENTATION AND STAGING

Similar to HL, NHL must be considered in any child with
lymphadenopathy. However, most children with B-cell lymphoma present with a palpable abdominal tumor or a mediastinal tumor.88,89 Frequently, the lymphoma presents as a mass
Staging laparotomy is not performed in non-Hodgkin lymphoma, because all patients require systemic chemotherapy.
However, patients may require surgical intervention because
of abdominal complications, such as intussusception or bleeding or to obtain diagnostic tissue. In some cases, the disease is
localized and a total resection can be performed, in others the
524
PART III
TABLE 38-3
St. Judes Murphy Staging System for Non-Hodgkin Lymphoma
Stage
I
II
III
IV
Hematopoetic stem cell
Description
A single extranodal tumor or single anatomic nodal area with
exclusion of mediastinum and abdomen
A single extranodal tumor with regional nodal Involvement
greater than or equal to two involved nodal regions or
localized involvement of extranodal disease on the same side
of diaphragm
A primary gastrointestinal tract tumor that is completely
resected
Greater than or equal to two nodal or extranodal tumors on
opposite sides of the diaphragm
Any primary intrathoracic tumor
Unresectable primary intraabdominal disease
Any paraspinal or epidural disease
Involvement of central nervous system and/or bone marrow
disease is extensive with involvement of the mesenteric root

and retroperitoneum.
Although no staging system is entirely satisfactory, the most
widely used staging system for NHL is the St. Judes Murphy
system (Table 38-3).93 The Childrens Oncology Group divides NHL into two categories: limited and extensive. Limited
disease corresponds to stages I and II in the St. Judes system,
and extensive correlates with stages III and IV.
NHL SUBTYPES IN CHILDREN

AND ADOLESCENTS
A detailed review of all the different types of NHL is beyond
the scope of this chapter. The most common subtypes of NHL,
accounting for 90% of cases found in children, are presented.
T cell lineage
B cell lineage
Stage I
CD2, CD7, CD38, CD71
Pre-B cell
Stage II
CD1, CD2, CD4, CD7,
CD8, CD38
Pre-B cells
Stage III
CD2, CD3, TCR, CD4/8,
CD5, CD7, CD 38
Mature B
cell with
surface IGM
Activated
mature
B cell
FIGURE 38-6 A schematic of B-cell and T-cell lineages.
Mature B-cell NHL: Burkitt Lymphoma,

Burkitt-Like Lymphoma, and Diffuse Large
B-Cell Lymphoma
B cells originate in the bone marrow from totipotential stem
cells that differentiate through many intermediate cell types
to eventually become antibody-producing plasma cells
(Fig. 38-6). Malignant transformation can occur at any point
along the path of differentiation. The clinicopathologic subtypes of NHL are determined by the stage of differentiation
at which malignant transformation occurs. Because of their
appearance by light microscopy, tumors in this category are
also called small, noncleaved cell lymphomas. Because B cells
develop in the bone marrow and then migrate to secondary
lymphoid organs (lymph nodes, spleen, Peyer patches, liver),
one would expect clinical localization of the developing neoplasm in those anatomic sites. Alternatively, B-cell lymphoma
should not occur in the anterior mediastinum in the region of
the thymus, because normal B cells are not thymic dependent.
Usually, but not always, this anatomic distribution is consistent with clinical observations.
Burkitt Lymphoma and Burkitt-Like Lymphoma BL was
first described in 1958 in Uganda by a surgeon who observed
rapidly enlarging tumors involving the jaw in children.94 BL
and BLL account for about 40% of childhood NHLs
(Fig. 38-7).86 There are three variants of BL: endemic, sporadic, and immunodeficiency related. In the United States,
BL most frequently occurs in the abdomen; in western
FIGURE 38-7 A low-power hematoxylin and eosinstained slide of a

patient with Burkitt lymphoma. Shows the typical starry sky appearance
of the tumor.
equatorial Africa, it usually arises in the mandible, but abdominal lymphoma is also noted in up to 20% of these patients. BL
can also be found in the central nervous system and bone marrow. BL of the anterior mediastinum is extremely rare.95,96 The
gold standard for the diagnosis of BL is c-MYC rearrangement.97 This is based on a characteristic chromosomal translocation, usually involving chromosomes 8 and 14, that was
discovered in BL in 1976.98 In 80% of the translocations, this
involves the locus at 14q32, in 15% of cases it is 2p11, and in
5% it is 22q11. BL is the most rapidly growing tumor in
CHAPTER 38
children, with a doubling time of approximately 24 hours. The

rate of cell death or apoptosis is also high, with the dead cells
being taken up by pale histiocytic cells within the tumor that
punctuate the low-power view, giving a starry sky appearance (see Fig. 38-7).97 BL cells are mature B-cells that are positive for CD19, CD20, CD22, and CD79a and have a
monotypic surface IgM.99 BLL is an aggressive highly proliferative variant with features that overlap classical BL and
DLBCL; BLL is treated by BL regimens.97 The distinction
between BL and BLL is controversial. Because of the rapid
growth seen in BL and BLL, when the disease is suspected,
the pediatric surgeons are often asked to intervene immediately so that treatment can begin.
Diffuse Large B-Cell Lymphoma DLBCL accounts for 10%
of pediatric lymphomas. It is less common in young children
and becomes frequent in adolescents. DLBCL is derived from
transformed mature B cells of the peripheral lymphoid organs.100 DLBCL tumors have cells that are 4 to 5 times the size
of small lymphocytes. In adults, there is a genetic signature;
however, in children this is not the case. The tumors do express c-MYClike BL as well as genes from the NF kappa-beta
pathway, but there is no specific marker of DLBCL.101 The
tumors express CD19, CD20, CD22, and CD79a. The three
most common morphologic variants are centroblastic, immunoblastic, and anaplastic. Gene expression profiling has identified two subtypes: germinal center B-celllike (GC) and
activated B-cell (ABC). The most common subtype, GC, has
a more favorable outcome. A progressively enlarging mass is
the most common mode of presentation. Symptoms are based
on tumor location. About 20% of pediatric DLBCL present as a
mediastinal mass, but the tumors can occur anywhere. Increased lactate dehydrogenase (LDH), pleura effusions, and
ascites are less frequently observed than in other NHL. The
bone marrow (BM) and the central nervous system (CNS)
are rarely involved.87
T-Cell Tumors
Lymphoblastic Lymphoma Lymphoblastic lymphomas (LL)
make up approximately 30% of childhood NHL.12,82,83,86 In
pediatric patients with LL, 75% will have a T-cell immunophenotype. The remaining LL patients have a precursor B-cell
phenotype more commonly presenting as disease localized
in skin and bone rather than T-cell LL. Some oncologists
and pathologists feel LL is acute lymphoblastic leukemia in
an extramedullarly site. Whether the LL is a Tcell or B cell does
not affect prognosis. LL tumors have a precursor lymphoblast
phenotype (TdT [terminal deoxynucleotidyl] positive) and
express T-cell markers, including CD7 or CD5.12,102. Although
there is no genetic signature, T-cell rearrangements are
common, as well as several cytogenetic and molecular
changes.12,103,104 Because thymic residence is a necessary part
of T-cell development, most lymphomas presenting in the
anterior mediastinum originate from the T-cell lineage. Fifty
percent to 70 percent of patients with lymphoblastic lymphoma (T cell) present with an intrathoracic tumor. Abdominal involvement is uncommon and, when observed, usually
includes hepatosplenomegaly. Bone marrow infiltration is
common in this situation, making the distinction from acute
lymphoblastic leukemia difficult. In these cases, survival
may be better after treatment with a lymphoblastic leukemiatype regimen. Pleural effusions are often observed, and
525
patients may complain of dyspnea, chest pain, or dysphagia.

Superior vena cava syndrome with facial, chest, and upper
extremity edema and dilated cutaneous veins over the upper
torso and shoulders, or airway compression with severe
dyspnea or orthopnea (or both) can also occur. The central
nervous system is rarely involved at diagnosis.
Anaplastic Large T-Cell Lymphoma ALCL is a mature
T-cell cancer and accounts for 10% of NHL in children. Morphologically, ALCL are characterized by large cells with big cytoplasms and horseshoe- or kidney-shaped nuclei called
hallmark cells.105 More than 90% of ALCL cases are CD30positive (Ki- antigen) and have the translocation t(2;5)
(p23;q35). This results in production of a fusion protein
NPM/ALK, although variant ALK translocations have been
reported.106 The WHO divides ALCL into systemic (ALK
and ALK) and cutaneous lymphomas. ALK is predominantly found in adults with a poorer prognosis, with OS of
45%. ALK prognosis is good, with an 80% OS.107109 The
cutaneous form is extremely rare in children and only
accounts for 1.7% of ALCL; its OS is 90%. Clinically, ALCL
has a broad range of presentations, including involvement
of lymph nodes and a variety of extranodal sites, particularly
skin and bone. Involvement of the CNS and bone marrow is
uncommon. As opposed to other pediatric NHL, ALCL is often associated with B symptoms (e.g., fever and weight loss),
and a prolonged waxing and waning course can complicate
and often delay diagnosis.
Post-Transplant Lymphoproliferative Disorders The 2%
to 4% risk of developing cancer after solid organ transplantation (SOT) is about 5- to 10-fold greater than that of the general population. The risk correlates with the intensity and
cumulative exposure to immunosuppression.110 The lowest
frequency seen is in renal transplant recipients (1%), and
the highest is in heart-lung or liver-bowel allografts (5%).
EBV seronegativity at time of transplant and young age at
transplant are the two greatest risk factors for subsequent
PTLD. In children, post-transplant lymphoproliferative disorders (PTLDs) may occur early, because of their risk for a
primary EBV infection.111 Many of the tumors exhibit an
EBV-induced monoclonal or, more rarely, polyclonal B-cell
or T-cell proliferation as a consequence of immune suppression.112 The diagnosis can be difficult and patients tend to
present with nonspecific findings, such as episodic and unexplained fever, weight loss, and fatigue. A high index of suspicion is needed to diagnose PTLD. The tumors can occur both
within and outside the allograph, including lymphoid tissue,
gastrointestinal (GI) tract, lung, and liver. Involvement of the
GI tract may present with vomiting, diarrhea, bleeding, intussusception, or obstruction. Perforation may occur at presentation or immediately following initiation of therapy in the
presence of transmural necrosis of the lesion.
TREATMENT AND OUTCOMES

Chemotherapy for Non-Hodgkin Lymphoma
Non-Hodgkin lymphomas in childhood are in most cases disseminated at diagnosis. Chemotherapy is the primary treatment modality. Each regimen is divided into phases of
induction, consolidation, reintensification, and maintenance.
526
PART III
Historically, only 20% to 30% of patients with non-Hodgkin

lymphoma survived for 5 years until the pioneering work of
Wollner and colleagues in 1975, when the LSA2-L2 (cyclophosphamide, vincristine, methotrexate, daunorubicin, prednisone, cytarabine, thioguanine, asparaginase, carmustine,
hydroxyurea) regimen, adapted from the treatment of acute
lymphoblastic leukemia, resulted in a 73% salvage rate.113,114
At the same time, Ziegler and colleagues reported similar
success with treatment of these patients using the COMP (cyclophosphamide, Oncovin [vincristine], methotrexate, prednisone) regimen.115,116 A third important NHL treatment
regimen is the Berlin-Frankfurt-Munster (BFM). This is a similar regimen to LSA2-L2.The main difference is the earlier application of L-Asp and high-dose methotrexate (MTX) in the
BFM regimen.
The results of the Childrens Cancer Group (CCG) randomized trial CCG-551 is considered one of the main studies to
alter therapy for NHL in children. It compared LSA2-L2 with
COMP. This study stratified treatment modalities by biological
subgroups. The three main findings were (1) different chemotherapy regimens exert different effects in different NHL subtypes, (2) differences in treatment efficacy are seen mainly in
advanced-stage disease, and (3) in advanced-stage disease, the
differences in treatment efficacy are more pronounced in
patients with LBL (i.e., patients receiving LSA2-L2 had fewer
relapses) and BL (i.e., patients receiving COMP did better),
while event-free survival (EFS) rates were not significantly different between treatment regimens in patients with large cell
lymphoma.117 A Pediatric Oncology Group trial further
helped with the stratification issue by demonstrating that even
in patients with localized disease, different strategies had different effects in histologic subgroups.118 Despite the different
disease process, stages, and stratification, most treatment regimens are based on one of the three regimens described
previously with adjustment made for stage, histology, and
phases of therapy. For example, LBL protocols are continual
exposure to cytostatic agents over a long period of time; BL /
BLL and DLBCL are treated with rapid repeated short,
dose-intense chemotherapy courses. ALCL have a completely
different strategy.90
has a limited role (stage I disease) in the treatment of NHL.123

Radiation is used for CNS disease with limited effects and is
controversial.124
Burkitt Lymphoma and Burkitt-Like Lymphoma
and Diffuse Large B-Cell Lymphoma
Most BL and BLL regimens are derived from the LSA2-L2 or
BFM regimens with the use of methotrexate (MTX) for CNS
disease. Rituximab is currently being studied in clinical trials,
because it has shown good results in adult NHL. Because of its
high proliferation rate, BL therapy uses cytotoxic drug concentrations over a period and drugs with different mechanisms of
action with nonoverlapping toxicities that is sufficient to affect
as many lymphoma cells as possible during the active cell
cycle, using either fractionated administration or continuous
infusion.125 Treatments use high-dose intensity and short
treatment intervals. Although these regimens are effective,
they are toxic even with use of granulocyte colonystimulating
factor (G-CSF), because up to 3% can die from treatment complications.126,127 One particular threat is acute tumor cell lysis
syndrome (ATLS). Depending on the size of the tumor, the
acute lysis of many tumor cells places a tremendous metabolic
load on the kidneys, composed of phosphates, potassium,
purines, and protein. Patients may present with elevated
serum uric acid, lactate, and potassium levels. This syndrome
may be further aggravated during the initial massive cell lysis
caused by chemotherapy. ATLS can result in hyperuricemic
nephropathy and renal shutdown. Patients with localized
resected tumors have nearly 100% EFS with two 5-day therapy courses. Recent trials report overall survival rates of 98%,
90%, and 86% in stage I/II, III, and IV disease, respectively.128
DCLC also has excellent outcomes when treated on BL and
BLL protocols with event-free survival reaching 97%.
Lymphoblastic Lymphoma
Initial surgical management includes incisional biopsy for diagnosis, followed by intense, multiagent chemotherapy, except for small, easily resectable lesions.119 Resection of
massive retroperitoneal or mediastinal masses is not indicated.
In abdominal BL, the extent of disease is a more significant
predictive variable than is completeness of surgical resection.
The surgical committee of the Childrens Cancer Group (CCG)
evaluated the role of surgical therapy in 68 patients with nonHodgkin lymphoma in the CCG-551 study.60 Tumor burden
was the most important prognostic factor. However, in disease
that can be completely resected, it may improve EFS and prevent complications such as bowel perforation. In the setting of
localized disease, data do support a role for complete resection.120122
Event-free survival for children with LL ranges from 60% to

90%, with 5-year survival, with lower stages reaching
90%.129132 Most current treatments are based on one of
two protocols: the LSA2-L2 protocol (cyclophosphamide, vincristine, methotrexate, daunorubicin, prednisone, cytarabine,
thioguanine, asparaginase, carmustine, hydroxyurea) or the
BFM group strategy. Each uses similar drugs divided into
phases of induction, consolidation, reintensification, and
maintenance. The main differences between the protocols
are earlier application of L-Asp and high-dose MTX in the
BFM regimen. Treatment intensity is stratified according to
stages I and II versus stages III and IV. Children with stage
I/II are rare and achieve EFS rates higher than 90% with
reduced-intensity (omission of reintensification in the BFM
protocol) and full-length maintenance therapy. Most relapses
occur early. Radiation is used for CNS disease with limited
effects and is controversial.124
A current COG study is looking at the benefit of high-dose
MTX with added cyclophosphamide and anthracycline during
induction with the regimen from the BFM-95. The study is
still open and accruing patients.
Radiation Therapy
Anaplastic Large Cell Lymphoma
In the treatment of localized non-Hodgkin lymphoma, radiation therapy has been shown to add toxicity with no therapeutic benefit. Several studies continue to show that radiotherapy
ALCL uses different treatment for local and systemic disease.

Patients with localized disease show the best results with
pulsed multiagent chemotherapy similar to the regimen used
Surgery
CHAPTER 38
in mature B-cell NHL reporting overall survival of 93%. Children and adolescents with disseminated ALCL have a poorer
survival of 60% to 75%. It is unclear which strategy is best
for the treatment of disseminated ALCL. COG is testing the
replacement of vincristine with vinblastine in the maintenance
phase of the APO regimen (doxorubicin, vincristine, and
prednisone.)
527
TOXICITIES
The long-term toxicity profile for patients with NHL is very
similar to HL. Acutely, the NHL regimens, because of their intensity, tend to be more toxic as described previously.
expertconsult.com.
Epidemiology
------------------------------------------------------------------------------------------------------------------------------------------------
CHAPTER 39
Ovarian Tumors
Daniel Von Allmen and Mary E. Fallat
Ovarian Tumors Incidence

------------------------------------------------------------------------------------------------------------------------------------------------
Primary cysts and tumors of the ovaries are uncommon in

children. The majority of these masses are not malignant.1
Gynecologic malignant conditions account for approximately
2% of all types of cancer in children, and 60% to 70% of these
lesions arise in the ovary.2 The North American Association of
Central Cancer Registries released data from 1992 to 1997 regarding more than 1.6 million women and children diagnosed
with cancer.3 This report revealed that 1.2% of ovarian cancer
cases occurred in females between birth and age 19 years.3
Lindfors4 analyzed several large series of ovarian tumors in
children and estimated that the annual incidence of combined
benign and malignant lesions was 2.6 cases per 100,000 girls
younger than 15 years. Using the Surveillance, Epidemiology,
and End Results (SEER) registry, 1,037 pediatric patients with
malignant ovarian tumors were identified.5 The age-adjusted
incidence of malignant ovarian tumors in those less than
9 years was 0.102 versus 1.072 per 100,000 in those aged
10 to 19 years old. Malignancy is very rare in children less than
5 years old. The predominant pathology was germ cell tumors
in all age groups (77.4%) and 61.7% of tumors occurred in
patients 15 to 19 years old. The concept that the highest
incidence of malignant conditions occurs in the youngest
patients has been reassessed. Newer diagnostic imaging
techniques have increased the detection of all gonadal masses,
and the frequency of ovarian cancer has decreased.
A few syndromes or diseases are associated with ovarian

pathology. The Peutz-Jeghers syndrome is associated with
granulosa cell tumors, ovarian cystadenomas, and sex cord
stromal tumors with annular tubules.6 Juvenile granulosa
and Sertoli-Leydig cell tumors are detected with Ollier disease
(multiple enchondromatosis)7 and juvenile granulosa cell
tumors and fibrosarcoma with Maffucci syndrome (enchondromatosis and hemangiomas).8,9 Sclerosing stromal tumors
are associated with the Chediak-Higashi syndrome (oculocutaneous albinism, pyogenic infections, and leukocyte granule
abnormalities that result in deficient phagocytosis).10 The
presence of ovarian cysts had been noted in various dysmorphic syndromes, including those with craniofacial, laryngeal,
and digital malformations.11 The McCune-Albright syndrome
(triad of cafe-au-lait macules, polyostotic fibrous dysplasia,
and autonomous endocrine hyperactivity) is generally
characterized by gonadotropin-independent sexual precocity
resulting from recurrent ovarian follicle formation and cyclic
estradiol secretion.12 Fibromas are associated with the basal
cell nevus syndrome.13
Nulliparity and increased education are associated with a
greater risk of the development of ovarian cancer.14 Women
who have never used oral contraceptives have a greater risk
than women who have used them, and hormone replacement
therapy slightly increases the risk.15 Other potential but more
controversial risk factors include exposure to ovulationinducing drugs without successful pregnancy and diets high
in meat and animal fats, dairy products, and lactose. The risk
is not uniform across histotypes for most of these factors. Prior
tubal ligation and hysterectomy may reduce the risk of
epithelial ovarian cancer.1618 More recent reports suggest
that higher body mass index (BMI) may predict a higher risk
of ovarian malignancy in women presenting with adnexal
masses, and avoidance of obesity and smoking seem protective against development of benign serous and mucinous
epithelial ovarian tumors.19,20 Late age at menarche, earlier
age at menopause, the use of vitamin E supplements, and fish
consumption tend to be associated with a decreased risk of
some histologic subtypes. Occupational physical activity
seems protective against all histotypes.16
Approximately 5% to 10% of women with breast and ovarian cancer have a genetic predisposition. High percentages of
hereditary breast and ovarian cancers arise from mutations in
the tumor suppressor genes BRCA1 and BRCA2. Approximately 70% of familial ovarian cancer cases are caused by
BRCA1 mutations and 20% by BRCA2. These mutations are
inherited in an autosomal dominant fashion. If a woman is
a carrier of one of these gene mutations, she has a lifetime risk
of developing ovarian cancer as high as 60%.21,17 Genetic testing of adolescents is controversial.22 Kodish23 formulated the
argument that physicians should respect the rule of earliest
onset and defer testing until the age when the onset of disease
becomes possible. An alternative view proposed by Elger and
Harding24 is that some mature adolescents may obtain
significant psychological relief from knowing their mutation
status and may be capable of using this information for reproductive and health decisions. In most cases, surgical intervention is not indicated until age 35 years or older or
completion of childbearing. The use of oral contraceptives
529
530
PART III
has been shown to reduce the risk of ovarian cancer in the

general population. Whether the use of these agents in young
women with BRCA mutations is beneficial remains to be
determined.17
------------------------------------------------------------------------------------------------------------------------------------------------
The clinical presentation is variable and does not differentiate

a benign from a malignant tumor. Abdominal pain is the most
common symptom.22,25 With cysts and other nonneoplastic
conditions, the pain can be acute in onset, with a crescendo
pattern of severity because of torsion, rupture, or hemorrhage.
The clinical picture may mimic appendicitis. A more chronic,
insidious pattern of pain, increasing girth, and marked distention over several weeks to months may occur. Secondary
symptoms include anorexia, nausea, vomiting, and urinary
frequency and urgency. A palpable abdominal mass with or
without tenderness is the most frequent finding on physical
examination and is detected in more than half of patients with
ovarian tumors.22 These tumors are usually mobile and palpable above the pelvic brim. Bimanual palpation between
the lower abdomen and rectum may be helpful in detecting
smaller lesions. Vaginal examination is usually reserved for
sexually active patients, although vaginal inspection is of value
in all patients. An increasing number of ovarian lesions
are discovered incidentally by abdominal radiographs or
ultrasonography (US) done for other reasons.
Both neoplastic and nonneoplastic ovarian lesions demonstrate endocrine activity in approximately 10% of cases.13
Ovarian cysts of the simple, follicular, or luteal type may secrete
estrogen and can cause precocious isosexual development.
The lesions usually function autonomously, and the girls have
suppressed gonadotropin concentrations. As a result, they can
be distinguished from patients with central precocious puberty
(with accelerated skeletal maturation) or premature thelarche
(isolated breast development) by estrogen withdrawal and
vaginal bleeding after cyst involution or removal. Precocious
pseudopuberty may occur because of the production of human
chorionic gonadotropin in girls with germ cell tumors, including dysgerminomas, yolk sac tumors (YSTs), and choriocarcinomas. Ovarian tumors most commonly associated with
precocious puberty include the sex cordstromal tumors, such
as juvenile granulosa cell tumors or some Sertoli-Leydig cell
tumors, which cause elevated levels of circulating estrogen.
In the Grumbach syndrome, hypothyroidism presents with
precocious puberty and bilateral ovarian cystic masses that
resolve with thyroid replacement therapy.
Virilization resulting from androgen excess can occur with
Sertoli-Leydig cell tumors, and masculinization is occasionally
seen in older girls with dysgerminomas that contain syncytial
trophoblastic giant cells. Yolk sac tumors, steroid cell tumors,
and polycystic ovaries can be associated with virilization.
Diagnosis
------------------------------------------------------------------------------------------------------------------------------------------------
LABORATORY TESTS
Many ovarian neoplasms are associated with the secretion of
specific tumor markers or hormones. These are outlined in
Tables 39-1 and 39-2 and are discussed further in the sections
on individual tumors.
TABLE 39-1
Ovarian Tumor Markers
CA 125
AFP
hCG
Inhibin
Germ Cell Tumors

Dysgerminoma
Yolk sac tumor*
Choriocarcinoma
Embryonal carcinoma
Immature teratoma
Mixed germ cell tumor
/
/
/
/
/
/

/
/
/
/
/

/

Epithelial-Stromal Tumors
Serous carcinoma
Mucinous carcinoma
Endometrioid carcinoma
/

Sex CordStromal Tumors

Granulosa cell tumor
Thecoma-fibroma
Sertoli-Leydig cell tumor
/
/
/

/

*Endodermal sinus tumor

Comments: CA 125 levels may be slightly elevated in any of the ovarian
tumors. LDH levels are useful for staging and risk assessment in germ cell
tumors.
AFP, alpha fetoprotein; CA 125, cancer antigen; hCG, human chorionic
gonadotropin; LDH, lactate dehydrogenase.
Table courtesy Dr. Robert Debski, Assistant Professor of Pediatrics and
Pathology, University of Louisville.
Tumor Markers
Germ cell tumors are associated with various biologic markers
that are useful in identifying and managing this group of tumors.26 Protein markers, including alpha fetoprotein (AFP),
beta-human chorionic gonadotropin (beta-hCG), and lactate
dehydrogenase (LDH), are the most readily available. They
are measured with serum assays or immunohistochemical
staining of paraffin-fixed or frozen tumor.
Alpha Fetoprotein
Because the fetal yolk sac is the source of AFP early in human
embryogenesis, elevations of the marker occur with yolk sac
tumors.27 This is also true with hepatoblastoma, hepatocellular carcinoma, and teratocarcinoma.28 The elevation reflects
the presence of fetal tissue from which normal progenitor cells
arise. There is wide variability in normal levels of AFP from
birth through the first year of life,29 and AFP is significantly
elevated in premature and normal newborns. Its usefulness
in the diagnosis of yolk sac tumor or embryonal carcinoma
in the first month of life is limited. Its value in tumor identification begins when the AFP level is significantly elevated
over the normal range at any particular age. The normal serum
half-life of AFP is 5 to 7 days. Its decline after removal of an
AFP-producing tumor signifies a response to treatment. The
goal of any treatment is to return AFP to normal levels. Tumor
recurrence is marked by a sudden elevation of the AFP level.
Beta-Human Chorionic Gonadotropin
Beta-hCG is a glycoprotein produced by placental syncytiotrophoblasts. It comprises two subunits, alpha and beta; the
latter can be reliably assayed.30 Beta-hCG elevation in a patient
with a germ cell tumor suggests the presence of syncytiotrophoblasts, as seen in seminoma, dysgerminoma, choriocarcinoma,
CHAPTER 39
OVARIAN TUMORS
531
TABLE 39-2
Ovarian Tumors and Hormones
Histologic Subtype
Ovarian cyst
Simple
Follicular
Luteal
Sex cordstromal
Juvenile granulosa
Sertoli-Leydig
Luteinized thecomas
Sex cord tumors with annular tubules
Steroid cell tumor
Gonadoblastoma
Choriocarcinoma
Estradiol
Testosterone
Urinary 17-ketosteroid
Gonadotropin
MIS
#
#
"}
"
"
"
"
"*
"
"
"{
"
"{
"{
"{
"
"
"
"
#
#
"
*Functioning Sertoli cells predominate.

Functioning Leydig cells predominate, biologic marker for disease behavior.
Indicates rarer variants of the tumor.
}
May be useful tumor marker for diagnosis and follow-up.
MIS, Mullerian inhibiting substance.
{
{
and, occasionally, embryonal carcinoma.31 Elevations greater

than 100 ng/mL are unusual and suggest the diagnosis of
choriocarcinoma.32 Unlike the much longer half-life of AFP,
the beta subunit has a half-life of 20 to 30 hours.32 Its rapid
disappearance implies complete removal of a tumor.
Serum Lactate Dehydrogenase
Serum LDH is a nonspecific marker that is widely distributed
in human tissues and is therefore of limited value in establishing tumor type or response to treatment. However, elevated
LDH may indicate increased cell turnover and has been used
as a nonspecific indicator of malignancy.33 It is most useful as a
prognostic marker for lymphoid tumors and neuroblastoma.
The gene for this isoenzyme is located on 12p, and nonrandom structural changes in chromosome 12 have been seen
in all histologic subtypes of germ cell tumors, particularly
dysgerminoma.
CA 125
CA 125 is the best available marker for epithelial ovarian cancer,
although it lacks sensitivity for stage I disease and specificity
for early ovarian cancer. Levels greater than 35 U/mL may indicate malignant or borderline ovarian tumors. However, levels are
also occasionally raised in some benign conditions, including
endometriosis, uterine myomas, acute and chronic salpingitis,
and pelvic inflammatory disease.34 One small series showed a
low sensitivity and specificity of CA 125 for detection of epithelial ovarian malignancy in premenarchal girls.35
concluded that diagnostic accuracy rates were high for both

malignant and benign tumors but low in borderline tumors.36
This has relevance because a fertility-sparing approach can be
used in borderline tumors, but surgeons confronted with this
potential diagnosis during surgery should also use a standard
approach for staging (discussed later in this chapter), because
determination of extent of disease has implications for future
treatment and prognosis.36
IMMUNOHISTOCHEMISTRY
Immunohistochemistry (IHC) has had a major impact in
recent years as an aid to diagnosis in ovarian neoplasia. From
a practical standpoint, the time-honored approaches, including gross and microscopic features, thorough sampling, and
consideration of patient age and presence or absence of coexisting endometriosis, still take precedence. In general, IHC
panels should include markers which are expected to be positive (and negative) in the various tumors in the differential
diagnosis. Virtually no antibody is specific for any given tumor, and unexpected positive and negative immunoreactions
may occasionally occur. In ovarian pathology, IHC seems to be
most valuable in the evaluation of tumors with follicles or
other patterns that bring a sex cordstromal tumor into the
differential. The two most useful markers are alpha inhibin
and calretinin. Calretinin is a slightly more sensitive marker
of ovarian sex cordstromal tumors as a group, but alpha inhibin, produced by granulosa cells, is a more specific marker,
because most other ovarian neoplasms are negative.37,38
VALUE OF FROZEN SECTION FOR

INTRAOPERATIVE DIAGNOSIS
CANCER GENETICS
Benign, borderline, and malignant lesions have been identified within the same surgical specimen, suggesting evolution
from dysplasia to cancer in some cases, although frequency
and speed of this process remain unknown. A quantitative
systematic review performed to estimate the diagnostic
accuracy of frozen sections compared with paraffin sections,
including specimens from 3,659 women aged 1 to 95 years
Ovarian germ cell tumors are associated with sex chromosome

abnormalities. Although a few case studies suggest otherwise,
a large study examining 456 first- or second-degree female
relatives of 78 patients with ovarian germ cell tumors did
not identify an increased risk for occurrence.39 Some abnormal karyotypes are associated with abnormal gonads that
are predisposed to the development of germ cell tumors.
532
PART III
The application of new cytogenetic technologies has

increased our understanding of the genetics and molecular
mechanisms involved in the development of germ cell tumors.
Nonrandom changes in molecular structure have commonly
been reported in chromosomes 1 and 12, as well as in
others.4042 For example, the chromosomal aberration of trisomy 12 has been identified in many stromal tumors.43 An
isochromosome is a chromosome in which both arms are derived from one of the two arms by breakage at the centromere
and subsequent duplication. Isochromosome 12p [i(12p)] has
been identified in all types of germ cell tumors,4447 including
testicular germ cell tumors in men.46 The presence of three or
more copies of i(12p) has been associated with treatment
failure.41 Nonrandom endodermal sinus tumors in children
involve the deletion of segments of chromosome 1p and 6q.
Deletion of the terminal portion of 1p has been identified
in other tumors, indicating that it may be a locus of one or
more tumor suppressor genes not yet characterized. Endodermal sinus tumors in children may show cytogenetic differences from adults with no evidence of i(12p), but with
deletions involving 1p, 3q, and 6q.48 The c-MYC oncogene
has been found in a few endodermal sinus tumors, and
the current Childrens Oncology Group protocol will begin
to correlate amplification with survival and response to therapy.49 Further studies are required to determine the significance of these findings. Many germ cell tumors in children
express P-glycoprotein, a membrane-bound protein that
can decrease the response to chemotherapy; this may explain
why these tumors are frequently resistant to treatment.50
Role of Tumor Markers in the Incidentally
Identified Ovarian Mass
If a mixed cystic and solid ovarian mass is discovered
incidentally on an imaging study, a preoperative AFP, betaHCG, and CA-125 assay should be done. If normal, the mass
should be removed with ovarian sparing, if possible. If any of

these are elevated, a chest CT should be included in the evaluation to look for metastatic disease. Follicle-stimulating hormone (FSH), luteinizing hormone (LH), thyroid-stimulating
hormone (TSH), estradiol and lactate dehydrogenase (LDH)
serum levels should be added to the preoperative testing if
there are signs of precocious puberty.51
IMAGING TECHNIQUES
Various radiographic studies play an important role in the
clinical evaluation of pediatric ovarian lesions. Prenatal US
can usually differentiate ovarian lesions from intestinal duplication, hydronephrosis, duodenal atresia, choledochal cyst,
urachal remnants, hydrometrocolpos, and intestinal obstruction (Fig. 39-1). Mesenteric and omental cysts are more difficult to distinguish from simple ovarian cysts, because the
ovary is an abdominal rather than a pelvic organ in an infant.
US is the diagnostic study of choice for the initial evaluation
of potential ovarian pathology in all age groups. Adequate
urinary bladder distention is mandatory to displace gas-filled
intestinal loops out of the pelvis and to ensure adequate
sound wave transmission through the ovaries. Ovarian volume changes with age from less than 0.7 cm3 in girls younger
than 2 years to 1.8 to 5.7 cm3 in postpubertal patients.52 Morphologic characteristics also change. In children younger than
8 years, the ovaries are generally solid, ovoid structures with a
homogeneous echogenic texture. During and after puberty,
the ultrasonographic spectrum of the gonad undergoes cystic
changes that parallel ovulatory follicle activity in the organ.
Ovarian cysts are generally anechoic, thin-walled masses with
through transmission. With torsion, fluid debris or septation
may be present.53 Most benign tumors are complex masses
that are hypoechoic with peripheral echogenic mural nodules,
which may exhibit acoustic shadowing. Malignant tumors are
Axial view
Spleen
Coronal view
Amniotic fluid
Colon
Spleen
Ribs
Colon
Spine
Stomach
Liver
Umbilical vein
Small
bowel
Liver Stomach
FIGURE 39-1 A and B, Two views of an ultrasonogram of a fetus in the third trimester. A large, complex ovarian cyst containing fluid debris, internal
septation, and solid components can be seen (arrowheads). An ovarian neoplasm was identified during surgery after birth. (Courtesy Gary A. Thieme, MD,
Prenatal Diagnosis Center, University of Colorado School of Medicine.)
CHAPTER 39
often larger in diameter and appear as complex soft tissue

masses with ill-defined, irregular borders and central necrosis,
thick septations, or papillary projections on US. Doppler colorflow imaging and transvaginal US are also valuable in postpubertal patients to determine morphologic characteristics of
ovarian lesions.54,55 When vessels are located in the central,
septal, or papillary projections, together with a diffuse vascular
arrangement, the tumors are likely to be malignant.56 Other
discriminating factors include the presence and nature of solid
components and free intraperitoneal fluid. Based on the premise that angiogenesis is a neoplastic marker for malignancy,
newer methods of ultrasonography using high-resolution color
Doppler with extended flow (e-flow) have resulted in better
discrimination of malignancy because of higher sensitivity in
detection of blood flow in minute vessels.57
Computed tomography (CT) and magnetic resonance
imaging (MRI) are useful when the origin of the pelvic mass
cannot be established by US or when assessment of the full
extent of a noncystic lesion is necessary. The characteristic
finding of a benign tumor on CT is a fluid-filled mass with
fat and calcifications.58 Focal solid components arising from
the tumor wall are common (Fig. 39-2). Malignant lesions
are large and predominantly solid with occasional cystic areas
as well as fine or coarse calcifications. Direct extension of
tumors to adjacent pelvic structures or to the liver and
lungs can also be demonstrated by CT, which provides more
accurate staging of disease than US. Adnexal torsion in association with any tumor has a distinct appearance on CT, which is
demonstrated by dynamic scanning after the administration of
contrast medium. The appearance is generally characterized by
lack of enhancement of mural nodules, which indicates interruption of blood flow, and demonstration of thick, engorged
blood vessels that drape around the tumor and indicate
markedly congested veins distal to the site of torsion.
OVARIAN TUMORS
533
MRI is well suited for imaging pelvic lesions, because it is

not influenced by extensive subcutaneous fat and offers
superb soft tissue contrast resolution.59 The technique is
especially valuable in determining whether a mass is ovarian
or uterine in origin, and it contributes to the characterization
of adnexal masses based on criteria suggestive of benignity
(fatty components, shading on T2-weighted images) or malignancy (vegetations or solid portions within cystic masses).60
MRI accuracy can reach 91%. The long imaging times required
may cause peristalsis and respiratory motion to obscure peritoneal and intestinal surfaces, and sedation may be needed in
small children. Ovarian torsion with hemorrhagic infarction
can be detected on MRI by the finding of a high-intensity
rim at the periphery of the mass on the T1-weighted image.61
Positron-emission tomography (PET) scanning is a newer
modality that may play a role in the differentiation of malignant from borderline ovarian tumors.62 PET and PET-CT have
a potential role in evaluating patients for recurrent ovarian
cancer, particularly those with negative CT or MRI findings
and rising tumor marker levels. Fused PET-CT scans obtained
with combined scanners can help localize pathologic activity
and differentiate this activity from physiologic radiotracer
uptake.63
Disease Classification
and Staging
------------------------------------------------------------------------------------------------------------------------------------------------
Ovarian lesions are generally divided into nonneoplastic and

neoplastic entities; the former category includes functioning
cysts, and the latter includes benign and malignant tumors.
The clinical system presented here is modified from the most
recent version of the World Health Organizations proposal for
A
FIGURE 39-2 A, Plain abdominal radiograph of a 16-year-old girl with a unilateral ovarian teratoma; the pelvic mass contains toothlike calcifications.
B, Computed tomography scan of a large, calcified abdominal mass. The mass has a large cystic component, with solid, thickened walls that are eccentric
in appearance. The tumor was a thin-walled fibrous cyst with extensive hemorrhagic infarction throughout the entire cyst wall. Histology was consistent
with a benign cystic teratoma.
534
PART III
TABLE 39-3
World Health Organization Histologic Classification
of Nonneoplastic Ovarian Lesions
TABLE 39-4
World Health Organization Classification of Tumors
of the Ovary
000 Ectopic pregnancy

D27 Benign neoplasm of ovary
E28 Ovarian dysfunction
E28.2 Polycystic ovarian syndrome
N70-77 Pelvic inflammatory disease
N80 Endometriosis
N83 Noninflammatory disorders of ovary, fallopian tube, and
broad ligament
N83.0 Follicular cyst of ovary
N83.1 Corpus luteum cyst
N83.2 Other and unspecified ovarian cysts (simple cyst)
N83.8 Other noninflammatory disorders of ovary, fallopian
tube and broad ligament
1. Surface epithelialstromal tumors

1.1. Serous tumors
1.2. Mucinous tumors
1.3. Endometrioid tumors
1.4. Clear cell tumors
1.5. Transitional cell tumors
1.6. Squamous cell tumors
1.7. Mixed epithelial tumors
1.8. Undifferentiated and unclassified tumors
2. Sex cordstromal tumors
2.1. Granulosastromal cell tumors
2.1.1. Granulosa cell tumor group
2.1.1.1. Adult
2.1.1.2. Juvenile
2.1.2. Tumors in thecoma-fibroma group
2.2. Sertolistromal cell tumors
2.3. Sex cordstromal tumors of mixed or unclassified cell types
2.3.1. Sex cord tumor with annular tubules
2.3.2. Gynandroblastoma
2.4. Steroid cell tumors
3. Germ cell tumors
3.1. Primitive germ cell tumors
3.1.1. Dysgerminoma
3.1.2. Yolk sac tumor (endodermal sinus tumor)
3.1.3. Embryonal carcinoma
3.1.4. Polyembryoma
3.1.5. Nongestational choriocarcinoma
3.1.6. Mixed germ cell tumors (specify components)
3.2. Biphasic or triphasic teratomas
3.2.1. Immature
3.2.2. Mature
3.3. Monodermal teratomas
4. Germ cell sex cordstromal tumors
4.1. Gonadoblastoma
4.2. Mixed germ cellsex cordstromal tumor of
nongonadoblastoma type
5. Tumors of rete ovarii
6. Miscellaneous tumors
6.1. Small cell carcinomas, hypercalcemic type
6.2. Gestational choriocarcinomas
6.3. Soft tissue tumors not specific to ovary
7. Tumorlike conditions
8. Lymphoid and hematopoietic tumors
9. Secondary tumors
From WHO International Classification of Diseases (ICD), 2007. Available at

http://www.who.int/classifications/icd/en/. Accessed June 6, 2010. WHO
International Statistical Classification of Diseases and Related Health
Problems, revision 10, 2007.
the international histologic classification of diseases and its

adaptation for oncology (Tables 39-3 and 39-4).6466 Nonneoplastic and neoplastic lesions may arise from surface epithelium, germ cell components, or support stroma. Neoplastic
lesions are listed based on the tissue of origin.
Proper management of ovarian neoplasms requires accurate staging of the initial extent of disease. In malignant cases,
recent advances in therapy have resulted in increased survival
rates and preservation of fertility. Surgical staging with histologic confirmation must be done to supplement the clinical
assessment of disease status. Precise staging is based on clinical examination, surgical exploration, tissue histology, and
fluid cytology. In the United States, staging of epithelial
ovarian cancer is performed at the time of surgery using the
International Federation of Gynecology and Obstetrics
(FIGO) staging system of 1988 (which was evaluated and
not changed in 2009) (Table 39-5).67,68 This system is ideal,
because it accurately correlates clinical findings with survival
in a continuum. However, the FIGO staging protocol does not
describe the thoroughness of the lymphadenectomy required
for ovarian cancer staging, and it has been suggested that the
number of lymph nodes obtained at surgery has prognostic
and clinical significance.68
Because ovarian neoplasms are relatively uncommon,
evaluation and treatment protocols developed from multiinstitutional collaborative studies have been valuable. Stromal
and germ cell tumors have been assessed in studies from
the Childrens Cancer Group (CCG), the Pediatric Oncology
Group (POG), and the Gynecologic Oncology Group
(GOG).49,6971 In children, the intergroup POG 9048/9049
and CCG 8882/8891 studies used a system that incorporated
both surgical and pathologic findings.70 This concept has
been preserved by the Childrens Oncology Group (COG)
(Table 39-6). Uniform surgical guidelines that incorporate
standard approaches to these lesions have been formulated, although the approach to ovarian neoplasms has become more
conservative with time.72,73 Preoperative assessment should
try to exclude obvious malignancy by the collection of serum
tumor markers and carefully performed pelvic US to determine
whether the ovarian mass is complex in nature.
(From International Classification of Diseases for Oncology, ed 3 (ICD-O-3).

Creation date: 1976; last date change: 2000.
Elevated tumor markers and a complex mass on US strongly

suggest a malignancy, and an abdominal and pelvic CT scan
should be obtained. For potentially malignant lesions, an
adequate abdominal incision is used, and violation of the tumor capsule is avoided. Alternatively, if tumor markers are
negative and the mass is thought to be benign (e.g., a mature
cystic teratoma) a laparoscopic approach can be considered.
Initial resection in pediatric patients should virtually
always be conservative. Pelvic washings, unilateral ovarian
cystectomy, intraoperative frozen section, and careful visual
CHAPTER 39
TABLE 39-5
Staging of Carcinoma of the Ovary: International Federation
of Gynecology and Obstetrics (FIGO)
Stage
0
I
IA
IB
IC
II
IIA
IIB
IIC
III
IIIA
IIIB
IIIC
IV
Extent of Disease
Primary tumor cannot be assessed
No evidence of primary tumor
Tumor confined to ovaries
Tumor limited to one ovary, capsule intact
No tumor on ovarian surface
No malignant cells in ascites or peritoneal washings
Tumor limited to both ovaries, capsule intact
No tumor on ovarian surface
Tumor limited to one or both ovaries, with any of the
following:
capsule ruptured, tumor on ovarian surface, malignant cells in
ascites or peritoneal washings
Tumor involves one or both ovaries with pelvic extension
Extension to or implants on uterus or tubes or both
Extension to other pelvic organs
IIA or IIB with positive malignant cells in ascites or peritoneal
washings
Tumor involves one or both ovaries with microscopically
confirmed peritoneal metastasis outside the pelvis or regional
lymph nodes metastasis
Microscopic peritoneal metastasis beyond the pelvis
Macroscopic peritoneal metastasis beyond the pelvis 2 cm or
less in greatest dimension
Peritoneal metastasis beyond the pelvis more than 2 cm in
greatest dimension or regional lymph nodes metastasis
Distant metastasis beyond the peritoneal cavity
TABLE 39-6
Clinicopathologic Staging of Ovarian Germ Cell Tumors:
Childrens Oncology Group (COG)
Stage
Extent of Disease
Limited to ovary (peritoneal evaluation should be negative);

no clinical, radiographic, or histologic evidence of disease
beyond the ovaries (Note: The presence of gliomatosis
peritonei does not change stage I disease to a higher stage.)
Microscopic residual; peritoneal evaluation negative (Note:
The presence of gliomatosis peritonei does not change stage
II disease to a higher stage.)
Lymph node involvement (metastatic nodule); gross residual
or biopsy only; contiguous visceral involvement (omentum,
intestine, bladder); peritoneal evaluation positive for
malignancy
Distant metastases, including liver
II
III
IV
inspection of the contralateral ovary are appropriate in the initial management of benign lesions or tumors of low malignant
potential. Pelvic washings are part of the staging system for
ovarian tumors and should be performed immediately on
entry into the abdomen (by either laparoscopy or laparotomy)
in an attempt to avoid contamination in the event of intraoperative tumor rupture. Because the final pathology will not
be known until either frozen section or histologic evaluation
of paraffin-embedded tissue, peritoneal washings should be
OVARIAN TUMORS
535
performed in all patients with complex adnexal masses in case

of an unsuspected malignancy. If there is no evidence of free
fluid upon entering the abdomen, lactated Ringer solution
can be used to irrigate the pelvis and paracolic gutters, then
aspirated and sent as washings.
Malignant germ cell and stromal tumors are almost
never bilateral in early-stage disease; so, unilateral salpingooophorectomy with a staging procedure is adequate first-line
management. Excellent responses have been reported with
chemotherapy, even in children with extensive tumors, and
maintenance of childbearing capability is possible with this
approach. In bilateral or more advanced disease, the current
success of in vitro fertilization techniques has prompted the
consideration of uterus-sparing procedures during the initial
operation.74,75 The expected biologic behavior of the tumor
and its response to adjuvant therapy generally dictate the ultimate extent of surgery required. The value of laparoscopic
examination in the assessment of pelvic disease is well established, because it will allow identification and management
of ovarian masses as well as identification of nonovarian
lesions.76,77 The American Association of Gynecologic
Laparoscopists reviewed more than 13,000 procedures
performed for persistent ovarian masses.78 Stage I ovarian
cancer was detected in 0.4% of cases. Although these results
are encouraging in adult women, there is concern about the
difficulty of establishing the true nature of an ovarian tumor
by gross examination in children, because experience with
such an evaluation is so infrequent. Nevertheless, techniques
are being established to avoid tumor spillage that may expand
the use of this method. Experienced surgeons have performed
more extensive staging procedures and lymph node dissections using the laparoscope.79 Studies evaluating the laparoscopic approach have been retrospective and suggested that
staging is safe, feasible, and a valid alternative, but there has
been no prospective trial to date comparing the laparoscopic
to open approach.68
Treatment
------------------------------------------------------------------------------------------------------------------------------------------------
NONNEOPLASTIC OVARIAN TUMORS

Ovarian cysts are known to arise from mature follicles. Fetal
FSH, LH, estrogens (maternal, placental, and fetal), and
placental hCG all stimulate the ovarian follicle, and mature
follicles can be found in more than half of newborn ovaries.80
A postnatal decrease in hormonal stimulation often leads to a
self-limited process. Autopsy studies of prepubertal girls have
documented active follicular growth at all ages and in normal
oocytes, granulosa cells, and cysts in various stages of involution.81,82 By convention, physiologic follicles are differentiated from pathologic ovarian cysts on the basis of size, and
any lesion larger than 2 cm in diameter is no longer considered
a mature follicle.
Nonneoplastic cysts are benign and generally asymptomatic. Although surgical intervention is rarely indicated, these
lesions occasionally have clinical manifestations, based on size
or associated functional activity, that warrant differentiation
from true ovarian neoplasms. When an operation is necessary,
a conservative approach should be undertaken with the goal
of ovarian preservation.
536
PART III
Follicular Cysts
Follicular cysts represent about half of nonneoplastic ovarian
lesions. They are unilateral, unilocular, and histologically
benign and often have a thin, yellowish, clear liquid content.
Cohen and associates83 detected cysts in 84% of all imaged
ovaries in 77 patients from birth to 24 months of age. The
prevalence was similar in each 3-month age bracket. Parallel
findings were noted in premenarchal girls between 2 and
12 years of age,84 with a generally equal distribution across
the age spectrum. Occasionally, ovarian cysts persist and
enlarge and are capable of secreting estrogen, thereby leading
to precocious isosexual development.85
The size of an ovarian lesion has been a major factor in
determining clinical management.80 Simple cysts, regardless
of size, are more likely to regress. Larger cysts (>5 cm) have
a greater risk of torsion. Larger cysts in children have a greater
association with sexual precocity. Complex cysts may already
have torsed or may be neoplastic. Complex cysts should
be resected, rather than observed, in prepubertal children.
Complex cysts in adolescents are most often due to hemorrhage
into a functional cyst and can be managed conservatively with
symptom control. Operation is indicated for persistent cysts or
persistent symptoms despite conservative management.
Ovarian cysts noted in the prenatal period can be expected
to spontaneously regress during the first year of life, and in
utero therapy is seldom justified.86,87 Cysts that develop in
utero are most often lined by luteinized cells, whereas those
in older children are more often lined by granulosa cells.32
These lesions may occasionally be complicated by torsion, intestinal obstruction, or perforation and cyst rupture.80,88
Bagolan and colleagues89 and Giorlandino and colleagues90
confirmed that echogenic cysts with fluid debris, retracting
clot, or septation were associated with torsion and hemorrhage. In newborns, torsion is often a prenatal event, and viable ovarian tissue may not be identified, even with the most
expeditious neonatal surgical intervention (Fig. 39-3). Most
authors now advocate increasingly conservative measures
for neonatal ovarian lesions.80 Small, asymptomatic cysts are
generally observed for regression with serial US. Cysts 5 cm in
diameter or larger and those with a long adnexal pedicle are

more likely to undergo torsion and may be excised with ovarian preservation or aspirated.91 However, in one randomized
study of postmenarchal patients, cysts greater than 5 cm in diameter and those with a complex appearance on imaging studies were followed for a short time with serial pelvic US. High
regression rates were seen with those followed expectantly.35
Although practitioners often reflexively prescribe oral contraceptive pills (OCPs), hormonal therapy has not been shown to
improve the regression rates of ovarian cysts compared with
those followed expectently.92 Exploratory laparotomy or laparoscopy has been recommended for patients with cysts that
do not resolve or increase in size within 2 to 3 months92 and
for cysts associated with acute or severe chronic abdominal
pain or intra-abdominal complications.
In prepubertal children, the occurrence of acute symptoms
and endocrine activity are more problematic. Surgical intervention is recommended for any cyst that increases in size
or fails to regress on follow-up US or if there is evidence of
a neoplasm on imaging studies.
As many as 75% of girls with juvenile hypothyroidism have
large multicystic ovaries and may show varying degrees of sexual precocity and/or galactorrhea resulting from increased secretion of pituitary gonadotropins and prolactin.93 Multiple
follicular cysts should be distinguished from polycystic ovary
syndrome, which is the most common cause of delayed puberty and heavy anovulatory bleeding in adolescent females.94
In nonneoplastic ovarian cysts, surgical preservation of as
much normal ovarian tissue as possible is a high priority.95
A plane of dissection can usually be established between the
normal gonadal tissue and the cyst after injecting saline with
a fine-bore needle beneath the visceral peritoneum. If the
surgical manipulation necessary to completely remove the
lesion would threaten significant viable ovarian tissue, the cyst
should be unroofed and debulked, and the cyst wall excised to
the extent possible, while protecting the ovary. Unilateral
oophorectomy is indicated only if there is a reasonable
certainty that no viable gonadal tissue can be salvaged. The
ipsilateral fallopian tube should be spared, because fertilization is still possible from the contralateral normal ovary.
Corpus Luteum Cysts
FIGURE 39-3 This newborn female infant had a prenatal diagnosis of an

intra-abdominal cystic mass. Postnatal imaging showed a low-attenuation
cystic structure with a curvilinear calcification along one wall. Laparotomy
disclosed a torsed ovarian cyst and ovary, attached by only a small residual
stalk. The fallopian tube was preserved. Pathology showed a thin-walled
cyst containing dystrophic calcifications.
True functioning corpus luteum cysts develop only in adolescents who are actively ovulating. Although these cysts may be
bilateral and become quite large, they usually regress spontaneously with the cyclic decline in serum progesterone. The
gross appearance of the external surface is often bright yellow,
although it may take on a hemorrhagic appearance when filled
with bloody fluid. The cyst lining is composed of luteinized
granulosa and theca cells and is capable of actively producing
estrogen and progesterone. These cysts may cause acute pelvic
pain if they rupture or undergo torsion. Failure of the corpus
luteum to involute may cause menstrual irregularity and dysfunctional uterine bleeding. Surgical goals for corpus luteum
cysts parallel those for other follicular lesions. Surgical intervention is indicated in the presence of cyst accident or persistence, demonstrated by repeat pelvic US performed 4 to 6
weeks after the initial assessment. Hasson96 was able to treat
17 of 19 patients who had corpus luteum cysts with laparoscopic aspiration, fenestration, or cyst wall excision. Clinical
symptoms resolved in all but one patient. Cyst recurrence
was rare.
CHAPTER 39
OVARIAN TUMORS
537
Parovarian cysts are usually small and rarely symptomatic.

They do not arise from ovarian tissue but are usually considered with this group of lesions because of their proximity to
the gonad. These cysts originate from the epoophoron and
are located in the leaves of the mesosalpinx. Parovarian cysts
cannot be distinguished from ovarian follicular cysts using any
radiographic imaging technique. During an operation, their
gross features are virtually identical to those of follicular
lesions, but they can usually be accurately distinguished
because of their anatomic position. When surgical treatment
is required, both standard open and minimally invasive
techniques have been used.96,97 Large parovarian cysts
(>3 cm) should be completely enucleated from the mesosalpinx in such a way that the fallopian tube and ovary are not
damaged.98 Those less than 3 cm may be treated with puncture and bipolar coagulation of the cyst wall.98
because of either direct or indirect hormonal stimulation.106

Histologic and biologically intermediate forms between benign and malignant epithelial lesions have been identified
and designate tumors of low malignant potential.
Age influences the relative frequency of the various types of
ovarian neoplasms. In adults, most tumors are derived from
the epithelial line and adenocarcinomas predominate. In children, germ cell tumors are most common and represent
approximately 60% to 77% of cases.105 Epithelial lesions
account for approximately 15% of tumors in the younger
age group.95,107 Although germ cell tumors predominate in
each age group, the peak incidence of sex cordstromal tumors occurs in the first 4 years of life, and epithelial tumors
are more common in older teenagers. Neoplasms that are rare
in children include endometrioid and clear cell tumors (which
are usually malignant); Brenner tumors, which are usually
benign; disseminated malignant lymphoma; and metastatic
lesions to the ovary.
Endometriosis
Surface Epithelial-Stromal Tumors
Endometriosis is a disorder in which the endometrial glands

and stroma are implanted on the peritoneal surfaces of extrauterine sites. The proposed mechanisms for the pathogenesis
of this disease include menstrual flow obstruction with retrograde menstruation, mechanical transplantation and implantation of endometrial elements, and coelomic metaplasia.99101
The interval between the onset of menarche and the diagnosis
of endometriosis may be as short as 1 month, and the incidence
of disease in teenage girls may be far higher than previously
anticipated or described.102 Extensive disease and the presence
of endometriomas is uncommon in children and young adolescents unless it is associated with an obstructive mullerian
anomaly.103An endometrioma or endometrioid cyst may occur
in the ovary and can be diagnosed by ultrasonography. Endometrioid cysts are filled with dark, reddish-brown blood and
may range in size from 0.75 to 8 inches. Several surgical treatments are available for endometriomas, including simple puncture, ablation, removal of the cyst wall, or drainage and medical
therapy, followed by later removal. Complete removal is the
procedure of choice to decrease recurrence of disease. The revised American Fertility Society classification of endometriosis
is widely accepted as the staging system for the disease and was
developed as a prognostic tool for patients with infertility.104
For patients with pelvic pain and a suspected diagnosis of
endometriosis, medical therapy with nonsteroidal antiinflammatory drugs or oral contraceptives should be considered. Both
medications act to suppress prostaglandins, which are known
to be important in the pathophysiology of dysmenorrhea.
These drugs along with gonadotropin hormone antagonists,
used for a 6-month period, are the most commonly used
medications.
Epithelial tumors account for 70% of all ovarian neoplasms,

but they are much less common in children. In most series,
they account for approximately 15% of all surgically resected
ovarian masses.108 Norris and Jensen109 reported that 67 of
353 ovarian tumors (19%) in children were epithelial in origin
and 12% were malignant. The tumors are usually serous or
mucinous.13 Twenty percent of serous tumors are bilateral,
and very few are malignant.13,110 Mucinous tumors are usually
unilateral, and 10% are malignant.13 Deprest and colleagues111
calculated a 16% malignancy rate for ovarian epithelial
neoplasms derived from a collected series that reported more
than 1700 pediatric patients with various types of ovarian
tumors. Ovarian carcinoma is different in children than in
adults. The proportion of mucinous tumors in children was
40% compared with 12% in adults, and 30% were of borderline
malignant potential compared with the adult rate of fewer than
10% for these more favorable lesions. As previously discussed,
serum CA 125 is a useful tumor marker in malignant epithelial
ovarian tumors.68 However, in premenopausal patients, it may
also be raised in several benign gynecologic conditions, including endometriosis, pelvic inflammatory disease, fibroids, and
pregnancy.
Proper staging of epithelial tumors is important and differs
from the staging algorithm used in pediatric germ cell tumors,
which are far more common. Epithelial tumors are staged using
the adult FIGO system (see Table 39-5).67,68 Stage IA tumors
may be treated with unilateral salpingo-oophorectomy. The
opposite ovary should be examined externally and a biopsy
should be taken of any surface abnormalities. Most young
patients with stage IB tumors (tumors limited to both ovaries)
may be adequately treated by bilateral gonadectomy, but the
uterus should be preserved to allow future fertilization.74,112
In ovarian cancer of a more advanced stage, maximum cytoreduction is important and has been associated with an improved
outcome.113 Total abdominal hysterectomy and bilateral
salpingo-oophorectomy with omentectomy and resection of
as much gross intraperitoneal disease as possible is necessary.
Systemic chemotherapy after appropriate surgery has been beneficial in cases of advanced ovarian carcinoma. Combinations of
cisplatin, cyclophosphamide, and paclitaxel are standard agents,
while newer biologic therapies hold some promise to improve
the overall poor outcome in advanced-stage disease.114
Parovarian Cysts
NEOPLASTIC OVARIAN TUMORS

Most neoplastic ovarian tumors develop from cell lines
derived from one of three sources: the germinal epithelium
covering the urogenital ridge, the underlying stromal elements
of the urogenital ridge, or the germ cells that arise from the
yolk sac. Cells from each of these lineages may develop into
an ovarian neoplasm by de-differentiation, proliferation,
and eventually malignant transformation.105 Malignant ovarian tumors probably arise from their benign counterparts
538
PART III
Fortunately, advanced-stage disease is uncommon in pediatric

patients as tumor stage is the most important prognostic factor.68
Tumors of Low Malignant Potential
Ovarian epithelial tumors of low malignant potential or borderline ovarian tumors (BOTs) differ from epithelial cancer in two
major ways: They occur in younger patients, and they have a
better prognosis than ovarian cancer. They have been described
for all subtypes of ovarian cancer.111 The serous and mucinous
tumors are by far the most common and resemble their benign
counterparts. These borderline tumors are differentiated from
standard adenocarcinoma in that they lack stromal invasion by
neoplastic epithelial elements (Fig. 39-4). Up to 50% of these
tumors are bilateral, and they demonstrate a characteristic
indolent clinical course. However, recurrences may occur as

long as 10 to 15 years after surgery for the primary tumor,
and they may be in the form of invasive cancer.115,116
In adults, 91% of borderline mucinous tumors present
with stage I disease and have a 5-year survival rate of 98%. Serous tumors have a similar outcome. The extensive review of
Massad and colleagues117 noted an overall survival of 98% for
stage I tumors, 94% for stage II, and 79% for stages III and IV.
In children, Morris and colleagues118 noted that 75% of the
cases presented with stage I disease, and overall survival
was 100%. The combined 10-year survival rate for all stages
was 73%. In a more recent adolescent study, 26/28 cases were
stage I, two cases were stage II, and all patients were alive at
5 years.116
C
FIGURE 39-4 A, Ovarian tumor from a 17-year-old girl with massive bilateral ovarian lesions. The opened specimen shows a cavity filled with clear fluid,
and the wall is lined by numerous nodules and papillary protuberances. B, Histologic section of the lesion shows serous papillary tumor of low malignant
potential (hematoxylin-eosin stain). C, Higher-power photomicrograph of a section of the lesion shows mucinous tumor of low malignant potential
(hematoxylin-eosin stain).
CHAPTER 39
Surgery is the primary method of therapy. Unilateral

salpingo-oophorectomy is adequate for all low-stage tumors
and has been standard treatment; however, some studies have
shown that ovarian cystectomy can be performed in young
patients with careful follow-up.119 These patients require
close follow-up with pelvic exams, CA 125 assay, and ultrasonography every 3 to 6 months, because patients managed with
ovarian cystectomy have a higher risk of recurrence than those
managed more aggressively.120 Morice and colleagues have
demonstrated this to be 36.3%, 15.1%, and 5.7% after cystectomy, oophorectomy, and hysterectomy/bilateral oophorectomy, respectively.119 Despite the difference in recurrence
risk, there was no demonstrated impact on overall survival,
because all patients were salvaged with further surgery.
Conservative treatment should therefore be considered in
young patients who wish to preserve their fertility and will
comply with routine follow-up.121
Bilateral tumors will require bilateral oophorocystectomy
or salpingo-oophorectomy. Uterine-sparing procedures are
probably not appropriate for advanced-stage disease. The
pathologic features that identify poor prognosis are being
sought,115 but currently there are no clear candidates.
At present, surgery remains the most effective therapy for
these patients with the place of adjuvant therapy yet to be
established.122 No individual treatment strategy has led to
consistently superior outcomes, but the favorable biology of
this tumor minimizes the importance of the limited clinical
benefit from adjuvant therapy.
OVARIAN TUMORS
539
Sex CordStromal Tumors

Sex cordstromal tumors probably arise from uncommitted
mesenchymal stem cells that reside below the surface epithelium of the urogenital ridge.123,124 This totipotential tissue
may differentiate into several different cell lines, including
granulosa-theca cells in the ovary and the Leydig-Sertoli cells
in the testicular interstitium. Sex cordstromal tumors are
referred to as functioning ovarian tumors, because they
produce systemic hormonal effects. They account for 5.7%
to 17% of malignant tumors in series of ovarian neoplasms
in children.5 Before 9 years of age, most sex cordstromal
tumors are feminizing, and after 9 years of age, there is
a predominance of virilizing neoplasms.32
Granulosa-Theca Cell Tumors Granulosa-stromal cell
tumors are the most common type of sex cordstromal
neoplasms, and the most common type of functioning ovarian
neoplasm. The juvenile granulosa cell tumor is a specific
subclassification of these lesions; 44% of these occur in the
first decade of life and 97% are seen by 30 years of age.125
Isosexual pseudoprecocious puberty is the presenting sign
in the majority of premenarchal girls who have this tumor
(Fig. 39-5).123 Most patients have elevated serum and urinary
estrogen levels, whereas gonadotropin levels are low. This
profile assists in differentiating children with these tumors
from those with true sexual precocity, gonadotropin-secreting
lesions, or feminizing adrenal tumors. The peptide hormones
FIGURE 39-5 A, Three-year-old girl demonstrating isosexual pseudoprecocious puberty. B, Surgery revealed a benign juvenile granulosa cell tumor.
Unilateral salpingo-oophorectomy was performed to remove the tumor.
540
PART III
inhibin and antimullerian hormone are produced by ovarian

granulosa cells and may be useful tumor markers for diagnosis
and follow-up of granulosa cell tumors.126
Clinical findings include premature thelarche, vaginal
discharge or bleeding, labial enlargement, development of
pubic or axillary hair, increased somatic growth, and advanced bone age. Clitoral enlargement is a rare manifestation
of virilization and tumor androgen production. Postpubertal
girls may present with an abdominal mass, relatively nonspecific symptoms of abdominal pain, or increased girth.
Amenorrhea and other menstrual irregularities may occur.
In addition to differences in clinical presentation, juvenile
granulosa cell tumors demonstrate a pattern of histologic
features and biologic behavior that are very distinct from
the adult counterpart. The juvenile variety is usually a relatively large lesion that averages 12.5 cm in diameter.127 At
laparotomy, it appears as a yellow-tan or gray solid neoplasm
with cystic areas that often contain hemorrhagic fluid.
In contrast to the adult tumors, the juvenile type has abundant
eosinophilic or luteinized cytoplasm with atypical nuclei and
a higher mitotic rate. Deoxyribonucleic acid (DNA) content
and cell cycle kinetics analyzed by flow cytometry do not
necessarily correlate with the prognosis in children as they
often do in adults.128
Although the adult form is generally an indolent, slowgrowing lesion of relatively low malignant potential, the
biologic behavior of the juvenile tumor is more aggressive
and correlates well with tumor size, disease stage, presence
of rupture, and degree of nuclear atypia and mitotic activity.
The lesion was unilateral in 122 of 125 cases reviewed by
Young and colleagues.129 If the adult tumor recurs, it is usually
more than 5 years after diagnosis. Malignant granulosa cell
tumors in young patients tend to recur much more quickly.
Granulosa cell tumors are staged similarly to other ovarian
lesions (see Table 39-5). In children, these tumors are associated with a favorable prognosis, because more than 90%
of affected children present with stage I disease. In a German
series, 69% of patients were less than 10 years of age, and 82%
of patients less than 5 years of age presented with endocrine
symptoms. Survival of FIGO stage IA patients was 100%,
stage IC was 76%, and stage II/III was 67%. Platinumbased chemotherapy is recommended for tumors of stage IC
and above.93
Fibromas and Thecomas Fibromas and thecomas account
for 14% of sex cordstromal tumors in pediatric patients.13
Although they are extremely uncommon in females younger
than 20 years of age, fibromas are usually associated with
the basal cell nevus syndrome and are frequently bilateral,
multicentric, and calcified. Most ovarian thecomas occur in
menopausal women; however, two variants of this lesion have
been reported in the second decade of life. Calcified thecomas
invariably cause amenorrhea or other menstrual irregularities
and hirsutism.130 If these tumors contain a substantial
number of lutein cells, they are appropriately called luteinized
thecomas and can occur in younger girls associated with
androgenic manifestations.
On gross examination, fibromas are firm, solid masses with
a whorled, trabeculated appearance on cross section. The lipid
content of thecomas imparts a pale yellow to orange color on
sectioning the tumor. These lesions are benign, and unilateral
oophorectomy is adequate treatment. In the case of bilateral
fibromas, all gross tumor tissue should be removed with

particular attention to sparing normal-appearing ovarian
tissue.131 Tumor recurrence is rare and managed by reoperation. Virilizing symptoms usually resolve after resection of
the tumor.
Sclerosing Stromal Tumors Sclerosing stromal tumors have
recently been recognized as distinct tumors that are separate
from fibromas and thecomas. These tumors are seen in girls,
with 30% of documented cases occurring in the first 2 decades
of life. Estrogen secretion has occasionally been reported,
whereas androgen manifestations are quite rare. The typical
presentation includes the presence of a pelvic mass and pelvic
pain in a young patient with a history of menstrual irregularity.
This lesion has also been associated with the Chediak-Higashi
syndrome.10
Sclerosing stromal tumors are unilateral, usually larger
than 5 cm in diameter, and benign. At laparotomy, these
tumors are well-circumscribed, firm, whitish-yellow masses
with clearly demarcated areas of edema and cyst formation.
Histologically, the tumor is characterized by a pseudolobulated pattern with cellular foci clearly demarcated from the
edematous and collagenized areas.132 Gross tumor removal
is generally adequate for treatment.
Sertoli-Stromal Cell Tumors Sertoli-Leydig cell tumors account for less than 0.5% of all ovarian tumors but represent
10% of the sex cordstromal neoplasms.13 Although most
of these tumors are masculinizing, some are nonfunctional
or even associated with estrogenic effects. Therefore the older
terms, arrhenoblastoma and androblastoma are no longer
favored. One third of cases occur in patients younger than
20 years of age. These tumors are almost always unilateral
and present as stage IA at diagnosis. Survival is excellent, with
tumor-related deaths in only 5% of affected individuals.32
Similar to granulosa cell tumors, the gross appearance of
Sertoli-Leydig cell tumors varies widely, but these lesions
are less often filled with hemorrhagic fluid and rarely have
a unilocular thin-walled cystic appearance. Current classifications now recognize five histologic patterns based on the
degree of differentiation and presence of heterologous, endodermal, or mesenchymal elements. Tumor stage and histologic
appearance are important prognostic factors. Sertoli-Leydig
cell tumors with heterologous elements are more common
in younger patients and may be difficult to distinguish from
immature teratomas.32 There are two phases of the masculinizing effects of androgen overproduction. Initially, defeminization takes place with amenorrhea, breast atrophy, and loss of
female body habitus. This may be followed or overlapped by
masculinization characterized by hirsutism, clitoral hypertrophy, and deepening of the voice. In prepubertal girls, masculinization and accelerated somatic growth predominate.
Postpubertal girls usually have menstrual irregularities, acne,
body habitus masculinization, and hirsutism. The virilizing effects are caused by testosterone accumulation resulting from a
deficiency in catabolizing enzymes. Gonadotropin levels are
low, and excretion of urinary 17-ketosteroids and pregnanetriol is normal. Because the testosterone level is often directly
related to tumor tissue volume, this hormone is a biologic
marker for monitoring disease behavior.133 Tumor markers
most likely to be elevated are alpha fetoprotein (AFP) and
CA 125.134 LDH may be elevated or normal. The hormonal
CHAPTER 39
profile of these lesions assists in differentiating them from

exogenous androgen sources, adrenal tumors, true hermaphroditism, and polycystic ovaries. Similar to granulosa cell
tumors, the Sertoli-Leydig cell lesions may be associated with
multiple enchondromas caused by nonhereditary mesodermal
dysplasia (Ollier disease).124
Surgical therapy should be conservative for patients with
low-stage disease. Unilateral oophorectomy or adnexectomy
is adequate for such disease and will preserve later childbearing capacity. If tumors are bilateral, poorly differentiated or
have ruptured or demonstrate aggressive behavior, a more
aggressive approach similar to that used for granulosa cell
tumors is necessary. Oral contraceptives and gonadotropinreleasing hormone agonists may provide some ovarian protection both during and following chemotherapy.135
Sex Cord Tumors with Annular Tubules Sex cord tumors
with annular tubules (SCTAT) are rare but distinct variants
of sex cordstromal tumors. They have potential for bidirectional differentiation into granulosa or Sertoli cells.13 These
lesions are observed in patients with Peutz-Jeghers syndrome.6
When associated with this syndrome, the lesions are small,
multifocal, and usually bilateral. The tumors are often calcified and are invariably noted incidentally during autopsy or
in an ovary removed for reasons unrelated to neoplasia.
Although patients with these tumors occasionally have menstrual irregularities suggesting hyperestrogenism, surgical
therapy is rarely indicated. When these tumors occur in the
absence of Peutz-Jeghers syndrome, the clinical difference is
significant. Such lesions occur in older patients with a mean
age of 34 years, although cases have been reported in patients
from 6 to 76 years of age. In the younger patients, the tumor is
unilateral and almost always larger than 5 cm in diameter;
20% are malignant. Even with aggressive therapy, 50% of
patients with these tumors die.136
Steroid Cell Tumors Steroid cell tumor is the now preferred
name for lesions previously called lipid cell tumors. This name
is more appropriate because of the morphologic features of the
tumor, its propensity to secrete steroid hormones, and because
many such lesions contain little or no lipids. The group is
subclassified into three major categories according to the cells
of origin: (1) stromal luteoma is a small steroid cell tumor
contained in the ovary arising from the stromal lutein cell;
(2) Leydig cell tumor contains the classic intracytoplasmic
Reinke crystals and arises from histologically similar precursor
cells found in the ovarian hilus; (3) steroid cell tumors not
otherwise specified account for approximately 60% of cases
and typically occurs in younger patients.
The first and second categories of lesion are usually encountered in postmenopausal women and are only rarely
reported in patients in the first 3 decades of life. Most of
the cases in the third category and in prepubertal children
have been associated with androgenic, heterosexual pseudoprecocity. The tumors are rarely estrogenic, but isosexual
pseudoprecocious puberty has been reported.137 The androgenic tumors show elevated testosterone and androstenedione
levels, increased urinary 17-ketosteroid excretion, and decreased gonadotropin levels. In children, these lesions are
virtually always benign and of a low stage. Unilateral
salpingo-oophorectomy is adequate treatment, but close
follow-up is essential. Most of the hormonal symptoms should
OVARIAN TUMORS
541
progressively resolve after removal of the tumor, although

younger children may develop true precocious puberty after
resection, because chronic androgen exposure appears to
induce an early maturation of the hypothalamus.138
Germ Cell Tumors
The path of descent of the primordial germ cells is imperfect;
as a result, some of the cells may occasionally miss their destination and be deposited anywhere along this migration
route. Germ cells have been found in the pineal area of the
brain, mediastinum, retroperitoneum, the sacrococcygeal
area, and the ovary and testis. If malignant transformation occurs at any of these sites, a gonadal or extragonadal neoplasm
will develop. Because these nests of cells are totipotential in
nature, a wide variety of tumors are seen. The specific type
of tumor depends on the degree of differentiation that has occurred. This has been characterized by Telium.27 According to
this schema, if no differentiation occurs, a germinoma develops; with differentiation, embryonal carcinomas occur;
and with extraembryonic differentiation, these lesions become
choriocarcinomas or endodermal sinus tumors. If embryonal
differentiation occurs, then the teratoma or most mature of
these tumors is seen.
Germ cell tumors are rare in children and adolescents, but
when they occur, the gonad is the most frequent site. The
ovary is the site of origin for 30% of all germ cell tumors in
children.139,140 Epithelial and stromal ovarian tumors prevail
in adults; germ cell tumors predominate in children. Several
large series of ovarian neoplasms report an incidence of germ
cell tumors ranging from 67% to 77%.5,141 This group of tumors develops from the same totipotential primordial germ
cell, but each neoplasm has different behavioral characteristics, and will be presented individually and then as a group
relative to overall management decisions.
Germinoma The term germinoma is used to include a group
of tumors with common histologic characteristics. It is the
primary malignant tumor found in dysgenetic gonads. This
tumor may be referred to as a seminoma if found in the testis,
a dysgerminoma in the ovary, and a germinoma in an extragonadal site. Germinomas are believed to arise from the totipotential germ cells that were present at the undifferentiated
stage of gonadal development.142 Germinomas represent
the most frequent ovarian malignant neoplasm seen both in
children and adults.32 They account for 26% to 31% of malignant ovarian tumors in children.143,144
Germinomas are most often seen in prepubertal girls and
young women, with 44% of cases occurring before 20 years
of age and 87% by 30 years of age.145 The typical patient is genotypically and phenotypically normal. These often large tumors may reach massive proportions and lead to abdominal
pain and symptoms of pelvic pressure, or symptoms related
to obstruction of the gastrointestinal or urinary tract. Occasionally, girls with these tumors present with an acute abdomen as a
result of torsion, rupture, or hemorrhage into the tumor.
Ascites may be present. In pure dysgerminoma, LDH is elevated
in 95% of patients, but other markers are negative. In the mixed
form of these tumors, other markers may be positive, including
neuro-specific enolase, beta-hCG, and CA 125, depending on
which germ cell component is present.146,147 Ovarian dysgerminomas may also be associated with a paraneoplastic syndrome causing hypercalcemia, which typically resolves with
542
PART III
removal of the tumor but may persist for several days.148 On

gross examination, these tumors appear bulky, encapsulated,
solid, and yellowish in color (Fig. 39-6); they can be bilateral
in 5% to 30% of cases.142,149,150 Germinomas have a rather
uniform microscopic appearance consisting of large, round
cells that have vesicular nuclei and clear-to-eosinophilic cytoplasm. These cells resemble primordial germ cells. Lymphoid
infiltrates may be present.
The management of germ cell tumors begins with surgical

excision. Conservative surgery with a unilateral salpingooophorectomy, thorough inspection of the contralateral ovary
with biopsy of suspicious lesions, and careful staging (as
outlined in the section on surgical approach) is mandatory.
Although these tumors are very radiosensitive, surgery alone
is adequate treatment in stage I disease. In more advanced
disease, radiation has been abandoned in favor of effective
C
FIGURE 39-6 A, This encapsulated mass from a 5-year-old girl with acute abdominal pain proved to be a dysgerminoma. The childs contralateral tube
and ovary are seen to the left of the tumor. A small portion of the ipsilateral tube and uterus were in the surgical specimen but uninvolved with tumor.
B, The cut surface of the tumor is characterized by lobules divided by thin, fibrous septae. C, Micrograph of a dysgerminoma demonstrating polygonal,
clear tumor cells divided into small lobules by fibrous septae that contain scattered lymphocytes.
CHAPTER 39
multiagent chemotherapeutic programs that include platinum, etoposide, and bleomycin, which is now standard
therapy.146,151,152
Endodermal Sinus Tumors Endodermal sinus or yolk sac
tumors are aggressive malignant neoplasms that, either alone
or as a component of a mixed germ cell tumor, are the second
most common histologic subtype of malignant ovarian germ
cell tumors in children and adolescents.32,153 In neonates
and young children the primary location of these tumors is
in the sacrococcygeal area. In older children and adolescents,
it is found most frequently in the ovary. The origin of this
particular tumor has been debated, and many microscopic
patterns of this tumor have now been described. Nogales suggested that this tumor originates from the primary yolk sac, a
structure that develops very early in embryogenesis and
consists of multipotential primitive endoderm.154 This tissue
is capable of differentiating epithelial somatic tissues as well as
secondary yolk sac tissue (a terminal, temporary structure
with limited differentiating capacity) and mesenchyme. Yolk
sac tumors with pure endodermal sinus subtypes are less
mature than the differentiated glandular or hepatoid subtypes.155 Symptoms are generally present for less than a
month and are related to the presence of an intra-abdominal
mass. Sixty-three percent of patients present with abdominal
pain and/or abdominal distention.156 Elevation of the biologic
marker AFP is the hallmark of this tumor.
The gross appearance of these tumors during surgery is pale
yellow-tan and slimy, with foci of cystic areas and necrosis.157
The tumors are soft and friable when handled. Most tumors
show a distinct histologic subtype with differentiation toward
vitelline or yolk sac structures.158 Microscopically, the most
common papillary pattern has the so-called endodermal sinus
structures (Schiller-Duval bodies) or perivascular sheaths of
cells. Most well-differentiated yolk sac tumors also contain
extracellular and intracellular droplets that are resistant to
periodic acidSchiff diastase staining and positive for AFP.
Embryonal Carcinomas A relatively uncommon isolated
germ cell tumor is embryonal carcinoma, which may resemble
an anaplastic carcinoma with extensive necrosis. Embryonal
carcinoma is more often found in association with other germ
cell tumors and is referred to as a mixed germ cell tumor. One
subtype of this tumor, the polyembryoma, is capable of
producing both AFP and beta-hCG, resulting in clinical endocrinopathies, including menstrual irregularities and isosexual
precocious puberty. The histologic appearance is characterized by bodies that resemble tiny embryos.159
The workup and surgical approach to this tumor is similar
to that for an endodermal sinus tumor. Isolated, unilateral
disease is managed by unilateral salpingo-oophorectomy.
Advanced local disease necessitates hysterectomy for local
control along with multiagent chemotherapy.160
Choriocarcinomas Choriocarcinomas are extremely rare in
the pure form but may be present in mixed germ cell tumors
as well. They are endocrinologically active, highly malignant
germ cell tumors that occur in girls and women. Estrogen is
produced both by the tumor and by the ovary itself in response
to release of gonadotropin by the neoplastic chorionic tissue.
The beta-hCG level is elevated, and AFP is normal. The clinical
presentation is influenced by the age of the patient. In a review
OVARIAN TUMORS
543
of 30 cases, Goswami reported a mean age of 13.9 years, with

the predominant presenting symptom being abdominal pain.
Ten cases occurred in prepubertal girls, three of whom developed isosexual precocious puberty, and in one case a mature
teratoma was identified in the contralateral ovary.161 These
usually large, solid tumors generally adhere to surrounding
tissues, and distant metastatic disease is associated with this
tumor. Operative excision can be a formidable task, because
the tumor may be friable, quite vascular, and often invades
contiguous structures.162 If the lesion is localized, surgery is
limited to unilateral salpingo-oophorectomy. However, this
rarely is the case, and a more extensive extirpative procedure
is usually required that involves removing the tumor, the opposite ovary, the uterus, and as much metastatic tissue as possible.
These tumors appear grossly as nodular with a friable consistency. The tumor is purple with variegated areas of dark
brown and yellow secondary to hemorrhage and necrosis.
Microscopic evaluation of these tumors reveals cytotrophoblasts and syncytiotrophoblasts with evidence of extensive necrosis and hemorrhage. Metastatic implants are friable and
have a similar gross and microscopic appearance as the
primary lesion. Survival is based on stage at diagnosis and
treatment. Platinum-based and methotrexate-based multiagent chemotherapy are described treatment regimens, and
platinum-based (bleomycin, etoposide, and cisplatin) chemotherapy has improved survival. Goswami reports an 82%
survival in patients treated with chemotherapy versus 28%
in those treated with surgery alone.161
Teratomas Teratomas are a group of neoplasms composed
of tissue elements that are foreign to the organ or anatomic site
in which they are found.163 Classically, these tumors are defined as being composed of tissue derived from the three germ
layers: ectoderm, mesoderm, and endoderm. All three germ
layers do not have to be present in each tumor, but some embryonic tissues must be found in an abnormal location. These
tissues show elements of disorganization as well as various
levels of maturation. As such, teratomas are histologically classified as mature and immature tumors and those with monodermal components.164,165 The development of a somatic
malignancy within a teratoma is a rare event in childhood,
and is thought to occur within differentiated teratomatous
elements rather than from totipotent embryonal cells.32
Mature Teratomas Most teratomas in children are of the
mature type. The majority of mature ovarian teratomas have entered, but have not completed meiosis, suggesting that they arise
from germ cells arrested in meiosis I.32 There is little or no tendency to malignant degeneration of preexisting benign elements
or the coexistence of malignant cells in a benign teratoma.107 In
neonates, mature teratomas are found most commonly in the
sacrococcygeal area followed by the head and neck.146,164,166
The ovary becomes an important site later in childhood,
especially during adolescence. Ovarian teratomas are predominantly cystic in nature.107 Overall, benign cystic teratomas are
the most common ovarian neoplasms in children162 and can
be bilateral in as many as 10% of patients.107,165,167
Symptoms of mature teratomas can be acute or chronic. Acute
symptoms that mimic appendicitis are seen when torsion,
hemorrhage, or rupture of the mass occurs. Gradual onset of
symptoms may be related to the presence of an intra-abdominal
adnexal mass, which may cause pressure on adjacent organs.165
Rarely, a ruptured teratoma may lead to a chronic inflammatory
544
PART III
response with the development of a mass of intestine and omentum adhering to the anterior abdominal wall; this condition is
associated with pelvic adenopathy, which mimics a malignant
tumor.168 On examination, findings are primarily related to
the mass itself. These tumors are located in the abdomen in
infants and young children. They are found in the pelvis of
adolescents, although large tumors may be palpated in the
abdomen, and there may be associated tenderness.
Plain abdominal radiographs demonstrate calcifications in
up to 67% of cases.169 Ultrasonography is a commonly used
diagnostic test. The positive predictive ability of ultrasonography approaches 100% when two or more characteristic findings for mature cystic teratoma (MCT), such as shadowing
echodensity and regionally bright echodensity, are present.170
Magnetic resonance imaging has been reported to be more
useful than CT scan in the diagnosis of mature cystic teratoma
due to its ability to clearly define soft tissue components.171
Conservative ovarian surgery in childhood and adolescence is important for the development of normal puberty
and future fertility. This must be balanced with complete removal of the mature cystic teratoma. Traditional management
of children with mature cystic teratomas has been oophorectomy by laparotomy. However, laparoscopic removal, either by
cystectomy or oophorectomy affords a safe alternative option
when done by an experienced laparoscopist.73 Campo and
colleagues, in a randomized controlled trial, demonstrated
that the use of an endobag in the removal a mature cystic
teratoma at the time of laparoscopy decreased spillage from
46% to 3.7% of cases.172 Aspiration of a giant predominantly
cystic lesion in order to facilitate removal through a smaller
incision runs the risk of upstaging the patient by spillage
of the cyst contents if malignant components are identified.
Techniques have been described to minimize this risk while
allowing a less invasive approach to large cystic lesions.173,174
Every effort should be made to spare the ovary when a
teratoma is suspected based on radiographic findings and normal tumor markers. Very large or bilateral teratomas can be
successfully enucleated in an attempt to preserve hormonal
and reproductive functions (Fig. 39-7).165,175, If this is not
possible, the gonad and tumor alone should be removed, leaving the ipsilateral fallopian tube in place.
Miliary, intraperitoneal glial implants (gliomatosis peritonei) are occasionally encountered in association with mature
teratomas.176 These implants are rarely suspected before
surgery. They appear as white or gray nodules, usually 1 to
3 mm in diameter, and are usually confined to the omentum,
pelvic peritoneum, or adjacent or adherent to the tumor itself.
Several explanations have been offered for the development of
these implants.177 The most recent data using microsatellite
DNA analysis suggest that the glial implants arise from subperitoneal cells, presumably pluripotent mullerian stem cells and
not from the teratoma.178,179 Implants can have a disturbing
appearance and biopsy is necessary, but no specific treatment
is indicated when they are well differentiated, and their presence does not change management of the primary tumor.
However, if adjacent components are immature, the lesions
may progress and require adjuvant therapy.
Immature Teratomas Immature teratomas are germ cell
neoplasms that are composed of tissue derived from the three
germ cell layers. These teratomas are clinically distinct from
benign or malignant teratomas, because they also contain
immature, neuroepithelial elements (see Fig. 39-7). Immature
teratomas can coexist with the more mature solid or cystic

benign teratomas or with malignant teratomas, in which case
treatment is determined by the malignant component.180
Immature teratomas are graded based on the relative quantity
of immature elements and the presence and quantity of the
neuroepithelial components. The grade of the primary tumor
is significant and is one of the major determinants of the likelihood of recurrence following resection. Multiple grading
systems have been proposed based on the system developed
by Thurlbeck and Scully.181 The criteria outlined by
Gonazez-Crussi identified the percentage of incompletely differentiated (embryonal) elements in the tumor as follows:
grade 0, 0%; grade I, less than 10%; grade II, 10% to 50%;
grade III, greater than 50%.163,182
The treatment of immature teratomas has gone through an
evolution from aggressive treatment with surgery followed by
multidrug chemotherapy to conservative surgical approaches
with no adjuvant therapy. In a study of 58 pure immature teratomas published in 1976 by Norris,180 survival was 82% for
patients with grade I tumors, 62% for grade II, and 30% for
grade III. Based on this study, along with others, use of multiagent chemotherapy for grade III immature teratomas was
advocated. The protocol for extracranial nontesticular germ
cell tumors of the German Society for Pediatric Oncology
and Hematology (GPOH), which was initiated in 1983,
recommended adjuvant chemotherapy for grade II and III immature teratomas of all nontesticular sites.183 Using this
approach, the relapse rate was 13.3% for patients with mature
and immature lesions. In a follow-up study from the German
registry, immature lesions had a higher rate of recurrence than
mature lesions when completely resected (8/78 vs. 3/104), as
in the previous study, but the recurrence rate overall dropped
from 13.3% to 9.5%. Complete resection was associated with
a relapse rate of only 4.2% in both studies, and the malignant
relapses were explained by microfoci of yolk sac tumor present in the primary tumor as shown retrospectively in single
cases by reevaluation of the primarily resected teratoma.184
The hypothesis that recurrent tumor stems from microfoci
of malignant cells present in the original mass is supported
by an intergroup study in the United States in which yolk
sac tumor elements were detected in 29% of immature
teratoma specimens (73 immature teratoma, 21 with YSTmicrofoci). It was suggested that the true incidence of such
microfoci might be underestimated in these typically large
masses, as a result of sampling errors.
The combined report from the Childrens Oncology Group
and the Pediatric Oncology Group in 1999 included 31 patients with pure immature teratomas of the ovary treated with
surgery alone. Eighty-six percent of the tumors were grade I or
II and the 3-year event-free survival (EFS) was 97.8%, with only
one patient developing recurrent disease. That patient was salvaged with a combination of surgery and platinum, etoposide,
and bleomycin. The authors advocate surgical excision alone,
with close follow-up as appropriate therapy for all ovarian immature teratomas.185 Based on the excellent survival and avoidance of the risks of chemotherapy, immature teratomas are
treated in the United States with fertility-preserving surgery
and observation without adjuvant chemotherapy.185,186
Monodermal Teratomas A monodermal teratoma refers
to an ovarian tumor composed exclusively or almost exclusively of ectoderm or mesoderm or endoderm, for example,
neuroectoderm.187
CHAPTER 39
OVARIAN TUMORS
545
FIGURE 39-7 A, Large ovarian dermoid tumor in a 14-year-old girl with acute severe abdominal pain upon awakening. The fallopian tube is seen below
the tumor. B, Opened gross specimen of ovarian dermoid showing multiple tooth- and jawlike calcifications. C, Characteristic gross appearance of an
immature teratoma in a 5-year-old girl who presented with a left ovarian mass. The tumor is a solid and cystic globoid mass with a smooth, shiny surface.
D, Cut section of an immature teratoma shows a variegated, solid, cystic appearance with focal areas of hemorrhage.
Gonadoblastomas
Gonadoblastoma, a tumor first described by Scully188 in 1953,
is relatively rare and occurs most commonly in patients with
dysgenetic gonads. Most patients are virilized or nonvirilized
phenotypic females. In the only large series reported, Scully189
reviewed 74 cases and found that 89% were chromatin negative
and the most common karyotype was 46XY or 45X/46XY.
Troche, in a literature review of 140 cases of neoplasms arising
in dysgenetic gonads, found that 80% also had these karyotypes.190 Patients are usually older adolescents or in the third
decade of life with a history of primary amenorrhea. Androgen
production by the tumor causes virilization. When a workup
for amenorrhea or virilization is undertaken, an abnormal
karyotype with a Y chromosome or chromosome fragment
can be found in as many as 90% of patients.190 These often
small tumors may then be identified during examination or

exploration. They may also be found incidentally during excision of gonadal streaks or dysgenetic gonads.191,192 These tumors become invasive early and gonadectomy is recommended
as soon as 46XY gonadal dysgenesis is diagnosed.18,190
Gonadoblastomas are composed of germ cells and sex-cord
derivatives that are similar to granulosa and Sertoli cells, although immunohistochemical and ultrastructural findings
are more supportive of Sertoli-like differentiation.193 Lutein
or Leydig-like stromal cells occur in two thirds of cases and
probably reflect a stromal reaction to gonadotropin stimulation.193 These tumors are considered precursors to germ cell
tumors in dysgenetic or streak gonads, because they may coexist with dysgerminomas and other germ cell tumors in more
than half of the patients.190 The tumor may be difficult to
identify on gross examination because of overgrowth by the
546
PART III
malignant component and other changes, including calcification, fibrosis, or both. In fact, calcification may be the only
remnant of the gonadoblastoma, and the presence of calcification in a dysgerminoma should raise the suspicion of an underlying gonadoblastoma. The malignant potential of this
tumor is determined by the underlying malignant component
and should be treated accordingly. The outcome for patients
with these tumors may be improved, because abnormal sexual
development prompts early evaluation of the patient and
subsequent diagnosis of the tumor. The prognosis of nongerminomatous germ cell tumors has improved with the advent
of bleomycin, etoposide, and cisplatin protocols, and survival
rates of 70% to 90% have been reported.32
Mixed Germ Cell Tumors
Germ cell tumors in children are often composed of more than
one pure histologic type. Benign but questionably malignant tumors (i.e., immature teratomas) and frankly malignant tumors
(germinomas, choriocarcinomas, endodermal sinus tumors,
and embryonal carcinomas) may be present. Management of
mixed tumors is geared toward the most malignant component
of the mass.
Surgical Guidelines for Ovarian

Germ Cell Tumors
------------------------------------------------------------------------------------------------------------------------------------------------
The goal of surgery is to completely evaluate the extent of

disease, safely and completely resect the tumor, and spare
all uninvolved reproductive organs. Preservation of reproductive potential is a high priority during surgery for ovarian
lesions in children. Laparoscopic procedures are being increasingly performed for evaluation of pelvic masses, and
there are now data to demonstrate that the benefits of a faster
recovery time and shorter hospital stay seen in adults are also
applicable to children.73,194 If a suspected ovarian malignancy
is detected at the time of laparoscopy, complete surgical
staging and resection by conventional laparotomy is recommended. Benign lesions require only tumor resection by
ovarian cystectomy or unilateral oophorectomy.
Benign tumors, frankly malignant tumors, and those with
mixed histologic characteristics often cannot be distinguished
based on gross appearance alone. If in doubt, staging is recommended, because treatment and prognosis of malignancies depend on accurate staging. The current intergroup COG
protocol includes thorough inspection, palpation, and biopsy
of any suspicious peritoneal and liver nodules (including the
subphrenic spaces).49 Both ovaries are inspected. If a tumor is
found in an ovary and malignancy is suspected, it should be
removed by unilateral oophorectomy if the fallopian tube is
not involved. A salpingo-oophorectomy is indicated if the
fallopian tube is involved. The contralateral ovary should
be inspected, and nodules or suspicious areas should be
biopsied. A contralateral salpingo-oophorectomy should be
avoided unless malignancy is confirmed.
Staging procedures for malignancies differ somewhat for
different cell types, which can result in inadequate staging of
unsuspected epithelial tumors. Staging guidelines for germ cell
tumors proposed by the Childrens Oncology Group (see
Table 39-6) include peritoneal fluid aspiration/washings, inspection of the omentum and contralateral ovary with biopsy
of suspicious lesions, biopsy of clinically suspicious lymph

nodes, and removal of the primary tumor. Epithelial tumors
are staged by the FIGO system (see Table 39-5), which requires
peritoneal biopsies, peritoneal washings/aspiration, omentectomy, removal of the primary tumor, and an ipsilateral lymph
node dissection. The need for a lymph node dissection is not
based on the gross appearance of the nodes, because up to
30% of clinically normal nodes can be positive for metastatic
disease.
Chemotherapy for Ovarian

Germ Cell Tumors
------------------------------------------------------------------------------------------------------------------------------------------------
Forty years ago, no effective therapy for germ cell tumors

existed. Based on the early success of management of testicular
germ cell tumors using multiagent platinum-based chemotherapy, ovarian tumor treatment evolved along similar lines.
The addition of chemotherapy reduced the risk of recurrent
disease for adult patients with completely resected ovarian
germ cell tumors.195 Current regimens for ovarian germ cell
and sex cordstromal tumors is platinum-based therapy,
and the regimen of cisplatin, etoposide, and bleomycin
(PEB) has become the preferred protocol. An 8-year study
from the GOG that closed in 1992 evaluated PEB, and 91
of 93 patients were free of recurrent germ cell tumors, with
a median follow-up of 38.6 months.196
Several chemotherapeutic regimens were also historically
tried in children, and the best results were achieved with
PEB.197,198 In a pilot study, Pinkerton and colleagues199 demonstrated the effectiveness of substituting cisplatin with carboplatin, a less toxic drug; carboplatin was then combined with
bleomycin and etoposide. Eight of eight patients with ovarian
germ cell tumors survived with this regimen. Using a
platinum-based regimen, only 1 of 17 girls with resected ovarian
nonseminomatous germ cell tumors in FIGO stage IA relapsed in
an analysis of European trials by Gobel and colleagues.200
In 1991, the Childrens Cancer Group (CCG) experience of
93 children with malignant germ cell tumors included 30 ovarian tumors.151 By study design, immature teratomas and dysgerminomas were not included. Using a cisplatin-based
regimen, the 4-year, event-free survival rate was 63%. Tumor
size affected prognosis. If the tumor was larger than 16 cm in
diameter, the outcome was worse. Patients in whom complete
tumor resection could not be done during the original procedure were more likely to have subsequent adverse events than
if the tumor was completely removed (P 0.08). In 1994, Nair
and colleagues201 reported their findings in 107 children with
germ cell tumors, including 43 girls with ovarian tumors. Of
these, 22 received multiagent chemotherapy. A complete response was seen in 6 of 11 patients treated with platinum, vinblastine, and bleomycin, compared with 10 of 11 patients who
completely responded to treatment with PEB (with etoposide
replacing vinblastine). The risk for chemotherapy-related
complications is low relative to the effectiveness of the PEB
regimen and compared with prior regimens that included
vinblastine.202 Others have shown that the PEB regimen is
superior to other chemotherapy regimens.160
Current efforts in the United States are geared toward
reduction of therapy for low- and intermediate-risk tumors.
A phase III study undertaken by the Childrens Oncology
CHAPTER 39
Group (COG-AGCT0132) stratified malignant germ cell tumors into three risk groups (low, intermediate, and high risk)
defined by stage and primary site. Based on data from the POG
9048/CCG 8891 study, demonstrating that patients with stage
I ovarian and extragonadal immature teratoma with malignant
elements appeared to do well following complete surgical resection,69 all patients with stage I ovarian tumors were categorized as low risk and were initially treated with surgery,
followed by close observation and monitoring. That arm of
the study has subsequently been amended to include stage I
ovarian tumors in the intermediate-risk group because of a
higher-than-expected failure rate with observation alone.
Overall survival remains greater than 95%. The intermediaterisk group will consist of patients with stage I to III gonadal
tumors. Such patients have been shown to have a 3-year EFS
of about 90% with standard-dose PEB.49,71 These patients will
be treated with a modified standard PEB regimen, consisting of
three cycles of compressed PEB every 21 days. Saxman and
colleagues203 reported that long-term survival was equivalent
for men treated with germ cell cancer for three or four cycles
of PEB. Patients who are partial responders (PR) may then
have surgical resection of residual tumor. Therapy is discontinued upon pathologic complete response and normal
markers, or continued for an additional three cycles in children
who remain PR. High-risk patients, defined as those with stage
IV disease, showed some improvement in survival with a highdose platinum regimen that was offset by increased toxicity.
Patients with recurring germ cell tumors may be salvaged
using high-dose chemotherapy with autologous stem cell
transplantation.
Miscellaneous Tumors
------------------------------------------------------------------------------------------------------------------------------------------------
Small cell carcinoma of the ovary is an extremely rare condition with a very poor prognosis.204 These tumors are very aggressive and are the most common undifferentiated ovarian
carcinoma in young patients. They have been encountered
in patients from 9 to 44 years of age, with a mean age of
23 years.205 Paraendocrine hypercalcemia occurs in two
thirds of cases, but patients rarely have clinical manifestations
of this abnormality. Serum parathormone levels are normal.
Virtually all tumors are unilateral, although only 40% have
been detected at stage 1A. Only one third of patients with stage
1A tumors survive long-term, and survival of patients with
more widespread disease is rare.13 Unilateral salpingooophorectomy has been associated with long-term survival in
some patients with stage 1A tumors. Asynchronous appearance
of tumor in a contralateral conserved ovary has been encountered, and bilateral adnexectomy may be a more appropriate surgical option. Despite various treatment modalities including
resection, radiation therapy, and intensive chemotherapy, the
average life expectancy remains low at 18 months.205
Primary ovarian sarcomas are a heterogenous group of aggressive tumors associated with poor survival. Most cases occur in older women; however, a recent review of 151 cases
described 10 of 29 patients with rhabdomyosarcoma who
were younger than 20 years of age.206 These patients presented with nonspecific symptoms of abdominal discomfort
or swelling with occasional urinary or gastrointestinal complaints secondary to mass effect. Accurate staging is critical.
Hysterectomy with bilateral salpingo-oophorectomy and
OVARIAN TUMORS
547
debulking of as much diseased intra-abdominal tissue as possible has been done. Radiation therapy was administered for
residual pelvic disease, and several chemotherapeutic regimens have been used. In contrast to rhabdomyosarcomas arising at other sites, the outcome for patients with ovarian lesions
has generally been poor, perhaps because of the advanced
stage of disease at diagnosis. Nevertheless, the most recent
chemotherapeutic regimens used in cooperative group studies
have been highly effective, and it is reasonable to assume that
more conservative surgical resection will provide adequate
treatment for these rare tumors.
Stromal sarcomas and low-grade endometrial stromal sarcomas of the ovary have been occasionally reported in the second decade of life. These lesions are believed to arise from
ovarian endometriosis, coelomic mesenchyme, or neometaplasia of stromal cells. Lesions are usually discovered because
of nonspecific pelvic discomfort, although early infiltration
into adjacent tissues may cause intestinal or ureteral obstruction. Tumor infiltration may not be grossly apparent, so initial
surgical resection should be aggressive with total hysterectomy and bilateral salpingo-oophorectomy. Progesterone administration may provide effective adjunctive therapy,
although this has to be continued indefinitely because stromal
sarcomas have been reported to reappear and spread dramatically when the medication is stopped. Radiation therapy has
been used for local residual disease, although recurrence is
common. The role of chemotherapy for these tumors has
not been defined.
Cases of genuine ovarian fibrosarcoma in children are extremely rare. Patients present with pelvic pain and a palpable
mass. Fibrosarcoma has been associated with Maffucci syndrome.8 Although the outcome has been uniformly poor in
older patients, survival of younger patients who have undergone aggressive surgical resection, including hysterectomy
and bilateral salpingo-oophorectomy, has been reported.
Success with subsequent radiation or chemotherapy has not
been reported.
Primary leiomyosarcoma of the ovary is extremely rare in
children. These tumors may arise de novo from any of the
smooth muscle sites in the ovary or may represent malignant
degeneration of leiomyoma, a benign counterpart.207 As with
most of these rare tumors, presenting symptoms are nonspecific and discovery may occur in the advanced stage of disease.
Aggressive surgical therapy is recommended, because no
adjuvant therapy has proven to be effective.
Secondary Tumors
------------------------------------------------------------------------------------------------------------------------------------------------
Although secondary ovarian malignancy is rare, the ovaries are

a potential metastatic site for a wide variety of childhood malignancies (Table 39-7).127,208 Distinguishing primary neoplasms from secondary neoplasms is important to prevent
inappropriate therapy or adverse sequelae. Metastatic spread
to the ovary occurs through four main pathways: (1) hematogenous spread, (2) lymphatic spread, (3) transcoelomic dissemination with surface implantation, and (4) direct
spread.208 Recently described highly malignant tumors that
have a predilection for the pelvic region are intra-abdominal
desmoplastic small round cell tumors.209
Lymphoma can occur in the ovary in children either as a
primary tumor or a manifestation of systemic disease. Most
548
PART III
TABLE 39-7
Secondary (Metastatic) Tumors Occurring in the Ovary
in Children
Colorectal
Breast
Gastric carcinoma
Carcinoid tumors (liver, lung)
Malignant melanoma
Burkitt lymphoma
Rhabdomyosarcoma
Wilms tumor
Neuroblastoma
Retinoblastoma
Ewing sarcoma
Rhabdoid tumor of the kidney
Medulloblastoma
Osteogenic sarcoma
Chondrosarcoma
Leukemia
of the reported cases have been of the small, noncleaved cell

type (Burkitt or non-Burkitt category), although T-cell nonHodgkin lymphoma and anaplastic large cell lymphoma have
also been reported.210 Pais and colleagues211 reviewed 23
cases of ovarian involvement in patients with relapsing leukemia. Abdominal pain was the most common symptom, and a
mass could usually be palpated. Although most patients in
whom leukemia treatment failed had systemic and not local
disease, ultrasonography revealed a characteristic appearance
and was effective in detecting ovarian involvement.212 Survival was based on aggressive systemic multiagent chemotherapy and not on the degree of surgical resection of the ovarian
lesion. Routine pelvic radiation therapy was of no benefit.
Reports have noted granulocytic sarcoma of the ovary occurring in patients with acute or relapsed acute myelogenous
leukemia.213 Although aggressive systemic chemotherapy is
critical to survival, an ovarian mass should be investigated immediately to determine its nature (i.e., benign or malignant
and exact cell type). In this instance, surgical resection of
the ovary and any other involved gynecologic organs or pelvic
tissue must be done. Radiation therapy has been used for residual disease in the pelvis. Although the ultimate outcome of
granulocytic sarcomas is probably more related to effectiveness of chemotherapy, local measures of tumor control cannot
be overlooked when this tumor is detected.
Unclassified Benign Tumors

------------------------------------------------------------------------------------------------------------------------------------------------
Although the ovary is highly vascularized, hemangiomas are extremely rare; a recent review found only 40 published cases.214
Their occurrence is relatively evenly distributed between infancy
and postmenopausal age groups. The lesions are usually quite
small, asymptomatic, and discovered incidentally. Bilateral
occurrence is rare, and the tumors are almost always cavernous.
Benign-appearing ultrasonographic features have been described.215 When the tumors are large, associated symptoms
include abdominal pain, distention, and bloody ascites. Torsion
or rupture may cause an acute surgical emergency. No malignant
tumors of this type have been described, and oophorectomy or

adnexectomy is curative if needed.
Primary ovarian leiomyomas are also extremely rare, although they have been reported in teenage girls.216 Most
reported cases are clinically silent; however, the lesion may
be large enough to cause increased abdominal girth and pelvic
pain. Tumor markers are normal, and imaging studies are generally unable to differentiate this benign solid tumor from a
malignant process. Unilateral salpingo-oophorectomy is curative. The ovarian myxoma is a rare benign tumor characterized
by conspicuous vascularity and mesenchymal proliferation
that requires only a conservative surgical procedure.217
Struma ovarii is a benign variant of a germ cell tumor that
typically occurs in older women but has been reported in teenagers. It is composed of more than 50% benign thyroid tissue,
which is functional in 5% to 12% of cases. Rarely, the tumor
contains malignant components and, in some cases, represents the patients only functioning thyroid tissue. CA 125
levels may be elevated, but other markers are usually normal.
Treatment is resection of the mass.218 In another thyroidrelated condition known as the Van Wyk and Grumbach
syndrome, long-standing hypothyroidism can lead to large
ovarian cysts. TSH levels are extremely high and several
theories hypothesize a crossover hormonal effect on FSH or
direct stimulation of the ovary by TSH. CA-125 and LDH
levels may be elevated. The ovarian cysts resolve with thyroid
replacement therapy.219
Summary
------------------------------------------------------------------------------------------------------------------------------------------------
The diagnosis and management of ovarian lesions in infants

and children remains a challenge because of the wide variety
of possible pathologies, some of which are extremely rare.
Nonneoplastic lesions are being detected more commonly
as imaging techniques continue to improve. Neoplastic lesions
are more readily diagnosed and completely characterized with
advances in biochemical, immunohistologic, and cytogenetic
technology.
Because of the relative rarity of ovarian tumors in children,
clinical approaches may be based on experience with similar
adult lesions. However, it is critical to recognize the differences exhibited by the juvenile forms of many of these entities,
which often present at a less advanced stage and have a more
favorable natural history and response to therapy. Preservation
of reproductive and endocrine function is of paramount importance in the treatment of ovarian lesions in infants and
children. Careful observation or nonoperative therapies may
be appropriate for many nonneoplastic conditions. Most benign neoplasms are adequately managed with conservative
surgical approaches. Even frankly malignant tumors increasingly yield to multimodal therapy, which can include less radically ablative surgery and still result in long-term survival and
possible preservation of fertility for young patients.
expertconsult.com.
CHAPTER 40
Testicular Tumors
Bryan J. Dicken and Deborah F. Billmire
Historically, prepubertal malignant testicular tumors were

managed using the same treatment protocols as their adult
counterparts, with radical orchiectomy and retroperitoneal
lymph node dissection (RPLND) weighing heavily in the treatment pathway.1,2 However, with growing clinical evidence, it
became clear that prepubertal tumors differed from the postpubertal population not only in presentation but differed in
terms of clinical behavior, incidence, histologic diagnosis,
and prognosis.16 In recognition of the differences in this
population of patients, the Prepubertal Testis Tumor Registry
(PTTR) of the urologic section of the American Academy of
Pediatrics was established in 1980 to better delineate the natural history of these lesions and to document their response to
therapy.1,2,4 Since its inception, several important features
have emerged that have significantly altered the management
of testicular tumors in the pediatric population. The management has been further clarified by a series of recent multicenter clinical trials of the most common malignant tumors of the
prepubertal testis.712
The results of the PTTR confirmed that testicular tumors
in children are rare, making up approximately 1% to 2%
of all pediatric solid tumors, with an incidence of 0.5 to
2/100,000 among whites, while African-American males
appear somewhat protected, with an incidence of 0.25/
100,000.13,13 Asian/Pacific Island males have a 1.4-fold
increased risk of testicular tumors compared with whites.14
In contrast to adult testicular cancer, which has experienced
a marked increase in incidence, pediatric testicular tumor

incidence has been stable during the past 30 years.15 Testicular tumors are 10 times more frequent in the postpubertal
cohort compared with boys younger than 12 years of age.1
Furthermore, epidemiologic data from several sources suggest
a bimodal distribution, with a small distinct peak in the first
3 years of life, followed by a large peak in adolescents (15 to
18 years).2,3 The majority of testicular tumors in the postpubertal age group are malignant, with 90% to 95% demonstrating histologic features of either seminoma or mixed germ
cells.2,6 Initially, the PTTR reported that the yolk sac tumor
(62%) was the most common prepubertal testicular tumor,
with benign tumors occurring much less commonly.1,2,13
However, a landmark paper by Metcalfe and colleagues13 suggested that the PTTR registry and the Armed Forces Institute of
Pathology American Tumor registry are subject to reporting
bias, with overreporting of malignant tumors and failure to
capture the benign tumors. In a series of articles that followed,
74% to 87% of tumors identified were benign, with teratoma
making up 43% to 48%, while malignant yolk sac tumors
constituted only 15%.1,5,6,13 Recognition of this fact led to a
marked reassessment of the management of testicular tumors
in the prepubertal population.
Risk Factors for Testicular Cancer

------------------------------------------------------------------------------------------------------------------------------------------------
Although a number of risk factors have been proposed regarding the occurrence of testicular tumors, to date only a few may
be considered as established based upon a sufficient level of
evidence.16 Other associations that have historically been considered important etiologically have since been refuted. Only
four factors have sufficient evidence that links them highly
with testicular cancer: (1) undescended testis (cryptorchidism), (2) contralateral testicular germ cell tumor (GCT),
(3) familial testicular germ cell tumor, and (4) gonadal dysgenesis.16,17 Associations that may be considered likely include
infertility, twin-ship, and testicular atrophy. Clinical factors
with equivocal/low association include scrotal trauma, inguinal hernia, mumps orchitis, testicular torsion, maternal estrogen exposure, and occupational exposure. Parameters that
have historically drawn attention but have since been shown
to be irrelevant include obesity, vasectomy, smoking, hydrocele, varicocele, alcohol, and circumcision.16
Cryptorchidism occurs in 2% to 5% of term infant males;
however, by 12 months of age, this number is reduced to
1%.18 To date, cryptorchidism is the only factor that has level
I evidence linking it with testicular cancer. A meta-analysis of
20 case control studies showed a strong association between
undescended testis (UDT) and testicular cancer, with an overall relative risk of 4.8.16 Similarly, Walsh and colleagues19
showed boys who underwent orchiopexy after 10 years of
age had a 3.5-fold increased risk of testicular cancer, compared with those that had the procedure at an earlier age.
In a population-based prospective observational study, Pettersson and colleagues followed 16,983 men treated for
UDT for a mean period of 12.4 7.4 years.20 This study demonstrated two important findings. There was an increased risk
of testicular cancer for the entire cohort (relative risk [RR]
2.23) versus normal population figures, and the incidence of
cancer was significantly higher (RR 5.4) in those who were
549
550
PART III
treated after the age of 13 years.20 The lowest incidence of

cancer was seen in children who underwent orchiopexy before the age of 6 years (RR 2.02). Orchiopexy before the
age of 10 to 12 years results in a twofold to sixfold relative risk
decrease in testicular cancer in children with unilateral UDT.21
Because of the increased risk of malignancy, patients with
UDT seen after age 10 may still be candidates for orchiopexy
with close surveillance; however, consideration of testicular
biopsy may be useful in directing therapy. The decision of
timing for orchiopexy should include consideration not
only of an effort to reduce the incidence of testicular cancer,
but also a consideration of the possibility of spontaneous
descent and the evidence regarding preservation of fertility.
Canavese and colleagues demonstrated an inverse relationship
between age at orchiopexy and total sperm counts and sperm
motility, and they recommended orchiopexy during the first
year of life.22 Taking all factors into account, consideration
should be given to orchiopexy in all children if complete
descent has not occurred by 12 months of age.21
------------------------------------------------------------------------------------------------------------------------------------------------
The most common presentation of a testicular tumor is a

nontender scrotal mass, accounting for 50% to 85% of
cases.5,6,13,23 The presentations of children that were subsequently diagnosed with a prepubertal tumor have included
trauma and persistent swelling (3%), hydrocele (10%), epididymitis (13%), incidental discovery during surgical repair
of a congenital or acquired disorder (53%), testicular pain/
torsion (21%), and bruising.5,6,13,23 Tumors may also be diagnosed by ultrasonography during investigations for UDT or
nonresolving acute hydroceles.
Physical examination should differentiate between those
problems arising from the cord (varicocele, spermatocele, epididymitis) and those arising from the testicle (trauma, orchitis,
tumor). There may be bruising or a hydrocele present that may
confound the diagnosis, because both of these findings may
coexist with a tumor. This is particularly true in cases where
preceding trauma draws attention to the scrotal area. Careful
evaluation of the childs pubertal status relative to their chronologic age is important, because stromal cell tumors may present with precocious puberty (Leydig cell) or gynecomastia
(Sertoli cell).
Diagnosis
with intervening solid components with calcifications (bone

or psammoma bodies).13 This contrasts with malignant lesions, which tend to be more solid in appearance. In cases
where malignancy is suspected, computed tomography (CT)
of the chest, abdomen, and pelvis should be obtained to exclude metastatic disease to the most common siteslung
and retroperitoneum.2,4
Tumor Markers
------------------------------------------------------------------------------------------------------------------------------------------------
Serum tumor markers are essential in the workup and postoperative monitoring of children with testicular tumors. Human
chorionic gonadotropin (HCG) and a-fetoprotein (AFP) are
important markers for certain malignant germ cell histologies.2 AFP is secreted by yolk sac tumors in up to 90% of cases,
and b-HCG is secreted by choriocarcinoma. HCG has a halflife of 24 hours, whereas AFP has a half-life of 5 days. In the
prepubertal age group yolk sac tumors are the most common
malignant histology, and AFP is very important, whereas HCG
is rarely elevated. An important consideration is that AFP is
normally very high in infancy, and remains elevated for up
to 8 months, decreasing to adult levels around 1 year of
age.2,23 Older boys are more likely to have malignant germ cell
tumors of mixed histology, and both AFP and HCG may be
elevated. For those patients with elevated tumor markers at
diagnosis, serial AFP and HCG should be monitored monthly
in the first postoperative year, then every other month in the
second year to follow current recommendations.25 Patients
presenting with precocious puberty and a testicular mass
should prompt assessment of a urinary 17-ketosteroid, serum
luteinizing hormone (LH), follicle-stimulating hormone
(FSH), and testosterone. Unlike precocious puberty induced
by a pituitary lesion, in which the LH, FSH, and testosterone
are high, testicular tumors display a low LH and FSH and a
high testosterone.
Classification and Stage

------------------------------------------------------------------------------------------------------------------------------------------------
Table 40-1 lists the histologic diagnoses for prepubertal testicular tumors from several institutions, and the diagnoses are
compared with the 2002 AAP tumor registry.1,13,26 This table
demonstrates the reporting bias of the national tumor registry
and the population-based distribution of all testicular tumors.
Table 40-2 outlines the Childrens Oncology Group (COG)
testicular tumor staging system.7
------------------------------------------------------------------------------------------------------------------------------------------------
In addition to a history and physical examination, all boys

with testicular masses, and those with a tense hydrocele or
with a suspicious examination, should undergo a scrotal ultrasound. Although preoperative ultrasound is highly sensitive
for distinguishing intratesticular from extratesticular tumors,
it has poor specificity to distinguish between benign and malignant lesions.5,13,24 Tumor size (volume) on ultrasonography has not been shown to be indicative of benign or
malignant tumors.5 Sonographic features suggestive of benign
tumors (epidermoid cyst) include intratesticular cystic lesions
with a hypoechoic center, representing central keratinizing
debris, and an outer hyperechoic rim. Features of a teratoma
may include an entirely intratesticular cystic, septated mass
PRIMARY TESTICULAR TUMORS

Epithelial-Based Tumors
Epidermoid Cysts The epidermoid cyst is a benign tumor,
accounting for 2% to 14% of testicular tumors in the prepubertal population.1,5,13 They are hormonally inactive and
typically present as a smooth, firm intratesticular mass. The
tumor consists of a cystic structure filled with keratinizing
squamous epithelium, contributing to a characteristic ultrasound appearance: central hypoechoic mass, a surrounding
echogenic rim, or a mixed internal echogenicity.27,28 Epidermoid cysts are rare, making up only 1% of testicular tumors.
CHAPTER 40
TABLE 40-1
Differences in Distribution of Testicular Tumors Based on
Tumor Histology among Study Sites
Tumor Type
Benign
Teratoma
Epidermoid cyst
Leydig cell
Sertoli cell
Juvenile granulosa
cell
2002
Registry %
(N 395)
Pohl %
(N 98)
Metcalfe %
(N 51)
Ciftci %
(N 51)
23
3
1
3
3
48
14
4
3
5
43
10
0
4
0
18
6
6
0
N/A
Malignant
Yolk sac
62
Mixed germ cell
0
Rhabdomyosarcoma
4
Gonadoblastoma
1
15
8
0
8
Excluded 25
2
2
45
6
19
0
N/A, not available.
TABLE 40-2
Staging of Testicular Malignant Germ Cell Tumors
Testicular
Stage
I
II
III
IV
Limited to testis, completely resected by high inguinal

orchiectomy; no clinical, radiologic, or histologic
evidence of disease beyond the testis; tumor markers
normal after resection
Transscrotal orchiectomy; microscopic disease in
scrotum or high in spermatic cord; retroperitoneal
node involvement (<2 cm) and/or increased tumor
markers after resection
Gross residual disease, retroperitoneal lymph node
involvement (>2 cm), or malignant cells in pleural or
peritoneal fluid
Distant metastases involving lung, liver, brain, bone,
distant nodes, or other sites
From Cushing B, Giller R, Cullen JW, et al: Randomized comparison of

combination chemotherapy with etoposide, bleomycin, and either highdose or standard-dose cisplatin in children and adolescents with high-risk
malignant germ cell tumors: a pediatric intergroup studyPediatric
Oncology Group 9049 and Childrens Cancer Group 8882. J Clin Oncol
2004;22:2691-2700.
These cysts lack atypia and mitotic activity. Although some

epidermoid cysts show loss of heterozygosity for certain chromosomal loci, there is currently debate as to whether they represent a true neoplasm.29
Stromal Tumors
Sex Cord-Stromal Tumors The stromal tumors consist
of three subtypes: Leydig cell, Sertoli cell, and juvenile
granulosa cell tumors. This group of tumors accounts
for 8% to 11% of pediatric tumors.1,13 The vast majority
of stromal tumors are benign, compared with a 10% rate
of malignancy in postpubertal males.
Leydig tumors and granulosa cell tumors are universally
benign in children. Leydig cell tumors tend to present in boys
5 to 10 years of age and with precocious puberty.2 The precocious puberty is a peripherally driven etiology; therefore, the
TESTICULAR TUMORS
551
hormone profile consists of a low luteinizing hormone (LH),

low follicle-stimulating hormone (FSH), and elevated testosterone. Granulosa cell tumors are rare in children and occur
almost exclusively in the first 6 months of life. Chromosomal
anomalies of the Y chromosome are common, and granulosa
cell tumors have occurred in association with ambiguous genitalia.2 Because of the benign nature of both Leydig and granulosa cell tumors, both can be treated with either orchiectomy
or tumor enucleation in the prepubertal population.30
Approximately 10% of adult Sertoli cell tumors are malignant, whereas malignancy is rare in prepubertal males. Review
of the PTTR showed a median age of presentation of 6 months,
with no cases of malignancy reported in children less than
5 years of age.31 Therefore complete excision of the tumor
is adequate treatment in infants and young children. Presently
there are no histologic criteria to predict tumor behavior in
older children; however, a full metastatic evaluation should
be considered if there is microscopic invasion of the spermatic
cord,31 or worrisome findings, such as a large tumor, necrosis,
vascular invasion, cellular atypia, or increased mitotic activity.2 Large cell calcifying Sertoli cell tumors are histologically
distinct tumors occurring in older children and adolescents.
One third of these patients have an associated genetic syndrome or endocrinopathy, most commonly, Peutz-Jeghers
and Carney syndromes (myxoma of the skin, soft tissue, heart
or breast, lentigines of the face and lips, cutaneous nevi, pituitary adenoma, and schwannoma).2 They have been universally benign in patients less than 25 years of age and may
be treated with testis-sparing procedures. Bilateral or multifocal disease is present in 25% of cases, increasing the need for
this approach.
Germ Cell Tumors
Teratoma Testicular teratoma is the most common germ cell
tumor in prepubertal males according to recent literature.1,5,6,13 These tumors are invariably benign, unlike the
adult population, where 90% to 95%6 of germ cell tumors
are malignant.32 Teratomas are typically pure; derived from
ectoderm, mesoderm, and endoderm; have diploid DNA;
and a normal 46 XY karyotype.29,32 The tissue arrangement
is often organized with a gross solid cystic appearance. Dermoid and epidermoid cysts analogous to the prepubertal teratomas occur in the postpubertal testis. The testicular
dermoid, like the ovarian dermoid, contains hair within a cystic tumor, and microscopic replication of skin without cellular
atypia or widespread mitotic activity. The adjacent testis has
normal spermatogenesis.29 The finding of pilosebaceous units
in an epidermal surface, occasionally with a lipoid reaction
resulting from leakage of oil from the sebaceous glands, is a
prerequisite for diagnosis of a testicular dermoid.
Yolk Sac Tumor The yolk sac tumor comprised approximately 60% of the tumors historically reported in the AAP registry.2 However, recent population-based studies, including
benign testicular tumors, now report the incidence of yolk
sac tumors to be 8%13 to 15%.1 Most of the yolk sac tumors
occur in boys less than 2 years of age, but they are rare in the
first 6 months of life. This is important in differentiating this
tumor from the juvenile granulosa cell tumor (see previous
section).32 The majority of patients (84.5%) identified in
the PTTR presented with localized stage I disease.33 Prepubertal patients are less likely than adults to have metastasis limited
552
PART III
to the retroperitoneum. In a review of the PTTR of the

American Academy of Pediatrics, 15.5% of boys with yolk
sac tumors presented with metastatic disease. The reported
sites included retroperitoneum (27%), retroperitoneal and
hematogenous spread (18.8%), chest (24%), lung and an
additional hematogenous site (12%), scrotum (3%), and 2%
were not documented.33
Elevated AFP levels in excess of age-adjusted levels in the
context of a testicular mass should raise suspicion of a yolk sac
tumor, and the child should be managed with a standard
radical inguinal orchiectomy (see later).
Grossly, the tumor is a soft solid, white to grey, or pale yellow mass with cystic degeneration containing areas of necrosis
and hemorrhage. Microscopically, the yolk sac tumor characteristically contains solid papillae with a connective tissue
core containing a central vessel projecting into cystic spaces;
these structures are referred to as Schiller-Duval bodies.32
The tumor invariably stains positive for AFP and placenta-like
alkaline phosphatase.
Embryonal Carcinoma Embryonal carcinoma is a relatively
common testicular germ cell tumor after puberty; 10% are
pure embryonal tumors, and a substantial number of tumors
will have a mixed embryonal component.29 This tumor demonstrates distinctive sheets, glands, and papillary structures
composed of primitive epithelial cells with crowded pleomorphic nuclei. In poorly differentiated tumors, positive immunostains for CD30 and OCT3 with a c-KITnegative profile
are helpful in confirming an embryonal carcinoma.32 Embryonal carcinoma is treated with orchiectomy. Tumors composed
of more than 80% embryonal cell carcinoma or with elevated
preoperative AFP (>10,000 mg/mL), vessel invasion in the
primary tumor, and tumors of stage T2 or greater are considered high risk and are treated with postoperative chemotherapy and close follow-up.34
Gonadoblastoma Gonadoblastoma has classically been
identified in patients with mixed gonadal dysgenesis
(45,X/46,XY), and is likely related to the presence of the
testis-specific protein-Yencoded gene (TSPY).35 The ectopic
location of the testis adds to this risk. The most commonly
encountered invasive tumor in the intersex gonad is the seminoma. The development of these invasive tumors is always
preceded by the presence of an in situ neoplastic lesion
intratubular germ cell neoplasia unclassified (ITGNU) or
gonadoblastoma.35,36 ITGNU is commonly referred to as
carcinoma in situ (CIS). Because gonadectomy is performed
prophylactically in early childhood in patients with gonadal
dysgenesis, most of the encountered germ cell tumors are
benign or CIS lesions. The overall prevalence of germ cell
tumors in dysgenetic gonads is 15%, which is much lower
than the previously reported prevalence of 33%.35 The tumor
presents with virilization of a phenotypic female harboring an
XY karyotype.37
Gonadoblastoma typically arises from an intraabdominal
testis in a young patient with gonadal dysgenesis. It is usually
small, bilateral in 30% of cases, malignant in 10%, and histologically resembles a seminoma. Available data suggest the
gonad of origin to include dysgenetic testis in 20%, streak
gonad in 26%, and an undifferentiated gonad in 54%.36
Extension beyond the testis has not been reported. Management has traditionally involved bilateral gonadectomy because
of the risk of degeneration into an invasive seminoma. However, the recognized role of testosterone in gender differentiation has led to a more conservative approach to the
contralateral gonad, which may involve a contralateral orchiopexy to allow gender development, followed by annual scrotal
examinations and ultrasonography after age 10 years until puberty. At puberty, testicular biopsy should be carried out to
evaluate for CIS in the remaining testicle.38 If no evidence
of CIS is identified, annual follow-up with testicular ultrasonography until age 20 is recommended. If CIS is identified
at puberty, orchiectomy should be considered.38
Choriocarcinoma Choriocarcinoma is among the rarest of
the gonadal germ cell tumors, representing 0.3% of testicular
tumors.29 These tumors elaborate b-HCG, and may be associated with a number of hormonal manifestations. These include precocious puberty from b-HCGinduced Leydig cell
stimulation, gynecomastia, and hyperthyroidism because of
the similarity of the b-HCG subunits to thyroid-stimulating
hormone.29 Testicular choriocarcinomas frequently have
distant metastasis at the time of presentation rather than a
scrotal mass. Histologically, they are composed of syncytiotrophoblastic cells with mononucleated cells around foci
of hemorrhage. They stain positive for b-HCG and placental
lactogen.32
Rhabdomyosarcoma
Although technically a paratesticular tumor, rhabdomyosarcoma should be included in the differential diagnosis of
scrotal tumors. It is the most frequent tumor of paratesticular origin, accounting for 4% to 25% of scrotal masses.13 The
tumor has a bimodal distribution, peaking between 3 to
4 months of age and 15 to 19 years of age. The infant tumor
has a more indolent behavior than the tumor presenting in
the adolescent age group (90% vs. 63% failure-free survival).39 Despite its aggressive behavior, the prognosis of
paratesticular rhabdomyosarcoma has improved dramatically from 10% to 77% overall survival with the introduction
of vincristine, dactinomycin, and cyclophosphamide (VAC)
chemotherapy.39,40 The most common subtype is embryonal
rhabdomyosarcoma, which accounts for 97% of paratesticular tumors.
The tumor consists of small round blue cells and presents
as a scrotal mass in 80% of patients. Ultrasonography is highly
effective in demonstrating its paratesticular location and distinguishing it from the tumors of testicular origin.13 CT or
MRI of the retroperitoneum should be performed prior to surgery for staging purposes. Thirty to 40 percent of boys will
have micrometastasis to the retroperitoneum. The tumor
should be resected by a radical inguinal orchiectomy. A retroperitoneal lymph node dissection (RPLND) is recommended
for all patients 10 years of age or older for accurate staging,
and in patients less than 10 years with radiologic evidence
of retroperitoneal involvement.39 A metastatic workup should
include a chest CT, liver function tests, bone scan, and bone
marrow biopsy.
SECONDARY TESTICULAR TUMORS

Lymphoma and leukemia are the dominant secondary
tumors of the testis. Acute lymphoblastic leukemia (ALL) is
a common cause of a prepubertal testicular mass. Microscopic
CHAPTER 40
involvement of the testis has been found at autopsy in

66% of patients with ALL.32 Malignant lymphomas account
for 5% of testicular tumors; 10% to 15% are bilateral at
presentation.32
The management of leukemia and lymphoma are the
same. The presence of a palpable mass in a patient with
newly diagnosed leukemia/lymphoma should prompt a scrotal ultrasonography. This usually demonstrates a homogeneous hypoechoic mass. Current literature discourages
testicular biopsy in patients prior to initiating chemotherapy,
because there is no survival advantage.41 In contrast, a patient with persistent or newly enlarged testis undergoing
chemotherapy, particularly in leukemia, implies a relapse
while on therapy. This should prompt a biopsy to direct
subsequent therapy. This typically involves additional chemotherapy to eradicate residual disease in sanctuary sites
and possible systemic residual disease and radiation to the
affected testis.
In 25% of cases, testicular lymphoma is a manifestation of
widespread systemic involvement, another 25% present with
Ann Arbor stage II disease (involvement of lymph nodes below the diaphragm), and the remaining 50% have disease confined to the testis (Ann Arbor stage I).32
Metastasis to the testis in children is rare. The most frequent metastasis has been carcinoma from the prostate, colon,
kidney, stomach, pancreas, and malignant melanoma in
adults, while neuroblastoma and Wilms tumor predominate
in children.32 Most of these tumors have distinctive features
that allow easy identification.
TESTICULAR TUMORS
553
Surgical Management
------------------------------------------------------------------------------------------------------------------------------------------------
TESTIS-SPARING SURGERY
In the last 2 decades, multiple reports have confirmed that many
testicular tumors in the prepubertal population can be managed
more conservatively than in adults, because the distribution of
prepubertal tumors favors a benign histology. This realization
has confirmed the safety and feasibility of testis-sparing surgery,
especially when the lesion is evaluated preoperatively by
ultrasonography and serum AFP and intraoperatively by frozen
section analysis. Metcalfe and colleagues13 have provided a
practical treatment algorithm incorporating the common benign
tumors for nonradical surgery (Fig. 40-1).
In general, before puberty, teratoma, gonadal stromal tumors (Leydig cell and Sertoli cell) and epidermoid cyst can
be managed with a testis-sparing approach (Fig. 40-2). Postpubertal patients with teratoma or stromal tumors should be
treated as adults, with radical orchiectomy because of their
more malignant behavior.
Testis-sparing surgery is carried out through an inguinal incision. The cord is mobilized after opening the external oblique aponeurosis to the level of the internal ring. The
cremasteric fibers are dissected from the cord structures to
allow circumferential control of the cord. The cord should
be occluded at the level of the internal ring with a noncrushing
clamp. The testis is then delivered through the inguinal incision, and the wound is protected. The tunica vaginalis is
opened directly over the mass, and an excisional biopsy of
Prepubertal patient with

painless scrotal mass
Scrotal ultrasound + serum AFP
Paratesticular
rhabdomyosarcoma
Normal AFP
Confirms/suggests
malignancy
Elevated AFP
Discrete mass, benign appearance

Normal AFP
Testis-sparing surgery with

intraoperative frozen section
Radical orchiectomy
No to either
Benign histology on frozen section

and technically feasible?
Yes to both
Testis-sparing procedure
FIGURE 40-1 Proposed treatment algorithm for prepubertal patients presenting with a painless scrotal mass. AFP, a-fetoprotein. (From Metcalfe PD,
Farivar-Mohseni H, Farhat W, et al: Pediatric testicular tumors: Contemporary incidence and efficacy of testicular preserving surgery. J Urol 2003;170:24122415; discussion 2415-2416.)
554
PART III
than 90% of cases. These cells are the precursors for germ cell
tumors and are felt to represent a risk for recurrent neoplasia
if the residual testicular parenchyma is left in situ. The progression through puberty evolves over a period of time and
sequential histologic changes. The testes go through a maturation process starting from simple tubules without lumen
and with interstitial Leydig cells in the neonate. The Leydig
cells then regress and the tubules become more tortuous.
As puberty begins, the Leydig cells become more prominent,
and the basal germ cells begin to divide. There are multiple
layers of spermatocytes and the tubule lumens form, followed
by the appearance of mature sperm. The appearance of mature
sperm or ITGCN would be indicative of completion of pubertal changes.
RADICAL INGUINAL ORCHIECTOMY

AND RETROPERITONEAL LYMPH
NODE DISSECTION
C
FIGURE 40-2 A, Intraoperative photograph of a child with painless
swelling of testicle. B, Wedge resection of epidermoid cyst. C, Suture closure of testicular capsule. (Courtesy Dr. P. Metcalfe, personal file.)
the mass is performed without violating the tumor capsule.

Frozen section evaluation is obtained. Hemostasis is achieved
with electrocautery. If a benign testicular tumor is diagnosed,
the tunica vaginalis is closed with fine absorbable sutures (see
Fig. 40-2), and the testis is replaced in the scrotum. Normal
tissue adjacent to the tumor must be assessed by a pathologist
to exclude pubertal changes. It is commonly assumed that the
postpubertal testis with a tumor will behave in a similar fashion to the adult testis, although specific data are lacking. Adult
testicular tumors are associated with intratubular germ cell
neoplasia (ITGCN) in the surrounding parenchyma in more
A radical inguinal orchiectomy is performed through a standard inguinal incision, with clear demarcation of the external
oblique aponeurosis and external ring and opening of the external ring back to the level of the internal ring. The cremasteric fibers are once again dissected from the cord, and the
cord is fully mobilized from the inguinal canal, followed by
vascular control at the internal ring. The cord is then clamped
and divided at the level of the internal ring, after which the
stump is suture ligated. After ligation, dissection proceeds distally with mobilization of the testis from the scrotum and division of the gubernaculum. If the tumor is too large to deliver
through the scrotal canal, the incision may be carried onto the
superior aspect of the scrotum.42,43 Once the tumor is excised,
the wound is closed in standard fashion.
Current pediatric testicular tumor protocols do not include
a RPLND. Postchemotherapy masses are treated with local resection. Postpubertal patients will often be managed with
adult protocols, although data regarding adolescents is lacking. The indications for and the extent of RPLND are a matter
of some controversy even in adults. Prechemotherapy RPLND
is no longer employed, and postchemotherapy RPLND is eliminated in some centers if residual disease is less than 1 cm in
dimension by imaging.44 In the event that a RPLND is required, a midline abdominal incision is made and a thorough
laparotomy performed to identify retroperitoneal low-volume
metastasis not appreciated on preoperative imaging. There is
also controversy regarding the extent of dissection. Because of
the morbidity of bilateral RPLND (40%), a variety of unilateral
templates have been developed in addition to the concept of
nerve-sparing dissection.45,46 In low-stage disease, lymphatic
spread is typically unilateral, and therefore a full bilateral
RPLND is not used in some centers.47 For the unilateral template, dissection for patients with right-sided disease involves
removal of the lymphatics in the interaortocaval, precaval, and
right paracaval distribution (Fig. 40-3, A).43 For left-sided lesions, this includes the left paraortic and preaortic lymphatics
(Fig. 40-3, B). This dissection strategy is important, because it
preserves the contralateral sympathetics important for emission and ejaculation.48 Preservation of efferent sympathetic
fibers maintains emission and ejaculation rates at 99%.48
The finding of viable tumor outside of the template distribution has led to the recommendation for bilateral dissection
C
FIGURE 40-3 A, Right modified nerve-sparing retroperitoneal lymph node dissection. B, Left modified nerve-sparing retroperitoneal lymph node
dissection. C, A full retroperitoneal dissection involves left and right combined. (From Marshall FF [ed]: Operative Urology. Philadelphia, WB Saunders,
1996, p 368-369.)
556
PART III
in all patients undergoing RPLND in other centers.46 In advanced-stage/high-volume disease bilateral RPLND is always
used, as shown in Figure 40-3, C.45 With either technique,
the nodal packets are split at the 12 oclock position over
the vessels and rolled laterally away. The sympathetic fibers
are carefully identified and preserved as they cross the iliac
bifurcation.
CHEMOTHERAPEUTIC STRATEGIES
AND SURVIVAL IN CHILDREN WITH
MALIGNANT GERM CELL TUMORS
Prior to effective chemotherapy, children with malignant
germ cell tumors (MGCT) had 3-year survival rates of 15%
to 20% with surgery and radiation.7,49 The introduction of
cisplatin-based regimens has dramatically improved outcomes
(Table 40-3).11 In patients with low- and intermediate-risk
(<15 years of age) MGCT, the 6-year overall survival rates
for advanced gonadal tumors (stages III and IV) are now
greater than 94%.7 Standard chemotherapy for children
with MGCT of the testes includes standard-dose cisplatin, etoposide, and bleomycin (PEB) for 4 to 6 courses.2 The current
protocol under investigation examines stages II to IV with
three courses of chemotherapy. Management of patients
is based upon risk groups as proposed by the Childrens
Oncology Group (COG) as follows (see Table 40-3)11:
1. Low risk: Stage I immature teratoma and MGCT of the
testis. Recommend surgery and close follow-up observation to document normalization of tumor markers following resection
TABLE 40-3
Standard Treatment for Children Younger Than 15 Years
with Testicular Germ Cell Tumors by Histology and Stage
Histology
Stage
Treatment
Overall
Survival
(6-Year)
Mature
teratoma
Immature
teratoma
MGCT
Localized
Surgery observation
100%
Localized
Surgery observation
100%
Stage I
Stage II-IV*
Surgery observation
Surgery PEB
100%
94%
From Cushing B, Giller R, Cullen JW, et al: Randomized comparison of

combination chemotherapy with etoposide, bleomycin, and either highdose or standard-dose cisplatin in children and adolescents with high-risk
malignant germ cell tumors: a pediatric intergroup studyPediatric
2004;22:2691-2700; Rogers PC, Olson TA, Cullen JW, et al: Treatment of
children and adolescents with stage II testicular and stages I and II ovarian
malignant germ cell tumors: A Pediatric Intergroup Study-Pediatric
2004;22:3563-3569.
*Patients greater than 15 years old with stage IV testicular tumors should
be discussed in a multidisciplinary oncology group for more intensive
therapy.
MGCT, malignant germ cell tumors; PEB, cisplatin, etoposide, and bleomycin.
2. Intermediate risk: Stages II-IV gonadal tumors (excluding patients >15 years with stage IV testicular
tumors).
expertconsult.com.
as well as embryologically, structurally, and functionally. The

adrenal medulla is derived from ectodermal cells from the
neural crest. These precursors form the chromocell system
and the neuronal system, accounting for the potential development of two distinct medullary neoplasms: pheochromocytoma and neuroblastoma. Preganglionic sympathetic neural
cells innervate the secretory chromaffin cells, which synthesize norepinephrine and epinephrine.
The cortex comprises the outer portion of the adrenal gland
and secretes sex hormones, mineralocorticoids, and glucocorticoids. It is divided into three separate zones that have
distinct synthetic functions. The zona glomerulosa is the outermost cortical zone and produces aldosterone and related
mineralocorticoids. The zona fasciculata lies beneath the zona
glomerulosa and secretes cortisol and the adrenal sex
hormones. The inner zona reticularis maintains cholesterol
stores as a precursor for steroidogenesis and secretes cortisol,
androgens, and estrogens.
Embryology
CHAPTER 41
Adrenal Tumors
Michael G. Caty and Mauricio A. Escobar, Jr.
Anatomy
------------------------------------------------------------------------------------------------------------------------------------------------
The adrenal glands are found anteromedially to the superior

pole of the kidneys, are covered by perirenal fat, and enclosed
by Gerota fascia. In adults, the glands weigh approximately 5 g
each. The right gland abuts the inferior vena cava and liver and
lies on the posterior extension of the diaphragm. The left gland
lies next to the splenic vessels and the tail of the pancreas.
Although the blood supply to the adrenal glands is variable,
it generally comes from three sources: the inferior phrenic artery superiorly, the aorta medially, and the renal arteries inferiorly. The venous drainage does not parallel the arterial
supply; instead, a single large adrenal vein provides the majority of the venous drainage for each gland. The right adrenal
vein empties into the inferior vena cava, and the left adrenal
vein joins the left renal vein.
The adrenal lymphatics arise from one plexus beneath the
capsule and from a second plexus in the medulla. The right
adrenal lymph vessels drain into the periaortic lymph nodes
near the diaphragmatic crus, and the left adrenal lymphatics
empty into lymph nodes near the origin of the left renal artery.
The innervation of the adrenal glands arises from the celiac
plexus and the greater thoracic splanchnic nerves. The
preganglionic sympathetic fibers enter the hilum and end in
ganglia within the medulla.
The cortex and medulla form two distinct regions of the
adrenal gland. These regions are distinct on gross examination
------------------------------------------------------------------------------------------------------------------------------------------------
The primordium of the adrenal cortex becomes visible as early

as the fourth week of gestation and is clearly seen by the sixth
week. On prenatal ultrasonography (US), the adrenal glands
may be visible as early as 20 weeks gestation and are identifiable in the majority of fetuses by 30 weeks gestation.1
During the fourth to sixth weeks of gestation, the mesodermal
cells of the posterior abdominal wall at the adrenogenital ridge
become more columnar and invade the mesenchyma beneath
the epithelial surface, ultimately forming the fetal adrenal cortex. Another proliferation of epithelial cells subsequently
forms a cap over these primitive cortical cells, becoming the
zona glomerulosa of the definitive cortex. The ectodermal
chromaffin cells of the adrenal medulla arise from the neural
crest as early as the fifth week, with primitive cells from the
thoracic ganglia from the 6th to 12th segments invading the
gland and forming the medulla. Differentiation of these primitive medullary cells into chromaffin cells begins at the third
month of gestation, ultimately leading to the cells production
of epinephrine and norepinephrine.
The fetal zone of the adrenal cortex begins to appear
around the sixth week of gestation. This zone continues to enlarge and occupy the majority of the gland. In fact, because of
the large size of the fetal cortical zone, the fetal adrenal gland is
4 times the size of the kidney during the fourth month of gestation. This fetal cortex subsequently decreases in size, disappearing in the first year of life.
During fetal development, ectopic rests of medullary and
cortical tissue may remain and persist after birth. Extraadrenal
medullary rests are usually found along the aorta and its
branches. The organ of Zuckerkandl is an example of a
chromaffin mass at the origin of the inferior mesenteric artery.
Most extraadrenal chromaffin rests involute after birth; the
chromaffin cells in the medulla differentiate.
Extraadrenal cortical rests are common in children and
are found in the kidney or liver or along the migratory path
of the gonads, in hernia sacs, or in the gonads themselves.
Approximately 50% of newborns have adrenocortical rests,
but these rests typically atrophy and disappear within a
few weeks after birth.1
557
558
PART III
Physiology
------------------------------------------------------------------------------------------------------------------------------------------------
ADRENAL MEDULLARY FUNCTION

The adrenal medulla synthesizes and releases catecholamines:
dopamine, epinephrine, and norepinephrine. Catecholamine
synthesis begins with tyrosine, a nonessential amino acid.
Tyrosine hydroxylase converts tyrosine into dihydroxyphenylalanine (DOPA) and is the rate-limiting step in the synthetic
pathway. DOPA decarboxylase converts DOPA into dopamine.
Phenylamine beta-hydroxylase converts dopamine into norepinephrine. Finally, phenylethylamine N-methyltransferase
converts norepinephrine into epinephrine.
The chromaffin cells within the medulla contain cytoplasmic granules that store the catecholamines. Preganglionic
sympathetic nerve endings release acetylcholine, which causes
calcium-dependent exocytosis of these cytoplasmic storage
granules and release of the catecholamines. Regulation of
adrenal medullary catecholamine release is accomplished
through inhibitory feedback mechanisms involving norepinephrine. Norepinephrine inhibits acetylcholine release
from the presynaptic alpha2 receptors and also inhibits tyrosine hydroxylase activity when present in high concentrations.
ADRENAL CORTICAL FUNCTION

The adrenal cortex synthesizes three types of hormones: glucocorticoids, mineralocorticoids, and sex hormones. Regulation
of these is accomplished by the hypothalamic-pituitary-adrenal
axis. The hypothalamus produces corticotropin-releasing hormone (CRH); this is transported to the anterior pituitary gland
where it stimulates the release of adrenocorticotropic hormone
(ACTH). ACTH then stimulates the production of hormones
(glucocorticoids, mineralocorticoids, and sex hormones) from
the adrenal cortex.
The physiologic diurnal variation in CRH release leads to a
cyclic variation in ACTH and the hormones regulated by it.
Serum concentrations peak shortly before or at the time of
awakening and decline throughout the remainder of the
day. Both cortisol and ACTH inhibit CRH release, creating a
negative feedback loop.
Adrenocortical production of glucocorticoids begins with a
cholesterol substrate and is regulated by ACTH. The majority
of serum cortisol is bound by cortisol-binding protein (90%)
and albumin (6%), leaving only a small percentage (4%) free
and physiologically active. As with most steroids, the unbound cortisol fraction is lipophilic and therefore readily
crosses the plasma membrane of target cells. Specific receptors
then bind with cortisol and act in the cell nucleus to regulate
messenger RNA synthesis.
Cortisol affects metabolism primarily by opposing insulin.
It causes hyperglycemia by increasing the proteolysis necessary for gluconeogenesis and inducing hepatic gluconeogenic
enzymes. Cortisol also decreases the use of glucose by peripheral tissues; it inhibits glucose uptake into fat cells and
decreases the amount of insulin bound by insulin-sensitive
tissues.
Cortisol also decreases inflammation and immune function, affecting wound healing. Cortisol lowers both the
lymphocytic and the granulocytic cellular immune response
by decreasing the lymphocyte response to antigenic stimulation and impairing chemotaxis and phagocytosis of leukocytes.
These two immune functions are an important part of early

wound healing; thus wounds have decreased tensile strength
and impaired healing in the setting of excess cortisol.
Aldosterone, a mineralocorticoid, is synthesized in the
zona glomerulosa and metabolized primarily by the liver.
The renin-angiotensin system controls the majority of aldosterone regulation, with ACTH playing only a small role. The
macula densa of the renal juxtaglomerular apparatus releases
renin in response to a drop in renal perfusion or hyponatremia. Renin converts angiotensinogen, which is produced by
the liver, to angiotensin I. Angiotensin-converting enzyme,
found in the lung, converts angiotensin I to angiotensin II.
Angiotensin II stimulates the synthesis of aldosterone by directly acting on the cells of the adrenal zona glomerulosa; it
also acts as a vasoconstrictor. By increasing the renal retention
of sodium, aldosterone increases blood pressure and corrects
hyponatremia, thus reducing the release of renin.
The serum potassium concentration also provides a small
amount of aldosterone regulation. Hyperkalemia leads to increased aldosterone production by directly acting on the zona
glomerulosa cells, as well as increasing renin release from the
juxtaglomerular cells. Aldosterone promotes an increased
renal excretion of potassium, thus lowering aldosterone
production and providing another feedback mechanism.
Adrenal androgens are synthesized in the zona reticularis
and are regulated primarily by ACTH. These hormones are
released in a cyclic manner, correlating with the release of
cortisol and ACTH. The adrenal androgens are only weakly
active but are converted by peripheral tissues into more active
forms such as testosterone and dihydrotestosterone. Metabolism of these hormones occurs in the liver.
Lesions of the Adrenal Medulla

------------------------------------------------------------------------------------------------------------------------------------------------
PHEOCHROMOCYTOMA
In 1886, Frankel of Freiburg, Germany, published the first description of bilateral pheochromocytomas found during the
postmortem examination of an 18-year-old woman who had
presented with symptoms of anxiety, palpitations, and headache.2 In 1912, Pick named the tumor for its predominant cell
type, the pheochromocyte, but it was not until 1922 that
Labbe and colleagues first described a clear relationship
between pheochromocytoma and paroxysmal hypertension.
In 1927, Mayo performed the first successful removal of a
pheochromocytoma in a patient with paroxysmal hypertension who underwent surgical exploration without a preoperative diagnosis. In 1929, Pincoffs made the first correct
preoperative diagnosis, and the successful operation was
performed by Shipley.3 Since that time, the behavior of
pheochromocytomas has become better understood, particularly with respect to children.
Pheochromocytoma is an uncommon tumor of childhood,
and there are several characteristics that distinguish its presentation between adults and children. The incidence of
pheochromocytoma in childhood is 10% of the adult incidence, occurring in approximately 1 in 500,000 children compared with 1 in 50,000 adults.4 Approximately 10% of
childhood pheochromocytomas are familial, which is about
4 times the frequency in adults. Whereas only 7% of pheochromocytomas are bilateral in adults, the reported incidence of
CHAPTER 41
TABLE 41-1
Comparison of Pheochromocytoma in Children and Adults
Incidence
Familial pattern (%)
Bilateral (%)
Extraadrenal site (%)
Malignant (%)
Pediatric
Adult
1:500,000
10
24-70
30
3
1:50,000
2-3
10
10
10
bilateral pheochromocytomas in children ranges from 24% to

as high as 70%. Extraadrenal pheochromocytomas are approximately twice as prevalent in children as in adults (Table 41-1).5,6
Pheochromocytomas originate from medullary chromaffin
cells, which produce the catecholamines that cause the associated symptoms. These cells migrate along the aorta, usually
remaining near the branches of the aorta.
SYMPTOMS
In children with pheochromocytoma, the average age at presentation is 11 years, although the tumor can occur at any age.
Over half the children present with headaches, fever, palpitations, thirst, polyuria, sweating, nausea, and weight loss, but
the most common presentation is sustained hypertension.4,6,7
In children, most causes of hypertension are secondary, with
renal abnormalities being most common (78%), followed by
renal artery disease (12%), and coarctation of the aorta
(2%).8 Pheochromocytoma accounts for 0.5% of children
with hypertension and must be considered once other causes
are eliminated. In children with pheochromocytoma, hypertension is sustained in up to 70% to 90% of cases, with only
a small minority presenting with paroxysmal hypertension.
In contrast, up to 50% of adults with pheochromocytoma have
paroxysmal hypertension.6
ADRENAL TUMORS
559
After establishing the chemical diagnosis of pheochromocytoma, the tumor must be localized. Although large masses
such as a neuroblastoma can be seen on plain abdominal films,
most adrenal masses cannot be visualized without the use
of other imaging methods. Almost all pheochromocytomas
occur in the abdomen or pelvis, and although the adrenal
gland is the most common site, up to 43% of children may
have multifocal disease.6 The initial study in infants and
children is often US, which can be useful in distinguishing
between solid and cystic masses while determining their vascularity and avoiding ionizing radiation, but it may not
visualize small adrenal lesions. Additionally, it may be difficult
to identify the adrenal gland as the organ of origin for large
masses, because of compression from adjacent organs such
as the kidney. Computed tomography (CT) and magnetic
resonance imaging (MRI) offer the advantage of much better
resolution and sensitivity (Fig. 41-1). Although CT is an accurate method of diagnosing adrenal lesions, it is less accurate
in younger children because of the absence of retroperitoneal
fat. Other disadvantages of CT are the need for intravenous
contrast material and exposure to ionizing radiation. Simultaneous scanning of the chest to rule out pulmonary metastases in patients suspected of having adrenal carcinoma is a
benefit of CT. Currently, both CT and MRI offer multiplanar
imaging. Coronal imaging is a useful modality to distinguish
DIAGNOSIS
The diagnosis of pheochromocytoma relies on the demonstration of elevated levels of blood and urinary catecholamines
and their metabolites. A 24-hour urine measurement of
catecholamines, metanephrine, and vanillylmandelic acid
is the best diagnostic test.9,10 Urinary metanephrine levels
are increased in about 95% of patients, and urinary vanillylmandelic acid and catecholamine levels are increased in
approximately 90% of patients.10 There is also a linear relationship between the amount of vanillylmandelic acid and
the size of the pheochromocytoma.11 The normal 24-hour
urinary secretion is less than 100 mg for free catecholamines,
less than 7 mg for vanillylmandelic acid, and less than 1.3 mg
for metanephrine. Plasma catecholamines can also be measured by radioenzyme assay. However, patients must remain
supine and calm during the blood draws, which can be difficult in children. Patients with normal plasma catecholamine
levels during a hypertensive episode probably do not have
pheochromocytoma, but levels greater than 2000 pg/mL
are diagnostic of pheochromocytoma. Plasma catecholamine
levels between 500 and 1000 pg/mL are suspicious for a
pheochromocytoma, and further testing is indicated.6 It must
be remembered, however, that neuroblastoma can in some
cases secrete significant levels of catecholamines.
B
FIGURE 41-1 A, Computed tomography of the abdomen in a 10-yearold girl with a left adrenal mass associated with hypertension. B, Magnetic
resonance image demonstrates a left adrenal pheochromocytoma. No
other masses were noted. No contrast agent was required.
560
PART III
adrenal masses from the adjacent kidney and vice versa.1214

Pheochromocytomas demonstrate low or intermediate signal intensity on T1-weighted images and enhance with
gadolinium-diethylenetriaminepentaacetic acid (DTPA).15
One significant disadvantage to MRI is that children often
require sedation or general anesthesia, given the length of time
the study requires, which may be a risk if children have not
been treated with blocking agents.
Another useful imaging technique is 131I-labeled metaiodobenzylguanidine (MIBG) scanning; this radioisotope accumulates where norepinephrine is taken up and allows detection
of the tumor. MIBG, which is structurally similar to norepinephrine, is taken up by the norepinephrine transporter
system into intracytoplasmic vesicles. Radionuclide imaging
is achieved by labeling MIBG with one of two iodine isotopes
at the meta position of the benzoic ring. The iodine isotope
131
I has a half-life of 8.2 days and emits high-energy radiation.
The iodine isotope 123I has a shorter half-life and emits
lower-energy radiation.16 Patients undergoing MIBG scanning
should be given a saturated solution of potassium iodide to
block thyroid uptake of the free iodine isotope. Scintigraphy
is performed at 24 and 48 hours. This technique can be particularly useful in localizing extraadrenal tumors or sites of
metastasis. It also confirms the adrenal location of a pheochromocytoma in patients with positive urine or serum catecholamine tests. The head and neck may be a more common site of
these tumors in children compared with adults, followed by
the retroperitoneum.17
Positron emission tomography (PET) may be a useful
imaging study for pheochromocytoma in the near future.
PET scanning uses short-lived positron-emitting agents to
identify specific areas of uptake in the body. Because of the increased metabolism of tumors, labeled glucose can be used to
identify malignant tissue. The most common form of labeled
glucose in use for PET scanning is (18F)-fluorodeoxyglucose
(FDG). However, resolution of pheochromocytoma and
distinction between benign and malignant pheochromocytoma are not optimal with FDG PET. A more useful agent
may be 6-(18F)-fluorodopamine (DA). The similarity between
norepinephrine and DA allows selective uptake by sympathoadrenal tissue.18 One study found that FDG PET demonstrated metastases better than MIBG scanning did in
adults (one patient was 16 years old in this cohort of 29
patients).19 PET scanning results in lower radiation exposure
than standard scintigraphy. When specific agents, such as DA,
become generally available, PET scanning may prove to be the
imaging method of choice.
TREATMENT
The treatment of pheochromocytoma is surgical excision,
although medical management of the hypertension is an
essential part of the preoperative preparation. The high levels
of catecholamines increase the risk of sudden and severe
intraoperative hypertension, as well as profound hypotension
once the tumor is removed and catecholamine release has
ceased. In fact, these complications accounted for the
high mortality rate associated with surgical resection in the
past.6 Improvements in preoperative and intraoperative
management have reduced the operative mortality of 24% to
45% in the past to less than 10% today.20 Preoperative use of
alpha-adrenergic blockers, such as oral phenoxybenzamine
and phentolamine, reduces the effects of epinephrine and norepinephrine by blocking the alpha-adrenergic receptors. These
agents should be started at least 3 to 7 days before the procedure
and the dose increased until the pressures are well controlled to
minimize the intraoperative risks. Replacement of intravascular volume is often required as alpha blockade is achieved,
because patients with pheochromocytomas tend to be hypovolemic at baseline, with an average 15% reduction in plasma
volume. This volume re-expansion also helps minimize intraoperative blood pressure fluctuations and cardiac arrhythmias.
Beta-adrenergic blockade with agents such as propranolol
and labetalol may be used once an alpha-adrenergic blockade
is achieved, particularly if a resting tachycardia develops despite adequate volume replacement. If these agents are used,
it is crucial that alpha blockade be established first. Administration of a beta blocker before an alpha blockade can worsen
hypertension secondary to unopposed vasoconstriction.
Methyl-para-tyrosine (metyrosine) competitively inhibits
tyrosine hydroxylase, the rate-limiting step in catecholamine
biosynthesis. Treatment with metyrosine reduces tumor stores
of catecholamines, decreases the need for intraoperative antihypertensive drugs, lowers intraoperative fluid requirements, and
attenuates blood loss. It has not been tested in children less than
12 years of age. Metyrosine may not be necessary for patients
with minimal or no symptoms from a minimally functioning
pheochromocytoma.21
Despite good preoperative normalization of blood pressure, the anesthesiologist must be prepared for sudden fluctuations. The times of significant intraoperative risk are during
anesthetic induction and intubation, during surgical manipulation of the tumor, and immediately following ligation of the
tumors venous drainage.22 An arterial catheter and a central
venous line are crucial for monitoring intraoperative blood
pressure and fluid status. The anesthesiologist must also be
prepared to use fast-acting agents to raise or lower blood
pressure as needed. Sodium nitroprusside and nitroglycerin
are useful agents, as are vasopressors and intravenous fluids.
Cardiac arrhythmias can be managed with the use of propranolol, esmolol, and lidocaine. Adrenalectomy is described later.
An adrenal pheochromocytoma is typically encapsulated,
and although there may be small amounts of normal adrenal
tissue, the entire adrenal gland should be removed. It is rarely
necessary to perform a nephrectomy, because the tumor is
rarely adherent to the kidney.
As previously mentioned, once the adrenal vein is ligated
and the tumor is removed, the patient may become hypotensive because of the removal of the catecholamine excess. In
fact, it may be several days before the blood pressure normalizes. If hypertension returns postoperatively, one should
suspect a second pheochromocytoma. All patients should undergo follow-up to confirm normalization of catecholamine
levels. Long-term follow-up is indicated because of the
possibility of a metachronous occurrence of a multifocal pheochromocytoma or occult metastasis.6,22
ASSOCIATED DISORDERS
Familial pheochromocytomas may occur in the setting of
several syndromes. The most common syndromes are multiple endocrine neoplasia type 2 (MEN-2) and von HippelLindau disease. There is a smaller incidence of familial
CHAPTER 41
pheochromocytomas in patients with neurofibromatosis

type 1 and in patients without any other abnormalities.
Traditionally, a 10% incidence of familial cases of pheochromocytoma was expected. However, a germline mutation
has been identified in up to 59% of apparently sporadic
pheochromocytomas presenting at 18 years of age or younger
and in 70% of those presenting before 10 years of age in one
series.21 The inherited predisposition may be attributable to a
germline mutation in the von Hippel-Lindau gene, the genes
encoding the subunits B and D of succinate dehydrogenase,
the RET proto-oncogene predisposing to multiple endocrine
neoplasia type 2, or the neurofibromatosis type 1 gene. Of
these, the von Hippel-Lindau gene is the most commonly mutated gene in children presenting with a pheochromocytoma.
A mutation of the von Hippel-Lindau gene on chromosome
3 leads to von Hippel-Lindau disease. This condition is characterized by retinal angiomas, hemangioblastomas of the
central nervous system, renal cysts, renal cell carcinoma, pancreatic cysts, and pheochromocytomas. These pheochromocytomas are often multifocal and are frequently extraadrenal.
Multiple endocrine neoplasia type 2 is an autosomal
dominant disorder caused by a mutation of the RET protooncogene on chromosome 10. These patients are at risk for
medullary thyroid carcinoma, and up to 50% will develop
adrenal pheochromocytoma. These tumors are almost always
bilateral and are almost never malignant. Patients with
MEN-2A are also at risk for hyperparathyroidism, and patients
with MEN-2B may have a marfanoid habitus or mucosal
ganglioneuromas.
Malignancy has been reported to occur in up to 10% of
children with pheochromocytoma.7 The diagnosis of malignancy is generally based on the tumors clinical behavior, because the histologic examination is not an accurate predictor.
A malignant pheochromocytoma may have local infiltration or
distant metastasis, which most commonly occurs in bone,
liver, lymph nodes, lung, and the central nervous system. Synchronous or metachronous pheochromocytomas may present
anywhere along the sympathetic chain. Although surgical
resection remains the treatment of choice, long-term palliation
may be obtained through a multimodal approach, including
local excision, radiation, and chemotherapy.23
Lesions of the Adrenal Cortex

------------------------------------------------------------------------------------------------------------------------------------------------
Adrenocortical neoplasms are rare in the pediatric population,

accounting for less than 0.2% of all pediatric tumors and
6% of all adrenal tumors in children.24 The incidence of these
neoplasms has been reported to be approximately 25 cases per
year in the United States, of which about 75% are adrenocortical carcinomas.25,26 Adrenocortical tumors occur more
frequently in girls, with a male to female ratio of approximately 1:2 to 1:3.27 Like pheochromocytomas, adrenocortical
neoplasms behave differently in children than in adults.
Approximately 85% to 95% of these tumors are hormonally
active in children, compared with less than 50% in adults.28,29
Further, whereas there are clear pathologic criteria for malignancy in the adult population, these guidelines are not reliable
in the pediatric population. Because the clinical behavior of
these tumors does not always correlate with the pathologic
appearance, the diagnosis of malignancy should be based
ADRENAL TUMORS
561
on clinical behavior. Age less than 3.5 years at the time of

diagnosis and symptom duration of less than 6 months
before diagnosis are favorable prognostic indicators in adrenocortical carcinoma. Early detection is essential in these
children, because a delay in diagnosis adversely affects clinical
outcome.1
Adrenocortical tumors are associated with several congenital anomalies, including hemihypertrophy; other tumors
associated with hemihypertrophy include nephroblastoma
and hepatoblastoma. Patients with Beckwith-Wiedemann syndrome (exomphalos, macroglossia, and gigantism) also have a
higher than expected incidence of adrenocortical carcinoma.30
Most adrenocortical tumors, however, occur sporadically.1
CUSHING SYNDROME
In 1932, Cushing first described the syndrome that bears his
name in a patient with a pituitary adenoma. Since that time,
the understanding of the pathophysiology and cause has
expanded considerably. Endogenous Cushing syndrome is a
rare condition in the pediatric population. In general, the
incidence of spontaneous Cushing syndrome is approximately
5 per million persons; it occurs primarily in young adult
women, with a female to male ratio of 9:1. Ten percent of cases
occur in children and adolescents.31
The typical manifestation of Cushing syndrome in children
is generalized obesity and long bone growth retardation.31
Other symptoms include hypertension, weakness, thin skin
with striae and easy bruising, acne, menstrual irregularity,
osteoporosis, and glucose intolerance. Unlike in adults with
Cushing syndrome, muscle weakness, sleep disturbances,
and mental changes, such as emotional lability, irritability,
or depression, are rare in children.31 Cushing syndrome can
be divided into ACTH-dependent and ACTH-independent
types. In the former condition, the inappropriately high
ACTH levels stimulate the adrenal cortex to produce excessive cortisol. In the ACTH-independent type, abnormal
adrenal tissue produces excessive cortisol irrespective of
ACTH levels.
Cushing disease refers to Cushing syndrome caused by
pituitary tumors that lead to excessive ACTH production.
Typically, these tumors are microadenomas and are less than
1 cm in diameter; however, large, invasive pituitary adenomas
may develop. These tumors lead to bilateral adrenocortical
hyperplasia, with a corresponding glucocorticoid excess. As
the age of the patient increases, there is a greater likelihood
of a pituitary cause of the syndrome. In patients younger
than 6 years, the most likely cause of endogenous Cushing
syndrome is an adrenal tumor. Although adrenocortical carcinomas represent only 0.2% of all childhood malignancies
and 6% of adrenal cancers, approximately 60% to 80% of pediatric Cushing syndrome cases are caused by adrenocortical
carcinomas.28
The clinical diagnosis of hypercortisolism must be
confirmed biochemically to diagnose Cushing syndrome. In addition, the specific source of the syndrome must be localized
(Fig. 41-2). Cortisol production is normally suppressed at
night, but in Cushing syndrome, this suppression does not
occur. The normal circadian rhythm of cortisol secretion is lost
in Cushing syndrome. Random serum cortisol levels are of limited value. The three most common tests used to diagnose
Cushing syndrome are the 24-hour urinary free cortisol test,
562
PART III
Diagnostic Studies to Localize Hypercortisolism
<100 g/day
24-hour
urinary free
cortisol
Repeat test
>100 g/day
>5 pg/mL
<5 pg/mL
ACTH
ACTH-independent
causes
ACTH-dependent
causes
Ratio < 2
(No suppression)
Ectopic ACTH-secretion
Inferior petrosal
sinus
sampling
or
High-dose
dexamethasone
suppression
Ratio >2
(Suppression)
Cushing disease
CT of
adrenal
glands
No
mass
or
nodules
Nodular
hyperplasia
Dominant
mass
Adenoma
Adenocarcinoma
FIGURE 41-2 Algorithm to localize the cause of hypercortisolism in children with suspected Cushing syndrome. ACTH, adrenocorticotropic hormone;
CT, computed tomography.
measurement of midnight plasma cortisol or late-night salivary

cortisol, and the low-dose dexamethasone suppression test.
The dexamethasone-corticotropin-releasing hormone test
may be needed to distinguish Cushing syndrome from other
causes of excess cortisol. The 24-hour urinary free cortisol level
has a sensitivity of approximately 98%.32 In children, this value
must be corrected for size. A normal value is less than 70 mg/m2
per day; this is elevated with Cushing syndrome. Another useful
test is the 24-hour urinary 17-hydroxysteroid excretion; this is
an indirect measure of cortisol secretion and is elevated with
hypercortisolism. Once it is corrected for creatinine excretion,
the normal value is between 2 and 7 mg per gram of creatinine
per day. The overnight dexamethasone suppression test is administered as follows: After the administration of 1 mg (or
0.3 mg/m2 in children) of dexamethasone, a morning cortisol
level greater than 5 mg/dL indicates unsuppressed cortisol secretion consistent with Cushing syndrome.
Once the diagnosis of Cushing syndrome has been established, the next step is to determine the underlying cause of
the hypercortisolism. As shown in Figure 41-2, measurement
of the ACTH level can distinguish between ACTH-dependent
and ACTH-independent causes. If the ACTH level is greater

than 5 pg/mL, the source is ACTH dependent; if the level is
less than 5 pg/mL, it is ACTH independent.
ACTH-dependent causes of hypercortisolism include
both pituitary and ectopic ACTH-secreting neoplasms. Although ectopic production of ACTH is rare in children,
Wilms tumors and tumors of the thymus, pancreas, or
neural tissue can produce ACTH. Most patients with
ACTH-secreting tumors have Cushing disease (Cushing syndrome caused by a pituitary tumor). Although a high-dose
dexamethasone suppression test or an inferior petrosal sinus
sampling can distinguish a pituitary source from an ectopic
source, MRI can also show a pituitary tumor. An ectopic
tumor producing CRH is another ACTH-dependent source
of Cushing syndrome, but this condition has not been
reported in a child.32,33
In both adults and children, the treatment of choice for
Cushing disease is a transsphenoidal resection of the pituitary
adenoma. In patients with no postoperative improvement or
with recurrence, some response may be obtained with pituitary irradiation using cobalt 60.
CHAPTER 41
If an ectopic ACTH-secreting tumor is indicated by the

workup, the patient must undergo screening for medullary
carcinoma of the thyroid (serum calcitonin levels) and screening for pheochromocytoma (24-hour urine measurement of
catecholamines, metanephrine, and vanillylmandelic acid).
Other ectopic locations, such as a bronchial, thymic, or intestinal carcinoid tumor, may be seen on CT of the chest and
abdomen. Ectopic ACTH-producing tumors should be resected
if possible. If resection is not possible, bilateral adrenalectomy
can offer an effective treatment of Cushing syndrome.
ACTH-independent causes of Cushing syndrome include
adrenal neoplasms and nodular adrenal hyperplasia. ACTHindependent Cushing syndrome is relatively more frequent
in children than in adults.32 In children, an adrenocortical
tumor most frequently occurs in the setting of a virilizing
syndrome, and the majority of children present with virilizing
symptoms. Approximately 33% of these patients have
Cushing syndrome; less than 10% present with isolated
Cushing syndrome without any virilizing signs.29,25
Nodular adrenal hyperplasia is a rare condition that occurs
in children and young adults. This disease usually presents in
the first 2 decades of life, predominantly in girls. Although this
entity can occur sporadically, many cases are familial and appear in an autosomal dominant fashion.32 The adrenal glands
contain multiple nodules approximately 3 to 5 mm in size.
Histologic examination reveals lymphocytic infiltration of
the cortex, suggesting an autoimmune cause of the disorder.
The treatment of this cause of Cushing syndrome is bilateral
adrenalectomy.32 This procedure is associated with significant
morbidity and requires permanent postoperative mineralocorticoid and glucocorticoid replacement.
SEX HORMONEPRODUCING TUMORS

An adrenocortical lesion may lead to either a virilizing or a
feminizing tumor. As previously mentioned, most adrenocortical tumors in children are hormonally active. Virilization
with or without hypercortisolism is the most common presentation.26,29,34,35 These virilizing tumors may be more difficult
to recognize in boys than in girls. Boys may present with
precocious puberty, including penile enlargement, acne, and
premature development of pubic, axillary, and facial hair. Girls
may develop clitoral hypertrophy, hirsutism, and acne. The
treatment of choice is adrenalectomy.
Although feminizing adrenocortical tumors are rare in children, they are usually malignant. In the normal adrenal gland,
very small amounts of estrogens may be secreted. With adrenocortical tumors, however, overproduction of estrogens, particularly estradiol, may occur. In girls, these tumors present
with precocious isosexual development, including early breast
enlargement, accelerated growth, and advanced bone age. In
boys, these tumors cause bilateral gynecomastia, accelerated
growth rate, and delayed pubertal development; there is also
an absence of spermatogenesis.
TREATMENT OF ADRENOCORTICAL TUMORS

Surgical resection is the mainstay of treatment for adrenocortical tumors. The treatment of choice for a benign adrenal
adenoma is adrenalectomy. Adrenocortical carcinomas, however, require a wide excision with adequate abdominal exploration for metastatic disease. In either case, postoperative
ADRENAL TUMORS
563
steroid replacement is typically required until the contralateral

gland can recover from its suppression.
Computed tomography or magnetic resonance imaging
can help distinguish between adrenal hyperplasia and an
adrenal tumor. A 131I-iodomethyl-1-19-norcholesterol (NP59) scintiscan may aid in the evaluation of an adrenal lesion.
This cholesterol analogue is taken up as cholesterol into the
steroid pathways of the adrenal cortex. Adrenal adenomas usually have an increased uptake of NP-59, whereas adrenocortical
carcinomas typically do not take up the isotope. Bilateral uptake
of NP-59 indicates bilateral adrenal hyperplasia, which can be
the result of ACTH oversecretion.
The most common sites of metastatic adrenocortical carcinomas are the lung, liver, lymph nodes, contralateral adrenal
gland, bones, kidneys, and brain. If complete resection is not
possible, tumor debulking may be of some benefit to control
symptoms. Medical therapy with mitotane may also play a role
in treating patients with unresectable disease. Mitotane acts as
an adrenolytic agent by altering mitochondrial function,
blocking adrenal steroid hydroxylation, and altering the extraadrenal metabolism of cortisol and androgens. The success of
chemotherapy has not been clearly shown, however, and
complete surgical resection is the primary determinant of
survival.36
Hyperaldosteronism
------------------------------------------------------------------------------------------------------------------------------------------------
Overproduction of aldosterone, or hyperaldosteronism, may

be due to either adrenal dysfunction or overproduction of
renin. Primary hyperaldosteronism refers to adrenal dysfunction, such as an aldosterone-secreting tumor or bilateral
adrenal hyperplasia. Secondary hyperaldosteronism refers to
an overproduction of renin, which can be caused by cirrhosis,
congestive heart failure, a renin-producing juxtaglomerular
cell tumor, or renovascular abnormalities, such as renal artery
stenosis.
The symptoms of hyperaldosteronism include headaches,
fatigue, weakness, lethargy, poor weight gain, polyuria, polydipsia, and nocturia. Hypertension develops as a result of increased sodium and water reabsorption. Weakness occurs
because of hypokalemia, which is the most common laboratory finding, although metabolic alkalosis may be observed
from the loss of hydrogen ions in the urine. The biochemical
diagnosis of hyperaldosteronism is demonstrated by excessive
aldosterone secretion in the setting of suppressed renin secretion. Once the diagnosis of primary hyperaldosteronism has
been established, patients with aldosterone-secreting adrenal
tumors must be distinguished from those with the more
common condition of bilateral adrenocortical hyperplasia.
In patients with bilateral adrenocortical hyperplasia, dexamethasone administration normalizes the abnormally high
aldosterone level and low renin level.10
In the pediatric population, the incidence of aldosteronoma, or an adrenal adenoma causing primary hyperaldosteronism, is extremely low, with only a handful of reported cases
in the literature. As previously mentioned, the more common
cause of primary hyperaldosteronism is bilateral cortical hyperplasia.37 An aldosteronoma is best treated by unilateral adrenalectomy. Patients with bilateral adrenocortical hyperplasia
do not respond well to surgical treatment and are best
564
PART III
managed with medical therapy using spironolactone and

amiloride.10 Adrenal insufficiency resulting from bilateral adrenalectomy is more difficult to manage than hyperaldosteronism.
Addison Disease
------------------------------------------------------------------------------------------------------------------------------------------------
Insufficient production of steroid hormones (either glucocorticoids or mineralocorticoids) can lead to Addison disease.
Children with Addison disease present with a variety of symptoms, including weakness, anorexia, weight loss, fatigue, nausea, vomiting, and diarrhea. If the child has an elevated ACTH
level, hyperpigmentation will develop, because melanocytes
are stimulated by ACTH. Seizures may also occur in the setting
of the hypoglycemia, which occurs with adrenal crisis.
There are many causes of adrenal insufficiency in children.
Congenital adrenal hypoplasia can result from either an autosomal recessive disorder or an X-linked disorder that occurs in
boys. Errors in steroid metabolism can also lead to adrenal
insufficiency. The most common group of inborn errors involves defects in glucocorticoid synthesis and is collectively
known as congenital adrenal hyperplasia. Acquired lesions
involving the hypothalamus or pituitary can also lead to adrenal
insufficiency through a reduction in CRH or ACTH secretion.
Destruction of the adrenal glands can also lead to adrenal
insufficiency. Conditions causing adrenal demise include
hemorrhage, infection, adrenoleukodystrophy, and autoimmune diseases. In older patients, overwhelming infection
can lead to adrenal hemorrhage. Tuberculosis used to be a
common cause of infectious destruction of the adrenal; however, the incidence of this condition has fallen in modern
times. One of the more common causes of acute adrenal insufficiency is cessation of chronic exogenous glucocorticoid
administration.
In newborns, adrenal hemorrhage is not an uncommon
event. In fact, the adrenal gland is the second most common
source of hemoperitoneum in the newborn period.38 The
pathogenesis of adrenal hemorrhage in newborns is not
fully understood. Associated factors include traumatic delivery,
asphyxia, maternal hypotension, overwhelming infection, or
hemorrhagic disorders.35,39 The incidence of adrenal hemorrhage is almost 2 cases per 1000 live births,1 but as the
sensitivity of imaging technology improves, this number may
increase. Adrenal hemorrhage occurs 3 to 4 times more frequently in the right adrenal gland than the left and is bilateral
in 8% to 10% of patients.39 This bias toward the right
side may be due to the direct drainage of the right adrenal gland
into the inferior vena cava, making the right gland more
susceptible to changes in venous pressure. The left gland
remains somewhat protected by its drainage into the left renal
vein. The fetal cortex contributes to fetal and neonatal adrenal
hemorrhage because of both its size and its later involution.
The large size of the fetal cortex makes the adrenal glands
relatively large, increasing their vulnerability to trauma. The
normal adrenal gland is easily visualized by US during
the first week of life. The adrenal soon involutes, and the distinction between the cortex and the medulla is lost. The physiologic
involution of the fetal cortex may occur quite rapidly, tearing the
unsupported central adrenal gland vessels.38
On prenatal US, adrenal hemorrhage appears as an echogenic mass. This mass becomes increasingly hypoechoic
and usually involutes on subsequent sonograms.40 The lesion

may completely resolve, leaving only residual calcifications.
Adrenal hemorrhage may be confused with neuroblastoma.
Patients with normal urinary catecholamine levels and the appropriate risk factors for adrenal hemorrhage can be observed
and undergo repeat US. Differentiation of adrenal adenoma
and carcinoma by US is difficult; in addition, both resemble
an adrenal pheochromocytoma. An ultrasonographic characteristic that suggests malignancy is central necrosis from rapid
growth. Biochemical testing and the use of CT, MRI, and nuclear medicine studies narrow the diagnostic possibilities.
The treatment of Addison disease is replacement of the deficient steroid hormone. This may be accomplished with a
mineralocorticoid, such as fludrocortisone, or a glucocorticoid, such as hydrocortisone or prednisone. During periods
of acute stress, such as infection or operation, increased doses
of glucocorticoids are needed.
Incidental Adrenal Mass

------------------------------------------------------------------------------------------------------------------------------------------------
The incidental discovery of adrenal lesions on imaging studies

performed for other reasons has been increasing in both children and adults, perhaps because of the increased frequency
of imaging studies being performed and the increased sensitivity of those imaging modalities. In adults, the current recommendation is to remove all hormonally active tumors
regardless of size. In the case of nonfunctional adrenal masses,
it is considered safe to observe a mass less than 4 cm in
size.4143 In the pediatric population, however, there are no
clear guidelines about incidental, nonfunctional adrenal
masses. Because of the higher incidence of both functional tumors and malignant tumors in the pediatric adrenal gland,
many surgeons recommend adrenalectomy in this setting.43
Adrenalectomy
------------------------------------------------------------------------------------------------------------------------------------------------
The objective of adrenal surgery is to attain complete tumor

resection, resulting in normalization of endocrine function
and cure of malignancy. Perioperative planning includes correction of potential electrolyte abnormalities, establishing
alpha and beta blockade in the case of pheochromocytoma,
and performing localizing studies to guide the surgical approach. The surgical approach is based on the probable histology of the adrenal mass, the presence of bilaterality, and the
surgeons preference. The introduction of laparoscopic adrenal resection has provided an attractive alternative for the
resection of many adrenal masses in children.
Traditional approaches to adrenal resection have included
anterior, posterior, and thoracoabdominal approaches. The
anterior approach uses a transabdominal incision, usually
subcostal, which permits resection of either the left or the right
adrenal gland. It also allows bilateral resection through a
single incision, as well as visualization of the periaortic
sympathetic ganglia, the small bowel mesentery, and the
pelvis. More than 95% of pediatric pheochromocytomas are
located in the abdomen, and this approach reveals the majority of tumors. The surgeon must make a conscious effort to
minimize direct manipulation of the tumor during dissection.
CHAPTER 41
ADRENAL TUMORS
565
the right adrenal vein and the greater risk of tearing this vessel.
Multiple veins may be present and should be identified to prevent accidental avulsion. During the anterior approach to the
left adrenal gland, the initial maneuver is to mobilize the
splenic flexure of the colon. The pancreas and spleen are
retracted superiorly, and the Gerota fascia is opened, exposing
the left adrenal gland. Alternatively, the surgeon can divide the
gastrocolic ligament, mobilizing the stomach superiorly and
the transverse colon inferiorly. The posterior peritoneum
along the inferior pancreatic border can then be incised,
allowing mobilization of the pancreatic tail and exposure of
the adrenal vein. The left adrenal vein enters the renal vein superiorly and can be ligated in this plane. Several arteries enter
the medial surface of the adrenal gland from the lateral side of
the aorta; these arteries need to be divided before adrenal
Early control and ligation of the adrenal vein limit the release
of catecholamines as the tumor is removed.
During the anterior approach to right adrenalectomy,
the duodenum is mobilized by the Kocher maneuver
(Fig. 41-3) by reflecting the transverse colon inferiorly and
mobilizing the duodenum medially. This exposes the upper
portion of the right kidney as well as the right adrenal gland.
The Gerota fascia is opened, and the right lobe of the liver is
retracted in a cephalad direction. The most important element
of the procedure is the dissection between the medial border
of the adrenal mass and the lateral wall of the inferior vena
cava. This plane is developed in a cephalad direction until
the relatively short right adrenal vein is identified entering
the vena cava. There is a greater risk of hemorrhage on the
right side than on the left, because of the shorter length of
Live
r
Adrenal gland
Duodenum
Sto m
ac
o
col
Tran s
vers e
Rt. kidney
Vena cava
Adrenal tumor
Spleen
Stomach
and
pancreas
L. kidney
B
FIGURE 41-3 Transabdominal approach to tumors of the adrenal glands. A, The right adrenal gland is exposed by reflecting the transverse mesocolon
inferiorly, mobilizing the duodenum medially with a Kocher maneuver, and incising the posterior fascia to expose the diaphragm, adrenal gland, and
superior pole of the right kidney. B, The left adrenal gland is exposed by dividing the gastrocolic ligament and elevating the stomach. The colon is retracted
inferiorly, and the pancreas is elevated, exposing the adrenal gland and left adrenal vein that enters the renal vein.
566
PART III
removal. The posterior approach to the adrenal gland is accomplished most commonly through the bed of the 11th
rib. This strategy avoids intraperitoneal dissection, eliminates
postoperative adhesions, and decreases postoperative ileus.
The posterior approach is not useful for bilateral adrenal
lesions, malignancies, or large vascular tumors. The thoracoabdominal approach to adrenalectomy is best applied to
very large unilateral lesions. Although this approach provides
optimal exposure of large vascular tumors; postoperative pain
and impairment of ventilation limit its application.
The first laparoscopic adrenalectomy was reported in
an adult in 1991.44 Since then, a number of studies involving laparoscopic adrenalectomy in children have been
published,45,46 demonstrating the feasibility and safety of this
approach. Most commonly, laparoscopic adrenalectomy is
performed with the patient in the lateral position. A kidney
rest elevates the flank opposite the adrenal lesion. Four
or five trocars are placed in a subcostal position on the side
of the adrenal gland to be resected. Exposure is improved
on the right side by dividing the right triangular ligament of
the liver. Division of the lienocolic ligament on the left
improves exposure of the left adrenal gland. When possible,
the adrenal vein is ligated with clips at the initial point of

dissection. The adrenal specimen should be removed in a
specimen bag because of the potential for malignancy. Most
adrenal lesions in children are small and benign, making
laparoscopic resection an appropriate choice in the majority
of cases. Although no absolute contraindications to laparoscopic resection exist, an open approach should be
considered in patients with large tumors, malignancies
with potential lymph node involvement, and highly vascular
pheochromocytomas.
Partial adrenalectomies (termed cortical-sparing or adrenalsparing) have been described for bilateral pheochromocytomas, wherein a portion of a single gland or portions of bilateral glands are retained. Preliminary reports indicate few
recurrences and maintenance of corticosteroid independence.
Children are included in these cohorts but not individually
evaluated as a sub-group. Reports are surfacing of successful
laparoscopic cortical-sparing adrenalectomies as well.47 Longterm follow-up and continued surveillance are essential.
expertconsult.com.
CHAPTER 42
Tumors of the Lung

and Chest Wall
Stephen J. Shochat and Christopher B. Weldon
The majority of pulmonary neoplasms in children are due to

metastatic disease; however, primary pulmonary tumors of the
lung do occur in the pediatric age group. The approximate
ratio of primary pulmonary tumors to metastatic neoplasms
and non-neoplastic lesions of the lung is 1:5:60.1 Although
primary pulmonary tumors are rare in children, the majority
of these tumors are malignant. In a review of 383 primary pulmonary neoplasms in children by Hancock and colleagues,2
76% were malignant and 24% were benign. This incidence
is similar to that previously reported by Hartman and
Shochat.3 Table 42-1 demonstrates the variety of primary pulmonary neoplasms seen in children. This chapter addresses
the more common benign and malignant primary pulmonary
tumors in children and discusses the treatment of pulmonary
metastatic disease in the pediatric population.
Benign Tumors of the Lung
adults, occurs rarely in children younger than 10 years

(approximately 8% of cases). However, plasma cell granuloma
is the most common benign tumor in children and accounts
for slightly more than 50% of all benign lesions and approximately 20% of all primary lung tumors.3 These tumors usually present as peripheral pulmonary masses but occasionally
present as polypoid endobronchial tumors.5,6 The pathogenesis of plasma cell granuloma is not well understood, but an
antecedent pulmonary infection has been reported in approximately 30% of cases. The mean age at presentation in children
is 7 years of age, and 35% of the children are between 1 and
15 years of age.57 Many children are asymptomatic at the time
of presentation, but fever, cough, pain, hemoptysis, pneumonitis, and dysphagia may be present. The natural history is that
of a slow-growing mass, starting as a focus of organized pneumonia with a tendency for local invasion. However, rare cases
of rapid growth have been reported.8 Extension of the tumor
beyond the confines of the lung is common. At least four
deaths have been reported resulting from tracheal obstruction
or involvement of the mediastinum by massive lesions.
Treatment consists of a conservative pulmonary resection
with removal of all gross disease if possible. Primary hilar adenopathy may be present, and local invasion with disregard for
tissue planes mimics malignancy. A frequent problem is identifying the benign nature of these masses. However, the diagnosis can usually be confirmed by frozen section. Malignant
fibrous histiocytoma of the lung, an extremely rare tumor in
children, can mimic plasma cell granuloma and must be considered in the differential diagnosis.9 Recurrences following
resection are rare but have been reported. Nonsteroidal antiinflammatory drugs have been used to treat large inoperable
lesions, with encouraging results.10
HAMARTOMA
Pulmonary hamartoma is the second most frequent benign lesion seen in children. These lesions usually present as parenchymal lesions and can be quite large. Approximately one
quarter are calcified, and popcorn-like calcification is pathognomonic.11 Two endobronchial lesions have been reported.
Four tumors occurring in the neonatal period were quite large
and were associated with significant respiratory distress; all
were fatal. An interesting triad is the combination of pulmonary hamartoma, extraadrenal paraganglioma, and gastric
smooth muscle tumors; the majority of these patients are
young women. Carney triad, in addition to its female predilection, is seen in young patients, is associated with multifocal
gastrointestinal stromal tumors (GISTs) and has an unpredictable biological behavior.12 Conservative pulmonary resection
is the treatment of choice; however, lobectomy, or even pneumonectomy, may be required, especially for large lesions and
endobronchial lesions when sleeve resection is not possible.
Malignant Tumors of the Lung
------------------------------------------------------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------------------------------------------------------
PLASMA CELL GRANULOMA

(INFLAMMATORY PSEUDOTUMOR)
BRONCHIAL ADENOMA
Plasma cell granuloma has also been called inflammatory

myofibroblastic tumor, fibroxanthoma, histiocytoma, and
fibrohistiocytoma.4 This lesion, which is seen frequently in
The most frequently encountered malignant primary pulmonary tumor is bronchial adenoma. These tumors are a heterogeneous group of primary endobronchial lesions. Although
adenoma implies a benign process, all varieties of bronchial
567
568
PART III
adenomas occasionally display malignant behavior. There are

three histologic types: carcinoid tumor (most common),
mucoepidermoid carcinoma, and adenoid cystic carcinoma.
Carcinoid tumors account for 80% to 85% of all bronchial adenomas in children.13 The presenting symptoms are usually
due to incomplete bronchial obstruction, with cough, recurrent pneumonitis, and hemoptysis. Because of diagnostic difficulties, symptoms are often present for months; occasionally,
TABLE 42-1
Primary Pulmonary Neoplasms in Children
Type of Tumor
No. of Patients (%)*
Benign (n 92)
Plasma cell granuloma
Hamartoma
Neurogenic tumor
Leiomyoma
Mucous gland adenoma
Myoblastoma
Benign teratoma
48 (52.2)
22 (23.9)
9 (9.8)
6 (6.5)
3 (3.3)
3 (3.3)
1 (1.1)
Malignant (n 291)
Bronchial adenoma
Bronchioalveolar carcinoma
Pulmonary blastoma
Fibrosarcoma
Rhabdomyosarcoma
Leiomyosarcoma
Sarcoma
Hemangiopericytoma
Plasmacytoma
Lymphoma
Teratoma
Mesenchymoma
Myxosarcoma
118 (40.5)
49 (16.8)
45 (15.5)
28 (9.6)
17 (5.8)
11 (3.8)
6 (2.1)
4 (1.4)
4 (1.4)
3 (1.0)
3 (1.0)
2 (1.7)
1 (0.3)
Modified from Hancock BJ, DiLorenzo M, Youssef S, et al: Childhood primary

pulmonary neoplasms. J Pediatr Surg 1993;28:1133-1136.
*Percent of benign or malignant tumors.
children with wheezing have been treated for asthma, delaying

diagnosis for as long as 4 to 5 years. Metastatic lesions are
reported in approximately 6% of cases, and recurrences occur
in 2%. There is a single report of a child with a carcinoid
tumor and metastatic disease who developed the classic
carcinoid syndrome.14 Bronchial adenomas of all histologic
types are associated with an excellent prognosis in children,
even in the presence of local invasion.15
The management of bronchial adenomas is somewhat controversial, because most are visible endoscopically. Biopsy in
these lesions may be hazardous because of the risk of hemorrhage, and endoscopic resection is not recommended.
Bronchography or computed tomography (CT) may be helpful to determine the degree of bronchiectasis distal to the
obstruction, because the degree of pulmonary destruction
may influence surgical therapy.16 However, Tagge and colleagues17 described a technique for pulmonary salvage despite
significant distal atelectasis. Conservative pulmonary resection with removal of the involved lymphatics is the treatment
of choice. Sleeve segmental bronchial resection is possible in
children and is the treatment of choice when feasible.1820 Adenoid cystic carcinomas (cylindroma) have a tendency to
spread submucosally, and late local recurrence or dissemination has been reported. In addition to en bloc resection with
hilar lymphadenectomy, a frozen section examination of the
bronchial margins should be carried out in children with this
lesion.
BRONCHOGENIC CARCINOMA
Although bronchogenic carcinoma is rare in children, this tumor was the second most common malignant lesion reported
by Hancock and colleagues.2 Interestingly, squamous cell carcinoma was rare, with the majority of tumors being either
undifferentiated carcinoma or adenocarcinomas. The term
bronchioalveolar carcinoma has been used in most cases.21
These tumors are associated with both cystic adenomatoid
malformations and intrapulmonary bronchogenic cysts
(Table 42-2).4,11,2138 Only rare survivors have been reported,
TABLE 42-2
Bronchioalveolar Carcinoma Associated with Congenital Cystic Lung Malformations
Year of
Publication
Type of Lung
Cyst
Age at Diagnosis
(Year)
Author Comments
Prichard22
Hurley23
Benjamin24
1984
1985
1991
CCAM type 1
CCAM type 1
CCAM type 1
30
Died of metastatic disease
19
BAC diagnosed in same lobe with segmental resection 19 years earlier;

died at 23 years of age
Morresi21
Ribet26
Kaslovsky27
Granata28
Endo29
1995
1995
1997
1998
1982
20
42
11
11
37
De Perrot30
2001
MacSweeney31
2003
Sudou32
2003
CCAM type 1
CCAM type 1
CCAM type 1
CCAM type 1
Bronchogenic
(intrapulmonary)
Bronchogenic
(intrapulmonary)
CCAM type 1, 0.5,
13, 18, 30, 36
CCAM type 1
79
Incomplete resection of CCAM in neonatal period

Lobectomy for recurrent infection; BAC was finding
Abnormal CXR noted 10 years earlier; presented with dyspnea, BAC was
incidental finding
Long-standing history of cyst infections
17
1 BAC in a recurrent cyst; one other patient with a typical adenomatous

hyperplasia (both patients underwent segmental resection)
Abnormality seen on CXR from 10 years earlier
Adapted from LaBerge JM, Puligandla P, Flageole H: Asymptomatic congenital lung malformations. Semin Pediatr Surg 2005;14:16-33.
BAC, bronchioalveolar carcinoma; CCAM, congenital cystic adenomatoid malformation; CXR, chest radiograph.
CHAPTER 42
TUMORS OF THE LUNG AND CHEST WALL
569
and mortality exceeds 90%. The majority of children present

with disseminated disease, and the average survival is only 7
months. Localized lesions can be treated by complete resection, followed by adjuvant therapy. Mucoepidermoid carcinoma of the bronchus has also been described in children
as young as 4 years (Fig. 42-1).39
PULMONARY BLASTOMA
Pulmonary blastoma is a rare malignant tumor that occurs
primarily in adults and arises from mesenchymal blastema.
This tumor is an aggressive lesion, with metastatic disease at
presentation in approximately 20% of cases.40,2 They may
arise from the lung, pleura, and mediastinum.41 These tumors
are classified into three types: type I (purely cystic), type II
(cystic and solid), and type III (completely solid).42 Type I
tumors may be difficult to distinguish from cystic adenomatoid malformation.43 Occasionally, these tumors may arise
in an extralobar sequestration or in a previous lung cyst
(Table 42-3).22,25,29,36,41,4473 The majority of cases occur in
the right hemithorax (Fig. 42-2). Frequent sites of metastases
FIGURE 42-1 Anteroposterior view of a right upper lobe lesion in a

4-year-old girl. The tumor was resected by right upper lobectomy
and was shown to be a mucoepidermoid carcinoma. (Courtesy Jay
L. Grosfeld, MD.)
TABLE 42-3
Mesenchymal Malignancy and Cystic Lung Malformations
Author
Year
Type of Lung Cyst
Type of Malignancy
Stephanopoulos44
Ueda45
Martinez46
Valderrama47
Sumner48
Weinberg49
Krous50
Weinblatt51
Holland-Moritz36
Morales25
Williams52
Allan53
1963
1977
1978
1978
1979
1980
1980
1982
1984
1986
1986
1987
Myxosarcoma
RMS
Pulmonary blastoma
Pulmonary blastoma
Pulmonary blastoma
Mixed mesenchymal sarcoma
Embryonal RMS
Pulmonary blastoma
PPB
Pulmonary blastoma
Embryonal RMS
RMS
Hedlund54
Cairoli55
Domizio56
Senac57
Murphy58
Bogers59
Calabria60
McDermott61
Seballos62
Tagge63
Adirim64
DAgostino65
Federici66
Ozcan67
Papagiannopoulos68
Stocker69
1989
1990
1990
1991
1992
1993
1993
1993
1994
1996
1997
1997
2001
2001
2001
2002
Cystic hamartoma
CCAM
Polycystic disease
Extralobar sequestration
Peripheral cyst
Congenital lung cyst
Bronchogenic cyst (intrapulmonary)
Cystic lung disease
Pneumatocele
Congenital cyst
CCAM
Congenital origin of cysts not
confirmed
Cystic hamartoma
CCAM
Congenital cyst
Bronchogenic cyst, CCAM (2)
Lobar emphysema
Pneumatoceles
Congenital cyst
CCAM
Bilateral pneumatocele
CCAM type 1
CCAM type 2
CCAM type 1
CCAM
CCAM type 4
CCAM type 4
RMS
RMS
Malignant mesenchymoma
PPB
Embryonal RMS
RMS
Pulmonary blastoma
Embryonal RMS
Pulmonary blastoma
PPB
Pulmonary blastoma
Embryonal RMS
PPB
Embryonal RMS
PPB
PPB
Adapted from LaBerge JM, Puligandla P, Flageole H: Asymptomatic congenital lung malformations. Semin Pediatr Surg 2005;14:16-33.
CCAM, congenital cystic adenomatoid malformation; CPAM congenital pulmonary airway malformation; PPB, pleuropulmonary blastoma;
RMS, rhabdomyosarcoma.
Age at Diagnosis
(Months)
18
24
48
108
30
30
48
21
21, 30
18, 22
36
48
24, 36, 42
18
36
22
45
22
36
13
30
48
570
PART III
FIGURE 42-2 A, Computed tomography scan of the chest shows a cystic lesion in the right hemithorax. B, The tumor was resected (lobectomy), and the
histology showed findings consistent with a pleuropulmonary blastoma. (Courtesy Jay L. Grosfeld, MD.)
Although children with primary lung tumors represent a heterogeneous group of patients, analysis of the reported cases suggests
that evaluation and treatment are similar in the majority of
patients. Many children are asymptomatic, especially those with
benign tumors; however, cough, recurrent pneumonitis, and
symptoms of atypical bronchial asthma may be the initial presentation. Radiographic findings usually indicate a solitary mass lesion or evidence of airway obstruction with resultant atelectasis
and pneumonitis. Because many of these tumors can be visualized by bronchoscopy, a bronchoscopic examination should be
performed. Flexible bronchoscopic techniques may be helpful
for diagnosis, but the use of rigid bronchoscopy with modern
magnification, along with general anesthesia, is necessary if endoscopic biopsy is contemplated. Preparation for emergency
thoracotomy should be made at the time of bronchoscopy in
the event of life-threatening hemorrhage.
Bronchoscopic removal of some isolated lesions may be
attempted, but because of the high incidence of recurrence
and the possibility of severe hemorrhage, this technique should
be used selectively. Conservative surgical resection is the procedure of choice for benign pulmonary tumors to achieve histologic diagnosis and preserve maximum functioning lung
tissue. Thoracoscopic resection is an option in these children.83
CT and magnetic resonance imaging should be performed in
children with large space-occupying lesions to determine resectability. Fine-needle aspiration for cytology or core needle biopsy
may be performed as the initial procedure for diagnosis in selected cases. Treatment of malignant lesions varies, depending
on location and histology. Sleeve resections should be considered for bronchial adenomas. Resection of involved lymphatics
should be considered with malignant lesions. Combinedmodality therapy with adjuvant chemotherapy and possibly radiation therapy may be helpful in children with large primary
malignancies or dissemination.
An important consideration is the association of primary
lung tumors with congenital cystic pulmonary malformations.
These lesions may be asymptomatic and are often discovered
incidentally. In some instances, the natural history of the lung
cyst is unknown, and a few may regress.80 Although some authors recommend simple observation, most pediatric surgeons argue against prolonged observation of cystic lesions
because of an increased risk of infection, pneumothorax,
sudden cyst enlargement with potential respiratory compromise, and associated malignancy.* As mentioned previously,
* References 3, 23, 25, 26, 31, 36, 4469, 7782.
* References 46, 57, 60, 63, 64, 68, 80.
are the liver, brain, and spinal cord. Local recurrences are frequent, and the mortality rate is approximately 40%.2,7476 The
majority of children present before 4 years of age, and symptoms include persistent cough, chest pain, episodes of pneumonia that are refractory to antibiotics, and hemoptysis.
Diagnosis is achieved by CT of the chest, bronchoscopy, and
biopsy. Because most of these tumors are located peripherally,
resection is usually possible by segmental or lobar resection.
The use of multimodal neoadjuvant chemotherapy and radiation following surgical resection has shown promising results
in a few patients with extensive disease and dissemination.41,75
Chemotherapeutic agents that have been used include actinomycin D, vincristine, cyclophosphamide alternating with
courses of doxorubicin, and cisplatin. Histologic evaluation
of the tumor shows an exclusive mesenchymal composition,
including primitive tubules, immature blastema, and spindle
cell stroma. Some demonstrate elements of embryonal rhabdomyosarcoma (RMS) arising within a multicystic lesion.
RHABDOMYOSARCOMA
RMSs of the lung are rare and account for only 0.5% of all
childhood RMSs (see Chapter 35).45,77 Many of the lesions
are endobronchial in origin (Fig. 42-3); however, several
cases apparently originated in congenital cystic anomalies.
(see Table 42-3).* This is an important issue because 4% of benign tumors and 8.6% of malignant tumors enumerated in
Table 42-1 were associated with previously documented cystic
malformations.2 Tumors that developed in these malformations included 11 sarcomas, 9 pulmonary blastomas, 3 bronchogenic carcinomas, and 2 mesenchymomas.
COMMENTS
CHAPTER 42
571
FIGURE 42-3 Patient with complete atelectasis of the left lung (A) and obstruction of the left main bronchus secondary to rhabdomyosarcoma (B).
Pulmonary metastases occur much more frequently than primary tumors in children, and the surgical approach depends
on the histology of the primary tumor and the response of the
primary site to combined-modality therapy.72,85 Pulmonary
metastases should not be considered for resection until the
primary tumor is eradicated, without evidence of recurrence
and other sites of metastatic disease ruled out. Tumors most
frequently considered for pulmonary metastasectomy are
osteosarcoma (OS), soft tissue sarcoma, and Wilms tumor.86
pulmonary metastases.87,88 Roth and colleagues73 showed that

patients with fewer than four pulmonary nodules had an improved survival versus those with more than four lesions.
According to Goorin and colleagues,89 a complete resection
of all pulmonary lesions is an important determinant of
outcome, and penetration through the parietal pleura is associated with an adverse outcome. Although somewhat controversial, the outlook seems to be somewhat improved, even in
patients presenting with pulmonary metastases, if complete
resection of all metastatic lesions can be accomplished.90 Harris
and colleagues91 reported a 68% survival rate in 17 patients
with fewer than eight pulmonary nodules at presentation
following chemotherapy, resection of the primary tumor, and
pulmonary metastasectomy. The data in Table 42-4 suggest that
an aggressive attempt at surgical resection of pulmonary metastases is indicated in OS, possibly irrespective of the number of
lesions or the interval to the development of metastases.*
A number of recent studies have shown a survival advantage
in patients with repeated metastasectomy, including patients
with as many as five recurrences.74,78,79
OSTEOSARCOMA
SOFT TISSUE SARCOMA
Children with OS should be considered for resection of pulmonary metastases once the primary lesion is controlled. The
overall disease-free survival is approximately 40% in children
who develop metachronous pulmonary metastases. Multiple
factors, such as number of pulmonary nodules and time of recurrence, play an important role in children with OS and
The usefulness of resecting pulmonary metastases in patients

with soft tissue sarcoma depends on the histologic subtype.
Rarely is pulmonary resection of metastatic lesions required
in RMS, and resection of pulmonary metastasis in Ewing
there is evidence suggesting a relationship between type IV

cystic adenomatoid malformation and type I pulmonary blastoma. Although complete lobectomy with negative margins is
adequate treatment for these patients, close observation is
recommended.31,35,84 If patients with asymptomatic cystic
malformations are observed without resection, they should
be followed closely and evaluated frequently.
Treatment of Metastatic Disease

------------------------------------------------------------------------------------------------------------------------------------------------
* References 34, 38, 70, 82, 89, 9295.
572
PART III
TABLE 42-4
Pulmonary Metastasectomy for Osteogenic Sarcoma
Martini92
Spanos93
Telander82
Giritsky34
Rosenberg94
Marion70
Schaller38
Goorin89
Carter95
Average Interval
to Relapse
No. of Procedures
(Months) (Range)
Disease-Free Survival,
(No. of Lesions)
Median Follow-up for Survivors/

Author No. (%) (mo)
No. of Patients
(Range)
22
29
28
12
18
12
17
32
43
10 (2-25)
15.7 (4-30)
9/6 (2-34)
9 (1-21)
13 (2-20)
12.5 (4-59)
13 (1-83)
59 (113)
52 (124)
60 (173)
19
9
34
26 (>63)
7 (32)
11 (37)
13 (46)
6 (50)
7 (39)
5 (42)
7 (41)
9 (28)
4 (10)
33 (15-234)
36 (9-234)
25 (6-48)
17 (9-39)
(36-72)
(12-192)
55 (19-101)
69 (59-80)
From LaQuaglia MP: The surgical management of metastases in pediatric cancer. Semin Pediatr Surg 1993;2:75-82.
sarcoma has not been found to be efficacious.71,72 Several

European protocols are being developed to better define the
role of pulmonary resection in Ewing sarcoma. The remaining
sarcomas should be considered for resection if complete excision is possible and the patients primary tumor is under control. The time to development of pulmonary metastases,
number of lesions, and tumor doubling time are all significant
prognostic factors in soft tissue sarcomas. Historically, approximately 10% to 20% of these patients can be salvaged by resection of pulmonary metastases.37
is not responsive to chemotherapy or radiation.97 Musclesparing techniques are available in those children requiring
posterolateral thoracotomies, and thoracoscopy may be appropriate in certain cases.37 New localization techniques
are being developed to aid in the thoracoscopic resection of
lung lesions.81 However, port site recurrences have been
reported following thoracoscopic resection of pulmonary
metastatic disease.98,99
Tumors of the Chest Wall

------------------------------------------------------------------------------------------------------------------------------------------------
WILMS TUMOR
EPIDEMIOLOGY
Rarely is pulmonary resection of metastatic disease required in

children with Wilms tumor. In a review of the National
Wilms Tumor Study by Green and colleagues,96 no advantage
of pulmonary resection was found compared with chemotherapy and radiation therapy alone. In an attempt to avoid pulmonary radiation, de Kraker and colleagues33 suggested a
protocol using primary pulmonary resection after chemotherapy for pulmonary metastases. Only 5 of 36 patients
ultimately required resection of pulmonary metastases following chemotherapy, because most patients had a complete
response with chemotherapy alone. One encouraging finding
was that only 4 of 36 children required whole-lung irradiation. Because the results of chemotherapy and whole-lung
irradiation are excellent for children with Wilms tumor and
pulmonary metastases, pulmonary resection of metastases
should be reserved for selected cases (see Chapter 30).
Tumors of the chest wall are rare entities in the pediatric population with an incidence of no more than 2%,100,101 and up
to two thirds of these lesions are malignant.102 The majority
arise from the bony structures of the chest wall (55%), as
opposed to soft tissue (45%).103 Collectively, a 60% 5-year
overall survival rate for all tumors has been reported, with a
recurrence rate of 50% (local and distant) and subsequent
5-year survival rate of only 17%.104
COMMENTS
Operation for pulmonary metastases in children depends on
the histology of the primary tumor, the extent of the metastatic
disease, and whether the metastatic disease is responsive to
chemotherapy. The surgical approach varies, depending on
the disease process and the age of the patient. No difference
in survival has been demonstrated with sequential lateral thoracotomy versus sternotomy, but the latter is preferable in
older patients with OS. Complete resection of all metastatic
disease is an important consideration, and the use of
automatic stapling devices can be helpful. Wedge resection
is usually possible in children with OS. However, formal lobectomy or segmentectomy may be required to remove all
of the tumor completely, especially when the primary tumor
PRESENTATION
Masses of the chest wall typically present as lumps bulging underneath the skin, and the majority of malignant lesions have
pain as a presenting symptom as well. In young children and
infants, they are often found incidentally by caregivers, while
older children and young adults may present with larger
masses that have been present and growing for some time. Incidental discovery on routine chest imaging has been reported
to be as high as 20%.105 They can be found anywhere on the
thorax, and the tissue of origin is generally mesenchymal in
nature, regardless of whether the tumors are malignant or
benign. Hence, sarcomatous variants are the most common
malignant tumors, while carcinomas are almost nonexistent.
The minority of patients present with nonspecific symptoms
of respiratory compromise or dysfunction (tachypnea,
hypoxia, cough, dyspnea on exertion), and these symptoms
may have been present for quite a while before seeking
medical advice. Symptoms stem from parenchymal compression from the mass intruding into the pleural space and onto
the lung or from malignant effusions, both of which interfere
with normal respiratory mechanics. Regardless of the presentation, a full history and physical exam, including a family
CHAPTER 42
history, travel history, injury history, and extensive review of

systems, is warranted to document other etiologies or associated conditions. Finally, depending on the degree of respiratory embarrassment, pulmonary function tests may be
indicated prior to proceeding with any intervention.
DIAGNOSTIC ADJUNCTS
Once the initial evaluation has been performed in the office,
basic laboratory evaluations for complete blood count, coagulation profile, and baseline chemistries are needed. Imaging
studies should consist first of erect, posterior-anterior, and lateral chest radiographs to evaluate the location, size, presence
of calcifications, osseous involvement, and the presence of
pulmonary parenchymal disease. Next, an ultrasound exam
to determine the echo features (solid versus cystic, degree of
homogeneity) and vascularity of the mass is recommended.
Axial imaging (computed tomography or magnetic resonance
imaging [MRI]) is performed afterward. The advantages of CT
reside in its ability to clearly define the lung parenchyma and
pleural space in relation to the osseous, vascular, and soft
tissue components of the thorax (and hence mass), and the fact
that it is a fast technique requiring minimal to no sedation even
in the youngest of patients. The negative aspects of CT are the
radiation exposure with subsequent risk of a secondary malignancy.106 The benefits of MRI include better definition of the
soft tissue components versus CT, as well as enhanced evaluation of the osseous and neural structures to determine the extent of central or peripheral nerve involvement and/or the
presence of skip lesions or metastases. Unfortunately, this
technique is time consuming and generally requires sedation
or even general anesthesia to adequately acquire the data. Motion artifact from the heart and lungs can also interfere with
this technique, limiting its utility, but this obstacle is being
overcome with the use of cardiac-gated, respiratory-triggered
protocols.107,108 Determination of the precise entity from
radiology studies alone is impossible, but the accurate construction of a differential diagnosis is readily possible, including the differentiation of malignant versus benign
lesions.107,108 Finally, other imaging studies may also be
indicated to determine the presence of metastases (brain and
abdominal CT, bone scan, positron emission tomogram [PET]
scan) depending on the type of lesion, especially if malignant.
Recent reports have suggested that the combination of PET
and CT scans yields more accurate data in assessing the primary tumor, local and regional lymph node basins, evidence
of recurrence, and for response to ongoing therapies.109,110
Once initial studies have been performed, retrieval of tissue
for histopathologic evaluation and diagnosis is warranted.
DIAGNOSIS
Biopsy options include small or large specimen approaches. If
a mass is small (less than 3 centimeters) or thought to be benign, then an upfront excisional biopsy may be warranted.
However, the incision should be oriented so that a future reexcision, if needed, can be performed without compromising
oncologic principles. Excising a normal rim of tissue circumferentially around the mass is also something for which the
surgeon should opt. If the mass is large (greater than 4 to 5
centimeters), fixed to surrounding structures, involving many
structures in the thorax, or if it is considered malignant by
573
imaging, then either an incisional biopsy or core needle biopsy

is warranted. Placing the incision in-line with any future resection is of paramount importance, regardless of the technique
used, and either approach will yield enough tissue for histopathologic and cytogenetic analyses.111 Once a diagnosis is
confirmed, then disease-specific treatment algorithms may
be initiated.
THERAPEUTIC PRINCIPLES
Though treatment regimens are tumor specific, there are certain general principles that apply. For malignant lesions, multimodality therapy is the accepted paradigm for the majority of
lesions, while simple extirpation is the rule with benign entities. With surgery, the most important concept to emphasize is
that of the need for negative margins to decrease the risk of
recurrence and subsequent therapy. Surgical extirpation also
mandates wound reconstruction, which must be considered
prior to the initiation of operative therapy. Large defects
(greater than 5 centimeters, except for posterior and superior
lesions where the defect will be buttressed by the scapula) will
require the use of prosthetic materialsrigid (silicone, Teflon
[DuPont, Wilmington, Del.], methyl methacrylate) or flexible
(Prolene mesh [Ethicon, Cincinnati, Ohio], PTFE mesh, Marlex mesh [Chevron Phillips Chemical, Bartlesville, Okla.],
Gore-Tex [WL Gore & Associates, Newark, Del.])and/or autologous tissues (pedicle or free flaps [latissimus dorsi, rectus
abdominis, or pectoralis major]) to reconstruct the chest wall
and thus ensure normal chest wall mechanics and prevent respiratory embarrassment.
TUMOR TYPES
Chest wall tumors are separated into benign and malignant cohorts (Table 42-5), as well as primary and secondary lesions.
Specific tumors and their treatment will be outlined in the
subsequent sections, but a discussion concerning secondary
tumors is beyond the scope of this work.
Benign Chest Wall Tumors
Aneurysmal Bone Cyst Aneurysmal bone cysts (ABCs) can
be found anywhere on the chest wall, and they generally
arise in the ribs. They have characteristic patterns of appearance
on both chest radiographs and MRI,107 and they can grow to be
quite large, producing local destruction to the adjacent tissues.
Surgical extirpation with complete excision is the treatment of
choice, and recurrence is rare. Histologically, the lesions are
blood-filled cysts composed of fibrous tissue and giant cells.
Chondroma Chondromas are slow growing, painless masses
that usually arise in the costal cartilages. On imaging studies,
they are lytic lesions with sclerotic margins, and unfortunately,
they are difficult to distinguish radiographically from their malignant brethren, chondrosarcomas. Hence, complete resection
with a wide margin of normal tissue is advocated.112
Desmoid Desmoid tumors are fibrous neoplasms that can be
found anywhere in the body. They are thought to be benign,
but they have also been reported to undergo malignant degeneration.112 Desmoid tumors infiltrate adjacent and surrounding tissues, and they are known to travel down fascial planes
and to encase neurovascular structures in the mediastinum or
574
PART III
TABLE 42-5
Pediatric Chest Wall Tumors
Benign
Aneurysmal bone cyst
Chondroma
Desmoid
Fibroma
Fibrous dysplasia
Lipoblastoma
Lipoma
Mesenchymal hamartoma
Osteochondroma
Osteoma
Vascular malformations
Malignant
Chondrosarcoma
Ewing sarcoma family
Fibrosarcoma
Langerhans cell histiocytosis
Leiomyosarcoma
Leukemia
Liposarcoma
Lymphoma
Neuroblastoma
Rhabdomyosarcoma
Osteosarcoma
the thoracic inlet. MRI is the radiologic procedure of choice to

best define the extent of involvement and the structures involved. Treatment is wide local excision with negative margins, but recurrence rates from 10% (negative margins) to
75% (positive margins) have been described by some authors.113115 If a complete resection is not possible, or if vital
structures are meant to be sacrificed during operative extirpation, then multimodality therapy consisting of radiation (50 to
60 Gy), and cytotoxic (vinblastine and methotrexate) and cytostatic (tamoxifen and diclofenac) chemotherapy is recommended, though the exact regimen is not well defined.116119
B
FIGURE 42-4 Axial (A) and coronal (B) images of a computed tomography scan of the chest in an infant with a mesenchymal hamartoma.
Fibrous Dysplasia Fibrous dysplasia is a benign condition

where normal bone is replaced by fibrous tissue. These lesions
are generally not large, and patients present with pain, generally from a pathologic fracture. On plain radiographs, these
lesions are described as lytic in nature with a characteristic
soap bubble appearance.120 Treatment is based on symptoms and concerns for possible fracture secondary to the inherent structural weakness the lesion produces in the bone.
Simple excision is the recommended procedure.
interspersed with osteoclastic giant cells. Treatment strategies

have traditionally consisted of complete resection with subsequent chest wall reconstruction, but considering the large size
of these lesions and the small volume of the chest cavity in the
infants in which they are discovered, concern over the future
complications of scoliosis and respiratory compromise from
this approach has been considerable. In light of the fact that
they are not known to undergo malignant degeneration,122
observation123,124 or other less morbid approaches (radiofrequency ablation125) have been described and recommended.
Mesenchymal Hamartoma Mesenchymal hamartomas

(MH) are masses found in infants or young children that
can also be discovered antenatally. The lesions are generally
well circumscribed, and though emanating from the chest wall
(one or several ribs), they abut or compress, as opposed to invade, thoracic structures (Fig. 42-4). Hence, presenting symptoms are primarily from respiratory embarrassment. These
lesions are well defined by radiographic features on crosssectional imaging, including mineralization and hemorrhagic
cystic structures.121 Histopathologically, these lesions consist
of chondroid tissue with blood-filled, endothelial-lined spaces
Osteochondroma Osteochondromas are tumors composed of bony and cartilaginous elements more commonly
found in males (3:1 ratio).112 The lesion can present
with pain from a pathologic fracture or compression of
nearby nerves, or it can be asymptomatic if it grows inward
into the thoracic cavity. The lesion is well characterized on plain radiographs, and it arises from the cortex
of the rib at the metaphysis and has a cartilage cap.120
Malignant degeneration has been documented,107 and resection
is warranted in all postpubertal patients, with symptoms, or if
the mass is growing.
CHAPTER 42
575
Malignant Chest Wall Tumors

The majority of clinically prevalent malignant tumors in the
pediatric population are sarcomatous lesions, and a select
sampling of these tumors will be addressed individually in
the following sections.
Chondrosarcoma Chondrosarcomas (CSs) are derived from
cartilaginous elements (costal cartilages) that are the most
common primary malignant bone tumor of the chest wall in
adults,126 and they are more common in males.112 CSs
have been associated with a prior history of trauma,127 as
well as being known to form from malignant degeneration of
the benign counterpart discussed previously.126 Some 10% of
patients will present with metastatic disease,103 especially in
the lungs and brain. Primary therapeutic intervention is
complete surgical extirpation with a margin of normal tissue
of at least 4 centimeters112 secondary to the high risk of
local recurrence (up to 75% with positive margins), even
with negative margins at the initial operation (10%).128 These
tumors are not chemotherapy responsive, and the role of
radiation is only for those lesions that are unresectable or
have known positive margins. Five-year survival has been
reported to range from 60% to 90%,128,129 and beneficial
prognostic factors are the absence of metastases at presentation and a complete resection.103,128
Ewing Sarcoma Family/Primitive Neuroectodermal
Tumors Ewing sarcoma family/primitive neuroectodermal
tumors (EWS/PNETs) are the most common malignant chest
wall lesions in the pediatric population.123 They are aggressive
tumors requiring multimodality therapy, but survival is still
poor despite these interventions. The tumors often present
as painful masses with frequent metastases (25%) to the lung,
bone, or bone marrow.103 EWS/PNET lesions are characterized by a balanced gene translocation (EWS/FLI1) (t11:22
[q24:q12]),130 and these tumors are defined histologically
as sheets of small, round cells with scant cytoplasm. On imaging studies, they have characteristic bony destruction described as lytic or sclerotic lesions.108 Treatment involves an
initial biopsy followed by neoadjuvant chemotherapy (four
cycles) with vincristine, actinomycin, cyclophosphamide,
and Adriamycin (Adria-VAC) alternating with etoposide
and ifosfamide. This regimen has demonstrated a great deal
of success in shrinking the tumor to improve survival
and facilitate complete resection (Fig. 42-5).131,132 In fact,
with the use of neoadjuvant chemotherapy, complete surgical
extirpation with negative margins was possible in 71%
of patients versus 37% who underwent primary surgical
intervention.132 The extent of surgery should include all
involved structures and a soft tissue or osseous margin. Postoperative adjuvant therapy uses the same preoperative
chemotherapy regimens, but not radiotherapy if complete
resection is achieved. This should be the goal, despite the
known radiosensitivity of this tumor,133 because of the
concern over the late effects (scoliosis, pneumonitis, cardiotoxicity, secondary malignancy, growth retardation, and breast
hypoplasia or aplasia) radiotherapy poses.132 The use of
radiotherapy is for residual and unresectable disease and for
patients who present with a malignant pleural effusion, where
it is an accepted therapeutic intervention. A recent European
consensus conference advocated for surgery rather than
B
FIGURE 42-5 Axial images of a computed tomography scan of the chest
in a child with a Ewing sarcoma family/primitive neuroectodermal tumor
(EWS/PNET) of the chest wall before (A) and after (B) neoadjuvant
chemotherapy.
irradiation in all cases.134 Five-year survival using the previously mentioned protocol was around 70% for nonmetastatic
disease,135 and the 8-year survival was roughly 30% with
metastatic disease.136 In patients presenting with metastatic
disease, the European Intergroup Cooperative Ewings
Sarcoma Studies Group demonstrated improved survival
with the use of myeloablative chemotherapy followed by stem
cell rescue at the conclusion of conventional treatment
protocols.137
Fibrosarcoma Fibrosarcoma (FS) (also known as infantile
or congenital fibrosarcoma) are malignant tumors found
throughout the body in infants who present with large masses
that often involve, invade, and surround adjacent structures.
FS have been found in the chest wall, and several reports have
documented the success of multimodality therapy in combating these tumors.138,139 FS can be distinguished from other
myofibrous and sarcomatous lesions by the presence of a
unique gene rearrangement between the TEL gene (12q13)
and TRKC gene (15q25).138 FSs are chemotherapy sensitive,
and reports demonstrating the effectiveness of neoadjuvant
chemotherapy with vincristine, actinomycin, cyclophosphamide, and Adriamycin, followed by surgical extirpation, are
well accepted.138,139 A recent report139 from Europe
576
PART III
demonstrated that 5-year overall and event-free survival rates

were 89% and 81%, respectively. The authors reported that in
their series complete surgical extirpation was rarely feasible
and that conservative surgical approaches should be adopted.
Furthermore, 71% of patients responded to alkylating agentfree and anthracycline agent-free regimens, and hence, this
regimen should be started first to limit toxicity.
Osteosarcoma OS of the chest wall can be primary or secondary tumors (prior sites of irradiation or from preexisting osseous lesions [Paget disease]).112 Primary lesions are primarily
of the ribs, and on imaging, they can be confused with chondrosarcomas.140 Chest radiographs will demonstrate a sunburst
pattern, and axial imaging concentrating on regional (bony
skip lesions) and distant (lung, liver, brain) metastases must
be sought.112 Pretherapy biopsy is the rule, and neoadjuvant
therapy precedes extirpative procedures. Overall survival rates
are poor (15% to 20%),103 but in the presence of nonmetastatic
disease, 5-year survival rates can exceed 50%.103 Prognosis is
related to the presence of metastases, the degree of tumor
burden, and the response to chemotherapy.141
Rhabdomyosarcoma RMS of the chest wall is a rare tumor
and encompasses no more than 7% of all RMS in Intergroup
Rhabdomyosarcoma Studies (IRS).142144 The chest wall site
is deemed an unfavorable site, and therefore this is an adverse
prognostic factor.142,143 Other adverse prognostic factors have
been reported to be histopathologic findings (alveolar versus
embryonal), tumor burden and size, incomplete resection,
and presence of metastatic disease (including lymph node metastases).143,145 Despite advances in the treatment of RMS over
the last 40 years, unfavorable sites carry an overall survival of
only 55% (versus 90% for favorable sites),143 and those with
truncal RMS have been reported to have a failure-free survival
rate of no greater than 67%.146 These tumors require multimodality therapy, and neoadjuvant chemotherapy followed
by surgical extirpation is the norm. Radiation is reserved for
lesions with positive margins following surgery, or in unresectable tumors. A report from Saenz and colleagues documented
the utility of radiation (median dose of 44 Gy) to salvage some
patients with residual disease.146 However, the necessity for
complete surgical resection has been called into question by
a recent report from the Childrens Oncology Group
(COG),147 where the outcome of patients enrolled in IRS
I-IV with chest wall RMS were analyzed. The report documents that regardless of clinical group (I-III) and other
tumor-specific factors (histologic subtype, tumor size), the
only critical factor to influence failure-free and overall survival
was the presence of metastatic disease. In the face of metastases, patients with chest wall RMS had an overall and failurefree survival of 7% and 7% versus 49% and 61%, respectively,
in the cohort without metastases (P < 0.001). Therefore the
authors suggest where gross total surgical resection will
produce significant morbidity or physical debilitation, less
aggressive operative approaches should be entertained.
expertconsult.com.
of plain radiographs can never be overstated. They facilitate

the initial workup and allow these patients to be referred to
specialized centers with multidisciplinary expertise. Although
the subsequent imaging modalities are important, the radiographs form a key part of surgical planning.
It is with the pediatric surgeon in mind that this chapter is
written. Lengthy discourse on the pathology is avoided,
and several excellent references exist.36 Instead, the format
adopted is a practical approach to the management of these
conditions. Where prudent, insights and controversies are
highlighted to spur interest in specific areas.
General Considerations
------------------------------------------------------------------------------------------------------------------------------------------------
PATHOPHYSIOLOGY
CHAPTER 43
Bone Tumors
Saminathan S. Nathan and John H. Healey
Bone tumors are rare. In the United States, there were 166,487
cases of breast cancer and 164,753 cases1 of prostate cancer in
2000. By comparison, there were only 2,051 cases of all types
of bone sarcomas that year. A large proportion of these tumors,
26.8% in one published database, occur in the pediatric population. There are no population-based benign bone tumor
registries; so, it would be impossible to establish their true
incidence. Most databases of this nature derive from tertiary
referral institutions, and so, benign conditions, which are
often asymptomatic, would be grossly underrepresented.
Nevertheless, one study has shown that up to 43% of children
have a bone lesion that mimics or is a true neoplasm during
skeletal development.2 This implies that the overwhelming
majority of lesions are benign.
The pediatric surgeon will often be called into the management of the patient with bone tumors for a number of
reasons. The very young child on follow-up for an unrelated
condition may manifest with a bone lesion secondary to
osteomyelitis or leukemia. The older child with a metastatic
osteogenic sarcoma may require the expertise of the pediatric thoracic surgeon for the resection of pulmonary nodules.
The teenager with a pathologic fracture through a unicameral bone cyst or nonossifying fibroma may present first to
the pediatric surgeon on call in the pediatric emergency
department.
The diagnosis of these rare conditions is readily attained
through a careful clinical evaluation. In that regard, the utility
The main aim of this section is to illustrate the specific issues

of the pathophysiology of bone tumors that distinguish them
from tumors of soft tissue. Bone tumors should be
approached initially from the standpoint of whether they
are benign or malignant. Whereas traditional approaches
regarding the treatment of most nonskeletal benign lesions
have been ones of benign neglect (if these lesions are not
perceived to be causing problems), the management of
benign bone lesions is complicated by the potential compromise of skeletal structural integrity. Cortical deficiency
weakens bones and can mandate treatment to prevent
fracture. The prudent, if rare, consideration is one of syndromic presentation and malignant transformation. Many of
these principles are applicable to malignant lesions as well.
However, malignant lesions have, as the cornerstone of consideration, their implications on survival, which will be elaborated. Metastatic lesions to bone are uncommon in the
pediatric age group. Their pathophysiologic implications tend
to be structural or diagnostic.
In the pediatric age group, benign lesions far outnumber
primary malignant lesions, which in turn outnumber metastatic lesions. Because of the protean manner in which benign
lesions behave, some are not evident in the physicians office.
Conclusions about their natural history and malignant potential are therefore difficult to ascertain.4 This is obviously not
the situation with malignant and metastatic lesions. In this
section, we discuss pathologic conditions of the bone that occur most commonly in the pediatric age group. In the pediatric
population, the commonly occurring benign lesions are the
unicameral bone cyst, aneurysmal bone cyst, enchondroma,
osteochondroma, nonossifying fibroma, and osteoid osteoma.
The common malignant bone tumors are osteogenic sarcomas and Ewing family tumors (Table 43-1). Here we highlight specific features of each tumor. For a more thorough
understanding of the pathology, the reader is directed to
any of a number of fine books on the subject.36
Benign Lesions
The typical benign lesion in the pediatric age group
(Table 43-2) is identified incidentally, because they rarely
cause symptoms. They are often diagnosed when a parent
notices a lump or deformity (e.g., osteochondroma) or a radiograph is obtained for an unrelated condition (e.g., nonossifying fibroma). The two main surgical issues are diagnosis
577
578
PART III
TABLE 43-1
Commonly Occurring Tumors by Age Group
Age
Benign Tumors
Birth to
5 years
Eosinophilic
granuloma
5 to 15
years
Unicameral bone
cyst
Osteochondroma
Aneurysmal bone
cyst
Osteoid osteoma
Enchondroma
Nonossifying
fibroma
Chondromyxoid
fibroma
Chondroblastoma
Unicameral bone
cyst
Osteochondroma
Osteoid osteoma
Aneurysmal bone
cyst
Nonossifying
fibroma
Giant cell tumor
Enchondroma
Chondroblastoma
Chondromyxoid
fibroma
15 to 20
years
TABLE 43-2
Incidence of the More Commonly Diagnosed Bone Tumors
Malignant
Tumors
Tumor-like
Conditions
Leukemia
Metastatic
neuroblastoma
Ewing sarcoma
Osteogenic
sarcoma
Osteomyelitis
Nonaccidental
injury
Fibrous
dysplasia
Osteomyelitis
Osteofibrous
dysplasia
Stress fracture
Osteogenic
sarcoma
Ewing sarcoma
Fibrous
dysplasia
Stress fracture
Bone Tumors
All Bone
Tumors (%)
Bone Tumors in the First

Two Decades (%)
7.86
2.60
2.99
1.13
3.02
5.10
1.07
0.41
4.69
1.96
1.94
0.99
0.98
0.80
0.66
0.14
Unknown
Unknown
14.9
4.6
7.53
3.50
Benign
Osteochondroma
Aneurysmal bone cyst
Osteoid osteoma
Nonossifying fibroma
Enchondroma
Giant cell tumor
Chondroblastoma
Chondromyxoid
fibroma
Unicameral bone cyst
Malignant
Osteogenic sarcoma
Ewing sarcoma
In using this table, a number of caveats need to be remembered. Most benign

lesions are often asymptomatic, and only symptomatic ones will present. Of
these, most will be managed in the primary care setting. Malignant lesions
will, however, usually present at a referral center. Hence, in terms of
population incidence, these figures are unreliable. In relative terms,
however, they have some utility in indicating their prevalence. Unicameral
bone cysts are left in this list as a reminder of their frequency.
By considering the factors of age, frequency, and location in the long bones
(see Fig. 43-3), a diagnosis can be proposed in the majority of cases. The
possibility of trauma should always be borne in mind, and in the
noncommunicative child younger than 5 years old, nonaccidental injury
may be the cause.
through a biopsy and stabilization of bones that have fractured

or are at risk to fracture, especially through a precarious
location. For example, a bone cyst in the neck of a femur
should be seriously considered for surgical stabilization,
because a fracture at this site may result in avascular necrosis
of the femoral head. The biopsy itself cannot be undertaken
lightly, because it can weaken the bone, mandating surgical
or external splinting. The challenge is to use a high-yield
biopsy with minimal morbidity.
Size of the Tumor Size is an important consideration for
surgical approach. For example, cartilaginous rib tumors
larger than 4 cm were found to have increased likelihood of
malignant behavior.3 Hence they should be resected widely
despite their relatively bland histologic appearance (Fig. 43-1).
Large tumors can also grow into neighboring compartments
and cause mechanical compromise to joints. Although this
is less critical in joints of the upper limb, it is important in
the spine and in the lower limbs, where they cause mechanical
impingement and pain. The disruption of a tubular bone by
growth of a neoplasm weakens the bone. Lesions that involve
more than 50% of the cross section of a bone are at risk of fracture and should be treated from a mechanical standpoint.79
Fracture of a malignant lesion may require amputation
rather than a limb-sparing operation.
FIGURE 43-1 A, Chondrosarcoma in the proximal humerus of a 13-yearold boy. This is an exceedingly rare diagnosis in this age group.
B, A proximal humeral resection with allograft reconstruction was performed. In children, the available prostheses may be too large, and hence
bulk allografts may be the only choice.
Fracture Through a Benign Lesion The fractured benign

lesion is typified by the unicameral bone cyst. These lesions
may appear radiographically to be aggressive, but a careful history and physical examination with appropriate
imaging modalities will usually establish their benign nature
CHAPTER 43
Benign lesions
BONE TUMORS
579
Malignant lesions
Enchondroma
Aneurysmal
bone cyst
Osteogenic
sarcoma
Osteochondroma
Unicameral
bone cyst
Non-ossifying fibroma
chondromyxoid fibroma
Ewings
sarcoma
Chondroblastoma
FIGURE 43-2 A, Large unicameral bone cyst of the proximal humerus

that had fractured. The aggressive appearance may lead one to suspect
a malignant process, but a careful evaluation of the margins of this lesion
and absence of periosteal reaction reaffirms the management decision of
observation before surgery. B, This cyst was curetted and packed with an
allograft 1 month after the fracture. Treatment with an intramedullary
fibular graft provided stabilization, and supplemental bone graft healed
the lesion.
(Fig. 43-2). Unicameral bone cysts that fracture may resolve

spontaneously. However, the vast majority continue to fracture
throughout a childs lifetime and prove to be disabling.10
In general, they should be treated surgically, especially if they
are symptomatic.
The timing of surgery is critical. An early biopsy after
fracture would show callus formation difficult to distinguish
from a malignant process. Therefore these lesions should be
observed during healing of the fracture for about a month,
after which a biopsy and definitive procedure are performed.
Location in Relation to the Physis Location in relation to
the physes is an important consideration distinguishing tumor
assessment and management of children versus adults
(Fig. 43-3). The term diaphyseal aclasis was coined to highlight a condition in which multiple osteochondromas, a condition primarily of the growth plate, caused disordered linear
growth of the long bone.6 These cases are often familial, and
children are rarely compromised by their condition. Joints of
the upper limb generally have a high tolerance for the resultant deformity. However, occasionally, degenerative arthritis
develops, especially in the lower limb, then requiring early
surgery.
Multiplicity of Bone Tumors Multiple bone lesions in an
individual are often syndromic and may confer a higher
incidence of malignant degeneration than when they occur
singly.46 Multiple osteochondromas occur in multiple
Osteoid osteoma
osteoblastoma
FIGURE 43-3 The location of lesions in relation to the physis gives a clue
to the diagnosis. In most cases, the diagnosis can be made on radiographs,
leaving further imaging to plan for surgery.
hereditary exostosesan autosomal dominant condition

caused by abnormalities of the EXT1, EXT2, and EXT3 genes
on chromosomes 8, 11, and 19.1113 Although each osteochondroma has a low probability of malignant transformation,
the cumulative risk is high. Children with this condition have
an increased incidence of 10% to 27.6% for malignant degeneration of an osteochondroma into a chondrosarcoma. By
comparison, isolated osteochondromas have a malignant
degeneration rate of about 1%.35 Because only symptomatic
lesions will present to the physician, the true incidence of malignant degeneration in isolated lesions is impossible to ascertain with certainty. Multiple enchondromatoses is a sporadic
condition that confers an increased incidence of malignant
transformation of up to 50% in the involved bones.4 Limblength inequality and malalignment are also common. Ollier
disease, as this condition is termed, has another counterpart
classically affecting one limb anlage. A variant, Maffucci syndrome, involves widespread enchondromas associated with
hemangiomas of the hand. The occurrence of multiple nonossifying fibromas, associated with mental retardation, cafe-aulait spots, endocrine disorders, cardiovascular malformations,
and ocular abnormalities has been termed Jaffe-Campanacci
syndrome, but this entity has no malignant implications.4,14
Site of Involvement The site of benign cartilaginous lesions
has important implications for malignant potential. Peripheral
lesions in the hand rarely turn malignant, while those closer to
the axial skeleton have important malignant potential even if
they appear benign histologically.36,14,15 Lesions in bones
adjacent to weight-bearing joints should be regarded with special concern. In the pediatric group, these lesions are usually
chondroblastomas. They grow epiphyseally and in so doing
can cause weakening of the subchondral bone and, ultimately,
an intraarticular extension or fracture that may even mimic
580
PART III
osteochondral defects. In the case of sarcomas, a relatively

conservative resection in this context would have to be
deferred to an extraarticular resection.
Metastatic Potential A unique feature of benign bone
tumors is that there is a small incidence of metastasis in these
lesions. Accordingly, 1.7% of chondroblastomas and 3% of
giant cell tumors5,1618 do metastasize. There is a controversy
about whether some of these lesions were, in fact, malignant
from the outset.19 However, the truly benign lesions that do
metastasize are atypical lesions that have had surgical manipulation, which may have embolized tumor cells. When
followed, some of these metastatic lesions, primarily in the
lung, may remain dormant and not progress. The possibility,
therefore, is that they represent a transport phenomenon more
akin to a mechanical embolism and not a true metastasis.3,19
MALIGNANT LESIONS
Epidemiology
The main histologic types of bone tumors are osteogenic
sarcoma, Ewing family tumor, chondrosarcoma, and other sarcomas. They affect children at a rate of 6:3:2:1, respectively.1,5
Osteogenic sarcomas (also known as osteosarcomas) are
malignant bone-forming tumors of the bone. They occur at
any age but most frequently present in an extremity in the
middle teenage years. There are various subtypes with varying
implications for survival. In general, the subtypes behave
similarly, except perhaps for telangiectatic osteogenic sarcoma, which bears special mention. In the prechemotherapy
era this was regarded as the tumor with the worst prognosis.20
Presently, however, it has the best prognosis.21 The lytic nature of these sarcomas weakens bone, resulting in the highest
rate of pathologic fracture. Increasingly, rarer forms of osteogenic sarcoma are described. Two variants of note are the small
cell sarcoma and giant cell-rich osteogenic sarcoma. The
former can be confused with a Ewing family tumor and thus
is often treated by similar chemotherapy protocols.22,23
The latter can be confused, in the appropriate setting, with a
giant cell tumor of the bone, which is a benign condition.2426
The Ewing sarcoma occurs at a younger age (see
Table 43-1) and may affect any bone, particularly, the femur,
pelvis, and humerus. It is the most common cancer in the
pelvis, ribs, foot, and fibula. It was once considered to be distinct from peripheral neuroectodermal tumors but has been
shown to be genetically identical to this entity. It is presently
considered to be in the same family of neoplasms also known
as Ewing family tumors.3,6
Chondrosarcoma is less prevalent in the pediatric age
group. It is more widely distributed in the body compared
with its occurrence in adults.
Genetics There have been few consistent genetic or syndromic associations with osteogenic sarcoma. Patients with the
Li-Fraumeni syndrome27 have a TP53 germline mutation28,29
on 9p21 and are predisposed to osteogenic sarcoma, breast
cancer, and leukemia (Fig. 43-4). Two to 3 percent of patients
with osteogenic sarcoma will be the proband for Li-Fraumeni
families.30 Another germline mutation of 13q14, hereditary
retinoblastoma (RB), predisposes to osteogenic sarcoma.31
Children who received radiation therapy for retinoblastoma,
B
FIGURE 43-4 A, Osteogenic sarcoma in the left scapula of a female
patient with Li-Fraumeni syndrome. This patient had a family history of
osteogenic sarcoma in a first-degree relative. At the time of staging for
the osteogenic sarcoma in the scapula, a lesion in the breast was
discovered on computed tomography (CT) of the chest. This was subsequently found to be an adenocarcinoma. B, The patient underwent
a scapular replacement. A latissimus dorsi flap was used for skin cover.
Hodgkin and non-Hodgkin lymphoma, Ewing family tumor,

and other cancers are at a 5% to 10% risk of developing
osteogenic sarcoma. Patients with an RB gene deletion and a
history of alkylating agent exposure from a prior malignancy
are predisposed to this complication as well. About 5% of all
osteogenic sarcomas occur as postradiation sarcomas.
The Ewing family tumor is a malignancy associated with a
number of translocations. The 11 to 22 translocation, resulting in an EWS-FLI1 fusion transcript, is the most common
variant, and type 1 is associated with the best prognosis.32
Other translocations include type 2 EWS-FLI1, EWS-ERG
from a 21,22 translocation, and EWS-ETV1 from a 7,22
CHAPTER 43
translocation. These rarer variants have not been as well studied but appear to confer a poorer prognosis.32 Further additive
mutations involving cell-cycle genes reduce the prognosis of
these tumors still more. The Ewing family tumor is the most
common solid tumor to metastasize to the brain.33
DIAGNOSIS AND STAGING

Bone tumors are diagnosed based on the well-recognized triad
of history, physical examination, and investigation. After a
clinical diagnosis, it is imperative that imaging and staging
procedures are done before biopsy. Preoperative imaging
allows for planning of the definitive procedure and hence
placement of the biopsy incision. In addition, changes that
would occur in the lesion after biopsy would be difficult to
distinguish from changes resulting from tumor growth on
imaging. Furthermore, changes in the lung after general
anesthesia (e.g., atelectasis) are difficult to distinguish from
metastatic deposits.
Clinical Evaluation
Although it is not possible to be comprehensive in this section,
the history and physical examination are important parts of
the assessment of a patient with a bone tumor. Patient demographics and tumor location narrow the differential diagnosis
and focus the workup efficiently.
The patients age is important (see Table 43-2). Most malignancies occur in the second decade of life.36 Among children,
subtle variation occurs in the prevalence of disease with
respect to age (see Table 43-1). Demographically, it is exceedingly rare for patients of African descent to have a Ewing
family tumor.6
Pain at rest is an important sign that occurs in tumors and
in other organic conditions, such as infection and bone infarction. It distinguishes these conditions from mechanical pain,
which occurs with activity. Most malignant tumors will
present with pain. Pain relieved by nonsteroidal antiinflammatory drugs (NSAIDs) is pathognomonic of osteoid osteoma.34 This lesion can occur at any age and is characterized
by painful scoliosis when it occurs in the spine.
A family history of malignancy should be discerned, especially in possible sentinel cases of the Li-Fraumeni syndrome.2729 Such patients should have systemic evaluation
in the form of radioisotope bone scans or positron emission
tomographic scans, to rule out other sites of involvement.
As described earlier, the surgeon should be alert to any
dysmorphism that the patient may have. Cutaneous stigmata
are evident in patients with neurofibromatosis, fibrous dysplasia, and Jaffe-Campanacci syndrome.14 Limb length discrepancies are seen in patients with multiple enchondromatoses
and multiple hereditary exostoses.35
Infection should be considered in the differential diagnosis
in almost every case seen. Tumor epidemiology is very telling.
For example, childhood leukemia is nearly 10 times as
common as Ewing family tumor, and so, rare manifestations
of leukemia are more common than routine presentations of
Ewing family tumor.
The nature of bony reconstruction also requires that the
method chosen be matched with the demands of the patient.
As such, an idea of the patients expectation should be sought
at this time.
BONE TUMORS
581
Radiology
The minimal radiologic assessment at the first visit should be
two orthogonal radiographic views of the area in question.
A radiograph remains the most specific diagnostic imaging test
and is the only one that gives the gestalt of overall assessment
of skeletal biology and mechanics. By analyzing the location
of the tumor (see Fig. 43-3), as well as whether it is benign
or malignant, the diagnosis can be made in the majority of
cases.36
Benign lesions are well circumscribed, with a good sclerotic
border, and produce no soft tissue edema. Malignant lesions
have lucent or variegated matrices and permeative borders.
Edema is often apparent with the presence of fat lines.
The often-quoted eponymous phrases are not specific to
distinct malignancies. The Codman triangle refers to the lifting
and ossification of periosteum at the periphery of an osteogenic sarcoma. The sunburst appearance is due to the ossification of fibers and vessels subperiosteally, as the tumor
expands out of the cortex. Onion skinning refers to the periodic ossification and expansion of periosteum from the cortex.
Any of these conditions can be seen in tumors or infections
that are sufficiently fast growing.
In Figure 43-3, epiphyseal lesions are typical of chondroblastoma or giant cell tumors; physeal lesions are typical of
osteochondromas; metaphyseal lesions are typical of osteogenic sarcomas, unicameral bone cysts, aneurysmal bone
cysts, and nonossifying fibromas; and diaphyseal lesions
are typical of Ewing family tumor, fibrous dysplasia, or
enchondromas.
Laboratory Evaluation
The main blood parameters of importance are lactate dehydrogenase and alkaline phosphatase.3638 Lactate dehydrogenase
levels have been used as a surrogate for tumor load and have
been correlated with survival in the case of Ewing family tumor.36 Serum alkaline phosphatase elevation is characteristic
of osteogenic sarcoma and is correlated with poor survival in
this condition.37,38 Glucose intolerance is associated with
chondrosarcoma of the bone.39,40 Erythrocyte sedimentation
rates, C-reactive protein, and white blood cell and differential
counts should be sought to rule out infection.
Preoperative Planning
Magnetic resonance imaging (MRI) of the lesion offers an assessment of compartmentalization of the tumor. A compartment is an abstract concept and refers to any plane that
offers a fascial or cortical bone barrier to contiguous spread.
It has implications for the extent of surgery, which by definition must be outside the compartment to be radical (see
later).41 Also, by forming a baseline assessment, one is able
to make an assessment of response to chemotherapy in the
case of neoadjuvant treatment.42 It has secondary importance
in providing the actual diagnosis. In specific examples it is
useful in histologic diagnosis. The aneurysmal bone cyst
shows fluid-fluid levels on an MR image. Pigmented villonodular synovitis is hypointense (dark) on T1- and T2-weighted
imaging because of hemosiderin deposition. Cartilaginous
lesions are hyperintense (light) on T2-weighted imaging.
Mineralized and dense fibrous tissues are dark on T1- and
T2-weighted imaging.43,44
582
PART III
Staging
Staging studies are meant to assess the degree of spread of
the disease. In the case of bone tumors two systems are used:
the Enneking system or surgical staging system (SSS),45
as adopted by the Musculoskeletal Tumor Society and the American Joint Committee on Cancer (AJCC) system, which at the
time of writing is in its sixth revision.46 In the case of Ewing
family tumor, a different classification than Enneking is used.47
In the SSS, tumors are designated G0, G1, and G2 for benign,
low-grade, and high-grade lesions, respectively. Benign lesions
(G0) are classified as latent, active, or aggressivedesignated by
Arabic numerals 1, 2, and 3, respectively. Malignant lesions are
designated with the Roman numeral I if low grade and II if high
grade. The further designation A or B denotes intracompartmental or extracompartmental disease. Stage III disease
is metastatic disease. Therefore in this classification, grade,
compartmentalization, and metastases are the fundamental
prognostic factors.
In the AJCC system, I and II similarly designate low- and
high-grade lesions. The letters A and B designate tumors
smaller or larger than 8 cm, respectively. The Roman numeral
III denotes multicentric disease, and IV denotes metastatic
disease. The designation IVA denotes pulmonary metastases,
and IVB denotes extrapulmonary metastases. Therefore
this classification considers grade, size, multicentricity, and
metastases as prognostic factors.
In the Enneking staging system of Ewing family tumor, stage
I tumors are solitary intraosseous lesions, stage II are solitary
lesions with extraosseous extension, stage III are multicentric
lesions, and stage IV are metastatic. It is unclear how to stage
patients who have independent sites of bone marrow involvement versus those who have circulating tumor cells identified by light microscopy (i.e., Enneking stage III or IV).
Modern pathology analysis extends these concepts to include
immunohistochemistry or reverse transcriptase polymerase
chain reaction (RT-PCR) of recombinant gene products.
The modalities used for staging are bone scans and computed tomography (CT) of the chest.45 Positron emission tomography scans are presently being evaluated, but have
fundamental utility in the management of recurrent or metastatic disease.48 In the case of Ewing family tumor, bone marrow
biopsies are obtained to try capturing cases that are multicentric
at presentation. The utility of this approach is being evaluated.49
BIOPSY
The biopsy is a critical procedure that can complicate management severely if not performed appropriately. Misplaced
incisions continue to be an important cause of resectable
tumors being rendered nonamenable to limb salvage surgery.41,50 A good pathologist who is comfortable handling
bony tissue is critical to this process. In the appropriate case,
extra tissue may be needed for cytogenetic studies. Ewing
family tumors are particularly fragile, and biopsy specimens
should be handled carefully to allow for processing.
Presurgical Considerations
As a general rule, all imaging and staging should be completed
before biopsy. The lesion that warrants biopsy should be given
consideration for a primary wide excision. This approach is
typically applicable to small lesions that are less than 3 cm,

lesions in expendable bones (e.g., distal phalanx), distal
lesions of the ulna, and proximal lesions of the fibula, where
there is a risk of common peroneal nerve contamination
(Fig. 43-5).
The lesion should preferably be sampled in the institution where the definitive procedure will be performed
and by the same surgeon. It has been shown repeatedly,
that when this approach is not used, the results are
compromised.50,51
Consideration should be given to needle biopsies in the
case of lesions in the pelvis or the spine, where the exposure
necessary for an open biopsy may be extensive and obliges
commitment to a definitive procedure.
A pathologist familiar with processing bone tissue should
be on hand to evaluate the biopsy. If tumor tissue can be
cut with a knife, then it can be cut with a microtome.
Frozen-section analysis is required primarily to ascertain
the adequacy and representativeness of the specimen and
secondarily for the definitive diagnosis.
Antibiotics should be withheld before the biopsy to improve the yield of cultures. The biopsy may be done with
use of a tourniquet, to prevent bleeding and dissemination
of the tumor locally. When the tourniquet is applied, simple
elevation should be used for exsanguination. Compressive
exsanguination should be avoided, because this could rupture
the tumor. At all times, the limb should be protected from
fracturing, because this would cause extensive local dissemination of disease.
Surgical Considerations
The planned incision for the definitive surgery should be
marked. This should generally follow extensile exposures
and be longitudinal along the line of the definitive incision.
The incision should be placed directly over the lesion. Flaps
and dissection should be avoided.
The incision is developed directly into the tumor. If there
is a soft tissue component of the tumor, then this alone
needs be sampled. If a bone biopsy is necessary, then
the edges of the biopsy specimen should be rounded to
minimize a stress riser. Frozen-section analysis will confirm
the adequacy of the biopsy. In the meantime, a culture is
taken, the tourniquet is released, and antibiotics are
given. Absolute hemostasis is needed at the conclusion
of the procedure to minimize spread of tumor cells in the
hematoma.
The wound is closed in layers. If a drain is necessary, this
should be brought out in the line of the incision so that it can
be excised at the time of definitive surgery.
Postsurgical Considerations
The patient should be limited to protected weight bearing, at
least until some healing of the biopsy or ossification of the
tumor as a response to neoadjuvant chemotherapy occurs.
This typically takes up to 6 weeks.
Fractures through osteogenic sarcomas have traditionally precluded limb salvage surgery. Recent studies
have shown that limb salvage may still be possible in
selected cases.5255 Special surgical consideration is needed
in these cases.
CHAPTER 43
BONE TUMORS
583
B
A
FIGURE 43-5 A and B, An aneurysmal bone cyst of the

right proximal fibula in a 17-year-old boy. C, In this instance, a primary wide resection was done, because the
bone was expendable and it prevented contamination
of the common peroneal nerve (arrow).
ADJUVANT THERAPY
This section concentrates on the use of radiation and chemotherapy. In general, these modalities are not used in the treatment of benign conditions. Up to 10% risk of malignant
transformation occurs when benign lesions are irradiated.36
Both chemotherapy and radiation therapy can be used in
the neoadjuvant (preoperative) or adjuvant (postoperative)
setting in the treatment of malignant conditions. The neoadjuvant approach has the advantage of shrinking the tumor
and provides a more discernible margin, theoretically improving local control of the disease. In the case of chemotherapy,
before the era of modular prostheses, the neoadjuvant route
was necessary while the custom prostheses were manufactured. This technique has been shown to be as efficacious
as primary surgery. Even so, the one randomized trial of preoperative and postoperative chemotherapy versus only
postoperative chemotherapy failed to show any difference in

survival. Therefore in selected cases, it is reasonable and
may be prudent to perform surgery first.56
SURGERY
In bone tumors, resection and reconstruction are two aspects
of management that have largely complementary but occasionally conflicting goals (e.g., cryotherapy is good for extending the margins of resection of a tumor but results in
weakening of the bone). Therefore, while the goals of resection are generally quite clear (i.e., cure), the goals of reconstruction are often compromised, especially in malignant
conditions. In benign conditions, reconstruction usually
restores more function. In this section, we present a general
list of considerations that will be elaborated further in the
section on specific considerations.
584
PART III
Minimally Invasive Options

The minimally invasive option is reserved for benign conditions. It is born of two management philosophiesthe desire
to effect local control and the hesitation to cause more morbidity than the primary lesion. Whichever modality is chosen, it is
imperative that a histologic diagnosis be obtained a priori.
Resections
Amputations
Radical
Radical
Radiofrequency Ablation Radiofrequency ablation uses

high-intensity heat in proximity to a lesion, to effect thermal
necrosis. It has wide utility in the ablation of various solid tumors. In bone tumors, it has been used principally in the ablation of osteoid osteomas. This condition is a painful one,
marked by increased night pain and is promptly relieved by
the use of NSAIDs. Otherwise, it is relatively benign. It can
be found most commonly in the proximal femur. In these locations, surgical ablation in the form of a resection can incur
high morbidity. Hence, an option such as radiofrequency ablation is ideal, although it incurs a 10% to 15% recurrence
rate57,58 compared with surgery, which has a near 0% recurrence rate.59 It has limited utility in the spine because of the
indiscriminate high heat generated.
Wide
Wide
Marginal
Marginal
Intralesional
Intralesional
Injection This technique is principally used in the treatment

of unicameral bone cysts. Clinically apparent bone cysts have
a tendency to recurrent fracture and need to be treated.10
However, they have no malignant potential and have been
known to regress.4,10 There is controversy about whether corticosteroid injection is a necessary element of treatment; it has
been shown that simple decompression of a cyst is sufficient to
induce a regression.60 Rates of cure up to 50% are reported,
with a median injection rate of three and a range of one to nine
injections.61,62 Each of these sessions requires the child to be
under anesthesia. Therefore it has not been widely embraced.
As alluded to earlier, various forms of decompression have
been advocated in the literature with varying success. One approach involves the injection of bone marrow.6367 Rates of
cure of up to 50% to 70% may be achieved. However, with
this technique, repeated injections may be necessary, incurring multiple episodes of anesthesia and donor-site morbidity.
Curettage, widely regarded to be the gold standard treatment, has a recurrence rate of 5% to 50%.10 Thus there is
no clearly superior modality in the treatment of this condition.
Resection
Surgical decisions are based on the concept of compartments
in relation to a tumor (Fig. 43-6). The compartment is bound
by a barrier, which naturally limits the expansion of a tumor.
When first described, it was useful in teaching the principles
of wide resection or a resection with a margin of healthy tissue:
If a resection was performed outside a compartment, it
resulted in a margin that was free of malignant involvement.45
This idea was useful in drawing parallels to conventional cancer surgery of that time. We realize now that this theory is
flawed at many levels. For example, most osteogenic sarcomas
present with tumors that have breached the cortex, and so,
their distinction from a contained osteogenic sarcoma is
moot. In the lower limb, a tumor that has involved the rectus
femoris has involved a compartment extending from the anterior inferior iliac spine of the pelvis to the tibial tubercle.
Clearly, it would not be practical, in this setting, to perform
a hindquarter amputation. Finally, especially in the region
FIGURE 43-6 Surgical margins in relation to the compartments involved.

At left are the resections, and at right are the amputations. These classifications are largely academic, because in the strictest terms, most of the
resections, except radical resections and only wide or radical amputations,
are performed. Radical resections involve the compartment bearing the
tumor, and hence, in this case, would amount to removing the tibia
(arrows). Marginal amputations may be used in the spine and pelvis,
whereupon local adjuvants assume significant roles in disease control
(see Fig. 43-7). Intralesional amputations are obviously not therapeutic
applications in tumor surgery but are included here for completeness.
Of interest, intercalary amputations in the pediatric population can be
problematic, when the remnant stump elongates through appositional
growth. To avoid this complication, it may be necessary to use a
through-joint (e.g., through-knee) amputation.
of the linea aspera, there are numerous perforating vessels,

which penetrate the lateral intermuscular septum; clearly
these do not form a continuous barrier to tumor spread.
Still, the concept of compartmentalization is useful when
one describes the surgical procedures as intralesional, marginal, wide, and radical.41 Although not often used in the
context of amputations, the concept of compartmentalization
applies here as well. Intralesional procedures, as the name
implies, are procedures that leave macroscopic residual tissue.
A biopsy or injection of a lesion is an intralesional procedure.
A marginal procedure stops at the level of the extent of maximal expansion of a tumor. Curettage is a marginal procedure.
A wide procedure goes beyond the reactive zone of the tumor.
When first described, the reactive zone referred to the zone
of reaction around the tumor, marked by inflammatory change
(i.e., hyperemia and edema).41,47 This assessment was made
predominantly at the time of surgery. With the advent of more
sophisticated imaging modalities, it can now be demonstrated
that this zone may extend further than previously appreciated. Therefore it appears that the description of a reactive
zone is rather more abstract than real. As a general rule, resecting a tumor beyond its capsule, where vessel tortuosity and
CHAPTER 43
edema is seen, is a wide resection, and hence this appreciation,

while strongly influenced by newer imaging, remains largely
surgical. Most malignant tumors are resected widely. A radical
resection is an excision of the compartment in which a
tumor resides. An above-knee amputation for a tibial lesion
is a radical resection.
There are a number of surgical adjuvants that may be used.
This can be in the form of heat (e.g., argon beam coagulator) or
cold (e.g., liquid nitrogen cryotherapy).68,69 In addition, chemical measures may be used (e.g., phenol, polymethylmethacrylate cement).70,71 In the occasional case, specialized forms
of radiation (e.g., brachytherapy, intraoperative radiation
therapy) may be used, especially in the pelvis (Fig. 43-7).
The purpose of these surgical adjuvants is to extend the margins
of resection beyond what can be mechanically removed by
the surgeon. These improve local control of the tumor.
Benign Lesions It is useful at this juncture to recall the staging system for benign lesions. These are classified as benign,
active, and aggressive. It is evident in these entities that, even
within this group, specific nuances of the condition warrant
special considerations. In benign bony conditions, the procedures available are curettage, high-speed burring of lesion
walls, adjuvant procedures, and wide resection.68,70 It is helpful to describe these procedures from most to least aggressive.
In benign conditions, wide resection may occasionally be
used, when the involved bone is expendable (e.g., rib or terminal phalanx of the little toe) or at the end of a bone (e.g.,
distal ulna or proximal fibula). In these situations, reconstruction provides little value and can, in fact, be the source
of considerable morbidity. Additionally, it may be used in
the context of a recalcitrant recurrent benign or aggressive
lesion. Typical lesions that are resected in this manner are giant
cell tumors, aneurysmal bone cysts, or fibrous dysplasia.
Marginal excision is typified conceptually by the technique
used to excise a soft tissue lipoma. Such a procedure is not
technically feasible in most bony lesions. Osteochondromas
and periosteal chondromas may be removed in such a fashion.
BONE TUMORS
585
Intralesional procedures are more commonly performed in

benign tumors. This typically involves curettage of a lesion
with high-speed burring of the wall. In general, this is the
typical procedure for most latent or active benign bony
conditions (e.g., unicameral bone cyst). The use of heat, cold
(Fig. 43-8), or chemical modalities serves to extend this
margin of clearance further and is typically used in active or
aggressive tumors (e.g., giant cell tumor, chondroblastoma).
Malignant Lesions
The sine qua non of the resection of a malignant bone lesion is
that, at minimum, a wide resection must be performed. In certain situations, however, this may not be possible (e.g., a tumor that has expanded into the spinal canal or a tumor that
has invaded the pelvic cavity). In these instances, the outcome
tends to be suboptimal.
With newer imaging modalities, it is now often possible to
perform a physeal-sparing procedure in growing children
(Fig. 43-9). Although the physis was thought to be an effective
barrier to tumor spread, it has been shown that up to 80%
of tumors abutting the physis have, in fact, breached it.7275
Physeal-sparing procedures must therefore be carefully
balanced with the response to chemotherapy, to determine
if this is feasible.
Occasionally, a variation on this theme is to save the epiphysis, and hence the neighboring joint, by performing a distraction procedure through the growth plate. This effectively
increases the margin of normal tissue proximal to a tumor.
A resection may then be performed through this nowlengthened segment.76
Another approach to retaining a joint would be to perform
a Van Nes rotationplasty (Fig. 43-10).77 This procedure, generally undertaken for high-grade tumors near or involving the
knee, involves wide extraarticular resections, whereupon the
distal leg and foot are joined to the remaining proximal femur.
In the process, the sciatic nerve is retained, and a segmental
resection of the femoral artery with a true femoral-popliteal
arterial anastomosis is performed. The foot is rotated with
Linear accelerator
Celiotomy
with shielded
adjacent
organs
Electron
beam
applicator
FIGURE 43-7 Intraoperative radiation therapy in a 19-month-old girl

who underwent a wide resection with nodal clearance for a rhabdomyosarcoma of the pelvis.
FIGURE 43-8 Cryosurgery in a patient with chondrosarcoma. Liquid

nitrogen is poured into a funnel that directs the agent into the lesion, while
avoiding contact with the surrounding skin. The effect of freezing extends
the margins of necrosis beyond that which can be felt by the surgeon,
effectively extending the surgical margins from an intralesional or marginal
excision to a wide resection.
586
PART III
FIGURE 43-9 A, Ewing sarcoma of the tibia in an 11-year-old boy. The

lesion extended to 1 cm from the growth plate. It responded well to chemotherapy, with virtually no remaining soft tissue involvement. A physealsparing resection was done along a resection plane (double-headed arrow),
carefully performed under image intensifier guidance. B, The use of a pin
fixator, in this regard, is extremely advantageous, because it allows
stabilization of the small proximal tibial segment that precludes routine
pin fixation. The remaining gap was reconstructed with a proximal tibial
allograft (thick arrow) and vascularized fibular graft (broken arrow) harvested with a paddle of skin, which provided skin cover of the construct.
the heel pointing anteriorly. Of practical interest, the distal

segment is rotated externally, bringing the sciatic nerve and
vessels anteromedial. This should be documented in the
surgical note to facilitate further surgical procedures that
may be necessary. The ankle, therefore, functions as a knee
joint. This procedure has poor acceptance among patients because of their cosmetic abhorrence, but it is highly functional
and durable.78 A similar Winkelmann procedure may be
performed, where the proximal tibia is brought to the hip.
In children, it is remarkable to note the plasticity and remodeling of these disparate bones, which in time will accommodate each other in a stable fashion.79,80
Radical procedures and amputations have received poor
support, because they are regarded as being disfiguring. Studies have shown that patients with limb salvage procedures do
better in terms of function and cost savings.81,82 Although this
appears true at face value, in-depth analysis shows that these
studies are too heterogeneous to allow any firm conclusions.
With the aid of modern prostheses, patients with amputations
are able to achieve very high levels of activity. Furthermore,
complications are 3 to 4 times higher in limb salvage compared with limb ablative surgery. Although most series have
not shown a significant survival benefit comparing amputation and limb-sparing surgery, these studies are underpowered or include cases of amputation being used as salvage
procedures.56,83,84 The primary remaining question is
whether there is any survival and functional benefit in two-site
and stage-controlled groups with respect to amputation or
wide resection. This would require a case-controlled study
with amputation and wide resection arms, and it is a safe
assumption that this will never be performed.
Tumor
Above knee amputation
Acetabulum remodelling in
Winklemann procedure
Van Nes rotationplasty
Winklemann rotationplasty
FIGURE 43-10 A, Osteogenic sarcoma (arrow) with large soft tissue extension in an 8-year-old child. The small size of the child and high level of activity
precluded endoprosthetic reconstruction. B, A Van Nes rotationplasty was performed. C, Variants of the rotationplasty are compared with the above-knee
amputation. The bottom panel illustrates how the proximal tibia remodels and accommodates the acetabulum in the Winkelmann procedure.
CHAPTER 43
There is still a role for amputations, especially when the

tumor is in the distal extremity, adjuvant therapies are ineffective, or reconstruction is too problematic because of nerve,
vessel, or soft tissue problems.
Local recurrence in malignant lesions is a poor prognostic
factor and is associated with a 90% mortality rate. It is generally a reflection of compromised local control, although in one
study good chemotherapy response was associated with a low
local recurrence rate.83 Specifically, in this series, when intralesional procedures had been performed for osteogenic sarcoma, standard responders were 3 times as likely to get a
local recurrence as good responders. However, even among
good responders, local recurrence was 14 times more likely
if an intralesional procedure had been done rather than a wide
resection. This underscores the need both for good surgical
margins and effective chemotherapy.
Reconstruction
BONE TUMORS
587
resections, the ability to provide intercalary stability with overlying skin closure can be provided by a vascularized fibular
graft with a skin paddle. The skin paddle affords the additional advantage of monitoring the viability of the flap. Rotationplasties and their variants are remarkably functional
solutions to the problem but have poor acceptance among patients because of their appearance. Similarly, amputations are
often an instant solution to the problem, although, even here,
the occasional exception exists.82
Joint reconstruction is a challenging endeavor. Biologic
solutions include the use of bulk allograft (Fig. 43-11). They
have the advantage of becoming incorporated by the body. The
disadvantages89 are a high fracture rate of 19%, a nonunion
rate of 17%, and an infection rate of 11%. Osteoarticular allografts also become arthritic (16%) with time. Theoretically,
however, with good incorporation of the allograft, a conventional, less-constrained joint replacement can be performed
(Fig. 43-12). The endoprosthetic solution tends to be easier
In most instances, after the resection of benign lesions, small

defects result. These are easily dealt with through the use
of various gap fillers. With malignant lesions, large creative
solutions are needed. It becomes difficult to determine
which lesions are best treated by which technique because
of the relative paucity of cases and the high-risk nature
of these procedures. In this section, we will highlight the
various modalities available and the pertinent qualifiers for
each modality.
Benign Lesions Following resection of benign lesions, a
small defect usually remains. Thus the aim becomes reconstitution of bone. The modalities that have been used are bone
graft and bone graft substitutes. In general, autografts tend
to have better rates of incorporation but incur the risk of
donor-site morbidityor worse, donor-site tumor implantation. Allografts have a low risk of disease transmission and
immunologic response.85,86 Synthetic grafts tend not to incorporate as well as allografts or autografts.87,88
In the more aggressive lesions, the risk of recurrence increases. In these situations, bone substitutes could be
resorbed by the disease process and would increase the delay
before subsequent radiologic imaging is able to distinguish
between postoperative change and recurrence. In this setting,
bone cement becomes a good alternative.69,71 Furthermore,
radiopaque cement acts as a contrast agent. Recurrence at
the margin of the cemented defect can be identified readily
and treated.
Malignant Lesions The solutions that have been used to
solve the complex bone, joint, and soft tissue defects left after
tumor resections form a veritable cornucopia of techniques,
spanning all of orthopedic and plastic surgery. It is impossible
to reiterate all these solutions here. Instead, we present a list of
principal solutions pertinent to the specific reconstructive
option.
The paramount requirement of all solutions is to provide a
space filler and skin closure. Without meeting these two requirements, chemotherapy cannot resume, and the patient
will not survive. Most solutions will provide space-filling
ability if there is adequate skin for closure. If skin closure is
not possible, a local flap or vascularized pedicular graft may
be necessary. In some instances, especially with intercalary
C
FIGURE 43-11 A, Ewing sarcoma of the proximal tibia in an 11-year-old
child. B and C, This was widely resected and reconstructed with an osteoarticular tibial allograft. A gastrocnemius flap was raised to provide soft
tissue cover to the construct.
588
PART III
FIGURE 43-12 A, Resection and reconstruction of a Ewing sarcoma of

the pelvis in a boy. B, Two years later, degenerative changes developed
in the boys hip, and he required hip replacement surgery.
FIGURE 43-13 A, Osteogenic sarcoma in a 16-year-old girl. B, An endoprosthetic device was placed in the patient after resection of the lesion. C,
As a child grows, it occasionally becomes necessary to swap implants
with devices that can provide further extensibility.
but is less resilient, suffering from wear and loosening with

time.9092 With advances in technology, better designs will
lead to longer-lasting implants (Fig. 43-13). The allograft
prosthetic composite is another approach that appears to capitalize on the lasting nature of allografts and their soft tissue
capsular attachments and the simplicity of prosthetics
(Fig. 43-14). In very young children, the available endoprostheses may be too large, and this may be a relative indication for the use of bulk allografts instead (see Fig. 43-1).
Downsized pediatric implants are incapable of holding up
in adults and are destined for failure and revision
(Fig. 43-15). Prosthetic reconstruction has the distinct advantage of allowing immediate weight bearing, which is very important in patients who may have a reduced life expectancy.
In truth, the various modalities are complementary rather
than independent.
Growth is a complex issue in the management of patients
with bone resection. In the year that patients receive chemotherapy, growth is often stunted. After this, however, the child
resumes normal growth. There are various means to predict
this growth.93,94 As a rule of thumb, the distal femur grows
1 cm/year and the proximal tibia grows 7 mm/year. Girls

generally stop growing at 14 years of age and boys at 16 years.
Therefore a 10-year-old boy who has an extraarticular resection potentially would have 10 cm of growth to accommodate.
In general, a 2-cm length discrepancy is considered compensable and does not require treatment. Thus, in this example,
an additional 8-cm correction is needed.
The modalities available include contralateral epiphysiodeses. This method ablates the growth plate of the contralateral
knee. The procedure needs to be timed accurately and tends to
be practical only in the older child approaching the last few
centimeters of growth.
Bone transport is another option. This yields good results,
but the child must remain in the apparatus for long periods of
time. At an elongation rate of 1 mm/day, the child with an
8-cm defect must remain in the apparatus, at minimum, for
3 months for the elongation and a further 3 months for
consolidation of the regenerate (Fig. 43-16). This duration
is commonly doubled when distraction osteogenesis is done
during chemotherapy. Even in healthy individuals, the risk
of pin-tract infection during the procedure is greater than
FIGURE 43-15 A, Osteogenic sarcoma of the proximal femur in a 14-yearold girl. B, A wide resection and bipolar hemiarthroplasty with proximal
femoral replacement was performed. Of note, the femoral head matched
the acetabulum; so, an additional bipolar component was not added.
FIGURE 43-14 A, Osteogenic sarcoma in proximal humerus of a 16year-old boy. B, A proximal humeral
resection with allograft and prosthetic composite was used to reconstruct the defect.
FIGURE 43-16 Ewing sarcoma of the tibia. The patient underwent wide
resection and a planned bone transport procedure. The middle ring (arrow)
is secured to a segment of bone that has been osteotomized. This segment
of bone is allowed 5 days for a provisional callus to form. By progressively
advancing the ring distally at a rate of 1 mm/day, the segment of bone is
transported to fill the defect, while at the same time remaining connected
to the proximal tibia. This regenerate is weak and requires an equivalent
amount of time to consolidate. For example, an 80-mm defect would require 5 days to form a provisional callus, 80 days to lengthen, and 80 days
to consolidate before removal of the frame. This ungainly device needs to
be tolerated by the patient for the duration of the limb-lengthening
procedure.
590
PART III
FIGURE 43-17 A, A patient presented with osteogenic sarcoma of the

proximal humerus that was resected and reconstructed with a vascularized fibular graft shoulder arthrodesis at 6 years of age. He developed a
shortened humerus at maturity, which was lengthened. B, After lengthening, the regenerate was protected with a plate and hypertrophied
with time.
90%.95 In the patient with malignant disease who is to receive

chemotherapy, this would be an important consideration.96
In addition, the regenerate tends to be weak and is prone to
fracture (Fig. 43-17). Patients on chemotherapy are prone
to osteoporosis and are already at risk for fracture.
The extensible prosthesis is a marvel of modern science
that is presently undergoing teething issues.9799 The manual expansion designs require repeated surgical procedures
to periodically lengthen the limb to keep pace with normal
growth (see Fig. 43-13, C). The Stanmore implants (Stanmore Implants Worldwide, Elstree, United Kingdom) have
been used for nearly 20 years and have a 23% revision
rate.91 Survivorship analysis, however, shows a near-zero
survivorship at 10 years.100 Self-extending designs work
through electromagnetic couplers or heating coils that allow
motors or heat-release springs to extend the implant. The
Phenix device (Phenix Medical, Paris, France) is presently
undergoing evaluation in the United States.101 Preliminary
results show a complication rate of up to 44%, necessitating
revision. The Repiphysis system (Wright Medical Technology, Inc., Arlington, TN) uses an external electromagnetic
field to provide controlled released of a spring held in place
by a locking mechanism. This device is associated with an
implant revision rate of 44%.102 In general, the stems in
these devices are too narrow and mechanically insufficient,
and fixation techniques remain inadequate. Thus all these
designs have poor longevity but reduce immediate surgical
complications (e.g., infection). They are well tolerated by
patients and families.
There are many solutions to the problem of limb reconstruction in the skeletally immature child, but none is perfect.
Therefore it is apparent that the surgeon dealing with potential
limb length inequality after tumor resection and subsequent
growth must be able to perform, or at least facilitate, the reconstructive procedures previously discussed. Any one of these
procedures is applicable to an individual case, and they
remain complementary to each other.
expertconsult.com.
The development of immunohistochemical staining techniques allows pediatric tumors to be classified by histology.
Tumors can arise from any of the cell types of the central
nervous system. The brain is composed of neurons and glial
cells. Glial cells far outnumber the neurons, and provide a
nourishing and supportive role. The three main types of glial
cells are astrocytes, oligodendrocytes, and ependymal cells,
and the neoplasms they give rise to are gliomas. More specifically, they form astrocytomas, oligodendrogliomas, and ependymomas, respectively. Tumors involving both neuronal and
glial cells are called ganglion cell tumors and consist of gangliogliomas, desmoplastic infantile gangliogliomas, and gangliocytomas. Another mixed neuronal and glial tumor is a
dysembryoplastic neuroepithelial tumor (DNET). Finally,
there are embryonal tumors, which include medulloblastoma,
primitive neuroectodermal tumors (PNETs), medulloepithelioma, neuroblastomas, melanotic neuroectodermal tumors
in infancy, and atypical teratoid/ rhabdoid tumors (ATRTs).4
Other primary brain tumors include germ cell tumors, choroid plexus tumors, craniopharyngiomas, and meningiomas.
CHAPTER 44
Brain Tumors
Eamon J. McLaughlin, Michael J. Fisher,
Leslie N. Sutton, and Phillip B. Storm
With the exception of trauma, neoplasms are the most common

cause of death in children less than 19 years of age. Tumors of the
central nervous system are the most common solid neoplasms
found in the pediatric population, accounting for 20% of cancer
deaths, and are second only to leukemia in overall cancer frequency.1,2 Approximately 4030 brain tumors are diagnosed
each year in the United States, for an overall incidence of 4.71
cases per 100,000 person-years. Of these cases, it is estimated
that 2880 will occur in children less than the age 15 years.1,3
The important factors in diagnosing brain tumors are location, age, and cell type. Location is probably the most important factor radiographically, followed by the age of the patient.
The brain is divided into two compartments by the tentorium.
Above the tentorium (supratentorial) are the cerebral hemispheres, basal ganglia, and the thalamus. Below the tentorium
(infratentorial) are the pineal gland, the tectum, the pons, the
medulla, and the cerebellum. Adult brain tumors tend to be
supratentorial; however, pediatric tumors are evenly split
between supratentorial and infratentorial. This division of
location in the pediatric population is dependent on the age
of the patient. In children younger than 2 years of age, the
tumors are predominantly supratentorial, whereas children
between the ages of 3 and 15 years more often have infratentorial tumors (Table 44-1).1 The prognosis is usually poor in
children with brain tumors younger than the age of 1 year,
with choroid plexus papilloma being the main exception.1,4
Clinical Features
------------------------------------------------------------------------------------------------------------------------------------------------
The signs and symptoms of brain tumors in children vary considerably based on tumor type, location, and age of the patient. In the absence of a seizure or a focal neurologic
deficit (e.g., diplopia caused by sixth nerve paresis), the vast
majority of the symptoms are nonspecific and easily attributable to many more common and less serious causes. Common
symptoms may include headache, nausea, vomiting, lethargy,
subtle changes in personality, and worsening school performance. This constellation of symptoms can often be attributed
to gastrointestinal problems, depression, school anxiety, migraines, sinusitis, or the need for prescription eyeglasses. Even
a long-standing seizure disorder may ultimately be diagnosed
as a supratentorial brain tumor. Infants typically present with
failure to thrive, decreased intake, macrocephaly, or lethargy.
Because of the nonspecific nature of these symptoms, it is
common for a patient to present for neurologic evaluation
after having visited numerous other specialists without establishing a diagnosis.
Most pediatric patients with brain tumors are between
the ages of 2 and 14 years and typically present with a few days
to weeks of headache, nausea/vomiting, gait ataxia, and/or
diplopia. This constellation of symptoms is caused by hydrocephalus resulting from obstruction of the ventricles by tumor,
commonly located in the midline posterior fossa. Headaches
are common in children with viral infections, whereas
frequent, daily morning headaches should raise the clinical
suspicion of an intracranial mass lesion. This is especially true
in the absence of a fever or other viral sequelae. Patients with
elevated intracranial pressure often have an exacerbation of
their symptoms in the morning. Both lying in the recumbent
position overnight and sleep-induced hypoventilation (which
leads to an increase in Pco2) cause an increase in intracranial
pressure. Elevated intracranial pressure can also cause the cerebellar tonsils to herniate into the foramen magnum and result
in occipital headaches and neck pain.
There are two instances in which tumors cause nausea and
vomiting. One is the elevation of intracranial pressure, and the
other is direct irritation/infiltration of the vomiting center. The
591
592
PART III
TABLE 44-1
Brain Tumors in Children
Age
Tumor Histology
0 to 2 years
Teratoma
Primitive neuroectodermal tumor
Astrocytoma (high grade)
Choroid plexus papilloma
Supratentorial tumors (50%)
Astrocytoma (low grade)
Craniopharyngioma
Hypothalamic glioma
Primitive neuroectodermal tumor
Ependymoma
Choroid plexus papilloma
Infratentorial (50%)
Primitive neuroectodermal tumor: medulloblastoma
Cerebellar astrocytoma
Ependymoma
Brainstem glioma
2 to 15 years
vomiting center (area postrema) is located on the floor of

the fourth ventricle and is vulnerable to compression from
large posterior fossa tumors or from direct invasion of intrinsic
brainstem tumors. Given that an intrinsic tumor in the medulla
can cause vomiting in the absence of other neurologic symptoms, persistent vomiting should raise the possibility of a
posterior fossa tumor, which could be confirmed through a
detailed history and neurologic examination. Ataxia is commonly associated with tumors in the cerebellum and is often
described by the parents as clumsiness, walking like he is
drunk, walking with the head tilted to one side, or falling to
one side.
The visual complaints associated with posterior fossa tumors are frequently diplopia, difficulty looking up (sunsetting
eye or Parinaud syndrome), and occasionally decreased visual
acuity. As mentioned before, these symptoms are a result of the
hydrocephalus. A decrease in visual acuity can result from
papilledema. Loss of vision is a more common symptom of
supratentorial tumors, because of optic nerve atrophy from direct compression. Patients with posterior fossa tumors are
usually diagnosed with magnetic resonance imaging (MRI),
because their other symptoms occur long before any visual
defects. Therefore lack of visual signs and symptoms does
not exclude a brain tumor. However, patients with poor access
to health care can present with posterior fossa tumors and
accompanying visual deficits.
Supratentorial tumors are especially common in patients
younger than 2 years of age. These children often present with
a failure to thrive, hemiparesis, seizures, or a full bulging anterior fontanelle and a rapid increase in head circumference.5,6
At more than 2 years of age, supratentorial tumors present
similarly in both children and adults, most commonly with
headaches and/or seizures. When a patient presents with sudden onset of severe headaches or a rapid decline in mental status, it usually indicates a hemorrhage into their lesion. Rarely,
obstructive hydrocephalus can cause such a rapid decline in
mental status, but this is unlikely because of the slow growth
rate of most tumors.
Less commonly, brain tumors can present with endocrine
abnormalities. These can include weight gain or loss, diabetes
insipidus, short stature, truncal obesity, galactorrhea, and precocious or delayed puberty. These symptoms result from tumors affecting the hypothalamic-pituitary axis. Because of
the proximity of these tumors to the optic nerves and chiasm,
they often cause decreased vision and visual field deficits.
Radiographic Evaluation
------------------------------------------------------------------------------------------------------------------------------------------------
Patients suspected of having a brain tumor should be evaluated with an MRI with and without gadolinium. Although
MRI is the gold standard for evaluating tumors, many patients
presenting in the emergency department with progressive
clinical signs and symptoms of a brain tumor are evaluated
with a head computed tomography (CT) without instillation
of a contrast medium. CT is the ideal imaging modality to
use during emergent situations for a number of reasons. CT
is excellent in evaluating hydrocephalus and hemorrhage,
the two main causes of rapid neurological decline. Furthermore, CT can be performed in minutes, frequently does not
require sedation, gives excellent detail and information, and
is considerably less expensive. If the patients condition is rapidly deteriorating, a contrast agentenhanced head CT is occasionally performed to better characterize the lesion for the
radiologist and neurosurgeon when the patient requires emergent surgical intervention. If the patients condition is stable,
the contrast agent may be omitted, and MRI with and without
gadolinium should be performed, the timing of which is dictated by the clinical signs and symptoms.
Magnetic resonance imaging provides much better resolution of the brain and provides images in the sagittal, axial, and
coronal planes. Standard MR imaging combined with newer
imaging sequences and spectroscopy can even point to a
specific histologic diagnosis.7 Furthermore, it is difficult to
evaluate the lower brainstem with CT, because of the bony
artifact from the skull base. One limitation of MRI is that it
does not show intratumoral calcifications very well, and occasionally, patients require both studies to aid in establishing the
proper diagnosis.
Magnetic resonance imaging with and without gadolinium
can provide significantly more information about the patients
tumor. The bloodbrain barrier is made up of tight junctions
in the endothelial cells lining the capillaries in the brain,
which prevent most blood contents from entering the brain,
including gadolinium. However, certain brain tumors cause
breakdown of the bloodbrain barrier and permit the gadolinium to enter the tumor and then enhance the tissues (appear
bright on T1-weighted images). In general, in the adult population, enhancement in an intra-axial lesion means a more
aggressive brain tumor and a poorer prognosis. This is not
as consistent in pediatric tumors. There are many enhancing
pediatric brain tumors that are not aggressive and are curable
with total resection.
When viewing an MRI, the important factors to consider
are (1) the location of the tumor (e.g., supratentorial, infratentorial, pineal region, suprasellar), (2) whether the tumor is
intra-axial (within the brain tissue) or extra-axial (outside
the brain tissue), (3) the age of the patient, (4) whether the
tumor enhances, and (5) if there are single or multiple lesions.
By systematically assessing the scans and considering these
factors, the differential diagnosis can be narrowed considerably, which can be extremely helpful in preoperative planning.
CHAPTER 44
If there are multiple lesions in the brain, or the location

and enhancement suggest a tumor type associated with leptomeningeal metastases or drop mets to the spine, then a
spinal MRI with and without gadolinium is performed. It
is preferable to obtain the spinal MRI preoperatively, but this
is often dictated by the patients clinical examination. Postoperatively, brain tumor patients should have an MRI within
36 to 48 hours to evaluate the extent of the resection and rule
out hydrocephalus, bleeding, or ischemia. The timing is
important, because after 36 to 48 hours, expected postoperative changes can enhance and make it difficult to distinguish
scarring from residual tumor. If the patient did not get a
preoperative MRI evaluation of the spine and the histologic
diagnosis is consistent with tumors that can metastasize to
the spine, then the study should be done 2 weeks after
surgery, because postoperative debris and blood can be
mistaken for metastatic disease
Surgical Intervention
------------------------------------------------------------------------------------------------------------------------------------------------
The goal of a surgical intervention for brain tumors is to safely

debulk as much tumor as possible, to obtain a histologic diagnosis, to reestablish normal cerebrospinal fluid (CSF) pathways, or to divert CSF. The location of the tumor is often the
determining factor as to how aggressively the tumor is
debulked. In fact, some tumors, because of their location
and their ability to be diagnosed with MRI, are not biopsied.
For example, an intrinsic pontine glioma, which is an astrocytoma of the brainstem, cannot be debulked safely and has a
characteristic appearance on MRI. Therefore these patients
are referred to a neuro-oncologist for management without
a tissue diagnosis. Pineal region tumors are another example
of a lesion that may be diagnosed without surgical intervention. Patients with pineal region masses should have serum
b-human chorionic gonadotropin (b-HCG) and alpha fetoprotein (AFP) levels obtained. If these are negative, then
CSF markers are needed. If the serum or CSF markers are
positive, then a diagnosis of a germ cell tumor can be made
without the need for a biopsy.
However, most tumors require surgical intervention, consisting of either a stereotactic biopsy or an open craniotomy
to obtain tissue for a definitive diagnosis. The most important tool for preoperative planning is MRI. Diffuse intrinsic
tumors of the thalamus or basal ganglia typically undergo
stereotactic biopsy. This procedure involves rigidly fixing
an MRI-compatible frame to the patients skull. The patient
then has an MRI, and the x, y, and z coordinates are determined. These coordinates are then used to position the
frame and the arc so that the tip of the needle is exactly
where these three points intersect in the brain. Given the
improvements in frameless stereotaxy, all but the smallest lesions can be biopsied without a rigid frame.8 The advantages
of a stereotactic biopsy include a short procedure time, the
possibility of diagnosis in areas of the brain that carry an
unacceptable morbidity and mortality with an open craniotomy, and the patient is discharged on postoperative day 1.
The disadvantages are that only a small amount of tissue
is obtained, which may be nondiagnostic or result in the
wrong diagnosis, and if bleeding occurs it is difficult to treat,
or it may not be recognized until the patient deteriorates
neurologically after the procedure. Lastly, if the diagnosis
BRAIN TUMORS
593
cannot be made with a stereotactic biopsy or the diagnosis

requires aggressive debulking, the patient will require a second operative procedure.
Because of the fact that the prognosis of many pediatric
tumors is strongly influenced by the amount of postsurgical
residual tumor,9 the majority are approached with a craniotomy/craniectomy for open biopsy, with an attempt at maximal
microsurgical tumor resection. Cerebral hemispheric tumors
are approached through a craniotomy. Preoperative planning
consists of an MRI coupled with a frameless stereotactic navigation study. The navigation study allows the neurosurgeon to
view the tumor in the operating room in the sagittal, axial, and
coronal planes and can be used to find the tumor and plan the
incision and approach. However, the main limitation of this
technology is that it is not a real-time study, and actions such
as retracting the brain or draining cysts or CSF spaces may
cause the brain to shift position, thus compromising the accuracy of the intraoperative navigation system. When this
occurs, intraoperative ultrasonography is extremely helpful
in localizing lesions.
Intraoperative MRI aims to correct the limitations of the
navigation system by providing a real-time image. Previous
intraoperative MRIs were limited because of poor resolution;
however, newer intraoperative suites have 3-tesla magnets and
provide excellent resolution. The drawbacks of the intraoperative MRI suites are that they are prohibitively expensive for
many institutions, are helpful in only a small number of procedures, and significantly extend the time of the procedure.
Nevertheless, this is exciting technology, and as the expense
decreases and the efficiency improves, it will be an invaluable
tool to surgeons operating upon brain tumors. Functional MRI
(fMRI) techniques can localize speech and motor cortex.
When tumors involve these areas of eloquent cortex, fMRI
can aid in selecting the safest site to incise the cortex.10 In
the pediatric population, fMRI can prove challenging, because
it requires a cooperative non-sedated patient. Electrophysiologic recording and stimulation are sometimes helpful in locating the motor strip. Recently, magnetoencephalography
(MEG) is being used to help localize motor, sensory, and language cortex for both tumor surgery and epilepsy surgery.11
Such advances undoubtedly aid the neurosurgeon throughout the surgical procedure; however, there is still no substitute
for an outstanding understanding of the three-dimensional
anatomy of the brain. When choosing an approach, anatomic
planes, such as the interhemispheric fissure, the sylvian fissure,
and the cranial base are used, if possible, to avoid resecting normal brain. If there is no plane available, the approach is usually
through the least amount of tissue, while avoiding areas of
eloquent language, motor, and visual cortex.
Tumors of the midline (hypothalamus, thalamus, basal
ganglia, and brainstem) were once considered inoperable.
However, advances in microsurgical techniques and innovative instrumentation now make these tumors approachable.
At the same time, advances in chemotherapy and single-dose
and fractionated radiosurgery offer alternatives, and it is currently unclear which strategy or combination of strategies is
best for a particular tumor. Advances in surgical techniques
now allow for multiple options for the approach to tumors.
For example, pineal region tumors may be approached
through a posterior fossa route (retracting the cerebellum from
the underside of the tentorium), by a supratentorial route
between the hemispheres and through the posterior corpus
594
PART III
callosum, or through the tentorium itself. The relationship of

the pineal tumor to the tentorium dictates the approach.
Tumors of the cerebellum and the lower brainstem are
approached through a posterior fossa craniotomy or craniectomy.12 Midline tumors of the fourth ventricle usually present
with obstructive hydrocephalus. Some neurosurgeons prefer to
place a shunt before tumor resection; however, most now favor
giving the child corticosteroids and placing a ventriculostomy at
the time of the craniectomy. The ventriculostomy is either removed or converted to a shunt if needed in the postoperative
period. Between 20% and 40% of children will ultimately
require a shunt.13 Many neurosurgeons are performing an endoscopic third ventriculostomy (ETV) at the time of the resection. This procedure involves inserting an endoscope into the
lateral ventricle, passing it through the foramen of Monro
and making a small hole in the floor of the third ventricle. This
allows the CSF to bypass the distal obstruction and enter
directly into the cisternal system.14 One series of patients with
posterior fossa tumors showed a reduction in the postoperative
shunt rate of 26.8% to 6% in patients treated with EVT and
tumor removal versus tumor removal alone.15
To access the fourth ventricle, the patient is placed in the
prone position, and the bone overlying the cerebellum is
removed, occasionally including the posterior ring of the C1
vertebrae. After opening the dura, the cerebellar vermis is
vertically incised, providing access to the tumor and the fourth
ventricle. The tumor is removed with bipolar cautery, suction,
or an ultrasonic aspirator. Laterally placed tumors of the
cerebellopontine angle are reached by retracting the cerebellum medially. Electrophysiologic monitoring of cranial nerves
V, VII, VIII, IX, X, XI, and XII is often required throughout this
approach. Tumors of the brainstem may be debulked, if they
are dorsally exophytic. The dura is closed and covered with
DuraGen (Integra LifeSciences, Plainsboro, NJ), a collagen
product that augments dura integrity. Replacement of the bone
is not required, but we prefer to whenever possible. Postoperative problems include acute hydrocephalus, pseudomeningoceles, aseptic meningitis, mutism, pseudobulbar palsy,16
cranial nerve or brainstem dysfunction, and gastrointestinal
hemorrhage.17 Patients with swallowing dysfunction and aspiration may require tracheostomy and feeding gastrostomy.
Tumor Types
------------------------------------------------------------------------------------------------------------------------------------------------
CEREBELLAR ASTROCYTOMAS
These tumors are usually low-grade and curable with total surgical resection. The average age at presentation is 9 years, and
the patient normally presents with pernicious vomiting, intermittent morning headaches, and disturbances of balance, usually spanning a period of months. The classical CT appearance
of these tumors is a hypodense, cystic cerebellar mass (usually
around the vermis) with a brilliantly enhancing mural nodule.18 However, about one fourth will be entirely solid tumors. MRI is helpful in defining the surgical anatomy, such
as the relationship of the tumor to the brainstem, and the
nature of the cyst wall. Cerebellar astrocytomas are typically
of low signal intensity on T1-weighted MRI sequences, demonstrate increased intensity on T2-weighted sequences, and
show enhancement of the solid component with gadolinium
(Fig. 44-1). Because of their location and size, they cause
effacement of the fourth ventricle, resulting in obstructive

hydrocephalus.
Histologically, they consist of benign-appearing astrocytes.19 Subtypes are the juvenile pilocytic form (80% to
85%) and the fibrillary form.4 Detailed examination may
reveal cellular pleomorphism and tumor extension to the
subarachnoid space, but these tumors rarely disseminate.
High-grade astrocytomas in this location are rare and usually
follow radiation therapy given for a previous low-grade
tumor.20
Treatment for cerebellar astrocytomas is complete surgical
resection. In tumors with no brainstem involvement, this can
be accomplished in a high percentage of cases. If complete surgical excision can be demonstrated radiographically, these tumors rarely recur, and no adjuvant therapy is indicated.21
Therefore if there is residual tumor on the postoperative scan,
reoperation for total excision is recommended. Radiation therapy can be considered for multiple recurrent lesions or in
cases in which brainstem involvement precludes complete removal. However, even in these cases, residual tumor may
remain indolent for years without additional therapy. Regular
postoperative surveillance scanning is appropriate, especially
when there is suspicion for residual tumor. Recurrence is treated with reoperation if this is feasible.
PRIMITIVE NEUROECTODERMAL TUMOR

AND MEDULLOBLASTOMA
Primitive neuroectodermal tumor and medulloblastoma are
related tumors; and, in fact, the term medulloblastoma and
posterior fossa PNET are often used interchangeably. Medulloblastoma is the most common malignant brain tumor of
childhood. Histologically, the classical medulloblastoma is
composed of densely packed cells with hyperchromatic nuclei
and little cytoplasm, giving the histologic slides a blue color
when stained with hematoxylin and eosin. Tumors with identical histology can occur in the cerebral hemispheres and are
termed supratentorial PNETs. Children with medulloblastoma
typically present with headache, vomiting, and lethargy of
relatively short duration, and the mean age (3 to 4 years) is
typically younger than that seen with cerebellar astrocytomas.
Infants typically present with failure to thrive. Supratentorial
PNETs present with increased intracranial pressure and focal
neurologic deficits, depending on the location of the tumor.
On a CT scan, medulloblastomas typically appear as wellmarginated homogeneously dense masses filling the fourth
ventricle, causing obstructive hydrocephalus. They usually
enhance brilliantly with contrast. However, unlike ependymomas, they lack calcifications. On MRI, they can show variable
signal characteristics. The images are often slightly hypointense on T1 weighting, becoming brighter on fluid-attenuated
inversion recovery (FLAIR) sequences, and may be bright or
dark on T2-weighted studies. They usually enhance on MRI
(Fig. 44-2) and show restricted diffusion on diffusionweighted imaging (DWI). MRI of the spine is indicated
2 weeks postoperatively to evaluate for spinal metastases
(drop mets; Fig. 44-3).22
Treatment begins with biopsy and surgical excision. Medulloblastoma and PNET tumors are not curable with surgery
alone; and in cases with metastases at diagnosis or extensive
brainstem involvement, the major mass should be debulked,
CHAPTER 44
BRAIN TUMORS
595
C
FIGURE 44-1 A, Axial T1WI postgadolinium image of a cerebellar pilocytic astrocytoma, in a 3-year-old boy, showing a large cyst (white asterisk)
and enhancing mural nodule (white arrowhead). B, Axial T2WI image showing markedly dilated lateral ventricles and transependymal flow of cerebral spinal
fluid (CSF) out of the ventricles into the surrounding brain parenchyma (black arrows). The obstructive hydrocephalus is a result of the cerebellar
astrocytoma. C, Sagittal T2WI postoperative image showing resection of tumor and flow through the floor of the third ventricle (white arrow) after the
endoscopic third ventriculostomy done at the time of tumor resection.
FIGURE 44-2 A, Axial T1WI postgadolinium image, in an 8-year-old boy, showing an enhancing primitive neuroectodermal tumor (PNET) arising from the
roof of the fourth ventricle and involving the cerebellar vermis (white arrow). B, Axial T1WI postoperative image showing resection of tumor and partial
splitting of the vermis (white arrowhead). The patient suffered severe postoperative mutism.
596
PART III
of cure for recurrent tumors are low, high-dose chemotherapy

and stem cell rescue can salvage some patients at relapse.46,47
Late sequelae of therapy include pituitary dysfunction, hearing
loss, growth delay, cardiomyopathy,48 cognitive delay,49
psychosocial adjustment and family problems, and radiationinduced meningiomas, astrocytomas, and sarcomas.50
EPENDYMOMAS
FIGURE 44-3 Sagittal T1WI postgadolinium image of a 4-year-old with

metastatic primitive neuroectodermal tumor (PNET) to the spine (white
arrows) from her fourth ventricular tumor. The drop mets were present
at the time of her diagnosis.
but no attempt should be made to resect tumor in vital areas.23

After the operation, radiation therapy is usually administered
to the entire brain and spinal canal, with a boost to the tumor
bed. Younger children (less than 9 years old) suffer significant,
global cognitive problems as a result of whole-brain radiation
in an age- and dose-dependent fashion.24 They are chemotherapy sensitive, and various chemotherapy combinations
have been used to improve outcomes and allow for a reduction
in craniospinal radiation dose.2528 Chemotherapy alone has
been shown to have some success in treating these tumors and
can be used in the treatment of infants (for whom craniospinal
radiation is contraindicated); however, the long-term survival
of a chemotherapy-only approach is not as good as combined
modality treatment.2932 In determining the best treatment,
staging criteria are important to define risk groups. In the past,
the Chang system was used, which incorporated the surgeons
estimate of the tumor size at operation and the extent of metastatic disease based on postoperative imaging.33 In most
centers today, patients are assigned to a high-risk group based
on younger age (<3 years old), supratentorial tumor location,
postoperative residual disease greater than a volume of 1.5
cubic centimeters, or presence of disseminated disease.28,34
Molecular markers have been identified that have prognostic
significance, but are not yet being used to dictate therapy.3539
The rate of progression-free survival at 5 years ranges from
more than 80% in groups with standard-risk factors40,41 to
less than 70% in high-risk groups.28,4244 Infants treated with
chemotherapy alone historically have progression-free survival rates in the range of 20% to 40%,2931 although recent
studies suggest that intensification of therapy may improve
survival rates.32,45
Patients require long-term supportive care, preferably in the
setting of a multidisciplinary pediatric neuro-oncology clinic.
Surveillance scanning is standard practice, and although rates
Ependymomas occur in the region of the fourth ventricle or

cerebellopontine angle, spinal cord, or supratentorial compartment. Most are histologically benign, but despite this, they
have a tendency to recur in the local tumor bed and disseminate throughout the neuraxis. The median age at diagnosis is
between 3 and 5 years, although tumors in infants and adults
are not uncommon.51 Tumors typically arise in the posterior
fossa (60% of cases), and symptoms are similar to those of
other tumors in this region. Cranial nerve and brainstem involvement can occur. Vomiting may arise without hydrocephalus, which suggests infiltration of the region of obex, which is
characteristic of ependymomas. When the tumors do arise in
the supratentorial compartment in children, they are often extremely large, and despite their presumed ependymal origin,
may demonstrate no connection with the ventricle.
Computed tomography typically shows an isodense mass
with flecks of calcification and an inhomogeneous pattern
of enhancement. Posterior fossa lesions may extend through
the foramina of Luschka into the cerebellopontine angle
(Fig. 44-4). On MRI, ependymomas are usually isointense
to hypointense on T1-weighted images, hyperintense on
T2/FLAIR images, do NOT show restricted diffusion on DWI,
and often enhance inhomogeneously with gadolinium.52
FIGURE 44-4 Axial T2WI image of a fourth ventricular ependymoma, in a

5-year-old boy, growing out of the foramen of Luschka into the cerebellopontine angle (white arrows).
CHAPTER 44
Treatment for ependymomas primarily consists of surgery

and radiation. Prognosis is highly dependent on the extent of
surgical resection as determined by postoperative imaging.
The 5-year progression-free survival after complete resection
is 60% to 80%, compared with less than 30% after incomplete
resection.53 However, radical surgical resection may result
in permanent neurologic damage and may not be possible
in some cases. Unless the tumor has disseminated at diagnosis,
postoperative radiation is confined to the operative bed. Trials
of radiosurgery for unresectable tumors are ongoing at several
centers. Ependymomas are now being treated with proton
beam therapy because of the decreased radiation exposure
to the adjacent, normal uninvolved structures. Adjuvant chemotherapy has minimal impact on survival54; however, several chemotherapy agents have activity in this tumor,30,55
and chemotherapy is being evaluated in a neoadjuvant setting
to see whether giving chemotherapy after a subtotal resection
may shrink the tumor in such a way that a complete resection
can be achieved at a second surgery.
BRAIN TUMORS
597
presenting in the cisterna magna. They are often amenable

to aggressive surgical resection, and if the histology is benign,
adjuvant radiation therapy is usually deferred.
Dorsally exophytic brainstem tumors arise from the floor of
the fourth ventricle and present with symptoms of hydrocephalus. These tend to be pilocytic astrocytomas.58 Treatment is
primarily surgical. Gross total resection is difficult to achieve
without unacceptable neurologic risk; however, most patients
remain progression-free after resection because of the indolent
nature of the tumor. Radiotherapy is reserved for recurrence or
progression.59
Tectal gliomas are now recognized to be a not infrequent
cause of hydrocephalus.60 They typically present with symptoms referable to ventricular obstruction and are usually treated with either a ventriculoperitoneal shunt or endoscopic
third ventriculostomy. Biopsy is not required. They are usually
extremely indolent, and treatment of the tumor itself is
required only if it progressively enlarges.
BRAINSTEM GLIOMAS
HYPOTHALAMIC/CHIASMATIC
ASTROCYTOMAS
It is now recognized that there are several types of brainstem

gliomas, each associated with very different outcomes.56 The
most common variety is the diffuse intrinsic brainstem glioma,
which is not amenable to surgical resection. These tumors are
often centered in the pons and typically present with cranial
neuropathies rather than hydrocephalus. Patients tend to be
less than the age of 4 years, with sixth nerve palsies, facial
weakness, and ataxia. The diagnosis is established by MRI,
which shows a swollen pons with diffuse signal abnormalities
(Fig. 44-5). Surgery is not indicated. Radiation therapy can
provide symptomatic relief and prolong survival, but most
children die within a year.57 Chemotherapy has not been
shown to be effective.
Cervicomedullary astrocytomas are considered to be rostral
extensions of intrinsic spinal cord tumors and carry a better
prognosis. Signs and symptoms may include vomiting, torticollis, and slowly evolving motor weakness. MRI shows an
enlarged upper cervical spinal cord, with a rostral extension
Suprasellar astrocytomas are usually low-grade neoplasms,

which may occur in association with neurofibromatosis type
1 or as isolated tumors. The etiology of these tumors is not
well described, but the association with neurofibromatosis
type 1, which is localized to chromosome 17q, suggests a molecular genetic basis. They may present primarily with vision
abnormalities (visual field cuts, asymmetric loss of visual
acuity in association with optic atrophy, or nystagmus) or as
hypothalamic dysfunction (precocious puberty, diabetes insipidus, other endocrine dysfunction, growth failure, obesity, or
diencephalic syndrome, which consists of failure to thrive and
vomiting). Often both visual and hypothalamic complaints
coexist.61
Imaging studies usually cannot distinguish hypothalamic
tumors from those arising from the visual apparatus. The
tumors typically do not calcify, which helps distinguish them
from craniopharyngiomas, and appear as solid hypodense
lesions on CT or T1-weighted MRI sequences and enhance
FIGURE 44-5 A, Sagittal T2WI image in a 4-year-old girl showing an infiltrative, hyperintense tumor in the pons (white arrow). B, Sagittal T1WI postgadolinium image showing that the tumor does not enhance (white arrowhead). This tumor is an intrinsic pontine glioma, and the diagnosis is made by MRI
alone, a biopsy is not required.
598
PART III
C
FIGURE 44-6 A, Axial T2WI image in a 3-month-old girl who has a hyperintense lesion (black asterisk) in the left optic nerve and into the optic chiasm.
B, Axial T1WI postgadolinium image showing an enhancing tumor (white arrowhead). The tumor causes stretch on the internal carotid artery and middle
cerebral artery, putting the patient at risk of a postoperative stroke when this chiasmatic/hypothalamic glioma is resected. C, Axial T2WI postoperative image
showing a large left internal carotid stroke (white arrows) that occurred on postoperative day 4 and was a result of vasospasm in the stretched arteries.
after administering contrast. Extension to the intraorbital optic nerves or along the optic radiations is diagnostic and rules
out craniopharyngiomas, germinomas, or other tumors in this
location (Fig. 44-6).
Because most of these tumors will not progress significantly
(especially in the setting of neurofibromatosis type 1), initial
management is usually observation with serial imaging and
ophthalmologic screening. Tumors that progress significantly
and/or cause worsening vision are treated with chemotherapy.
The most common regimen used is vincristine and carboplatin62,63; however, thioguanine, procarbazine, lomustine
(CCNU), and vincristine (TPCV) are also effective.64 Radiotherapy is avoided if possible because of the high risk of
secondary effects, such as endocrine dysfunction, stroke,
secondary malignant neoplasms, and neurocognitive deficits.
Although radical surgical resection results in prolonged
disease stability in the majority of patients, it carries a higher
risk of stroke and injury to the optic pathways and the hypothalamic/pituitary axis.65,66 Surgery is reserved for cases when
the diagnosis is unclear, there is a unilateral optic nerve tumor
with severe visual impairment or painful proptosis, and
tumors are causing obstruction of the third ventricle or exerting mass effect on surrounding areas of the brain.
CRANIOPHARYNGIOMA
Craniopharyngiomas are histologically benign masses believed to arise from embryonic rests derived from the hypophyseal-pharyngeal duct. Symptoms result from optic
chiasm or nerve compression, hypopituitarism, hypothalamic
dysfunction, or increased intracranial pressure in association
with hydrocephalus.67 They also occur in adults, but the
childhood form represents a distinct entity characterized by
large size and extensive calcification. There are two varieties
of craniopharyngiomas, the papillary and adamantinomatous
types, the latter being the most common in the pediatric population. Histologically, they typically are composed of a
squamous epithelial cyst wall, with cystic fluid composed of
cholesterol crystals, and calcifications. They tend to be inseparable from the pituitary stalk and may have an interdigitating
gliotic interface with the hypothalamus above. This makes
complete surgical removal challenging, because small rests
of tumor may reside in the brain. This is also the reason for
hypothalamic dysfunction that may be seen after surgical
excision.68
Computed tomography can reveal either a cystic mass
with basal calcifications or an entirely solid tumor. MRI shows
CHAPTER 44
BRAIN TUMORS
599
employed with cortical tumors, but simple removal of the tumor usually provides good seizure control, and the value of
these strategies is uncertain.9,80
The outcome of low-grade astrocytomas,9 gangliogliomas,81(Fig. 44-8), and DNETs (Fig. 44-9) that are completely
resected is favorable, although surveillance scanning is
FIGURE 44-7 Sagittal T1WI postgadolinium image, in a 6-year-old boy,

showing a sellar/suprasellar craniopharyngioma growing down into the
sella turcica and up into the third ventricle (white arrow).
the sagittal anatomy well, but may miss the calcifications69

(Fig. 44-7). In some instances, imaging cannot distinguish a
craniopharyngioma from a hypothalamic glioma.
Controversy persists regarding the best treatment approach
for patients with this tumor. Gross total resection and subtotal
resection with adjuvant radiation have similar local control
rates.7072 Both are associated with potential post-treatment
problems, including panhypopituitarism, diabetes insipidus,
obesity, visual problems, stroke, behavioral difficulties, poor
school performance, and pseudoaneurysms of the carotid
artery.7375 Although aggressive resection of very large craniopharyngiomas is often associated with more significant
post-treatment complications, gross total resection of smaller
tumors in high-volume centers are more likely to be achieved
safely. Long-term survival is in the range of 90% at 10 years,
but local recurrences are not uncommon.76 Recurrences are
treated by reoperation,77 instillation of colloidal 32P into cysts,
or radiosurgery.
FIGURE 44-8 Axial T2WI image, in a 15-year-old girl, showing a small

hyperintense right temporal tumor (white arrow). The patient presented
with seizures and the tumor was a ganglioglioma.
LOW-GRADE SUPRATENTORIAL
ASTROCYTOMAS
Low-grade astrocytomas and gangliogliomas involving the
cortical regions and temporal lobes can often present with intractable seizures. CT may show masses of low density. MRI
usually shows a mass of decreased signal on T1-weighted images and increased signal on T2-weighted images that may or
may not enhance with gadolinium.
Complete resection is the goal of surgery, but this may
prove difficult because of problems in defining the tumor
margins and its proximity to eloquent areas. Adjuncts to aid
in this include language and motor mapping using implantable
grids or intraoperative electrophysiologic monitoring techniques,78 functional MRI techniques, and image-directed
tumor resection.79 Tumors of the temporal lobe are often
treated by formal temporal lobectomy to decrease the incidence of seizures. Seizure mapping techniques have also been
FIGURE 44-9 Axial T2WI image showing a hyperintense lesion involving

the white matter and overlying grey matter (white arrow). This 5-year-old
boy presented with a seizure, and the tumor was a dysembryoplastic
neuroepithelial tumor (DNET).
600
PART III
warranted. About 70% of children will remain recurrence free.

Recurrent tumors can be treated by reoperation alone or
reoperation followed by radiation therapy.82
PINEAL REGION TUMORS

Tumors of the pineal region encompass a wide range of histologic types. They can be divided into germ cell tumors (teratoma, germinoma, choriocarcinoma, embryonal carcinoma,
yolk sac tumor), pineal parenchymal tumors (pineocytoma,
pineoblastoma), tumors of surrounding structures (astrocytomas, meningiomas), and other benign conditions (cysts,
vascular malformations). The older term pinealoma is no
longer used.
Patients typically present with signs and symptoms of
hydrocephalus and Parinaud syndrome (upgaze paresis,
convergence nystagmus, and light-near dissociation). MRI
confirms the presence of a tumor, but is nonspecific regarding
histologic type. Specific germ cell tumors may secrete tumor
markers, which may be measured in CSF (obtained from a
lumbar puncture or ventriculostomy) or blood. Elevated
b-HCG is seen in choriocarcinomas, and elevated AFP is seen
in yolk sac tumors and embryonal cell carcinomas.
In the past, surgery in the pineal region was considered
prohibitively dangerous, and tumors were often treated
without histologic confirmation. Today, this region is now
readily approachable using supracerebellar/infratentorial or
interhemispheric-transcallosal routes with minimal morbidity, and in most centers, biopsy is performed if germ cell
markers are negative. As in the suprasellar region, pure
germinomas of the pineal gland carry an excellent prognosis
after radiation therapy. For focal disease, the radiation field
usually includes the whole ventricular volume with a boost
to the tumor bed. For disseminated disease, craniospinal
radiation is required.83 Initial treatment with chemotherapy,
followed by response-based radiation field and dose, is advocated in some centers. The nongerminomatous germ cell
tumors have a worse prognosis and require more intensive
chemotherapy along with radiotherapy.84 Pineoblastomas
are treated like PNETs in other regions of the brain. Pineocytomas may be simply observed if totally resected or given
focal radiation for residual tumor.
MALIGNANT SUPRATENTORIAL
ASTROCYTOMAS
Anaplastic astrocytomas and glioblastoma multiforme account
for roughly 9% of pediatric tumors, which is a smaller
incidence than in the adult population. Clinical signs and
symptoms are reflective of their location. Imaging features are
similar to those seen in adults, and the masses are often large,
with enhancing rings and necrotic centers (Fig. 44-10).
Dissemination occurs in about 10% of cases.86
Treatment includes maximal resection followed by radiation therapy. Unfortunately, the prognosis is still poor.
Although more extensive resection confers better outcome,
this may be due to the fact that more favorable tumors are
more amenable to aggressive surgery. Chemotherapy has a
modest impact on survival.87,88
CHOROID PLEXUS TUMORS

Tumors of the choroids plexus are divided into the benign
choroid plexus papilloma (CPP) and the malignant choroid
plexus carcinoma (CPC). In children, they tend to arise in
the trigone of the left lateral ventricle, and the patients often
present with hydrocephalus during infancy. On imaging,
the appearance is an intraventricular, homogenously enhancing, lobulated mass (Fig. 44-11). Carcinomas are typically
larger and may disseminate. The vascular supply is from the
choroidal arteries, which may be seen on high-resolution MRI.
Treatment is surgical excision, which is curative for papillomas. The procedure is hazardous, because of the highly
vascular tumors and the small size of the patients. Carcinomas
are particularly difficult to remove, because of extreme vascularity. This has led some surgeons to biopsy CPCs, followed by
chemotherapy and then second-look surgery.89 Otherwise,
ATYPICAL TERATOID/RHABDOID TUMORS

Atypical teratoid/rhabdoid tumors (AT/RTs) were previously
misclassified as PNET tumors, but have been shown to be a
distinct entity. They are highly malignant tumors with histologic resemblance to rhabdoid tumors of the kidney. Histologically, they can be distinguished from PNETs by larger cells
with pink cytoplasm that show immunohistochemical staining for smooth muscle actin, vimentin, and epithelial membrane antigen. AT/RTs typically occur in young children and
most commonly occur in the posterior fossa, but they may
be located in the spine or supratentorial space. Fluorescence
in situ hybridization (FISH) shows a deletion of the tumor
suppressor gene INI-1 in most cases.85 The prognosis of these
tumors is historically poor; however, treatment consisting of
surgical excision, intensive chemotherapy, and radiation in
older children has resulted in long-term survival in some
patients with localized disease.
FIGURE 44-10 Axial T1WI postgadolinium image showing a thalamic

enhancing tumor in a 14-year-old girl with headaches (white arrow). The
tumor was a glioblastoma multiforme (GBM).
CHAPTER 44
FIGURE 44-11 Axial T1WI postgadolinium image showing an avidly enhancing tumor in the atrium of the left ventricle in a 6-month-old girl with a
rapidly growing head circumference (black asterisk). The tumor was a choroid plexus papilloma.
the benefit of chemotherapy and radiotherapy is unproven.

With complete tumor removal, prolonged survival and even
cure are possible even in the case of CPCs.
MENINGIOMAS
Meningeal tumors are uncommon in childhood, accounting
for about 2% of intracranial tumors. Meningiomas can occur
in the orbit, sphenoid wing, or virtually any portion of the intracranial compartment, and do not necessarily need a dural
attachment. Radiographically, they typically enhance and
may be extremely large. Treatment is surgical resection. In
adults, a gross total resection is curative; however, in the pediatric population, it is less common to have a meningioma
with the typical benign histology seen in adults. Meningiomas
in pediatric patients are usually much more aggressive and
carry a worse prognosis than in adults.90
METASTASES AND DURAL-BASED MASSES

Sarcomas, particularly rhabdomyosarcoma and Ewing sarcoma,
are the most common, primary, dural-based non-CNS tumors
in children. Metastatic brain tumors are extremely uncommon
in the pediatric population and have been reported with
most tumor types, including neuroblastoma, Wilms tumor,
osteogenic sarcoma, and hepatoblastoma. Presentation is often
abrupt, with potential catastrophic neurologic symptoms
resulting from hemorrhage.
Tumor Genetics
------------------------------------------------------------------------------------------------------------------------------------------------
In the past 2 decades, there has been a rapid development in

imaging, navigational systems, and surgical instruments and
techniques. However, despite this rapid increase in surgical
BRAIN TUMORS
601
technologies, many tumors, especially high-grade lesions,

are still incurable with either surgery alone or in conjunction
with chemotherapy and radiation therapy. Like much
of medicine, the future in treating brain tumors lies in better
biological, molecular, and genetic understanding. For example, such techniques have given physicians a better understanding of neurofibromatosis type 2, which is associated
with the development of meningiomas and acoustic
neuromas in the pediatric population. The gene locus was
identified on chromosome 22,91 the same chromosome that
has been identified in pediatric meningiomas in patients
without neurofibromatosis type 2.92 These tumors have been
shown to arise from a loss of a tumor suppressor gene.
In contrast, neurofibromatosis type 1 is associated with
childhood gliomas, particularly of the optic pathway, hypothalamus, and brainstem. The affected gene locus, located at
17q11.2, encodes for the protein neurofibromin. This
protein acts as a negative regulator of the RAS signaling
pathway; therefore, a mutation in neurofibromin results
in dysregulated RAS signaling, leading to cell growth and
differentiation.93
Although molecular markers with prognostic significance have previously been identified for medulloblastoma,
only recently have pathways been identified that may be implicated in the pathogenesis of certain medulloblastoma
subtypes.3537,94 Approximately one third of medulloblastoma samples show increased signaling of the Sonic
Hedgehog (SHH) pathway, and constitutive activation of
SHH pathway promotes medulloblastoma formation in
mice. Based on the work implicating this pathway and
the preclinical efficacy of inhibition of this pathway on
medulloblastoma formation in mice, clinical trials of SHH
pathway inhibitors are underway.84,85
It has also recently been demonstrated that DNA from
sporadic (nonNF1-associated) pediatric low-grade astrocytomas contain a novel duplication at chromosome band
7q34.95 This duplication was identified in both juvenile pilocytic astrocytomas and fibrillary astrocytomas. This area of
duplication contained approximately 20 genes, one of which
was shown to be BRAF, which plays a regulatory role in the
mitogen-activated protein kinase (MAPK) pathway. Reverse
transcription polymerase chain reactionbased sequencing
reveals that this duplication results in a fusion product between KIAA1549 and BRAF. It is predicted that this fusion
gene would lack the N-terminal regulatory domains and
could result in constitutive BRAF kinase activity and subsequent unregulated activation of the MAPK pathway. Western
blot analysis revealed phosphorylated MAPK protein in tumor cells with this duplication.95 Further studies are required to determine the actual expression and function of
this KIAA1549-BRAF fusion protein. However, neuroscientists are already exploring BRAF as a potential tumor marker
or even a potential therapeutic target.
This discovery of a novel tumor pathway represents the ongoing trend in how the medical community is approaching
the treatment of brain tumors. The future will require continued
collaboration between neurosurgeons, oncologists, radiologists,
and molecular neuroscientists to continue improving outcomes
and diagnosis of patients with pediatric brain tumors.
expertconsult.com.
developed empirically rather than as a branch of classic immunology. This occurred in four distinct phases, each lasting
more than a decade. Only at the end was it possible to explain
organ engraftment and thereby eliminate the mystique of
transplantation.
PHASE 1: 1953 TO 1968
CHAPTER 45
Principles of
Transplantation
Jorge Reyes, Noriko Murase, and Thomas E. Starzl
The replacement of failing body parts with the transplantation

of organs, cells, and tissues has been a centuries-old dream,
fulfilled in the last 50 years. This success with both solid organ
and bone marrow cell transplantation has been established on
the following principles: histocompatibility matching, immunosuppression, tissue preservation, and techniques of implantation. However, neither kind of transplantation could have
emerged as a clinical service if not for the induction by the
graft itself of various degrees of donor-specific nonreactivity
(tolerance). Without this fifth factor, no transplant recipient
could survive for long, if the amount of immunosuppression
given to obtain initial engraftment had to be continued.
Enigma of Acquired Tolerance

------------------------------------------------------------------------------------------------------------------------------------------------
The variable acquired tolerance on which transplantation depends has been one of the most enigmatic and controversial
issues in all of biology. This was caused, in part, by the unexpected achievement of organ engraftment (the kidney) at an
early time (a decade before successful bone marrow transplantation) and in ostensible violation of the very principles that
would shape the impending revolution in general immunology. As a consequence, clinical organ transplantation was
The modern era of transplantation began between 1953 and

1956 with the demonstration that neonatal mice1,2 (with an
immature immune system) and irradiated adult mice3 (with
an immune system weakened by total-body irradiation) develop donor-specific tolerance after successful engraftment
of donor hematolymphopoietic cells. The key observation
was that the mice bearing donor cells (donor leukocyte chimerism) could now accept skin grafts from the original donor
strain but from no other strain (Fig. 45-1). The chimeric
neonatal mice and the irradiated adult mice were analogues
of todays bone marrow transplantation into immune-deficient
and cytoablated humans, respectively. But because a good
histocompatibility match was required for avoidance of
graft-versus-host disease (GVHD) and of rejection,4 clinical
application of bone marrow transplantation had to await discovery of the human leukocyte antigens (HLAs). When this
was accomplished,57 the successfully treated human bone
marrow recipients of 1968 were oversized versions of the
tolerant chimeric mice.
By the time of the clinical bone marrow transplant breakthrough of 1968, kidney transplantation814 already was an
established clinical service, albeit a flawed one.15 In addition,
the first long-term survivals had been recorded after liver16
and heart transplantation17; these were followed between
1968 to 1969 by the first prolonged survival of a lung18
and a pancreas recipient19 (Table 45-1). All of the organ transplant successes had been accomplished in the ostensible absence of leukocyte chimerism, without HLA matching, and
with no evidence of GVHD. By going beyond the leukocyte
chimerism boundaries established by the mouse tolerance
models, organ transplantation had entered unmapped
territory.
Pseudotolerant Organ Recipients
Two unexplained features of the alloimmune response had
made it feasible to forge ahead precociously with organ transplantation.14 The first was that organ rejection is highly reversible. The second was that an organ allograft, if protected
by nonspecific immunosuppression, could induce its own acceptance. Self-induced engraftment was observed for the
first time in 1959 in two fraternal twin kidney recipients,
first in Boston by Joseph Murray12 and then in Paris by Jean
Hamburger.8 These were the first successful transplantations
in the world of an organ allograft, in any species. Both patients
had been conditioned with 450-rad sublethal total-body irradiation before transplantation. The renal allografts functioned
for more than 2 decades without a need for maintenance
drug therapy, which was, in fact, not yet available.
A similar drug-free state was next occasionally observed
after kidney transplantation (and more frequently after liver
replacement) in mongrel dogs who were treated with a single
immunosuppressive agent: 6-mercaptopurine (6-MP),20,21
azathioprine,22,23 prednisone,24 or antilymphocyte globulin (ALG).25 After treatment was stopped, rejection in some
605
606
PART IV
TRANSPLANTATION
1953 Billingham, Brent & Medawar
1955 Main & Prehn
A
FIGURE 45-1 The mouse models of acquired tolerance described between 1953 and 1956. White cells (leukocytes) were isolated from the
spleen or bone marrow of adult donor mice (upper left) and injected into
the bloodstream of newborn mice (upper right) or of irradiated adult mice
(middle right). Under both circumstances, the recipient immune system
was too weak to reject the foreign cells (dark shaded). With engraftment
of the injected cells (i.e., donor leukocyte chimerism), the recipient mice
now could freely accept tissues and organs from the leukocyte donor
but from no other donor (bottom left).
TABLE 45-1
First Successful Transplantation of Human Allografts
(Survival >1 Year)
Physician/
Organ
City
Date
Surgeon
Reference
Kidney
Boston
Jan. 24, 1959
42, 48
Liver
Heart
Lung
Pancreas
Denver
Cape Town
Ghent
Minneapolis
July 23, 1967

Jan. 2, 1968
Nov. 14, 1968
June 3, 1969
Merrill/
Murray
Starzl
Barnard
Derom
Lillehei
72
5
18
34
animals never developed (Fig. 45-2, A). Such results were

exceedingly rare, less than 1% of the canine kidney experiments
done under 6-MP and azathioprine up to the summer of 1962.
However, the possibility that an organ could be inherently tolerogenic was crystallized by the human experience summarized
in the title of a report in 1963 of a series of live-donor kidney
recipients treated in Denver, The Reversal of Rejection in
Human Renal Homografts with Subsequent Development of
Homograft Tolerance.14 The recipients had been given azathioprine before as well as after renal transplantation, adding large
doses of prednisone to treat rejections that were monitored by
serial testing of serum creatinine (Fig. 45-3, A). Rejection
occurred in almost every case, and 25% of the grafts were lost
to uncontrolled acute rejection. However, the 1-year survival of
46 allografts, obtained from familial donors during a 16-month
period from 1962 to 1963, was an unprecedented 75%.
The development of partial tolerance in many of the
survivors was inferred from the rapidly declining need for
treatment after rejection reversal (see Fig. 45-3, A). Nine
(19%) of the 46 allografts functioned for the next 4 decades,
each depicted in Figure 45-4 as a horizontal bar. Moreover, all
FIGURE 45-2 A, Canine recipient of an orthotopic liver homograft,

5 years later. The operation was on March 23, 1964. The dog was treated
for only 120 days with azathioprine and died of old age after 13 years.
B, A spontaneously tolerant pig recipient described by Calne.29
immunosuppression eventually was stopped in seven of the

nine patients without rejection for periods ranging from 6
to 40 years (the solid portion of the bars). Eight of the nine
patients are still alive and bear the longest surviving organ
allografts in the world.26
What was the connection between the tolerant mouse
models, the irradiated fraternal twin kidney recipients in
Boston and Paris, the ultimate drug-free canine organ recipients
(see Fig. 45-2, A), and the unique cluster of pseudotolerant
human kidney recipients in Denver (Fig. 45-4)? What were
the mechanisms of engraftment and what was the relationship
of engraftment to tolerance? The mystery deepened with the
demonstration in 1966 in France,27 England,2831 and the United
States32 that the liver can be transplanted in about 20% of outbred pigs without any treatment at all (see Fig. 45-2, B). Because
graft-versus-host disease had yet to be seen (despite the use
of organs from HLA-mismatched donors) none of the animal
or human organ recipients, whether off or on maintenance
immunosuppression, was thought to have donor leukocyte
chimerism to explain organ engraftment.
False Premises of Phase 1
Organ transplantation became disconnected at a very early
time from the scientific anchor of leukocyte chimerism that
had been established by the mouse models and was soon to
CHAPTER 45
Immunosuppression (1962-1963)
Azathioprine
Azathioprine
Serum creatinine
Prednisone
Serum creatinine
607
Change in December 1963
Prednisone
PRINCIPLES OF TRANSPLANTATION
Rejection
Tolerance
Pretreatment
Tx
Tx
FIGURE 45-3 A, Empirically developed immunosuppression used for kidney transplant recipients from 1962 to 1963. Note the reversal of rejection with
the addition of prednisone to azathioprine. More than a third of a century later, it was realized that the timing of drug administration had been in accord
with the tolerogenic principles of immunosuppression (see text). B, Treatment revisions in immunosuppression made at the University of Colorado in
December 1963, which unwittingly violated principles of tolerogenic immunosuppression. Pretreatment was de-emphasized or eliminated, and high doses
of prednisone were given prophylactically instead of as needed. Although the frequency of acute rejection was reduced, the drug-free tolerance shown in
Figure 45-4 was no longer seen. Tx, treatment.
be exemplified by human bone marrow transplantation. The

resulting intellectual separation of the two kinds of transplantation (Fig. 45-5) was an unchallenged legacy of phase 1,
passed from generation to generation.
There was another dark legacy of phase 1 that began in
1964. This was a modified version of the treatment strategy
that had been developed with azathioprine and prednisone
(see Fig. 45-3, B). The principal change was the use of large
prophylactic doses of prednisone from the time of operation,
instead of the administration of corticosteroids only when
needed. In a second modification, the pretreatment was deemphasized (see Fig. 45-3, B). The incidence of acute rejection
was greatly reduced after these changes. However, no cluster
of drug-free kidney recipients, such as shown in Figure 45-4,
was ever seen again, anywhere in the world. More than
Recipient
Immunosuppression
35 years passed before the long-term immunologic consequences of the modifications were realized.
PHASE 2: 1969 TO 1979

Throughout the succeeding phase 2 that began in 1969, immunosuppression for organ transplantation was based on azathioprine and prophylactic high-dose prednisone to which ALG
was added after 196625,33 in about 15% of centers. Phase 2
was a bleak period. In the view of critics, the heavy mortality,
and particularly the devastating morbidity caused by corticosteroid dependence, made organ transplantation (even of kidneys)
as much a disease as a treatment. Most of the liver and heart
transplant programs that had been established in an initial burst
of optimism after the first successful cases closed down.
No Immunosuppression
Donor
CR
Sister
<1.5
Brother <1.5
Mother <1.5
Mother <1.5
Mother 2.5-3
Sister
Gr. Aunt <1.5
Father
Uncle
0
10
20
30
Years post transplantation
40
<1.5
<1.5 FIGURE 45-4 Nine (19%) of the 46 live-donor kidney recipients

treated at the University of Colorado during an 18-month period
<1.5
beginning in the autumn of 1962. The solid portion of the horizontal bars depicts the time off immunosuppression. Note that
the current serum creatinine concentration (CR) is normal in all
but one patient. *Murdered: kidney allograft normal at autopsy.
608
PART IV
TRANSPLANTATION
Solid Organ
Bone Marrow
1959
1968
Nonessential
Tissue match
Essential
Acceptance
Graft take
Tolerance
Uncommon
GVHD
Common
FIGURE 45-5 The developmental tree of

bone marrow (right) and organ transplantation
(left) after it was demonstrated that rejection
is an immunologic response. GVHD, graftversus-host disease.
But in the few remaining centers, patients, such as the one

shown in Figure 45-6, bore witness to what some day would
be accomplished on a grand scale. Four years old at the time of
her liver replacement for biliary atresia and a hepatoma in
1969, the patient depicted is the longest surviving recipient
of an extrarenal organ.
PHASE 3: 1980 TO 1991

In fact, what had appeared to be the sunset of extrarenal organ
transplantation was only the dawn of phase 3, which began
with the clinical introduction of cyclosporine,3437 followed a
decade later by that of tacrolimus.3841 The use of these drugs
Seed
was associated with stepwise improvements with all organs,

but their impact was most conclusively demonstrated
with liver and heart transplantation. The results with liver
transplantation shown in Figure 45-7 using azathioprine-,
cyclosporine-, and tacrolimus-based immunosuppression
were presented at the meeting of the American Surgical Association in April 1994.42 By then, intestinal transplantation
under tacrolimus-based immunosuppression had become a
service.43,44
As the new agents became available, they were simply incorporated into the modified formula of heavy prophylactic
immunosuppression that had been inherited from phases
1 and 2. Used in a variety of multiple-agent combinations from
FIGURE 45-6 Four years old at the time of liver replacement for biliary atresia and a hepatoma, but now in her 40th post-transplant year (shown here at
35 years post-transplant), the (former) patient is the longest-surviving recipient of an extrarenal organ.
CHAPTER 45
the time of surgery, the better drugs fueled the golden age of
transplantation of the 1980s and early 1990s. Acute rejection
had become almost a nonproblem. However, the unresolved
issues now were chronic rejection, risks of long-term immunosuppression (e.g., infections and de novo malignancies),
Patient survival (%)
609
and drug toxicity (e.g., the nephrotoxicity of cyclosporine

and tacrolimus).
PHASE 4: 1992 TO PRESENT

It was clear that relief from the burden of lifetime immunosuppression would require elucidation of the mechanisms of
alloengraftment and of acquired tolerance. An intensified
search for the engraftment mechanisms has dominated the
current phase 4, which began in the early 1990s. There was
a growing realization (particularly with recipients of liver
allografts), that immunosuppression could be withdrawn
successfully in selected cases, which sparked various prospective trials of immunosuppression withdrawal.45
100
80
60
TAC (n=1391)
CYA (n=1835)
AZA (n=168)
40
20
Historical Dogma
0
0
4
1
2
3
Time after transplantation (years)
FIGURE 45-7 Patient survival: results with orthotopic liver transplantation

at the Universities of Colorado (1963 to 1980) and Pittsburgh (1981 to 1993),
in periods defined by azathioprine (AZA)-, cyclosporine (CYA)-, and tacrolimus (TAC)-based immune suppression. Stepwise improvements associated
with the advent of these drugs also were made with other organs.
Until this time, organ engraftment had been attributed to

mechanisms that did not involve either the presence or a role
of leukocyte chimerism. Although it was known that organs
contain large numbers of passenger leukocytes, these donor
cells were largely replaced in the successfully transplanted
allograft by recipient leukocytes as shown in Figure 45-8, A.
The missing donor cells were thought to have undergone
Single response (organ)
Single response (bone marrow)

GVH
Proliferation
of host
antigraft cells
Proliferation
of host
antigraft cells
Defenseless recipient
HVG (rejection)
B
Double response (organ)
Double response (bone marrow)
Immunosuppression
GVH
GVH
Reciprocal clonal
deletion
Reciprocal clonal
deletion
Unconditioned recipient
Unaltered bone
marrow
HVG (rejection)
Conditioned recipient
HVG
FIGURE 45-8 Old (A and B) and new views (C and D) of transplantation recipients. A, The early conceptualization of immune mechanisms in organ transplantation in terms of a unidirectional host-versus-graft (HVG) response. Although this readily explained organ rejection, it limited possible explanations of
organ engraftment. B, Mirror image of A depicting the early understanding of successful bone marrow transplantation as a complete replacement of the
recipient immune system by that of the donor, with the potential complication of an unopposed lethal unidirectional graft-versus-host (GVH) response, that
is, rejection of the recipient by the graft. C, Our current view of bidirectional and reciprocally modulating immune responses of coexisting immune-competent
cell populations. Because of variable reciprocal induction of deletional tolerance, organ engraftment was feasible despite a usually dominant HVG reaction. The
bone silhouette in the graft represents passenger leukocytes of bone marrow origin. D, Our currently conceived mirror image of C after successful bone
marrow transplantation. Recipients cytoablation has caused a reversal of the size proportions of the donor and recipient populations of immune cells.
610
PART IV
TRANSPLANTATION
immune destruction with selective sparing of the specialized

parenchymal cells. As for bone marrow transplantation (see
Fig. 45-8, B), the ideal result had been perceived as complete
replacement of recipient immune cells (i.e., total hematolymphopoietic chimerism).
Discovery of Microchimerism
A flaw in this historical dogma began to be exposed in the early
1990s. The first puzzling observation in Seattle46 and Helsinki47
was the invariable presence of a small residual population
of recipient hematolymphopoietic cells in patients previously
thought to have complete bone marrow replacement (see
Fig. 45-8, D). This was followed in 1992 by the discovery of
donor leukocyte microchimerism in long-surviving human
organ recipients. Now it was evident that organ engraftment
(see Figure 45-8, C) and bone marrow cell engraftment (see
Fig. 45-8, D) were mirror-image versions of leukocyte chimerism,
differing in the reversed proportion of donor and recipient cells.
The discovery of microchimerism in organ recipients was
made with a very simple clinical study.4852 With the
use of sensitive detection techniques, donor hematolymphopoietic cells of different lineages (including dendritic
cells) were found in the blood, lymph nodes, skin, or other
tissues of 30 of 30 liver or kidney recipients who had borne
functioning allografts for up to 30 years. The donor leukocytes
obviously were progeny of donor precursor or pluripotent
hematolymphopoietic stem cells that had migrated from the
graft into the recipient after surviving a double immune
reaction that presumably had occurred just after transplantation, years or decades earlier.5356
It was concluded that organ engraftment had been the
result of responses of coexisting donor and recipient cells,
each to the other, causing reciprocal clonal exhaustion, followed by peripheral clonal deletion.48,50 The host response
(the upright curve in Fig. 45-9) was the dominant one in
most cases of organ transplantation but with the occasional
exception of GVHD. In the conventionally treated bone marrow recipient, host cytoablation simply transferred immune
dominance from the host to the graft (the inverted curve
in Fig. 45-9), explaining the high risk of GVHD. All of the
major differences between the two kinds of transplantation
were caused by the recipient cytoablation. After an estrangement of more than a third of a century, the intellectual
separation of bone marrow and organ transplantation was
ended (Fig. 45-10).
Seed
FIGURE 45-10 Unification of organ and bone marrow transplantation
(see text).
Immune Regulation by Antigen Migration

and Localization
But how was the exhaustion-deletion of the double immune
reaction shown in Figure 45-9 maintained after its acute
induction by the first wave of migratory leukocytes? Rolf
Zinkernagel, in Zurich (Fig. 45-11), had addressed this
question during the 1990s in experimental studies of the
nonresponsiveness that may develop to intracellular microorganisms, such as tubercle bacillus and noncytopathic
viruses.5760 The analogies between the syndromes caused
by such infectious agents and the events following transplantation were described in 1998 in a joint review with
Zinkernagel in the New England Journal of Medicine.61
The analogies between transplantation and infection had
been obscured by the characteristic double immune reaction
of transplantation and by the complicating factor of immunosuppression. Now, these analogies were obvious.
The antidonor response induced by the initially selective migration of the grafts leukocytes to host lymphoid organs
(Fig. 45-12, left)6265 is comparable to the response induced
by a spreading intracellular pathogen. The migration patterns
of the donor leukocytes were the same whether these cells
emigrated from an organ or were delivered as a bone marrow
Immunosuppression
FIGURE 45-9 Contemporaneous HVG (upright
curves) and GVH (inverted curves) responses
after transplantation. In contrast to the usually
dominant HVG reaction of organ transplantation, the GVH reaction usually is dominant
after bone marrow cell transplantation to the
irradiated or otherwise immunodepressed recipient. Therapeutic failure with either type of
transplantation implies the inability to control
one, the other, or both of the contemporaneous responses with a protective umbrella of
immunosuppression.61
Failure
HVG
Immune
reaction
Recipient
Success
GVH
Donor
Time after organ transplantation
Failure
CHAPTER 45
611
cell infusion. Cells that survived the antidonor response that

they had induced begin within a few days to move on (see
Fig. 45-12, right) to protected nonlymphoid niches, where
their presence no longer may be detected by the immune
system (immune ignorance61,6669). This was a survival tactic
of noncytopathic microorganisms.
The migration of donor leukocytes is shown schematically in
Figure 45-13, left by centrifugal arrows: first by hematogenous
routes to lymphoid organs and, after a few weeks, on to nonlymphoid sites (outer circle). A subsequent reverse migration
of donor cells from protected nonlymphoid niches back to host
lymphoid organs is depicted by the inwardly directed dashed
arrows in Figure 45-13, right. The retrograde migration is a
two-edged sword. On the one hand, these cells may sustain
the clonal exhaustion-deletion induced at the outset, usually
requiring an umbrella of maintenance immunosuppression.
But on the other hand, these cells can perpetuate alloimmunity
in the same way as surviving residual microorganisms perpetuate protective immunity. Not surprisingly, therefore, an alternative consequence of microchimerism may be the high-panel
reactive antibody (connoting sensitization to HLA antigens) that
commonly develops after unsuccessful transplantation.70,71
FIGURE 45-11 Rolf Zinkernagel. Swiss physician-immunologist whose
discovery, with Peter Doherty, of the mechanisms of the adaptive immune
response to noncytopathic microorganisms earned them the Nobel prize
in 1996.
Therapeutic Implications
How could the new insight be exploited clinically? The window of opportunity for the donor leukocyte-induced clonal
deletion that corresponds with collapse of the antigraft
response (Fig. 45-14, left) is open only for the first few
FIGURE 45-12 Initial preferential migration of passenger leukocytes from organ allografts (here a liver) to host lymphoid organs (left), where they induce
a donor-specific immune response. After about 30 days, many of the surviving cells move on to nonlymphoid sites (right).
TRANSPLANTATION
es
sit
d
oi
on
lati
cu hymus
T
cir
es
sit
mus
Thy
lee
Sp
n
lee
Sp
No
nly
mp
h
Lym
ph
d
oi
No
nly
mp
circLymph
h
ula
tion
Blo
od
PART IV
Blo
od
612
h
mp
Ly odes
n
h
mp
Ly odes
n
ne
Bo rrow
ma
ne
Bo rrow
ma
FIGURE 45-13 The migration routes of passenger leukocytes of transplanted organs are similar to those of infused bone marrow cells. Left, Selective
migration at first to host lymphoid organs. After 15 to 30 days, surviving leukocytes begin to secondarily move to nonlymphoid sites. Right, Establishment of
reverse traffic by which the exhaustion-deletion induced at the outset can be maintained.
Time
Tx
Delayed
rejection
Time
Too little
treatment
Immune response
Tolerance
Too much
treatment
Immune response
Tx
tio
Activ
atio
n
Exhaustion
le
De
Immune response
Tolerogenic
immunosuppression
Tx
Graft
loss
of the high risk of GVHD, this approach was too dangerous

and too restrictive to be practical for organ transplantation.
However, less drastic lymphoid depletion by ALG or
other measures (so-called nonmyeloablative conditioning)
had been repeatedly shown since the 1960s to be effective
without causing GVHD33 (see Fig. 45-15).
After pretreatment with one of todays potent antilymphoid antibody preparations, the preemptively weakened
clonal activation could proceed efficiently to clonal deletion
under minimalistic short- and long-term maintenance therapy (Fig. 45-16). In July 2001, we instituted the doubleprinciple strategy in adult organ recipients. The pretreatment
was with a single infusion of 5 mg/kg of Thymoglobulin.
Beginning in 2002, a single Campath dose of 30 mg was
substituted for Thymoglobulin in most adult cases. After
either kind of lymphoid depletion, treatment after transplantation was given with a conservative daily dose of a single
drug (usually tacrolimus), adding other agents only in the
event of breakthrough rejection and for as brief a period as
1. Irradiation
2. Thoracic duct
drainage
Immune responses
post-transplant weeks.55,7274 It was apparent that the window could be closed by excessive postoperative immunosuppression (Fig. 45-14, middle). With later reduction of the
initial overimmunosuppression, recovery of the inefficiently
deleted clone would be expected, leading to the delayed acute
rejection, or the chronic rejection, that was being seen in
the transplant clinics. Even in the best-case scenario, the
patients would be predestined to lifetime dependence on immunosuppression. However, too little immunosuppression
would result in uncontrolled rejection (Fig. 45-14, right).
The problem faced by clinicians was how to find just the
right amount of post-transplant immunosuppression. In
2001, it was suggested that this dilemma could be addressed
by successively applying two historically rooted therapeutic
principles: recipient pretreatment, followed by minimalistic
post-transplant immunosuppression.75 With pretreatment,
the recipients immune responsiveness would be reduced before exposure to donor antigen, thereby lowering the anticipated donor-specific response to a more readily deletable
range (Fig. 45-15). Clonal deletion by the kidneys passenger
leukocytes undoubtedly is what had been accomplished after
sublethal irradiation alone in the ground-breaking fraternal
twin (i.e., sublethal total-body irradiation or myelotoxic
drugs) cases of 1959.8,12 In fact, radical pretreatment by recipient cytoablation ultimately became the essential therapeutic
step for conventional bone marrow transplantation. Because
Irreversible
rejection
3. Other drugs
4. Antilymphoid
antibodies
Donor
specific
clonal
deletion
Pretreatment
Tx
Time
FIGURE 45-14 The effect of post-transplant immunosuppression on

the seminal mechanism of clonal exhaustion-deletion. Left, Just the right
amount. Middle, Too much. Right, Too little (see text). Tx, treatment.
FIGURE 45-15 Rather than producing rejection (thick dark arrow), the
donor-specific immune response to allografts may be exhausted and deleted, as depicted by the fall of the initially ascending continuous thin line,
when recipient immune responsiveness is weakened in advance of transplantation (the pretreatment principle). Tx, treatment.
CHAPTER 45
Prednisone
Organ Preservation
Tacrolimus
PROCUREMENT
Steroid bolus P.R.N.
HVG
------------------------------------------------------------------------------------------------------------------------------------------------
GVH
Immune responses
Lymphoid-depleting
pretreatment
Irreversible
rejection
With pretreatment
With pretreatment
and minimal
immunosuppression
Tx
FIGURE 45-16 Conversion of rejection (thick dark arrow) to an immune

response that can be exhausted and deleted by combination of pretreatment and minimalistic post-transplant immunosuppression. Tx, treatment.
possible. The strategy was extended to infants and children

for intestinal transplantation in 2002 and for all kidney
transplantations after April 2003.76
After 4 to 8 months, weaning from monotherapy to less-thandaily doses was begun in adults whose graft function was stable:
every other day, then three times per week, twice a week, and in
many cases to once a week by 1 year (Fig. 45-17). The strategy
has been used for the treatment of more than 1000 adult kidney,
liver, intestine, pancreas, and lung recipients.7779 This experience has demonstrated that the quality of life of transplant
recipients can be improved. For the first time, children are being
considered for spaced weaning.
These and other clinical trials have spawned definitions
of clinical or operational tolerance (normal graft function
without features of graft rejection and without the need for immunosuppressive drugs), and prope, or near tolerance
(the state of normal graft function in the presence of minimal
or undetectable levels of immunosuppression).80,81 However,
achieving this on a consistent basis may hinge on the development of standardized clinically applicable markers of immune
tolerance that can assess the appropriateness of this clinical
strategy on the individual patient. Prospective weaning trials
sponsored by the Immune Tolerance Network are currently
underway in stable adult and pediatric recipients of liver
allografts, incorporating the aforementioned strategies and
mechanistic studies.82
Every 2
weeks
2
613
Off immunosuppression
5
Daily multitherapy
1
EXTENDED PRESERVATION
Daily monotherapy
8
Every other
day
2
1 a week
10
2 a week
4
The breakthroughs of the early 1960s that made transplantation clinically practical were so unexpected that almost no
formal preparation had been made to preserve the transplanted
organs. Cardiac surgeons had used hypothermia for open-heart
operations from 1950 onward and knew that ischemic damage
below the level of aortic cross-clamping could be reduced
by cooling the subdiaphragmatic organs.83 In an early report,
Lillehei and colleagues84 immersed intestines in iced saline
before autotransplantation. In Boston, Sicular and Moore85
reported greatly slowed enzyme degradation in cold slices
of liver.
Despite this awareness, kidneys were routinely transplanted until 1963 with no protection from warm ischemia
during organ transfer. The only attempt to cool kidney
allografts until then was by the potentially dangerous practice
(used by thoracic surgeons for open-heart surgery) of
immersing the live donor in a bathtub of ice water (totalbody hypothermia).86 This cumbersome method of cooling
was quickly replaced by infusion of chilled solutions into
the renal artery after donor nephrectomy,87 exploiting a
principle of core (transvascular) cooling that had been
standardized several years earlier for experimental liver
transplantation.88
Core cooling in situ, the first critical step in the preservation of all cadaveric whole organs, is done today with variations of the technique described in 1963 by Marchioro and
coworkers,89 which permits in situ cooling to be undertaken90
(Fig. 45-18). Ackermann and Snell91 and Merkel and associates92 popularized in situ cooling of cadaveric kidneys with
simple infusion of cold electrolyte solutions into the donor
femoral artery or distal aorta. Procurement techniques were
eventually perfected that allowed removal of all thoracic
and abdominal organs, including the liver, without jeopardizing any of the individual organs (Fig. 45-19).93 Modifications
of this flexible procedure have been made for unstable donors
and even for donors whose hearts have stopped beating.94
During the 5 years between 1980 and 1985, such techniques
had become interchangeable in all parts of the world, setting
the stage for reliable organ sharing. After the chilled organs are
removed, subsequent preservation is possible with prototype
strategies: simple refrigeration or continuous perfusion (see
later).
3 a week
6
FIGURE 45-17 Diagram of 2.5-year follow-up.
Continuous Vascular Perfusion

Efforts to continuously perfuse isolated organs have proved to
be difficult. For renal allografts, Ackermann and Barnard95
used a normothermic perfusate primed with blood that was
oxygenated within a hyperbaric chamber. Brettschneider
and colleagues96 modified the apparatus and were able to preserve canine livers for 2 days, an unprecedented feat at the
time. When Belzer and associates97 eliminated the hemoglobin and hyperbaric chamber components, their asanguineous
hypothermic perfusion technique was immediately accepted
for clinical renal transplantation but then slowly abandoned
614
PART IV
TRANSPLANTATION
FIGURE 45-18 First technique of in situ cooling by extracorporeal hypothermic perfusion.

The catheters were inserted into the aorta
and vena cava through the femoral vessels
as soon as possible after death. Temperature
control was provided with a heat exchanger.
Cross-clamping of the thoracic aorta limited
perfusion to the lower part of the body. This
method of cadaveric organ procurement was
used from 1962 to 1969, before the acceptance
of brain death criteria. The preliminary stages
of this approach provided the basis for subsequent in situ infusion techniques.
Arterial inflow
Portal v.
Re-circulation line
Heat exchanger
I.V.C
To
femoral
a. & v.
Aorta
Inf. mesentric
a.
Pump
Preservation fluid
through splenic v.
Preservation fluid
through terminal
aorta
Static Preservation
With these slush techniques, special solutions, such as those
described by Collins and coworkers,98 were instilled into the
renal vascular system of kidneys or the vascular system of other
organs after their preliminary chilling and separation. The original Collins solution or modifications of it were used for nearly
2 decades before they were replaced with the University of Wisconsin (UW) solution that was developed by the team of Folkert
Belzer. Although it was first used for the liver,99101 the UW
solution provides superior preservation of kidneys and other
organs.102,103 The UW preservation permitted longer and safer
preservation of kidneys (2 days) and livers (18 hours), a higher
rate of graft survival, and a lower rate of primary nonfunction.
With the UW solution, national organ sharing was made
economical and practical. This success has refocused efforts
on understanding the mechanisms involved in the ischemia/
reperfusion injury (deprivation and then restoration of tissue
oxygen) that impacts organ function, and has resulted in the
development of other preservation solutions (Celsior, HTK:
histidine-tryptophan-ketoglutarate) and the inclusion of drugs
that act on the mediators of injury.104
Tissue Typing
------------------------------------------------------------------------------------------------------------------------------------------------
ANTIGEN MATCHING
FIGURE 45-19 Principle of in situ cooling used for multiple organ
procurement. With limited preliminary dissection of the aorta and of the
great splanchnic veins (in this case, the splenic vein), cold infusates can
be used to chill organs in situ. In this case, the kidneys and liver were being
removed. Note the aortic cross-clamp above the celiac axis.
in most centers when it was learned that the quality of 2-day

preservation was not markedly better than that of simpler
and less expensive infusion and slush methods (see later).
However, refinement of perfusion techniques may someday
permit true organ banking.
The human leukocyte antigen (HLA) system has an important

role in immune regulation and is thus a barrier that must be
avoided (with better matching) or modified (with immunosuppressive strategies). HLA class I A, B, and C, and HLA class
II DR, DQ, and DP molecules are expressed on various cell
types that, when bound to a specific repertoire of peptides,
present to CD8 or CD4 T cells. Patient and donor matching is of significant importance in bone marrow transplantation in order to prevent lethal GVHD but of variable
importance in solid organ transplantation.105
CHAPTER 45
Graft
Match
Rx
Host
615
the individual contributing centers see no such trend in their

own experience.83,109111 In a compelling study, Terasaki
and associates112 reported that early survival and the subsequent half-life of kidneys from randomly matched, living
unrelated donors was identical to that of parentoffspring
(one haplotypematched) grafts. The inescapable conclusion
is that more effective timing and dosage of immunosuppressive therapy, rather than refinements in tissue matching and
organ sharing, will be the primary method of improving the
results of whole organ transplantation.
CROSSMATCHING
Partial
mismatch
Total
mismatch
Rx
Rx
FIGURE 45-20 The nullification effect of simultaneous host-versus-graft

(HVG) and graft-versus-host (GVH) reactions when organs are transplanted
to recipients whose immune system has not been cytoablated. The reciprocal induction of tolerance, each to the other, of the coexisting cell populations is the explanation for the poor correlation of human leukocyte
antigen (HLA) matching with outcome after organ transplantation.
The first prospective antigen matching trials were begun in

1964 by Terasaki and associates106 in collaboration with the
University of Colorado kidney transplantation team. Although
the value of this serologic technology was demonstrable when
the kidney donor was a highly compatible family member (the
perfect match),107 lesser degrees of matching correlated
poorly with renal transplantation outcome.108 The reasons
for this paradox were inexplicable until the discovery of recipient chimerism (Fig. 45-20). However, the belief that matching should be a prime determinant of success resulted in its
use as an overriding factor for the allocation of cadaver kidneys in the United States.
The propriety of this kidney allocation policy has been repeatedly challenged on ethical as well as scientific grounds for
nearly a third of a century. Those in favor of perpetuating the
role of graded HLA matches cite multicenter case compilations
in the United States and Europe showing a small gain in allograft survival with histocompatible kidneys, whereas many of
None of the immunosuppressive measures available

today can prevent immediate destruction of kidneys and
other kinds of organ grafts in what has been called hyperacute rejection. This complication was first seen with the
transplantation of kidneys from ABO-incompatible donors
when they were placed in recipients with antidonor isoagglutinins.113 After the description by Terasaki and associates114 of hyperacute kidney rejection by a recipient with
antidonor lymphocytotoxic antibodies, Kissmeyer-Nielsen
and colleagues115 and others116119 confirmed the association of hyperacute rejection with these antigraft antibodies.
Although hyperacute rejection can usually be avoided with
the lymphocytotoxic crossmatch originally recommended
by Terasaki and associates, the precise pathogenesis of
such rejection remains poorly understood more than
30 years after its recognition as a complement activation
syndrome.116,117
Future Prospects
------------------------------------------------------------------------------------------------------------------------------------------------
The revisions in timing and dose control that encourage the

seminal mechanisms of clonal exhaustion-deletion and immune
ignorance should make it possible to systematically reduce exposure to the risks of chronic immunosuppression. Our prediction is that completely drug-free tolerance will be largely, but not
exclusively, limited to recipients of HLA-matched organs. But
variable partial tolerance will be more regularly attainable in
most of the others, not so much by developing better drugs
but by the mechanism-based use of drugs we already have in
hand. Xenotransplantation will have to be developed within
the same immunologic framework. Here, the problem, in
principle, is to create a better interspecies tissue match by transgenic modification. Although the a-1,3GT gene responsible for
hyperacute rejection of pig organs by higher primates has been
knocked out in pigs,120 it is not yet known what further changes
have to be made before porcine organs can be used clinically.
Where stem cell biology will fit remains unknown, but it also
will have to conform to the same immunologic rules.
expertconsult.com.
CHAPTER 46
Renal
Transplantation
John C. Magee
Transplantation is the preferred treatment for children with

end-stage renal disease (ESRD), because it provides the best
opportunity for health, growth, and development. Progress
continues in pediatric transplantation, and current patient
and graft survival is excellent.
Although single-center reports provide insight into many
issues, larger registry data provide a more substantive overview of the state of transplantation. The leading registry is
the North American Pediatric Renal Trials and Collaborative
Studies (NAPRTCS).1,2 In addition, data for all transplantations performed in the United States are collected through
the Organ Procurement Transplant Network (OPTN),
and they are regularly analyzed by the Scientific Registry of
Transplant Recipients (SRTR).3,4
End-Stage Renal Disease

in Children
------------------------------------------------------------------------------------------------------------------------------------------------
According to the United States Renal Data System (USRDS), 1343

individuals aged 19 years or younger began treatment for
ESRD in 2008.5 The incidence of ESRD in this age group is
15.5 per million per year. Because ESRD is much more
uncommon in children, this rate is well below the overall national incidence of ESRD of 362 per million per year.
The etiology of renal disease in the pediatric transplantation population is summarized in Table 46-1. According to
these NAPRTCS data, the five most common diagnoses are
renal aplasia/hypoplasia/dysplasia, obstructive uropathy, focal
segmental glomerulosclerosis (FSGS), reflux nephropathy,
and chronic glomerulonephritis.1 These diagnoses account
for just over half the transplantations performed. The causes
of renal failure are distinctly different from those in adults;
specifically, congenital abnormalities and obstructive uropathy are the leading causes for transplantation. In addition,
FSGS is the most common acquired renal disease and is much
more common in children compared with adults.
Within the pediatric population, the prevalence of causes
varies by age, sex, and race.1 Congenital causes are more
prevalent in younger children, whereas acquired diseases tend
to become manifest in older children. Overall, 59.4% of the
recipients are male, and males represent the majority of
the recipients with obstructive uropathy (85.2%), aplasia/
hypoplasia/dysplasia (61.8%), and FSGS (57.8%). Reflux
nephropathy, chronic glomerulonephritis, and lupus nephritis
are more prevalent in females, with females accounting
for 56.7%, 57.0%, and 83.3%, respectively. Regarding race,
for black children, FSGS was the most prevalent diagnosis
(23.1%), followed by obstructive uropathy (15%) and
aplasia/hypoplasia/dysplasia (13.5%). In white recipients,
obstructive uropathy was the most prevalent etiology
(17.0%), followed by aplasia/hypoplasia/dysplasia (16.9%)
and FSGS (9.0%).
Recipient Evaluation
------------------------------------------------------------------------------------------------------------------------------------------------
Any child with ESRD should be considered for transplantation. Absolute contraindications are rare and include
untreated malignancy or systemic sepsis. Relative contraindications include severe systemic disease that profoundly limits
the patients life span or a social situation that makes follow-up
with post-transplantation care and immunosuppression regimen absolutely impossible. At times, the decision whether to
transplant a child with a poor quality of life or significant
impairment can be extremely difficult. In such situations, a
thorough discussion focused on the expectations and goals
for that child is helpful.
All children with progressive chronic renal insufficiency
should be evaluated by a multidisciplinary pediatric transplantation team, including a pediatric nephrologist, a
transplantation surgeon, social worker, and nutritionist. In
addition, many teams include pediatric urologists and clinical
psychologists, with other experts included as indicated.
Ideally, the child would be fully evaluated before initiating
dialysis. This can facilitate evaluation of potential living donors and permit preemptive transplantation, obviating the
need for dialysis. Regarding infant size, although it is often
stated that approaching 10.0 kg is ideal, it is clear that transplantation can be performed successfully in smaller infants at
experienced centers.68 The guiding principle should be to
optimize the situation as much as possible but not let an
arbitrary weight target compromise the health of the child.
Our standard evaluation process is summarized in
Table 46-2. Every effort should be made to optimize the
617
618
PART IV
TRANSPLANTATION
TABLE 46-1
Primary Diagnosis for Renal Transplantation Recipients
(N 9854) Age 20 Years and Younger
TABLE 46-2
Evaluation of Pediatric Kidney Transplantation Candidate
Disease
Aplasia/hypoplasia/dysplasia
Obstructive uropathy
Focal segmental glomerulosclerosis
Reflux nephropathy
Chronic glomerulonephritis
Polycystic disease
Medullary cystic disease
Hemolytic-uremic syndrome
Prune-belly syndrome
Congenital nephrotic syndrome
Familial nephritis
Cystinosis
Pyelointerstitial nephritis
Membranoproliferative glomerulonephritis type I
Idiopathic crescentic glomerulonephritis
Systemic lupus erythematosus nephritis
Renal infarct
Berger (IgA) nephritis
Henoch-Schonlein nephritis
Membranoproliferative glomerulonephritis type II
Wegener granulomatosis
Wilms tumor
Drash syndrome
Oxalosis
Membranous nephropathy
Other systemic immunologic disease
Sickle cell nephropathy
Diabetic glomerulonephritis
Other
Unknown
15.9
15.6
11.7
5.2
3.3
2.9
2.8
2.6
2.6
2.6
2.3
2.0
1.8
1.7
1.7
1.5
1.4
1.3
1.1
0.8
0.6
0.5
0.5
0.5
0.4
0.3
0.2
0.1
9.8
6.2
From North American Pediatric Renal Trials and Collaborative Study

(NAPRTCS) 2008 Annual Report. Available at www.naprtcs.org. Accessed
March 20, 2010.
medical management of the child with ESRD, including management of bone disease, optimization of nutrition, and completing childhood immunizations. Several aspects of the
evaluation of the pediatric recipient are unique and deserve
special attention. One is the evaluation and management of
bladder function. Urologic anomalies are common, and many
will have also undergone previous urologic procedures.
Rarely, the bladder may be inadequate for transplantation.
Expertise in such issues, or a close working collaboration with
pediatric urology, is essential. Nutrition is also of paramount
importance to optimize growth and development. Finally, it
is important to evaluate and optimize issues related to the
psychological state of the child and caregivers.9 Adequate
social support is vital for all involved. The stress of a chronically ill child undergoing a complex procedure places a great
strain on all, and the need for ongoing education and reassurance is significant. In older children, it is important to ensure
they are actively involved. Adolescents can be particularly
challenging, because the risk of noncompliance appears to
be greatest in this group.
In addition to these factors, several other issues require special attention. One is the potential need to evaluate the patients
vasculature. As renal replacement therapy has improved, it is
History and physical examination

Laboratory tests:
Hematologic (complete blood cell count with platelets; prothrombin
time/partial thromboplastin time)
Biochemistry (renal function, electrolytes, liver function)
Serologic studies (hepatitis herpesvirus B and C, cytomegalovirus,
Epstein-Barr virus, varicella-zoster virus, human
immunodeficiency virus)
ABO blood typing
Tissue typing (human leukocyte antigen typing; alloantibody
screening)
Urinalysis
Chest radiograph
Electrocardiogram
Psychosocial assessment
As indicated evaluations:
Voiding cystourethrogram/urodynamic studies
Vascular imaging
Hypercoagulable workup
now possible to successfully hemodialyze smaller children,

including neonates. These therapies require indwelling
catheters, leading to increased rates of iliac and vena cava
thrombosis. The lack of adequate venous outflow can make
transplantation difficult and limit the standard options.10
Thomas and coworkers proposed a screening algorithm for
patients at risk, focusing on young children with a history of
femoral vein catheterization or history of any intra-abdominal
process associated with inflammation.11 In addition, patients
with a history of venous thrombosis, early graft loss, or recurrent vascular access thrombosis should be under evaluation for
a hypercoagulable state. Good results can be obtained in such
patients with anticoagulation prophylaxis.12
The potential need for native nephrectomy should be
addressed. Native nephrectomy is much more common in
children compared with adults. Nationally, 22% of children
have had all native renal tissue removed before transplantation.1 Potential indications for native nephrectomy include
recurrent severe infections because of reflux nephropathy,
uncontrolled hypertension, and congenital nephrotic syndrome. In the case of nephrotic syndrome, the concern centers
on these children being hypercoagulable because of the significant proteinuria. Children with polycystic kidney disease
may require native nephrectomies if there is recurrent bleeding, infections, or pain. In select cases, nephrectomy may be
warranted if the kidneys are so enlarged that the transplanted
kidney would be compromised. Some contend that massive
polyuria in small infants is an indication for nephrectomy, arguing that postoperative fluid management is made easier,
thus decreasing the risk of graft hypoperfusion. Although this
may have appeal, careful attention to postoperative management can often avoid this as the sole indication for nephrectomy. In addition, some believe that children with FSGS
should undergo native nephrectomy, suggesting it simplifies
the diagnosis of recurrent disease, because any proteinuria
reflects disease in the graft rather than persistent proteinuria
from the native kidneys. In such situations, an attempt at
medical nephrectomy with nonsteroidal therapy is worth
consideration.
CHAPTER 46
In addition to a rational consideration of the risks and benefits of native nephrectomy, the timing of the nephrectomy is
important. In children on renal replacement therapy, bilateral
native nephrectomies may be safer and easier to accomplish in
the weeks before transplantation. In children not on renal replacement therapy, the issue is more complex. We prefer not to
perform bilateral nephrectomies at transplantation because
this combines two major procedures. We also perform the
transplant through a retroperitoneal approach even in small
infants, which does not provide access to the contralateral native kidney. For children requiring bilateral native nephrectomy, and who are not yet on dialysis, some will have
sufficient renal reserve to tolerate a unilateral left nephrectomy
before transplantation and still not require dialysis. In this situation, at transplantation we extend the standard right retroperitoneal incision slightly cephalad and perform a right
native nephrectomy. In cases where unilateral native nephrectomy prior to transplantation would require initiation of renal
replacement therapy, we typically remove the ipsilateral native
kidney during the transplantation procedure. The remaining
contralateral native kidney can be removed several months after transplantation if still indicated.
In considering when to perform the transplantation, any
child currently on renal replacement therapy should undergo
transplantation as soon as a suitable living donor is identified
or a deceased donor organ becomes available. In children not
yet on dialysis, transplantation should be performed before
the onset of symptoms of uremia. It is important to be aware
of the impact of ESRD on growth and development. In patients
with FSGS or lupus nephritis, transplantation is typically
delayed until the disease is quiescent, which may preclude
preemptive transplantation. In most other situations, preemptive transplantation provides significant benefit by obviating
the need for dialysis. Unfortunately, only 33% of children
who receive a living donor transplant and only 13% of
deceased donor recipients are transplanted before initiation
of dialysis.13
Urologic Issues
------------------------------------------------------------------------------------------------------------------------------------------------
The high incidence of urologic issues in children requires

careful evaluation of bladder function before transplantation.14 In addition to dysplasia and bladder outlet obstruction,
bladder function may be abnormal because of neuropathy,
acquired voiding dysfunction, or acquired bladder pathology.
Any previous surgical bladder augmentation will impair
normal bladder function. A history of urinary incontinence,
frequent urinary tract infections, previous urologic procedures, and the need for bladder catheterization should prompt
further investigation. In patients with suspected bladder
dysfunction, a voiding cystourethrogram (VCUG) should be
obtained with urodynamic measurements. A pressure of less
than 30 cm H2O during the filling portion of the VCUG
generally indicates the bladder will be suitable.
The timing of any surgical intervention warrants careful
consideration. In some patients with anuria/oliguria, the bladder may not be functional, although it is often too early to tell if
it will eventually become suitable. Once bladder augmentation is performed, the patient will need to continue catheterizing long term, because the bladder will not have normal
RENAL TRANSPLANTATION
619
function. Urologic procedures that preserve native renal function for many years are clearly prudent, but interventions
before transplantation should be planned by carefully
considering the risks and benefits of the procedure and being
mindful of the impact on subsequent transplantation and
long-term management.
Dialysis Access
------------------------------------------------------------------------------------------------------------------------------------------------
For children who do not undergo preemptive transplantation

or who initially present with ESRD, establishment of adequate
dialysis access is of paramount importance. Proper dialysis access is necessary for adequate dialysis, which is directly linked
to the quality of life and health of the patient. According to
USRDS data, at the end of 2008, 60% of patients aged 19 years
and younger were on hemodialysis, whereas 40% were on
peritoneal dialysis.5 Both are suitable options, and the choice
is best made on an individual basis, considering the patients
and familys preferences and skill levels, as well as the treatment options available at the local site.
Regarding hemodialysis, all attempts should be made to
create a primary arterial venous fistula. For patients without
adequate veins, a polytetrafluoroethylene graft is required.
A native fistula is preferred because of superior patency rates,
but they require several weeks to mature before being
accessed. For patients in need of urgent hemodialysis, the only
option is a temporary catheter. Approximately three fourths of
pediatric patients have a temporary catheter at the time of initiation of dialysis.15 The use of these catheters is associated
with increased risks of infection and poor clearance with dialysis. In addition, catheters can lead to central venous stenosis and thrombosis, making future vascular access efforts more
difficult. Accordingly, the jugular vein is preferred rather than
the subclavian vein for catheter placement.
Peritoneal dialysis requires placement of a Tenckhoff catheter. A double-cuffed peritoneal dialysis catheter is inserted
with the loop of the catheter placed in the pelvis. During
the procedure, it is important to ascertain that fluid can instill and drain easily. The use of double-cuffed catheters
and orienting the catheter so that it exits the skin pointing downward are associated with a lower incidence of
infection.15
Donor Selection
------------------------------------------------------------------------------------------------------------------------------------------------
Living donor transplantation is the preferred option for all patients with ESRD. Living donor transplantation offers the best
outcomes, compared with deceased donor transplantation.4,16 In addition, living donor transplantation can be performed as soon as a suitable donor is identified, minimizing
exposure to ESRD. Living donors may be either genetically related or unrelated to the potential recipient. The results from
both types of living donors are equivalent, and both are superior to outcomes from deceased donors. Potential living donors should undergo a full evaluation by a transplant center
experienced in this process. The donor must be willing, be
in good health, and have two normal kidneys. In addition,
the donor and recipient must be ABO compatible. Although
there is a growing interest in strategies to cross this barrier,
efforts are relatively limited in the pediatric population.17
620
PART IV
TRANSPLANTATION
The recipient should also have a negative lymphocytotoxic

crossmatch with the potential donor. Crossmatching is done
to determine that the recipient does not have preformed antibodies directed against the donors human leukocyte antigens
(HLA), which would likely cause hyperacute rejection and
rapid graft loss. The most common causes of anti-HLA antibodies are blood transfusions, previous transplantation, and
pregnancy. Strategies to manipulate antidonor antibodies are
being investigated and include intravenous immunoglobulin,
plasmapheresis, and other agents designed to alter B-cell responses and/or complement.1821 Transplanting recipients
who are either ABO or anti-HLA antibody incompatible with
their donors requires additional immunologic manipulation,
with its attendant risks, and the long-term results appear
inferior compared with compatible transplantations. Accordingly, there is growing interest in paired kidney exchange
programs, which offer a larger living donor pool and the possibility of finding a more compatible donor.22
EVALUATION OF THE POTENTIAL

LIVING DONOR
Evaluation of potential living donors should occur independent of the recipients evaluation, giving donor safety the highest priority. Our standard evaluation process is summarized in
Table 46-3. Although HLA matching has traditionally played
an important role in choosing which living donor to evaluate,
current immunosuppression has minimized the impact of
matching, and we believe the best potential living donor is
the individual who is most motivated. Lacking that distinction, and all other factors being equal, we would choose the
donor with the best HLA match. It is also important to consider other issues unique to each donor, including psychosocial concerns (such as the need to care for other children), the
need to care for the recipient, and what options would be least
disruptive to the family unit. When discussing the situation
with the family, it is important to consider other siblings
who may also need renal transplantation in the future, because
this can play a role in deciding which donor donates to which
recipient.
Living-kidney donation appears to be safe and has been
practiced for more than 50 years. The risk of operative mortality
appears to be 3 in 10,000.23,24 After the procedure, living
TABLE 46-3
Evaluation of Living Kidney Donor
History and physical examination
Laboratory tests:
Hematologic (complete blood cell count with platelets; prothrombin
time/partial thromboplastin time)
Biochemistry (renal function, electrolytes, liver function)
Serologic studies (hepatitis herpesvirus B and C, cytomegalovirus,
Epstein-Barr virus, human immunodeficiency virus)
ABO blood typing
Tissue typing (human leukocyte antigen typing)
Urinalysis
Chest radiograph
Electrocardiogram
Psychosocial assessment
Helical computed tomography scan
kidney donors appear to do well over the long term as well.

The introduction of laparoscopic donor nephrectomy has been
a significant step forward for the individuals who consider
kidney donation. For the donor, the laparoscopic procedure
is associated with quicker recovery and appears as safe as
open-donor nephrectomy. Although there was some concern
that laparoscopic donation might result in inferior outcomes
compared with open-donor nephrectomy, particularly in small
infants, this concern has not been substantiated.2527
EVALUATION OF THE DECEASED DONOR

For children who do not have a living donor, deceased donor
transplantation is the only option. Deceased donors are individuals who have either suffered brain death or whose heart
has irreversibly stopped beating. The latter group has often
been referred to as DCD for donation after cardiac death.
In the United States, deceased donor organ allocation is governed by policies established by the OPTN. These policies undergo constant refinement as data support more rational and
fair allocation strategies. After a potential deceased donor is
identified, the blood type, HLA type, and other relevant donor
factors are entered into the national database maintained by
the United Network for Organ Sharing (UNOS). Kidney allocation is driven by a point system based on HLA matching, the
level of anti-HLA antibodies in the candidate, and waiting
time. Waiting times in many areas of the country for adults
are 5 or more years.
The OPTN has recognized the special needs of children with
end-stage organ disease and continuously reviews policy in an
effort to provide them optimal access to deceased donors. In
2005, the kidney allocation system was changed from a system
that provided pediatric priority points based on age and waiting
time to a system that provides relative priority to donors less than
35 years of age. At present, candidates who are listed at less
than 18 years of age are offered kidneys from donors less than
age 35 after any zero mismatch candidates, candidates with a
panel reactive antibody greater than 80%, or candidates receiving a kidney with a nonrenal organ. The rationale for this modification was based on the observation that pediatric waiting
times remained substantial despite the efforts to provide priority.
In addition, in an effort to optimize outcomes, centers were
waiting for the best donors. The donor age threshold was
based on an analysis demonstrating donor age between 5 and
35 years had the lowest relative risk of graft failure.28,29
The new policy has decreased pediatric waiting time while
maintaining access to the best deceased donors. As was
expected, the shorter waiting time has been associated with
lower HLA matching, reflecting the greater priority for transplantation compared with HLA matching. The decrease in
HLA matching is small, and there appears to be no adverse impact on outcomes.28 Waiting for a better-matched kidney is
not prudent, because there is no advantage, and it only delays
the benefit of transplantation. In addition, because of the relatively good access to the best deceased donors, the need to
use expanded criteria kidneys,30 DCD kidneys,31 and en-bloc
kidneys from small pediatric donors32 is less than in the adult
population. An important exception might be the highly
sensitized recipient who has waited for a long period of time.
In such situations, the decision needs to consider the risks and
benefits of the options available.
CHAPTER 46
The improved access to deceased donors has been associated with a decrease in the number of living donor transplantations performed in children. Prior to 2005, living donors
accounted for more than half of the pediatric transplantations.
This proportion has fallen to less than 40% in recent years and
is a source of concern. Although greater access to the best
deceased donors is appealing, it is important to note that
outcomes are significantly better with living donors compared
with deceased donors. Specifically, living donors up to
55 years of age provide greater long-term survival compared
with even the ideal deceased donor.16
Transplantation
------------------------------------------------------------------------------------------------------------------------------------------------
PREOPERATIVE PREPARATION
Deceased donor recipients are admitted once a kidney is
accepted. We typically also admit our living donor recipients,
although they can be admitted on the day of surgery if their
dialysis regimen is stable or if they are not on dialysis. On
admission, the need for dialysis is assessed. It is important
to ask about intervening health issues since the last visit, as
well as examining for any evidence of ongoing infection.
ANESTHESIA
Close coordination with the anesthesia team is vital to the conduct of any operation, and it is particularly important in kidney transplantation in small infants. Maintaining adequate
volume status is critical. Because a kidney from an adult donor
is typically used, blood flow to the graft often equals the entire
cardiac output of the recipient. Hypotension can be particularly problematic. Many children have an obligate polyuria
that can cause hypovolemia if not carefully monitored. After
reperfusion, the new kidney can also sequester several
hundred milliliters of blood, further aggravating hypovolemia.
OPERATIVE PROCEDURE
After induction of general anesthesia, adequate intravenous
access is established. In children larger than 20 kg, we do
not place a central venous line if adequate peripheral access
can be established. For smaller children, we find central
venous access useful, both for fluid administration as well
as for monitoring central venous pressure. In these smaller
children, we also place an arterial line to permit constant
blood pressure monitoring. The child is positioned supine.
A Foley catheter is inserted and connected to a three-way
irrigation system, using dilute providone-iodine (Betadine)
in saline. In other centers, an antibiotic solution may be preferred. This arrangement allows the bladder to be filled and
drained outside of the operative field as necessary. The childs
temperature should be monitored closely, especially with
small children who may become hypothermic with either
fluid resuscitation or the perfusion of the cold kidney. In
addition to routine monitoring, ongoing attention must be
directed to volume status. It is vital that the arterial blood
pressure and central venous pressure are adequate when
the kidney is reperfused. For infants and small children, the
central venous pressure is usually maintained in the range
of 12 to 18 cm H2O by administration of crystalloid and/or
621
colloid as necessary. Near completion of the vascular anastomoses, we typically give 0.5 mg/kg of mannitol intravenously.
We do not routinely employ a loop diuretic.
OPERATIVE TECHNIQUES
Small Children (<20.0 kg)
Historically, many have performed kidney transplantations
intra-abdominally in infants and small children using a midline incision. Since 1998, we have used a retroperitoneal approach similar to that used in adults, even in infants. Placing
the kidney on the right side is preferable, because this gives
the easiest access to the vena cava. A curvilinear skin incision
is made in the lower quadrant. The abdominal wall musculature is divided, and the preperitoneal space is entered. Attempts are made to stay extraperitoneal. The spermatic cord
is mobilized and preserved in males, whereas the round ligament is routinely divided in females. The dissection is carried
medially until the common iliac vessels, the distal aorta, and
vena cava are visualized. If a right native nephrectomy is necessary, this is performed at this point.
The site of the vascular anastomosis depends on kidney
size as well as the size of the child. In general, in small children, the renal vein is anastomosed to the vena cava, and
the renal artery is anastomosed to either the distal aorta or
the common iliac artery. The lymphatics overlying these
vessels are divided between ties in an effort to minimize the
risk of a lymphocele. When an aortic anastomosis is planned,
the aorta is mobilized from below the inferior mesenteric
artery to the bifurcation. Lumbar branches are controlled with
Pott ties rather than ligated. The common iliac arteries are
controlled just distal to the aortic bifurcation. The vena cava
is mobilized to allow placement of a side-biting vascular
clamp, which can require ligation of several lumbar veins.
Once the recipients vessels have been exposed, the donor
kidney is brought into the operative field. The kidney should
be inspected for any evidence of unsuspected pathology. The
renal vessels are examined. After preparing the kidney,
thoughtful consideration needs to be given for the fit of the
kidney in the recipients body cavity. Particular attention must
be focused on the length of the renal vessels as well as their
orientation. It is important to consider the final resting
position of the kidney after it is perfused, the retractor is
removed, and the fascia closed.
The venous anastomosis is performed first. The vena cava
or iliac vein is controlled with a side-biting clamp. A longitudinal venotomy is made along the anterolateral or lateral
aspect of the vein. The renal vein is cut to length, again after
considering the ultimate lie of the kidney, and mindful that a
redundant renal vein may predispose to thrombosis. We place
two corner sutures of 5-0 Prolene. The anastomosis is
performed in a running manner. An end-to-side arterial anastomosis is then performed to the recipients distal aorta or
common iliac artery. The recipient vessels are controlled using
vessel loops or gentle spring clips. A longitudinal arteriotomy
is made, mindful of the final orientation of the renal artery. We
enlarge the arteriotomy using a 4.0-mm aortic punch.
The renal artery is then sewn end-to-side using a running
6-0 Prolene suture. We typically perform the procedure with
loupe magnification.
622
PART IV
TRANSPLANTATION
If multiple renal arteries are present, they can either be

implanted separately or syndactylized before reimplantation.
When the vessels are syndactylized, it is important to consider
if this will allow the vessels to lie in good position, because
syndactylization will fix the vessels relatively firmly in two dimensions. This can limit the options of where the anastomosis
can be suitably performed or lead to kinking of one or both of
the donor arteries if the final position of the kidney is not
anticipated.
Before completion of the arterial anastomosis, the hemodynamic state of the patient should be considered. Intraoperative
assessment of the vascular volume by direct assessment of the
vena cava is possible. Mannitol is also given at this time. Because of the size of the adult kidney, it can be both slow to perfuse as well as sequester a significant volume of blood. The
anesthesiologist must be ready to give volume replacement
promptly as indicated. At this point, the clamp is removed
from the vein and bleeding assessed. Next, while the renal artery is gently compressed with vascular pickups, the arterial
clamps are removed, restoring distal blood flow. After a few
seconds, flow is established to the kidney. In small children,
we occasionally will briefly reclamp the recipients vessels distal to the arterial anastomosis to provide preferential flow to
the kidney. The field is carefully examined for bleeding. The
color and turgor of the graft are assessed. The renal artery
should have a good pulse, and a thrill can usually be appreciated as well. Both the lower and upper poles should be
assessed for perfusion. The renal vein should be full but not
tense, with a turgor similar to the vena cava. The lie of the kidney is again examined.
Attention is then directed to the ureteroneocystostomy. We
generally perform the ureteral anastomosis as an extravesical
ureteroneocystostomy,3335 although others routinely prefer
the transvesical Politano-Leadbetter approach.7,36 The Foley
catheter is clamped and the bladder is filled. A site on the
dome of the bladder is selected where the ureter will sit without any angulation. The muscle wall of the bladder is divided,
exposing the bladder mucosa. An opening is then made in the
mucosa. The donor ureter is trimmed to length. Care should
be taken to make sure it is sufficient to allow a tension-free
anastomosis, but excessive length should be avoided because
of the risk of ureteral obstruction or stricture resulting from
inadequate perfusion of the distal ureter. The end of the ureter
is spatulated, and a mucosa to mucosa anastomosis is performed using running 5-0 polydioxanone (PDS) suture. The
caveat with running suture is that care must be taken to avoid
cinching on the suture line, because this results in a pursestring effect causing stenosis. To prevent vesicoureteral reflux,
the bladder muscle wall is approximated over the anastomosis
using interrupted 4-0 PDS suture. This allows the ureter to
take a tangential course under the bladder wall so that during
micturition, the transvesical portion of the ureter is compressed by the overlying bladder wall. An adequate length
for this tunnel is essential to prevent vesicoureteral reflux.
In patients with a normal bladder and a good blood supply
to the distal ureter, we do not routinely place a stent. If there
is any concern regarding the ureteral anastomosis, either because of the donor ureter or the quality of the recipients bladder, we place a double-J ureteral stent, which is removed after
a few weeks as an outpatient procedure.
After completing the ureteral anastomosis, the kidney is
again inspected with attention to the renal vessels and the
lie of the kidney once the retractor is removed. Careful planning and attention to detail before performing the anastomosis
is usually rewarded at this point. The fascia is closed in one
layer with a running suture. The skin is closed using a running
absorbable suture. The urinary catheter is flushed with saline
to remove any clots that might obstruct the catheter. For small
infants, the volume resuscitation required to ensure excellent
renal perfusion, combined with the size of the kidney decreasing respiratory excursion, may make ventilatory support in the
immediate postoperative period necessary. If the patients oxygen saturation and pulmonary mechanics are satisfactory, the
patient can be extubated in the operating room.
Larger Children (20.0 kg)
The technique for transplantation in larger children is similar
to that in adults. We prefer to put the kidney on the right side
when possible. An incision is made in the right lower quadrant, extending from one to two fingerbreadths above the pubis to just lateral of the rectus sheath. As in smaller children,
the placement of the arterial and venous anastomoses depends
on the size of the child and the renal vessels. The venous anastomosis can be done to the vena cava, the common iliac, or the
external iliac vein. The arterial anastomosis is performed to the
distal aorta, the common iliac, or the external iliac artery. After
revascularizing the kidney, the ureteroneocystostomy is performed using an extravesicular technique. At the completion
of the operation, these larger children are extubated.
Ureteral Reconstruction in Patients with Previous
Urologic Procedures
The ideal urinary reservoir stores a reasonable volume at a low
pressure, does not leak, and empties nearly completely with
voiding.14 In the majority of cases, the ideal reservoir is the
patients bladder. If the bladder functioned normally before
development of oliguria, it is likely to function adequately
after transplantation. Nonetheless, up to 30% of pediatric recipients will not have normal bladder function, and frequently
a surgical augmentation or other urologic procedure has been
performed before referral for transplantation.
Drainage into an augmented bladder or urinary conduit is
an appropriate management strategy when the native bladder
is unsuitable or absent.37,38 When indicated, we prefer to have
the intended urinary reservoir created and suitable for use before the transplant procedure. Intraoperatively, when planning the ureteroneocystostomy to an augmented bladder, it
is important to consider the blood supply to the augmented
section so as not to compromise it during the transplant. It
is preferable to perform the ureteroneocystostomy to the native bladder, and this can be accomplished in most situations.
An antireflux ureteroneocystostomy is essential, and it is most
readily performed with the bladder wall.
Patients with an augmented bladder or urinary conduit are
at increased risk for urine infection, but compared with historical controls, graft survival is not adversely affected.39 The rate
of surgical complications related to the ureteral anastomosis is
higher in these patients, approximately 20%.3941 Regardless
of the etiology of the bladder dysfunction, these patients require regular clean intermittent straight catheterization after
transplantation.
Children with obstructive uropathy from posterior urethral
valves will not have normal bladder function, and this can
contribute to renal dysfunction after transplantation.37
CHAPTER 46
Awareness of these issues is vital, and evaluation with followup urodynamic studies is frequently indicated in children
with voiding disorders. Bladder dysfunction, such as hypocompliance and/or hyper-reflexia, requires medical or surgical
treatment.
623
children leave the hospital 5 to 7 days after transplantation,

assuming the family is comfortable with the immunosuppression regimen.
Postoperative Care
Evaluation of Early Allograft

Dysfunction
------------------------------------------------------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------------------------------------------------------
Attention to detail in the postoperative period is essential. Special care must be directed to the fluid and electrolyte status.
Many children are polyuric before transplant, and this obligate
urine loss will continue in the immediate postoperative period. Intravenous fluids are administered, taking into account
urine output as well as insensible losses. The composition of
these solutions is adjusted as needed, depending on regular
measurement of serum electrolytes. Serum sodium, potassium, and calcium levels are followed closely and replaced
as necessary. Heart rate, blood pressure, and central venous
pressure are carefully monitored. No single factor alone is
entirely reliable in assessing intravascular volume.
For patients who were oliguric or who had native nephrectomies before transplantation, monitoring urine output is an
excellent monitor of graft function. For patients who made significant urine before transplantation, evaluation of graft function is more difficult. The volume of urine production may be
suggestive. In addition, the serum creatinine concentration
should fall with time. Recipients with oliguria should be rapidly evaluated. The urinary catheter should be flushed with
small volumes of sterile saline. The volume status of the
patient should be carefully assessed. A fluid bolus is usually
warranted, both as a diagnostic test and as therapeutic intervention. Doppler ultrasonography will confirm adequate arterial flow and venous outflow. Ultrasonography will also show
evidence of fluid or blood around the kidney, as well as assess
for possible ureteral obstruction. In patients who appear to
be adequately volume loaded and hemodynamically stable,
a dose of diuretic can be given. It is important to do this
carefully, because sudden massive urine output can cause
significant intravascular volume depletion, which can then
lead to problems with renal perfusion. In patients who are
massively volume overloaded or have significant electrolyte
abnormalities, dialysis may be indicated.
If ventilated postoperatively, the smaller children are
weaned from the ventilator generally within the first 24 hours.
Enteral feedings can be started at a slow rate almost immediately after the extraperitoneal approach. Infants who were on
tube feedings before transplantation should resume these tube
feedings, because they usually will not feed orally in the
immediate post-transplantation period. Hypertension can be
problematic. The volume loading associated with the procedure as well as the use of calcineurin inhibitors (CNIs) for
immunosuppression can result in significant hypertension,
which can be severe and require aggressive therapy to prevent
seizures and other sequelae.
To monitor and replace urine output on an hourly basis, we
admit our children to the intensive care unit (ICU). If this can
be accomplished on a surgical floor unit, larger children could
be admitted to an area specializing in the care of renal transplant patients. Children who are admitted into the ICU are
typically transferred to the floor unit within 1 to 2 days. Most
Ideally, the donor kidney should begin to make urine shortly

after revascularization. The likelihood of this occurring depends on multiple factors, beginning with the quality of the
donor organ. Living donor kidneys will generally function immediately because of the healthy state of the donor as well as
the shorter cold ischemic time for the kidney. For deceased
donor kidneys, the cold ischemic time is generally longer.
In addition and more important, there are multiple factors
associated with the donor death, including hypotension, the
potential need for high doses of vasopressors, and other issues
related to the overall health of the donor.
Regardless of the donor source, the assessment of the graft
begins in the operating room, evaluating the graft for color and
turgor as well as vascular anastomoses. Particular attention
should be directed to considering how the kidney is positioned once the abdomen is closed and how this could impact
the vasculature. The renal artery will often have a thrill suggestive of excellent flow and low intrarenal resistance. Assuming
the technical aspects of the procedure appear satisfactory, additional volume for the kidney not making urine is the best
option. Once the patient is adequately volume loaded, loop
diuretics may be used to gently encourage a diuresis. In patients who were anuric before the procedure, continued failure to make urine in the postoperative period should prompt a
bedside Doppler ultrasound examination. For patients who
made urine before the transplantation, determining whether
the transplanted kidney is making urine is more difficult, although sometimes the amount of volume being produced will
give a sign that the kidney is working. During the first
24 hours, the serum creatinine level should fall as well. If there
is still concern about function, a Doppler ultrasound study
should be obtained, and any suggestion of problems with
the arteriovenous signal should initiate a prompt return to
the operating room. In general, the ultrasound evaluation will
be fine, or, occasionally, there will be a modest reduction in
flow suggestive of increased intrarenal resistance, most commonly because of acute tubular necrosis. This condition resolves without any specific intervention. Other diagnostic
studies are less frequently required. Renal arteriography is
rarely indicated. Radionucleotide scans are used by some centers, but we find them less helpful than ultrasonography.
A radionucleotide scan may be helpful in documenting a
suspected urinary leak.
Complications related to the ureteral anastomosis include
leaks and obstruction. The risk of ureteral complications is
approximately 7% to 9%.7,42,43 A leak at the ureteral anastomosis generally manifests in the first few days after the transplant. Leaks detected in the first 2 to 3 days may be repaired
operatively. Leaks detected later can generally be managed
nonoperatively with a percutaneous nephrostomy. This stent
is subsequently advanced across the anastomosis into the
bladder. The stent is usually left to external drainage for
624
PART IV
TRANSPLANTATION
several days and is then capped. If a large urinoma is present,

separate drainage of this collection may be required.
Obstruction of the urinary system can occur at any time. An
early obstruction is usually related to technical problems with
the anastomosis or other mechanical issues, such as torsion of
the ureter, while later obstruction often reflects ischemic stricture. Because the ureter relies on small arterial vessels from the
lower pole of the kidney, it should be no longer than necessary.
Late ureteral stenoses generally require operative intervention,
with resection of the stenotic segment and reconstruction.
Vesicoureteral reflux may cause recurrent urinary tract infections or graft dysfunction, and require operative intervention
in up to 5% of patients.44
Another complication after retroperitoneal kidney transplantation is lymphocele. Lymphoceles may produce discomfort or allograft dysfunction. The diagnosis is established by
ultrasound-guided percutaneous aspiration of clear fluid with
a creatinine concentration equivalent to serum. Percutaneous
drainage is associated with a very high incidence of recurrence, and the preferred treatment is creation of a peritoneal
window. This can be accomplished laparoscopically or
through a small open incision, with drainage of the lymphocele into the peritoneal cavity.
In instances of renal vein thrombosis, the graft is usually
not salvageable. The causes of renal vein thrombosis are several, and the exact mechanism may be difficult to ascertain but
include immunologic factors, a hypercoagulable state, and
technical issues.
Immunosuppression
------------------------------------------------------------------------------------------------------------------------------------------------
Significant advances have been made in understanding the immune response and several new immunosuppressive agents
have been developed. The introduction of new agents has permitted consideration of avoidance, conversion, and minimization strategies in an effort to minimize toxicities associated
with specific agents. There are several potential regimens,
but all require a balance between prevention of rejection
and unwanted side effects of immunosuppression. Most
centers use standardized protocols for recipients based on
immunologic risk. Immunosuppressive agents are used for
induction, maintenance, and treatment of rejection episodes.
ANTIBODY PREPARATIONS
Antilymphocyte Antibodies
Antilymphocyte antibodies include polyclonal preparations,
such as equine antithymocyte globulin (ATGAM) and rabbit
antithymocyte globulin (Thymoglobulin), and the monoclonal antibody preparations muromonab-CD3 (OKT3) and
anti-CD52 (Alemtuzumab). Of these antilymphocyte agents,
Thymoglobulin is currently predominant, and the use of the
others is either rare or of historic interest. Antilymphocyte
antibodies act by lymphodepletion, as well as by interactions
with cellular receptors. The use of antilymphocyte induction
regimens has declined precipitously with time.1,2 We
currently restrict the use of Thymoglobulin to recipients at
higher risk for immunologic graft loss, such as patients
requiring retransplantation, highly sensitized patients, or
black recipients. These agents are also effective in the
treatment of acute cellular rejection.
Antiinterleukin-2 Receptor Monoclonal

Antibodies
Two monoclonal antibodies have been developed that bind to
the alpha subunit of interleukin (IL)-2 receptor (CD25) and
inhibit IL-2mediated lymphocyte proliferation. Basiliximab
(Simulect) and daclizumab (Zenapax), received approval by
the U.S. Food and Drug Administration (FDA) in 1998,
though only basiliximab is currently marketed. Basiliximab
is a chimeric human/mouse monoclonal antibody that is effective in reducing the incidence of acute cellular rejection, with
good long-term results and no evidence of increased risk of
infection or malignancy.4547 The IL-2 receptor antibody is
used in induction regimens but is not effective in treating
rejection.
CALCINEURIN INHIBITORS
The introduction of cyclosporine after its FDA approval in
1983 was one of the most significant advances in transplantation. Tacrolimus received FDA approval in 1994. Both agents
act through inhibition of calcineurin activity. They first bind to
specific cytoplasmic proteins; cyclosporine binds to cyclophilin, and tacrolimus binds to tacrolimus binding protein (also
known as FK-binding protein). Both drug-protein complexes
then bind to calcineurin, a phosphatase that controls the
transport of transcriptional regulator factors across the nuclear
membrane. By inhibiting the translocation of these factors into
the nucleus, both drugs inhibit transcription of several early
T-cell activation genes, most significantly IL-2.
Both cyclosporine and tacrolimus are effective at preventing rejection. A randomized prospective open-label trial performed in Europe in pediatric renal recipients compared
tacrolimus with cyclosporine, along with azathioprine and
steroids. There was a significantly lower incidence of acute
rejection in the tacrolimus group (36.9%) compared with
cyclosporine therapy (59.1%).48 In contrast with this observation, a retrospective analysis of NAPRTCS data comparing
cyclosporine with tacrolimus, along with mycophenolate
mofetil and corticosteroids, showed equal rates of rejection
and graft survival. Although rejection rates were similar, tacrolimus therapy was associated with improved graft function at
1 and 2 years after transplant.49 Currently, approximately 70%
of pediatric recipients are reported as being discharged on
tacrolimus compared with 10% on cyclosporine.1,2
Cyclosporine side effects include hirsutism and gingival
hyperplasia, whereas tacrolimus is associated with increased
incidence of post-transplantation diabetes and neurotoxicity.
In children who develop a problematic side effect from one
agent, conversion to the other agent is appropriate. Both
calcineurin inhibitors have significant nephrotoxicity that
impact graft function with time.
MYCOPHENOLATE
Mycophenolate mofetil (MMF) (CellCept) and mycophenolic
acid (MPA) (Myfortic) inhibit purine synthesis. MMF is
converted in vivo to mycophenolic acid, a noncompetitive inhibitor of inosine monophosphate dehydrogenase (IMPDH),
which is a key enzyme in de novo purine biosynthesis.
Although most cells can synthesize purines by either the de
novo or the salvage pathway, B and T lymphocytes lack the
CHAPTER 46
salvage pathway. Mycophenolate is thus a selective inhibitor of

lymphocyte proliferation, and it has replaced azathioprine
(Imuran) as the primary antiproliferative agent.50 MMF has
been demonstrated to be safe and effective in pediatric patients.51 Experience with MPA in children is more limited,
though it appears equivalent.52,53 The primary side effects
are related to leukopenia and gastrointestinal intolerance.
PREDNISONE
Glucocorticoids have played an integral role in immunosuppression regimens since the earliest days of transplantation.
They act primarily through transcriptional regulation, diffusing across the plasma membrane and binding to cytoplasmic
steroid receptors. This complex is translocated to the nucleus,
where it binds to specific gene promoters and other regulatory
regions, inhibiting cytokine synthesis. Corticosteroids are also
lymphocytotoxic and possess significant anti-inflammatory
activity, inhibiting macrophage function and other nonspecific
aspects of the inflammatory response.
Long-term corticosteroid therapy is associated with increased risk of hypertension, hyperlipidemia, diabetes, bone
loss, cosmetic disfigurement, and cataracts. Attempts at minimizing corticosteroids have not had a significant effect on
these side effects, and efforts are being directed to corticosteroid avoidance. Although it is appealing to consider withdrawal of corticosteroids with time, late corticosteroid
withdrawal appears associated with increased risk of acute
and chronic rejection. Early corticosteroid withdrawal and
corticosteroid-free regimens, with and without antibody
induction, have shown promise. Sarwal and associates, at
Stanford University, have reported excellent results in a
corticosteroid-free protocol using an extended induction with
daclizumab, tacrolimus, and MMF.54 In their recent report,
129 recipients have been treated with a mean follow-up of
5 years. One-year graft and patient survival were 93% and
96%, respectively. The rate of acute rejection was 12% during
the first year. Significant improvements in post-transplantation growth and avoidance of steroid side effects were noted.
This experience led to a prospective, multicenter randomized
study that has been completed, though the results have not yet
been published. Similar results have also been reported by the
Stanford group in 13 recipients using Thymoglobulin induction in place of daclizumab.55
Another large randomized multicenter international trial
with 196 pediatric kidney recipients compared rapid steroid
withdrawal in children treated with daclizumab, tacrolimus,
and MMF with recipients maintained on steroids along with
tacrolimus and mycophenolate.56 Early steroid withdrawal
was associated with improved growth and metabolic profiles,
with similar acute rejection rates (10.2% vs. 7.1%, respectively) and equivalent graft and patient survival during the first
6 months.
Although encouraging, efforts to withdraw steroids while
maintaining acceptable rejection rates have also resulted in
regimens with a greater rate of complications of immunosuppression. One multicenter randomized trial of steroid withdrawal after 6 months in recipients treated with basiliximab
induction, cyclosporine or tacrolimus, sirolimus, and steroids
was halted early because of a high rate of post-transplantation
lymphoproliferative disorder (PTLD).57
625
It appears steroid avoidance or early withdrawal is possible

in selected patients with good short-term results. Longer-term
data are needed, and striking the correct balance of immunosuppression and other side effects remains critical.
PROLIFERATION SIGNAL INHIBITORS

Proliferation signal inhibitors are a relatively new class of immunosuppressants. Sirolimus (rapamycin) and everolimus are
macrolide agents that inhibit a protein, mammalian target of
rapamycin (mTOR), which is a critical regulatory kinase controlling cytokine-mediated proliferation. A potential role for
sirolimus in renal transplantation has been established.58
Everolimus also appears effective.59 There is interest in using
mTOR inhibitors to avoid or minimize calcineurin inhibitors
and/or corticosteroids, though the optimal strategy remains
elusive. Experience to date suggests that complete CNI avoidance is often associated with higher rejection rates, while late
CNI replacement or minimization may not offer any benefit
with respect to reversing CNI nephrotoxicity.
Sirolimus interacts with calcineurin inhibitors, particularly
cyclosporine, and careful monitoring is essential. Like many
other immunosuppressive agents, there is evidence to suggest
more rapid metabolism of sirolimus in children compared
with adults.60
TREATMENT OF REJECTION
Suspected rejection should be confirmed by biopsy. The
first-line therapy for acute cellular rejection is pulse corticosteroids. Typically, intravenous methylprednisolone is administered for 3 days, with doses ranging from 5.0 to
25.0 mg/kg/day (maximum dose, 1.0 g). We use 10.0 mg/kg
for children younger than aged 6 years and 5.0 mg/kg for children aged 6 years and older, with a maximum dose of 500 mg/
day. Severe rejection or rejection refractory to corticosteroids
is treated with Thymoglobulin. Treatment of acute rejection is
nearly always successful, although late episodes of rejection
are less likely to respond. After successful treatment, many
consider altering maintenance immunosuppression, including changing to the other calcineurin inhibitor, or substituting
sirolimus for MMF; however, there is little evidence to support
this approach. An assessment of adherence to the immunosuppression regimen should also be initiated. If the patients
creatinine level does not return to baseline, a follow-up biopsy
should be strongly considered. Although most acute rejection
episodes reflect primarily T-cellmediated processes, there is
growing recognition of the role of B cells and alloantibodies
in immunologically mediated graft injury.
Outcomes
------------------------------------------------------------------------------------------------------------------------------------------------
GRAFT AND PATIENT SURVIVAL

There are approximately 800 pediatric kidney transplantations performed annually in the United States. In 2009, there
were 866 transplantations, with 38.8% living donor and
61.2% deceased donor kidneys. Short-term graft and patient
survival after transplantation is excellent. Current graft survival for living donor and deceased donor kidney transplantations in children, stratified by recipient age, is summarized in
626
PART IV
TRANSPLANTATION
Living donor
1
5
year year
100%
98.7%
96.2%
100%
96.1%
96.0% 94.1%
Deceased donor
98.0% 96.7%
97.2% 93.9%
611
1217
91.6%
80%
Graft survival
75.6%
60%
Graft survival
85.4%
80%
60%
40%
20%
40%
0%
20%
0%
Recipient age
15
611
1217
Recipient age
1
5
year year
100%
94.9%
91.7%
Graft survival
80%
78.2%
74.5%
64.4%
60%
40%
0%
15
FIGURE 46-2 Five-year patient survival of living donor and deceased

donor kidney transplantations by recipient age. (From Organ Procurement Transplant Network/Scientific Registry of Transplant Recipients:
2009 Organ Procurement Transplant Network/Scientific Registry of
Transplant Recipients Annual Report: Transplant data 1999-2008. Health
and Human Services/Health Resources and Service Administration/
Special Pathogens Branch/Department of Transportation. Available at
www.ustransplant.org. Accessed September 29, 2010.)
93.4%
20%
15
611
1217
Recipient age
FIGURE 46-1 A, One- and 5-year graft survival of living donor kidney
transplantations by recipient age. B, One- and 5-year graft survival of
deceased donor kidney transplantations by recipient age. (From Organ
Procurement Transplant Network/Scientific Registry of Transplant
Recipients: 2009 Organ Procurement Transplant Network/Scientific
Registry of Transplant Recipients Annual Report: Transplant data
1999-2008. Health and Human Services/Health Resources and Service
Administration/Special Pathogens Branch/Department of Transportation.
Available at www.ustransplant.org. Accessed September 29, 2010.)
Figure 46-1. Recipient survival stratified by age range is summarized in Figure 46-2. The leading causes of death are infection (28.9%), cardiopulmonary (15.7%), and malignancy
(11.0%).2 Although patient survival is good, it is important
to realize that even with transplantation, these children face
a significantly increased risk of mortality compared with the
general population.61
With time there has been improvement in outcomes for all
pediatric age ranges. This improvement is particularly noteworthy in children younger than 2 years of age who previously
had the worst graft survival but now have outcomes that equal
the outcomes of any age group.13,50 In fact, the longest transplant half-lives of all recipients are now the youngest recipients,
especially if the pediatric recipient receives an adult kidney that
functions immediately.62 These improvements likely reflect
better donor selection, improvement in surgical techniques,
better immunosuppression agents, and a better understanding
of immunosuppression management in children.
Although short-term graft survival in children is excellent, it
is important to appreciate that long-term graft survival in the
adolescent group (ages 11 to 17 years) is poor (see Fig. 46-2).

Adolescent graft survival is less than in all recipient age groups
except adults older than 65 years.4 The reasons behind this
significant rate of graft loss are speculative, but noncompliance
likely plays a significant role.63 The higher incidence of recurrent
FSGS in this age group may also contribute to graft loss. Regardless of the cause, improving the long-term outcomes in this
patient population represents an important focus.
POST-TRANSPLANTATION OUTCOMES
AND RISK FACTORS ASSOCIATED
WITH GRAFT LOSS
Smith and coworkers reported that the most common causes of
allograft failure reported to NAPRTCS for transplantations performed between 2000 and 2005 are chronic rejection (41.3%),
vascular thrombosis (8.1%), recurrence of the primary disease
(7.9%), acute rejection (6.3%), and discontinuation of immunosuppression (6.3%).2 Analysis of large single-center experiences and registry data has revealed risk factors associated
with specific post-transplantation outcomes.2,13,64 For a given
child, some of these risk factors, such as their race, age, or
primary disease, are not modifiable. Other factors are potentially modifiable, and efforts should be made to mitigate risk.
The type and timing of the transplant affect outcomes.
Living donor transplantation is associated with better graft
survival compared with deceased donor transplantation.2,4
Preemptive transplantation is associated with better graft survival compared with patients on dialysis at the time of transplantation. For children on dialysis, the choice of dialysis
therapy does not impact graft survival, although graft loss
from vascular thrombosis is more common in children on
peritoneal dialysis compared with hemodialysis.65
DELAYED GRAFT FUNCTION

Delayed graft function (DGF) is defined as the need for dialysis
during the first week after transplantation and is a manifestation of acute tubular necrosis (ATN). DGF is more common in
recipients of deceased donor kidneys compared with living
donors, because of the impact of cold ischemic time and donor
CHAPTER 46
quality. Nationally, the incidence of DGF after deceased donor

renal transplantation is 12.5% in pediatric recipients compared with 23.4% in adult recipients,50 reflecting the differential donor selection made possible by the allocation system. An
analysis of more than 5000 pediatric transplantations has
demonstrated that DGF is an independent risk factor for subsequent graft loss.66 DGF is also associated with increased risk
for acute and chronic rejection, likely reflecting the impact of
renal injury on the subsequent immune response. DGF limits
the ability to use renal dysfunction as a sign of acute rejection,
potentially delaying the diagnosis of rejection. We believe that
all recipients with DGF longer than 5 to 7 days should be
biopsied, and even earlier in patients with increased risk of
rejection.
VASCULAR THROMBOSIS
Vascular thrombosis is currently the second most common
reported cause of graft loss.2 Risk factors include donor age
younger than 6 years, cold ischemic time greater than
24 hours, prior transplantation, and peritoneal dialysis before
transplantation.65 Careful consideration of donor quality,
along with efforts to ensure adequate perfusion to the graft,
may minimize the risk of thrombosis. Patients with ESRD have
a higher incidence of hypercoagulable conditions, and any
history of thrombosis, including recurrent or unexplained
thrombosis of hemodialysis access, should prompt further
evaluation.
ACUTE REJECTION
Acute rejection typically occurs between 1 week and 3 months
after transplantation, although it can happen at any time. A
rise in the serum creatinine level is frequently the first sign
of rejection. Findings such as low-grade fever, graft tenderness, hypertension, or decreased urine output are infrequent.
Any renal dysfunction should be promptly investigated. A
percutaneous biopsy should be obtained to confirm the diagnosis, because many other processes can lead to allograft dysfunction, including calcineurin inhibitor toxicity, ureteral
obstruction, infection, renal artery stenosis, and recurrence
of original disease. Acute rejection episodes are treated by
either pulse corticosteroids or antilymphocyte antibodies as
detailed previously.
Risk factors associated with acute rejection include AfricanAmerican race, delayed graft function, and a history of allosensitization. Acute rejection, and, in particular, late acute
rejection episodes occurring more than 1 year after transplant, are independent risk factors for graft loss because of
chronic rejection.67 One episode of acute rejection increases
the risk of graft loss from chronic rejection graft failure threefold, and two episodes of acute rejection increase the risk
12-fold. The incidence of acute rejection is decreasing with
time. In the 2003 to 2005 NAPRTCS cohort, 12.2% of living
donor recipients and 15.8% of deceased donor recipients
had a rejection episode in the first year after transplant.2
Acute rejection, even if successfully treated, impacts graft
survival and all efforts to minimize this risk are important.
Unfortunately, intensifying the immunosuppression regimen
is limited by the consequences of nonspecific systemic immunosuppression. Ensuring the patient remains on therapeutic
immunosuppression is vital, because noncompliance can
627
be disastrous. Prompt recognition and treatment of rejection

is important. Because serum creatinine is a relatively
insensitive indicator, particularly in small children with an
adult kidney, many advocate protocol biopsies to detect
subclinical rejection that may benefit from treatment.
CHRONIC ALLOGRAFT NEPHROPATHY

Whereas short-term results are excellent, progressive renal
dysfunction frequently occurs and is the leading cause of graft
failure. This process of chronic allograft nephropathy, often
called chronic rejection, involves both immunologic and
nonimmunologic factors. Although acute rejection episodes
are a major risk factor for chronic allograft nephropathy, it
is clear other processes can contribute as well. Evidence of
antibody-mediated injury is also present in 57% of patients
with late allograft dysfunction.68 Efforts to reduce chronic
allograft nephropathy are limited by our understanding of
the process. Aside from graft loss, the gradual renal dysfunction associated with chronic allograft nephropathy also
adversely impacts the recipients general health.
NONADHERENCE
Adherence with the medical regimen is essential for the success of transplantation. Nonadherence is believed to be largely
responsible for the poorer long-term graft survival seen in adolescent recipients. Shaw and coworkers reviewed 112 pediatric renal transplant recipients and found one third had
clinically significant periods of medication nonadherence.69
Nonadherence was significantly more common in adolescents
compared with younger recipients. Nonadherence was associated with both acute and chronic rejection, as well as graft
loss. The relative lack of reliable measures of adherence and
effective interventions has focused research in the field.70,71
Improved parental involvement and discussion of the child
parent relationship may improve adherence.
RECURRENT DISEASE
The recurrence of the patients primary disease is variable, and
recurrence may or may not lead to graft loss. Recurrent disease
is a more significant issue in the pediatric population, because
of the diagnoses leading to ESRD and their association with
higher rates of graft loss after recurrence. FSGS is the most
prevalent and clinically significant disease to recur after renal
transplantation. In children, the recurrence rate can be as high
as 40% to 50%.72 It can recur almost immediately after
transplant, and most recurrences are within the first month.
Patients with FSGS should be followed closely after transplantation with urine protein measurements. Graft survival is often
worse in adolescents with recurrent FSGS, with up to a 38%
risk of graft loss.73 A circulating permeability factor is believed
to play a critical role in the pathogenesis of FSGS. Plasmapheresis is the most frequently used therapy for recurrence,
although controlled trials supporting its efficacy are lacking.
Some have proposed a role for preoperative plasmapheresis
to decrease the risk of recurrence.74 Others have suggested
a role for intensifying the immunosuppression and potentially
rituximab. In addition to FSGS, other primary renal causes
associated with recurrent disease include membranoproliferative glomerulonephritis types 1 and 2 and IgA nephropathy.75
628
PART IV
TRANSPLANTATION
Again, the risk of graft loss is variable, and none constitute an

absolute contraindication to transplantation.
In addition to these primary glomerulopathies, other
recurrent diseases disproportionately affect the pediatric population. Hemolytic-uremic syndrome can recur after transplantation. Nearly all the risk of recurrence and subsequent
graft loss is in those with atypical nondiarrhea-associated
hemolytic-uremic syndrome.76 The risk of recurrence appears
related to specific defects of complement activation, and
screening for these defects is recommended pre-transplant.77
Henoch-Schonlein purpura can also recur. The overall risk
of renal recurrence after transplantation is 29%, and the
risk for graft loss appears equivalent to that observed in IgA
nephropathy.78
Oxalosis (primary hyperoxaluria type 1) is a metabolic disease caused by a defect in hepatic peroxisomal alanine:glyoxylate aminotransferase, which leads to increased synthesis and
excretion of oxalate. The excessive oxalate load leads to
urolithiasis, medullary calcinosis, and eventual ESRD. The
primary metabolic defect is not corrected by kidney transplantation, and the persistent oxalate load causes subsequent renal
graft loss. Simultaneous liver-kidney transplantation is
generally advocated as the primary treatment.79,80 Kidney
transplantation alone is uncommon, but has been advocated
in selected patients, most notably those who are pyridoxine
sensitive or those with lower oxalate burdens.81
MEDICAL COMPLICATIONS
Infection
Infection is a constant risk of immunosuppression and is
one clinical representation of the precarious balance between overimmunosuppression and underimmunosuppression.
Great vigilance should be maintained during periods of heaviest immunosuppression, as occurs immediately after transplantation or during treatment of rejection. Additional
prophylaxis is warranted during these periods of greatest risk.82
Post-transplantation infection accounts for more hospitalizations than acute rejection, even in the first 6 months after
transplantation.83 Post-transplantation infections are predominantly bacterial and viral. Fungal infections, although
accounting for 0.2% to 2.7% of infection-related hospitalizations, can be particularly dangerous. Pediatric recipients are
often at higher risk, reflecting the fact that they are more likely
to be nave to a particular pathogen than the general population. Younger age and the use of antibody induction immunosuppression are significant independent risk factors for
infectious complications.84
Cytomegalovirus
Cytomegalovirus (CMV) represents the most common viral infection after transplantation. CMV infection can occur in any
recipient, although the risk is highest when a seronegative recipient receives a kidney from a seropositive donor. Infection
occurs in seropositive recipients as well because of activation
of latent virus. The incidence and the severity of CMV have
declined with more effective prophylaxis. The severity of
CMV infection may range from asymptomatic to organ involvement and death. The typical presentation occurs 1 to
6 months after transplantation, with the patient feeling relatively well but having fevers or sometimes flulike symptoms.
Leukopenia is common. Patients with tissue-invasive CMV

disease will appear toxic, and there will be evidence of endorgan dysfunction. The diagnosis is confirmed using either
a CMV pp65 antigenemia assay or the CMV polymerase chain
reaction (PCR) assay. Both methods allow monitoring of
the response to therapy. Valganciclovir is effective for both
the prophylaxis and treatment of CMV disease.85 In more
severe cases, treatment with intravenous ganciclovir and
CMV hyperimmune globulin may be helpful.
Varicella-Zoster Virus
In pediatric recipients, there is high risk of a primary chickenpox infection. Treatment is with intravenous acyclovir until
the lesions crust over, then conversion to oral acyclovir. Primary infections can be severe. We immunize our candidates
who are seronegative for varicella-zoster virus (VZV) before
transplantation. For seronegative recipients who have a
defined exposure, we administer VZV immune globulin.
BK Virus
BK virus is a ubiquitous polyomavirus that is a significant concern in renal transplantation.86 There is a high incidence of
seroconversion by late childhood, and the virus is dormant
in the renal epithelium until reactivated. BK virus appears
to be an under-recognized cause of allograft dysfunction, with
BK interstitial nephritis resulting in a graft loss in 45% to 70%
of affected recipients.
The incidence of BK virus-associated transplant nephropathy is estimated to be 4% to 7%.86 Smith and coworkers evaluated a single-center cohort of 173 pediatric renal recipients
and identified BK nephropathy in 6 children (3.5%).87 The diagnosis was made on biopsy at a median of 15 months after
transplantation. There was a strong association between BK nephropathy and recipient seronegativity. In a subsequent analysis
of the NAPRTCS database from 2000 to 2004, the incidence of
BK nephropathy was 4.6% with a median onset of 10.1 months
post-transplantation. Graft failure occurred in 24% of patients
at a mean of 24 months after diagnosis. There was an association with polyclonal antibody induction therapy.88
BK nephropathy should be considered in the evaluation of
renal allograft dysfunction. BK nephropathy can be definitively diagnosed on biopsy using immunohistochemistry,
but the histology can be confused with acute rejection. Treatment with additional immunosuppression does not improve
renal function and often will cause further deterioration.
The initial treatment for BK nephropathy consists of decreasing immunosuppression. The addition of antiviral therapy
and intravenous immunoglobulin has been reported, but
there are no controlled trials and compelling data are lacking.89 Measurement of BK virus by PCR helps with diagnosis
and subsequent monitoring of the response to treatment. It is
hoped that improved awareness, prompt diagnosis, and treatment may reduce the risk of graft loss initially associated with
this disease process.
Malignancy
Transplant recipients face an increased risk of de novo malignancy related to their immunosuppression. Lymphomas,
specifically post-transplantation lymphoproliferative disorder
(PTLD), are the most common, with an incidence of 1% to
4% in renal transplantation.9092 PTLD actually represents
a spectrum of pathology, and the treatment and prognosis
CHAPTER 46
depends on the histology.90,93 Epstein-Barr virus (EBV) is believed to be the causative agent in the progression to PTLD,
especially in B-cell lymphomas. A wide variety of factors have
been proposed to be associated with an increased risk,
including the use of antilymphocyte induction therapy,
EBV-seronegative recipient, EBV infection, and era of transplant. Young white males appear to be at greatest risk.92 The
incidence of PTLD is associated with the overall intensity of
immunosuppression.91,92,94,95
The second most common cancer in pediatric recipients is
skin cancer. Squamous cell carcinoma accounts for the majority, followed by malignant melanoma and basal cell carcinoma.
The best strategy combines sunblock and sun avoidance. All
recipients should undergo regular dermatologic follow-up,
specifically focusing on this risk. Long-term immunosuppression is also associated with increased risks of cervical, vulvar,
and anal carcinoma.
Other Medical Issues
In addition to the risks of infection and malignancy, transplant
recipients face many other risks secondary to their history of
ESRD, their underlying renal disease, and the individual risks
associated with all their medications.
Renal transplant recipients are at high risk for cardiovascular disease.96 Preexisting renal insufficiency, time on dialysis,
and immunosuppressive medications after transplantation
all contribute to this risk. In addition, the prevalence of
hypertension in pediatric kidney recipients is 50% to
80%.97,98 The incidence of left ventricular hypertrophy at
initiation of renal replacement therapy ranges from 54% to
82%, though it generally improves after transplantation.99101
Many children and adolescents will have additional cardiovascular risk factors, including hyperlipidemia, hyperhomocysteinemia, anemia, malnutrition, and chronic inflammation.
Although there are few data examining the magnitude of
the risk in pediatric patients, young adult patients with ESRD
have a 1000-fold higher risk of cardiovascular death compared with the general population. Although the risk of cardiovascular death decreases after successful transplantation
629
compared with dialysis, it does not become normal.102

In addition to contributing to cardiovascular risk, hypertension is associated with a higher risk of graft dysfunction and
graft loss.98,103
Transplantation recipients also face significant problems
with bone metabolism and growth because of a history of
chronic renal insufficiency, malnutrition, graft dysfunction
after transplantation, and immunosuppressive medications.
Renal osteodystrophy is a substantial problem, but proper
calcium and vitamin D supplementation, along with other
agents, has improved overall bone health. The risk of osseous
complications has decreased with time, and the risk
decreases after transplantation compared with dialysis
therapy.104 The impairment in linear growth impacts final
adult height, but significant progress has been made.105,106
Height at transplantation is one of the best predictors of final
adult height, and better management prior to transplantation, including nutrition and the use of recombinant human
growth hormone, have improved height Z-scores at transplantation. Catch-up growth post-transplantation occurs,
but appears limited to recipients less than 6 years of age at
transplantation. Steroid avoidance protocols are also
associated with catch-up growth, adding a potential benefit
to consider with respect to immunosuppression regimens.
Cognitive and Psychosocial Development
The negative impact of ESRD on cognitive development has
diminished because of significant improvements in medical
management and renal replacement therapy. Children with
ESRD who undergo transplantation appear to have improvement in their level of cognitive function.107 Psychosocial
development tends to be below the healthy population,
though transplantation offers benefit compared with dialysis. Overall quality of life for the child and family appears
to be better after transplantation compared with dialysis,
although again, when compared with the normative population, there are disparities.108111
expertconsult.com.
available in clinical practice,5 and be accompanied by an increased frequency of severe hypoglycemia.4 Currently, the only
way to restore sustained normoglycemia without the associated risk of hypoglycemia is to replace the patients glucosesensing and insulin-secreting pancreatic islet beta cells6,7
either by the transplantation of a vascularized pancreas8,9 or
by the infusion of isolated pancreatic islets.1012 The tradeoff is the need for immunosuppression to prevent rejection
of allogenic tissue, and for this reason, most pancreas or islet
transplant recipients have been adults. However, the potential
for application earlier in the course of the disease exists, particularly in diabetic children already on immunosuppression
for other indications.13 Of the nearly 24 million people
estimated to have diabetes mellitus in the United States, 5%
to 10% have type 1 diabetes mellitus,14 and the prevalence
in children is increasing.15
Pancreas Transplantation
------------------------------------------------------------------------------------------------------------------------------------------------
HISTORY
CHAPTER 47
Pancreas and
Islet Cell
Transplantation
David E. R. Sutherland, Angelika C. Gruessner,
Bernhard J. Hering, and Rainer W. G. Gruessner
Type 1 diabetes is an autoimmune disease in which the pancreatic islet insulin-producing beta cells are selectively
destroyed.1 It most commonly presents in childhood and continues to represent a therapeutic challenge. Secondary diabetes complications, observed in 30% to 50% of patients who
live more than 20 years after onset of the disease, result in poor
quality of life, premature death, and considerable health care
costs.2 The principal determinant of the risk of devastating
diabetes complications is the total lifetime exposure to
elevated blood glucose levels.3 Therefore establishing safe
and effective methods of achieving and maintaining normoglycemia will have substantial implications for the health
and the quality of life of individuals with diabetes.
The Diabetes Control and Complications Trial (DCCT)
demonstrated that, given a qualified diabetes care team and
intensive insulin treatment control, near-normalization of glycemia could be achieved and sustained for several years.4
However, such a near-perfect level of treatment would
increase a patients burden of day-to-day diabetes management, be difficult to implement for many patients, require
more attention and medical services than are routinely
The first clinical pancreas transplantation was performed in

1966 by Drs. William Kelly and Richard Lillehei, simultaneous
with a kidney transplantation, in a uremic diabetic patient at the
University of Minnesota.16 Shortly thereafter, a few institutions
around the world began to perform pancreas transplantations, as
detailed in a comprehensive history in another book.17
The success rate (long-term insulin independence) with
pancreas transplantation was initially low but increased considerably in the 1980s, leading to increased application
(Fig. 47-1). Innovations in both surgical techniques and immunosuppression were responsible for the improvements.
The first pancreas transplantation was a duct-ligated segmental (body and tail) graft,16 but this approach was associated with multiple complications. In a series of 13 more
pancreas transplantations between 1966 and 1973 at the
University of Minnesota,18,19 Lillehei devised the whole
pancreasduodenal transplantation technique to the iliac vessels with enteric drainage through a duodenoenterostomy to
native small bowel, which is now routine at most centers.
The initial results, however, were not as good as today, and
several surgeons devised alternative techniques during the
1970s and early 1980s.17 Dubernard, in Lyon, France, introduced duct injection of a synthetic polymer as a method to
block secretions and cause fibrosis in the exocrine pancreas
of a segmental graft, with sparing of the endocrine component,20 and many pioneering centers adopted this technique,
although it is little used today. Gliedman introduced urinary
drainage through a ureteroductostomy for segmental grafts,21
and Sollinger later modified this approach with direct anastomosis of a duodenal patch of a whole pancreas graft to the recipient bladder.22 Drs. Dai Nghiem and Robert Corry did
further modification of urinary drainage,23 retaining a bubble
of duodenum for duodenocystostomy, as Lillehei had done for
duodenoenterostomy.18 From the early 1980s until the
mid-1990s, the bladder-drainage technique with duodenocystostomy was the predominant technique for pancreas
transplantations. The bladder-drainage technique had a low
acute complication rate and was helpful in monitoring for rejection by detection of a decline in urine amylase activity, but
chronic complications, such as recurrent urinary tract
631
632
PART IV
1600
1400
TRANSPLANTATION
U.S.
n = 23,850
Non-U.S. n = 11,365
1200
1000
800
600
400
200
Pr
e
-7
19 8
79
19
81
19
83
19
8
19 5
8
19 7
89
19
91
19
9
19 3
9
19 5
97
19
9
20 9
0
20 1
03
20
0
20 5
07
20
09
FIGURE 47-1 Annual number of U.S. and non-U.S. pancreas transplantations reported to the International Pancreas Transplant Registry (IPTR),
1978-2009.
infections or dehydration from fluid loss through the exocrine

secretions, were common. Thus in the mid-1990s, described
by Lillehei and colleagues,18 a change occurred and enteric
drainage, which was never totally out of fashion,24,25 overtook
bladder drainage as the predominant drainage procedure. In
addition, portal rather than systemic venous drainage was
used by some groups for enteric-drained whole pancreas
duodenal transplantations.26 Portal venous drainage was originally introduced by Calne in 1984 for segmental pancreas
grafts as a more physiologic technique26 and was applied by
several groups sporadically over the years.17
With advances in immunosuppression, including the
introduction of cyclosporine by Calne and coworkers in
1979,27 tacrolimus by Starzl and associates in 1989,28 and
mycophenolate mofetil by Sollinger and coworkers in 1995,29
bladder drainage had become less important for monitoring rejection. Furthermore, in recipients of simultaneous pancreas
and kidney transplants from the same donor, the kidney could
be monitored for rejection episodes (elevation of serum creatinine) as a surrogate marker for pancreas rejection before there
was sufficient pancreatic damage to cause hyperglycemia. However, in solitary pancreas transplants, serum creatinine could
not be used as a marker for rejection, and in such cases, bladder
drainage is useful and continues to be used.17
DETAILS OF SURGICAL TECHNIQUES

As mentioned in the history section, a variety of techniques
have been used for management of the exocrine secretions
and venous drainage of pancreas transplants.30,31 The majority
of pancreas grafts are procured from multiorgan deceased donors, and because the liver and pancreas share the origins of
their arterial blood supply, a whole organ pancreas graft usually
requires reconstruction.32,33 The blood supply to the tail of the
pancreas is supplied by the splenic artery, originating from the
celiac axis, and the head of the pancreas is supplied by the pancreaticoduodenal arcades, originating from the superior mesenteric artery and the hepatic artery. Because the latter goes with
the liver, along with the celiac axis, the usual approach is to attach an arterial Y-graft of the donor iliac vessels, with anastomosis of the hypogastric artery to the graft splenic artery and the
external iliac artery to the graft superior mesenteric artery,
leaving the common iliac artery segment of the Y-graft for anastomosis to the recipient arterial system, usually the right common iliac artery. The portal vein of the donor is usually divided
midway between the upper border of the pancreas and the liver,
leaving adequate length for transplantations of both organs, but
if necessary, an extension graft of donor iliac vein can be anastamosed to the pancreatic graft portal vein portion. In the recipient, the pancreas graft portal vein, with or without an
extension graft, can be anastomosed to the systemic venous system (usually the iliac vein or vena cava) or to the portal system
(usually the superior mesenteric vein).
When venous drainage is to the recipients iliac vein, the
whole pancreas graft can be oriented with the head directed
into either the pelvis or the upper abdomen. When directed cephalad, enteric drainage is the only option. When
directed caudad, the duodenum can be anastomosed to either
the bladder (Fig. 47-2) or bowel (Fig. 47-3). Figure 47-2
FIGURE 47-2 Simultaneous pancreas and kidney (SPK) transplantation

using a whole pancreas/duodenal graft from a deceased donor with systemic
venous drainage to the right iliac vein and bladder drainage of the pancreas
exocrine secretions through a duodenocystostomy. Both the pancreas and
kidney are placed in the peritoneum through a midline incision. The donor
splenic artery, supplying the pancreatic tail, and the donor superior mesenteric artery, supplying the pancreatic head, have been joined by a Y-graft constructed from the donor common external/internal iliac artery complex
during a bench procedure, and the base of the Y-graft is anastomosed to
the recipient common iliac artery. The mid-duodenum is anastomosed
to the posterior dome of the bladder, and the duodenal stumps are oversewn.
The kidney graft could be from a living donor or the same deceased donor as
the pancreas graft, but, in either case, is preferentially placed to the left iliac
vessels so that the right side, with its more superficial vessels, can be used for
the pancreas transplant. In this particular illustration, the donor ureter was
implanted into the bladder using the Politano-Leadbetter technique through
an anterior cystotomy, a technique that also allows the duodenocystostomy
to be performed with an end-to-end anastomosis (EEA) stapler, with internal
oversewing of the anastomotic line using an absorbable suture to cover the
staples, followed by closure of the cystotomy. However, when enteric drainage is used for an SPK transplantation, an external ureteroneocystostomy is
usually performed. (Reproduced from Gruessner RWG, Sutherland DER
[eds]: Transplantation of the Pancreas. New York, Springer-Verlag, 2004.)
CHAPTER 47
40 to 80 cm
FIGURE 47-3 Pancreas-duodenal transplantation using a deceased donor with systemic venous drainage and enteric drainage of graft exocrine
secretions to a proximal loop of recipient jejunum. In this particular case, an
end-to-side two-layer duodenojejunostomy, using the distal end of the
graft duodenum, is illustrated, and the anastomosis is located 40 to
80 cm distal to the ligament of Treitz (inset). Alternatively, a side-to-side
stapled or handsewn duodenojejunostomy, with or without a Roux-en-Y
loop, can be done. (Reproduced from Gruessner RWG, Sutherland DER
[eds]: Transplantation of the Pancreas. New York, Springer-Verlag, 2004.)
shows the bladder-drainage technique and also depicts a kidney transplantation to the left iliac vessels, but, as mentioned,
with a kidney transplantation, enteric drainage is more common than bladder drainage.
With the bladder-drainage technique, the anastomosis may
be handsewn or performed with an end-to-end anastomosis
(EEA) stapler brought through the distal duodenum (which
is subsequently stapled closed) for connection to the post of
the anvil projected through the posterior bladder by an
anterior cystotomy (see Fig. 47-2). The inner layer is then
reinforced with a running absorbable suture for hemostasis
and for burying the staples under the mucosa.
With enteric drainage/systemic venous drainage, the anastomosis may be handsewn in an end-to-side fashion (see
Fig. 47-3), or it can be done in a side-to-side fashion by handsewing or by using an EEA stapler.34 The barrel of the EEA
stapler is inserted into the end of the graft duodenum, and
the post is projected through the side wall. The anvil is
inserted into the recipient bowel through an enterotomy
secured around the connecting post by a purse-string suture.
The two posts are connected and the stapler is fired, creating
the anastomosis. The end of the duodenum is then closed with
a simple stapler. The enteric anastomosis can be done directly
to the most convenient proximal small bowel loop of the
recipient or to a Roux-en-Y segment of recipient bowel that
is created at the time. Outcome analyses do not show any
statistical advantage of a Roux-en-Y loop.
For portal drainage of the pancreas graft venous effluent
(Fig. 47-4), the head and duodenum of the graft is oriented
PANCREAS AND ISLET CELL TRANSPLANTATION
633
40 to 80
cm
FIGURE 47-4 Whole pancreas/duodenum transplantation using a

deceased donor with portal venous drainage with an end-to-side anastomosis to the recipient superior mesenteric vein, accessed below its confluence with the splenic vein. Drainage of exocrine secretions is through a
side-to-side duodenojejunostomy, 40 to 80 cm distal to the ligament of
Treitz. Note that the cephalad position of the pancreatic head, when portal
venous drainage is used, as opposed to the caudal orientation possible
with systemic venous drainage, is no different than that needed when
bladder drainage is done. In this particular illustration, the pancreas graft
overlies the root of the small bowel mesentery, with the duodenal segment
below the transverse colon, and the arterial Y-graft is anastomosed to the
recipient common iliac artery through a mesenteric tunnel. However, a retroperitoneal approach under the right colon is also possible, in which case
the arterial Y-graft can be anastomosed directly to the recipient iliac artery,
but the enteric anastomosis must be through a Roux-en-Y limb of recipient
bowel brought through the mesentery. If a kidney is simultaneously transplanted to the left iliac vessels, the ureter can be implanted into the bladder using the extravesical ureteroneocystostomy (Lich) technique, as
illustrated. (Reproduced from Gruessner RWG, Sutherland DER [eds]:
Transplantation of the Pancreas. New York, Springer-Verlag, 2004.)
cephalad, and the graft portal vein is anastomosed directly

to the recipient superior mesenteric vein. In the illustration,
the pancreas graft is ventral to the recipient small bowel mesentery so that the venous anastomosis is to the ventral side of
the vein, and the arterial Y-graft must be brought through a
window of mesentery for anastomosis to the recipients aorta
or common iliac artery. The graft duodenum is anastomosed to
recipients small bowel by the same techniques described for
systemic venous drainage, with or without (as depicted) a
Roux-en-Y loop of recipient bowel.
An alternative approach for portal venous drainage of the
pancreas graft effluent is to place the pancreas retroperitoneally
by reflecting the right colon to the left and exposing the dorsal
surface of the superior mesenteric vein, as described by Boggi
and associates.35,36 The arterial Y-graft can then be anastomosed
directly to the right common iliac artery, but this approach does
mandate creation of a Roux-en-Y limb of recipient bowel to
bring through the small bowel or transverse colon mesentery
for a graft duodenoenterostomy.
Other techniques can be used, including duct injection for
a segmental graft. Segmental grafts are rarely used, except in
the few cases of living donor pancreas transplantations,3742
and most of these have the exocrine secretions managed by
either a ductoenterostomy to a Roux-en-Y limb of recipient
bowel or a ductocystostomy to the recipients bladder
TRANSPLANTATION
1200
PTA
PAK
SPK
600
400
200
0
19
98
20
00
20
02
20
04
20
06
20
08
By the late-1990s, more than 2500 pancreas transplantations

were being done annually worldwide (see Fig. 47-1), as
reported to the International Pancreas Transplant Registry
(IPTR).45 By 2010, more than 35,000 vascularized pancreas
transplantations had been performed, more than half in the
United States, with very large series at some centers,25,4648
more than 2000 total at the University of Minnesota,49 and
more than 1000 simultaneous pancreas and kidney (SPK)
transplantations at the University of Wisconsin.47 The vast
majority were done to establish insulin independence in patients with de novo type 1 diabetes mellitus, but enteric-
800
19
88
GENERAL INFORMATION, PANCREAS

TRANSPLANTATION CATEGORIES,
AND IMMUNOSUPPRESSION
Number of transplants
1000
19
96
(Fig. 47-5). Segmental pancreas transplantations from living

donors, with or without a kidney transplantation, are particularly useful in candidates who would otherwise have a long
wait for a deceased donor organ, such as those with a high
level of human leukocyte antigen (HLA) antibodies but with
a negative crossmatch to a living volunteer.40 However, there
are circumstances where a segmental graft from a deceased donor is still appropriate for technical reasons, particularly for
retransplantations where conventional sites have been used
previously in such a way they cannot be reused and one
has to use unconventional sites, even orthotopically.43 For
more details concerning the variety of surgical techniques in
pancreas donors (deceased and living) and recipients, the
reader is referred to work by Benedetti and colleagues.44
19
94
FIGURE 47-5 Living donor segmental (body and tail) pancreas transplantation to right iliac vessels (systemic venous drainage) and bladder
drainage of exocrine secretions through a ductocystostomy by means of
an intraperitoneal approach. The donor splenic artery and splenic vein
are anastomosed end to side to the recipient external iliac artery and vein
after ligation and division of all hypogastric veins to bring the main vein as
superficial as possible. The splenic artery anastomosis is lateral and proximal to the splenic vein anastomosis. A two-layer ductocystostomy is constructed: The pancreatic duct is approximated to the urothelial layer (inner
layer) using interrupted 7-0 absorbable sutures over a stent (inset). If the
kidney is transplanted simultaneously, the donor ureter is implanted into
the bladder using the extravesical ureteroneocystostomy (Lich) technique. (Reproduced from Gruessner RWG, Sutherland DER [eds]: Transplantation of the Pancreas. New York, Springer-Verlag, 2004.)
drained pancreas transplantations have been used to correct

both endocrine and exocrine deficiency after total pancreatectomy in some patients5153 and to treat diseases such as cystic
fibrosis in others.54
Specialists in more than 150 institutions in the United
States, and nearly the same number elsewhere, have performed pancreas transplantations.55 The IPTR was founded
in 1980 to analyze the results.56 In 1987, reporting of U.S.
cases became obligatory through the United Network for
Organ Sharing (UNOS), and near-annual reports have been
made thereafter.5760
There are three categories of pancreas transplantation recipients: (1) uremic diabetic patients who undergo a simultaneous pancreas and kidney transplantation from either a
deceased or living donor61; (2) nephropathic patients who
already have had renal insufficiency corrected, usually by a
living donor kidney transplantation, and then undergo a
pancreas after kidney (PAK) transplantation6264; and (3) nonuremic diabetic patients who undergo a pancreas transplantation alone (PTA).65 The Pancreas Transplant Registry has
compared outcomes in the three categories spanning several
eras of data collection.5760,66,67
The majority of pancreas transplantations have been in the
SPK category, but in recent years, there has been an increased
emphasis in performing living donor kidney transplantations
to preempt the need for dialysis in diabetics with nephropathy.68 Thus until 2004, the number of PAK transplantations
increased, but the number of SPK transplantations did not
change (Fig. 47-6). Concomitantly, there has also been an increase in the number of PTA cases to treat diabetics without
advanced nephropathy who have diabetic management problems justifying immunosuppression, and to treat patients who
would also be candidates for islet transplantation, given the
conditions discussed later. Although most pancreas transplantation recipients have type 1 diabetes, insulin-treated type 2
diabetics also become insulin-independent after a pancreas
transplantation.6971
Immunosuppression management of pancreas transplantation recipients is similar to that of recipients of other solid organs, including kidneys, which the majority of pancreas
recipients also receive.72 Thus induction immunosuppression
19
91
PART IV
19
90
634
FIGURE 47-6 Number of pancreas transplantations performed annually

in the United States from 1988 through 2009 by recipient category. PAK,
pancreas after kidney transplantation; PTA, pancreas transplantation
alone; SPK, simultaneous pancreas and kidney transplantation.
CHAPTER 47
with antiT-cell monoclonal or polyclonal depleting or nondepleting agents may be used or reserved for rejection episodes.73 Maintenance immunosuppression usually consists
of a combination of a calcineurin inhibitor (cyclosporine or
tacrolimus), with the dose and blood levels adjusted to
minimize nephrotoxicity, and an antiproliferative agent
(mycophenolate mofetil or sirolimus), with or without prednisone. Corticosteroid-free regimens are now quite common
for all organ transplantations, including the pancreas.7483
Suspected pancreas allograft rejection episodes, based on
transient rise of serum amylase or lipase in enteric-drained
grafts or on a decline in urine amylase in bladder-drained
grafts, or by a rise in serum creatinine in SPK transplantations,
can be confirmed by biopsy of the graft.84,85
635
%
100
80
60
PAK
PTA
SPK
40
Category
20
1-yr fxn
SPK
4,146
85%
PAK
947
79%
PTA
465
78%
PANCREAS TRANSPLANTATION OUTCOMES
Current outcomes with deceased donor pancreas transplantations, according to recipient categories, surgical technique,
and immunosuppression protocol, for U.S. cases as reported
to UNOS from January 2005 to December 2009, are summarized here.55 During this period, 5567 primary deceased donor
pancreas transplantations were reported to UNOS, including
4155 SPK, 947 PAK, and 465 PTA transplantations.
The primary transplantation patient survival rates in the
three recipient categories are shown in Figure 47-7. At 1 year,
96% of the SPK, 97% of the PAK, and 97% of the PTA recipients were alive; at 3 years, 92%, 91%, and 92%, respectively,
were alive. The highest patient survival rate could be found in
PTA subgroups, presumably because this group had less
advanced complications before transplantation.
The primary pancreas graft survival rates in the three recipient categories are shown in Figure 47-8. At 1 year, 85% of the
SPK, 79% of the PAK, and 78% of the PTA recipients were
totally insulin-independent; at 3 years, 79%, 68%, and
62%, respectively, were insulin-independent (P < 0.001).
The highest pancreas graft survival rates are in the SPK category, presumably because the kidney graft (usually from the
same donor as the pancreas) can be used to detect rejection
episodes earlier than in the other categories, where only the
pancreas can be monitored. Support for this hypothesis comes
12
18
24
30
36
Months posttransplant
FIGURE 47-8 Pancreas graft functional survival rates (insulin independence) for 2005 to 2009 U.S. deceased donor primary transplantations
by recipient category. Fxn, function; PAK, pancreas after kidney transplantation; PTA, pancreas transplantation alone; SPK, simultaneous pancreas
and kidney transplantation.
from Registry data showing significantly higher early technical

failure rates and also large differences in rejection loss rates for
solitary transplants.
Of the 2005 to 2009 primary pancreas grafts, 6% of SPK
and 9% of solitary transplants failed for technical reasons,
with thrombosis being the highest risk for technical loss
(5%); infection, pancreatitis, and anastomotic leak made up
the rest. Technical graft loss was significantly higher in solitary
transplants than in SPK (P 0.0009).
The primary pancreas graft failure rates from rejection are
shown in Figure 47-9. At 1 year, 2% of the SPK, 4% of the
PAK, and 6% of the PTA recipients of technically successful
grafts had to resume exogenous insulin (significantly lower
in the SPK category; P 0.0001).
Regarding management of pancreatic duct exocrine secretions for the 2005 to 2009 cases, enteric drainage predominated
%
30
%
100
Category
25
PTA
PAK
SPK
20
90
n
413
811
3,687
1-yr loss
PAK
PTA
SPK
6.0%
5.5%
2.1%
15
80
Category
n
465
PTA
947
PAK
4,155
SPK
70
PAK
PTA
SPK
1-yr surv.
97.4%
96.8%
95.5%
10
5
0
60
0
12
18
24
30
36
FIGURE 47-7 Patient survival rates for 2005 to 2009 U.S. deceased donor
primary transplantations by recipient category. PAK, pancreas after kidney
transplant; PTA, pancreas transplant alone; SPK, simultaneous pancreas
and kidney transplantation; surv., survival.
12
18
24
30
36
FIGURE 47-9 Technically successful pancreas graft immunologic failure
rates (return to exogenous insulin) for 2005 to 2009 U.S. deceased donor
primary transplants by recipient category. PAK, pancreas after kidney
transplantation; PTA, pancreas transplantation alone; SPK, simultaneous
pancreas and kidney transplantation.
636
PART IV
TRANSPLANTATION
for SPK transplants (91%); for PAK and PTA, the proportion
that were enteric drained was slightly lower (86% and 79%,
respectively). Overall, the technical failure rate was significantly
higher with enteric-drained SPK than with bladder-drained
SPK (7% vs. 4%) transplants. No difference was found for
solitary transplants. Pancreas graft survival rates, however, were
not significantly different for enteric-drained versus bladderdrained transplantations in any of the categories: At 1 year,
the rates were 85% (n 3665) versus 86% (n 366) for
SPK, 79% (n 790) versus 82% (n 130) for PAK, and
80% (n 366) versus 75% (n 99) for PTA cases. No difference in the failure rate from rejection for technically successful
grafts for enteric-drained versus bladder-drained transplantations could be found anymore.
In the SPK category, bladder drainage and enteric drainage
would be expected to give similar results: In most cases, both
grafts come from the same donor, and monitoring of serum
creatinine serves as a surrogate marker for rejection in the pancreas transplant, allowing easy detection and reversal by treatment. In contrast, for solitary pancreas transplants (PAK and
PTA), serum creatinine cannot be used as a marker of pancreas
rejection, hyperglycemia is a late manifestation of rejection,
and exocrine markers must be used. Although serum amylase
and lipase may elevate during a rejection episode, this does
not occur in all cases, but for bladder-drained grafts, a
decrease in urine amylase eventually always accompanies
rejection (100% sensitive, even though it is not specific)
and nearly always precedes hyperglycemia so that a rejection
episode is more likely to be diagnosed in a bladder-drained
graft and lead to treatment and reversal.
Approximately one quarter of enteric-drained pancreas
grafts reported to UNOS were done with a Roux-en-Y loop;
in the past, the outcomes were not improved by this procedural addition,45 and that is still the case.55
Another variation in surgical technique is portal drainage of
the venous effluent for enteric-drained grafts.30,86 It establishes normal physiology and a theoretic metabolic advantage
versus systemic venous drainage, and some groups have
reported that portal venous enteric-drainage grafts are less
prone to rejection than systemic venous enteric-drainage
grafts,87,88 although others have not.89 The latest Registry
analysis shows that portal venous drainage was used for one
fifth of enteric-drainage transplantations, but there were no
significant differences in pancreas graft survival versus
systemic venous enteric-drainage transplantations in any of
the categories: at 1 year, 84% (n 718) versus 86%
(n 2896) for SPK, 79% (n 130) versus 79% (n 651)
for PAK, and 78% (n 51) versus 80% (n 305) for PTA cases.
Regarding immunosuppression, according to the latest
Registry analysis, antiT-cell agents were used for induction
therapy for more than 80% of the 2005 to 2009 U.S. pancreas
recipients in each category.55 The most frequently used regimen for maintenance immunosuppression (more than two
thirds of the recipients in each category) was tacrolimus and
mycophenolate mofetil in combination, with or without prednisone. In recipients of primary deceased donor pancreas
grafts given antiT-cell agents for induction and tacrolimus
and mycophenolate mofetil for maintenance immunosuppression (Fig. 47-10), the 1-year graft survival rates in the SPK,
PAK, and PTA categories were 86% (n 2737), 81%
(n 544), and 86% (n 271), respectively. A sirolimus-based
regimen was used as a maintenance immunosuppressive drug
%
100
80
60
PAK
PTA
40
SPK
Category
20
PAK
PTA
SPK
1-yr fnx
544
271
2,737
81.3%
85.6%
86.3%
0
0
12
18
24
30
36
FIGURE 47-10 Pancreas graft functional survival rates (insulin independence) for 2005 to 2009 U.S. deceased donor primary transplantations by
category in diabetic recipients given antiT-cell agents for induction and
tacrolimus (TAC) and mycophenolate mofetil (MMF) for maintenance immunosuppression. Fxn, function; PAK, pancreas after kidney transplantation; PTA, pancreas transplantation alone; SPK, simultaneous pancreas
and kidney transplantation.
in more than 13% of recipients in each category (Fig. 47-11),

with excellent outcomes: The 1-year pancreas graft survival
rates in the SPK, PAK, and PTA categories were 90% (n
407), 89% (n 94), and 89% (n 84), respectively. In contrast, the remaining recipients given alternative immunosuppressive regimens had distinctly lower pancreas graft survival
rates in each category: at 1 year, 74% in SPK (n 153), 61% in
PAK (n 71), and 29% in PTA (n 35) cases. A center effect
may play a role in the outcomes of the Registry analysis
according to immunosuppressive regimens.
Regarding the logistics of pancreas transplantation, the recent Registry data55 showed a significant increase in technical
100
80
60
PAK
PTA
SPK
40
Category n
20
PAK
PTA
SPK
95
87
407
1-yr fnx
89.3%
90.2%
89.6%
0
0
12
18
24
30
36
FIGURE 47-11 Pancreas graft functional survival rates (insulin independence) for 2005 to 2009 U.S. deceased donor primary transplantations by
category in diabetic recipients given sirolimus-based maintenance immunosuppression. Fxn, function; PAK, pancreas after kidney transplantation;
PTA, pancreas transplantation alone; SPK, simultaneous pancreas and
kidney transplantation.
CHAPTER 47
failure rates and a decrease in graft survival rates with increasing preservation time. The relative risk (RR) to lose the graft
doubled for SPK grafts with a preservation time greater than
24 hours compared with a preservation time of 12 to 24 hours.
Shorter SPK preservation time showed a decreased risk of one
third. HLA matching had virtually no impact on SPK graft
survival rates, but matching at least at the class I loci had a
beneficial effect in the PAK and the PTA categories.
Regarding pancreas recipient age, the recent Registry analysis of the 2005 to 2009 cases showed an effect on outcome
mainly in solitary recipients, with rejection more likely in
younger patients. In the SPK category, only 3 patients were
younger than 15 years of age, and 312 recipients (7%)
were between 15 and 29 years of age. In PAK, 5% (n 60)
were between the age of 15 and 29 years of age, and 15%
(n 75) in PTA. The relative risk for graft loss was not significantly increased for younger SPK recipients (P 0.21) but
clearly higher for PAK recipients (RR 1.75, P 0.003)
and PTA recipients (RR 1.99, P 0.009) compared with
recipients 30 to 45 years of age. Thus the young nonuremic
diabetic is highly immunocompetent and more prone to reject
a pancreas graft, consistent with an earlier analysis of outcomes in U.S. pediatric pancreas transplantation recipients
from 1988 to 1999.90 In that analysis, of slightly more than
8000 pancreas transplantations, only 49 were in recipients
younger than 21 years of age (<1%), 34 in the SPK, 2 in
the PAK, and 13 in the PTA category; all were deceased donor
pancreas transplantations, except for two PTA segmental
grafts from living donors. Less than half of the pediatric pancreas recipients were younger than 19. In the PTA recipients,
the 1-year graft survival rate was only 15%, with all but one
loss being from rejection in less than 1 year. The Registry data
do not include the indications for a PTA in the pediatric recipients, but presumably they had extremely labile diabetes, justifying placement on immunosuppression in an attempt to
gain control. In the pediatric SPK recipients, however, the
1-year patient, pancreas, and kidney graft survival rates were
96%, 78%, and 71%, respectively, which were outcomes comparable to those of adult SPK recipients for the entire period.
Of the pediatric SPK recipients, most had a renal disease other
than diabetic nephropathy.
Thus pancreas transplantations in the pediatric age group
are uncommon, and most are in diabetic children who also
have renal failure and thus need a kidney transplantation, obligating them to immunosuppression. At least in this group,
the outcomes are such that it seems reasonable to recommend
the addition of the pancreas so that the child can become insulin independent as well as dialysis free for the price of immunosuppression.13 For nonuremic diabetic children with
extreme lability, in whom a successful pancreas transplantation would be appropriate treatment, the antirejection strategies need to be optimized to improve the graft survival rates
versus what has been achieved in the past.
With respect to outcome measures other than insulin independence, prevention and reversal of secondary complications, improvement in quality of life, extension of life span,
and reduction of health care costs per quality-adjusted lifeyear have all been positively demonstrated in type 1 diabetic
pancreas transplant recipients.9197 In patients with labile diabetes and hypoglycemic unawareness, a pancreas transplantation can resolve an otherwise intractable and life-threatening
course.98100
637
Whether a pancreas transplantation has an effect on survival probabilities for the diabetic population selected for
the procedure is controversial. Two separate analyses of U.S.
data from UNOS and the Organ Procurement Transplant Network (OPTN) for pancreas transplantation candidates and recipients between 1995 and 2000 compared the survival
probabilities for patients who remained on the waiting list
with those receiving a transplant by category.103,104 In the first
analysis,103 SPK recipients had a significantly higher probability of survival than those remaining on the waiting list for the
procedure, but, for solitary (PAK or PTA) recipients, just the
opposite was the case. In the second analysis,104 the higher
survival probability for SPK recipients was confirmed, and,
in addition, the overall survival probabilities of solitary pancreas transplant recipients compared with those waiting,
and even after 1 year, were favorable for transplantation. In
the second analysis, patients who listed at multiple centers
were identified and were counted only once from the time
of first listing, corrections were made for patients who
changed categories, and longer follow-up was available. Thus
pancreas transplantation does not entail a higher risk than
staying on exogenous insulin for those on the waiting list
and may improve survival probabilities for solitary as well
as SPK recipients.105
In regard to secondary complications of diabetes,106 numerous studies show a beneficial effect on neuropathy,94,107113 retinopathy114118 and nephropathy92,119122
as well as on cardiovascular disease,123128 and quality of
life.129,130 In regard to nephropathy, specifically in PTA recipients, even though the diabetic lesions in the native kidneys
can improve,131 this can be offset by the nephrotoxicity of
the calcineurin inhibitors given for immunosuppression.132
It should be noted that pancreas retransplantation can be
done if the first graft fails, with only slightly lower graft survival rates than for primary transplants.66,67,133,134 Indeed,
even third transplants can do well.134
Also of note, pancreas allografts are often procured by one
center and transplanted at another, and there are studies
showing no difference in outcomes compared with locally
procured organs.135
Surgical complications of pancreas transplantation are numerous and are the subject of an extensive review recently
published.136 The most frequent complication leading to graft
loss is venous or arterial thrombosis (5% to 10%). Anastomotic leaks are also common, but if diagnosed early, graft salvage is possible.
Islet Transplantation
------------------------------------------------------------------------------------------------------------------------------------------------
Human islet transplantation, the less invasive islet beta cell replacement alternative to transplantation of the vascularized
pancreas, has been investigated for more than 3
decades137140 after the first clinical islet allograft was performed
in 1974.141 Since then, nearly 1000 islet allotransplantations
have been performed worldwide.142,143
Islet autotransplantations have had a relatively high success
rate in preventing diabetes after total pancreatectomy for more
than 2 decades; so, they are briefly described before reviewing
the current status of islet allografts for type 1 as well as for
surgical diabetes.
638
PART IV
TRANSPLANTATION
ISLET AUTOTRANSPLANTATIONS AFTER

PANCREATECTOMY FOR BENIGN DISEASE
Islet autotransplantations to prevent diabetes after a total pancreatectomy for benign disease, primarily painful chronic pancreatitis, have been successful since the first case was
performed in the 1970s144,145 but depend on the number
of islets transplanted.146151 Children with chronic pancreatitis and intractable pain who require pancreatectomy for relief
of the pain and resolution of narcotic dependence nearly
always have some beta cell function preserved by an islet autotransplantation, and they are either nondiabetic (more than a
third) or can maintain euglycemia with once-daily long-acting
insulin or near euglycemia with standard basal-bolus
insulin.152155 Total pancreatectomy is highly successful in relieving the pain of chronic pancreatitis in both adults and children, particularly if done early before years of narcotic
dependence.148,150,153 The islet yield is also better if the
pancreatectomy is not delayed, because the number of islets
isolated correlates with the degree of pancreatic damage in
both adults156 and children.157
The surgical technique of pylorus-sparing total pancreatectomy and duodenectomy is shown in Figure 47-12. The
procedure can be staged, but when the body and tail of the
FIGURE 47-12 Pylorus-sparing total pancreatectomy and partial duodenectomy technique for patients with chronic pancreatitis undergoing islet
autotransplantations. The bile duct is transected and reimplanted into the
duodenum, shown here proximal to a duodenoduodenostomy or duodenojejunostomy, but, more commonly, it is placed distal to the enteric anastomosis, with the site depending on the individual anatomy. When
possible, only the second portion of the duodenum is resected, and an
end-to-end duodenoduodenostomy is created; but if viability is not maintained, the entire distal duodenum must be resected and an end-to-end or
end-to-side duodenojejunostomy performed. The short gastric vessels are
preserved, as well as the gastroepiploic artery if possible, and the spleen is
not removed if its viability is maintained. (Reproduced from Gruessner
RWG, Sutherland DER [eds]: Transplantation of the Pancreas. New York,
Springer-Verlag, 2004.)
pancreas are removed, they should always be processed for islet isolation for an intraportal autotransplant (Fig. 47-13, A).
If a distal pancreatectomy is the primary procedure and a
Whipple (completion) pancreatectomy becomes necessary, diabetes will have been prophylactically prevented by the initial
islet autograft. If a Whipple procedure was the primary procedure, but pain persists and a distal (body and tail) completion
pancreatectomy is required, it should be done in an institution
capable of isolating islets from the excised gland for an
autotransplantation.158
ISLET ALLOTRANSPLANTATIONS
Islet allotransplantations have been performed for the treatment of surgical and type 1 diabetes. As with autotransplantations, islet allotransplantations are usually done with
embolization of the islets to the liver via the portal vein, where
at least some islets will survive by nutrient diffusion until
revascularization occurs (Fig. 47-13, B). A drawback of
islet allotransplantations, as compared with pancreas transplantations, is the reduced beta cell mass; much attention
has been given to compensating for the attrition that occurs.
Islet allografts in patients with surgical diabetes have been
associated with a very high success rate,159,160 possibly
because of the avoidance of diabetogenic steroids and the lack
of an autoimmunity.
Islet allograft transplantations in patients with autoimmune
type 1 diabetes have initially been performed simultaneously
with a kidney transplant or in patients with established kidney
transplants.142 Insulin independence in this recipient group,
even on an anecdotal basis, was not achieved until the early
1990s.161165
Islet allografts have also been performed in patients in
whom type 1 diabetes (T1D) is complicated by hypoglycemia
unawareness and defective hormonal glucose counter-regulation resulting in recurrent episodes of severe hypoglycemia.166 Today, the majority of human islet allografts are
performed in this recipient group.143 Acute complications
are frequent in the 12.5% of T1D patients who have become
aware of hypoglycemia 20 years after diabetes onset.167
Iatrogenic hypoglycemia is the most limiting factor in the
glycemic management of T1D;168 it causes recurrent physical
and psychological morbidity, including coma, seizures, and
significant social embarrassment, and 7% to 10% of all deaths
in patients with T1D are the result of hypoglycemia.169,170
Hypoglycemia-related problems have not abated during the
more than 18 years since they were first highlighted by the
landmark report of the DCCT in 1993.168,171 New glucose
monitoring technologies are being developed; however, continuous glucose monitoring did not lower the rate of severe
hypoglycemia in patients with T1D.173,174
A major milestone was reached in 2000 when Dr. Shapiro
and colleagues at the University of Alberta in Edmonton
achieved diabetes reversal in seven of seven recipients by
using islets from more than one donor pancreas and by using
corticosteroid-free, less diabetogenic immunosuppression.175
Since then, several groups around the world have reported
restoration of insulin independence after human islet allotransplantation in type 1 diabetic recipients.11,176194
Furthermore, remarkable additional progress has been
made in the past decade toward developing islet transplantation into a vital treatment option for T1D. First, new protocols
CHAPTER 47
639
Isolated native
islet of Langerhans
Liver
Islet isolation
Syringe
Patient with pancreatitis
Pancreas
Islet in portal vein

Donor
Pancreas
Islet isolation
Recipient
Liver
Isolated islet
of Langerhans
Syringe
Portal vein
Pancreas
B
Islet in pancreas
FIGURE 47-13 Islet transplantation using the portal vein for embolization to the liver where revascularization will occur, either as an autograft of islets
isolated from the excised specimen after pancreatectomy for benign disease (A) or as an allograft of cells isolated from a donor for treatment of a patient
with type 1 diabetes (B).
succeeded in achieving insulin independence with islets from

a single donor pancreas.179,180,187191 Many of these protocols
include adjunctive peritransplant anti-inflammatory and/or
cytoprotective therapy, likely facilitating improved islet
engraftment. Second, recent data indicate that islet allograft
survival in T1D can be sustained with calcineurin inhibitor
free protocols.188190 Two immunosuppressive regimens,
based on the costimulation blocker belatacept or the antileukocyte functional antigen-1 antibody efalizumab, were effective, well tolerated, and involved the first calcineurin
inhibitor/steroid-sparing islet protocols resulting in long-term
insulin independence. Although efalizumab is no longer
available for clinical use, these early results demonstrate that
calcineurin inhibitorfree regimens may be an effective alternative to improve graft function and longevity while minimizing renal and islet beta cell toxicity. Third, Berney and
colleagues reported on the first type 1 diabetic patient who
remained insulin independent for more than 10 years after
islet allotransplantation.195 Preliminary data now suggest that
long-term insulin independence (>5 years) can be achieved in
50% of recipients given T-celldepleting induction immunotherapy, matching insulin independence rates of solitary pancreas transplantation.196 Early studies examining long-term
islet function suggested a rapid loss of insulin independence
beyond 1 year in many patients.183,197 One possible factor contributing to islet loss is recurrent beta cell autoimmunity.198,199
640
PART IV
TRANSPLANTATION
Anti-CD3 antibodies and T-celldepleting therapies, including

anti-Thymoglobulin, have proved promising agents for minimizing autoimmunity in murine models of autoimmune diabetes and in clinical trials for new-onset T1D.200204 A recent
analysis of University of Minnesota data and data reported to
the Collaborative Islet Transplant Registry indicated that patients receiving an induction immunosuppression regimen that
includes T-celldepleting agents, either anti-CD3 antibody
alone or either antithymocyte globulin (ATG) or alemtuzumab
plus short-term TNF-a inhibition, for alloislet transplantation
are more than twice as likely to maintain long-term insulin independence for at least 5 years post-transplant.196 Three-year
and 5-year insulin independence rates in these recipients are
comparable with rates previously only attainable in recipients
of pancreas transplants alone. Fourth, numerous reports have
confirmed that human islet transplants are remarkably effective
in protecting recipients with full and even partial islet graft
function from severe hypoglycemia.166,205,206 Finally, a prospective clinical trial demonstrated reduced progression of diabetic microvascular complications after islet transplantation
compared with intensive medical therapy.207 These data extend
previous reports by the Milan group208,209 and highlight the
immense potential of cell-based diabetes therapy.
The unlimited and on-demand availability of xenogeneic
pig islets would boost access to islet beta cell replacement.
The quality of islet products from healthy, young, and living
donor pigs would be predictably high and not compromised,
as with human islet products, by comorbidity, brain death,
age, and cold ischemia. The actual risks of infectious disease
transmission from designated pathogen-free (DPF) source
pigs are lower compared with risks associated with the
use of deceased human donor organs.210 Finally, genetic
modification of source pigs would present opportunities
for minimizing recipient immunosuppression not available
to recipients of human islet allografts.211 Thus exploiting
the unique possibilities associated with porcine islet products would increase the availability and benefit-risk ratio
of islet replacement therapies when compared with human
islet products. The impact of such medical products on
addressing unmet clinical needs in diabetes would be
profound.
During the past few years, prolonged diabetes reversal
exceeding 6 months has been demonstrated after porcine
islet xenotransplantation in immunosuppressed nonhuman
primates (NHPs) and for up to 6 months in nonimmunosuppressed NHPs.212 Porcine C-peptide has been positive
in the plasma of these recipients, and their fasting and, in
some studies, also their non-fasting blood glucose levels have
been in the normoglycemic to near-normoglycemic range.
Perhaps most intriguing is that success has been achieved
by five independent groups involving the use of various tissue
sources (adult, neonatal, and embryonic pig islet tissue;
wild-type and transgenic), implantation sites (portal vein,
omental pouch, subcutaneous space), immunosuppressive
protocols (with and without anti-CD154 monoclonal antibodies or encapsulation, avoiding immunosuppression),
and animal models (streptozotocin-induced and surgical
diabetes; in cynomolgus and rhesus monkeys).213216 Collectively, the demonstration of prolonged functional islet xenograft survival in nonhuman primates with several distinct
xenotransplantation strategies suggests that porcine islet
products could potentially be developed into a more widely

available cellular therapeutic for T1D.
Although a transition from pancreas to islet transplantations as the dominant form of beta cell replacement therapy
may occur during the next few years, pancreas transplantations will not disappear entirely. Patients with high pretransplantation insulin requirements, in whom diabetes reversal
with islet transplantations is less likely, would best be served
with a vascularized pancreas transplant, at least as long as
other more unlimited sources of beta cells, such as porcine islets, are not available for clinical therapy. Furthermore, diabetic patients with exocrine deficiency would best be served
by an enteric-drained pancreas transplant. In addition, in patients who have very high insulin requirements or insulin resistance (type 2 diabetes), an intact organ may be needed to
obtain a sufficient islet mass to restore insulin independence
from a single donor in the presence of insulin resistance.
Tissue availability will be the limiting factor in determining
the magnitude of the impact of beta cell replacement therapy.
Six thousand deceased donors are available each year in the
United States, but it is estimated that only half have a pancreas
suitable for transplantation. Thus the maximal number of pancreas transplantations that could be done in the United States
is 12,000 per year, assuming that each deceased pancreas
could be split for use in two recipients,217 and that living donors would be used for segmental pancreas transplantations61
to the extent that they have been for kidney transplantations
(currently about 6000 per year in the United States). This scenario has not yet materialized, but the potential is there to
transplant at a rate approaching half of the annual incidence
of new-onset cases of type 1 diabetes (30,000/year in the
United States). The numbers could be increased further if
enough islets could be isolated from one donor for transplantations into more than two recipients. Although the efficacy of
islet transplantation protocols will continue to improve, and
the procedural and immunosuppressive risks now associated
with islet transplantations will continue to diminish, islet
transplantations will not be the ultimate approach to diabetes
care. Just as pancreas transplantations set the stage for islet
transplantations, the real value of islet transplantations will
be to create and build momentum for the development of xenogeneic and stem/precursor cellderived islet beta cell therapy218 that will then make cell replacement therapy routine
and commonplace in diabetes care.
Pancreas transplantations, and eventually islet transplantations, should be in the armamentarium of every transplantation center for the treatment of diabetic patients. Likewise,
every endocrinologist should consider beta cell replacement
in the treatment of patients in whom type 1 diabetes is complicated by hypoglycemia-associated autonomic failure219
and/or progressive microvascular complications. Continued
clinical research on pancreas and islet transplantations is
needed to identify the most appropriate recipient population,
the optimal timing in the course of diabetes, and the most
suitable donor tissue and transplantation protocol for a given
patient. Both pancreas and islet transplants need to be made
as economical as possible.220 Studies such as those done in
pancreas-kidney transplant recipients, showing the efficiency in the treatment of complicated diabetes,221 are
needed in islet recipients as well. Currently, beta cell replacement has a well-defined clinical role for adult patients with
CHAPTER 47
incapacitating hypoglycemic unawareness and is also appropriate in children and adults who otherwise need immunosuppression, such as for a kidney transplantation. As antirejection
strategies become safer and with fewer side effects, the indications for pediatric beta cell replacement therapy can be
liberalized.
641
Acknowledgments
We are indebted to Christine Johnson and Heather Nelson for assistance in
preparing the manuscript.

expertconsult.com.
age, only 25% of the pediatric cadaveric donor population

comes from this same age group (Fig. 48-1).2 As a consequence, achieving success in this arena requires technical
perfection, both from the standpoint of obtaining a suitable
graft and performing a meticulous transplantation operation.
With reduction in transplantation waiting time and with improvements in immunosuppression, surgical technique, and
long-term post-transplantation care, survival has markedly
improved and now exceeds 90% at 1 year and 80% at 5 years,
with many children surviving into adolescence and adulthood
with a good quality of life.3,4 In this chapter, we review the
major indications for liver transplantation in children, the basic pathophysiology and clinical presentation of liver failure,
operative strategies with emphasis on the unique surgical
options available to children, postoperative management with
emphasis on management of surgical complications, and outcome analysis.
Indications and Pretransplant

Care
CHAPTER 48
Liver
Transplantation
Bob H. Saggi, Douglas G. Farmer,
and Ronald W. Busuttil
The treatment of liver disease in children with transplantation

has its roots in the origin of liver transplantation itself, with the
initial cases performed by Thomas E. Starzl on two children in
1963 and 1968.1 Although the initial results were disappointing, during the ensuing 2 decades, liver transplantation developed into the standard therapy for decompensated cirrhosis,
certain malignancies of the liver and biliary tract, acute liver
failure, and many metabolic derangements. The National Institutes of Health Development Conference designated it as
such in 1983, and the National Organ Transplantation Act created a nationally regulated system of organ allocation in 1987.
The United Network for Organ Sharing (UNOS) was subsequently created and currently regulates the field through a
peer review process. In 2009, 6320 liver transplantation procedures were performed in the United States, and of these,
572 were in the pediatric population.2 This number of transplantation procedures has remained relatively stable since the
late 1990s, although an increasing number of partial liver
grafts from cadaveric and living donors are now being used.
Pediatric liver transplantation offers unique challenges because of size, perhaps enhanced immune responsiveness,
and the paucity of donor organs. Although nearly 65% of pediatric liver transplantation recipients are less than 6 years of
------------------------------------------------------------------------------------------------------------------------------------------------
INDICATIONS FOR LIVER

TRANSPLANTATION
Liver transplantation is currently indicated for children with
decompensated cirrhosis because of cholestatic and noncholestatic causes, acute hepatic failure, some metabolic liver
diseases, select tumors, and a variety of miscellaneous indications (Fig. 48-2). The general indication for liver transplantation in children is liver disease that limits long-term survival or
quality of life, or markedly impairs normal growth and development. Cirrhosis alone is not an indication for transplantation, because many patients can be medically managed for a
prolonged period prior to decompensation. In acute liver failure, the development of clear symptomatology, such as refractory coagulopathy, acidosis, and encephalopathy that correlate
with a poor prognosis for spontaneous recovery of liver function, is an indication for transplantation. Otherwise, medical
support in those with a better prognosis is provided until liver
function returns to normal.5,6 Finally, metabolic inborn errors
of metabolism are a unique and not uncommon indication for
transplantation in children, making up 9% of transplantation
in children compared with approximately 2% in adults. Parenteral nutritionassociated liver disease accounts for 8% of
liver transplantations, and less than 1% of adults are transplanted for this reason. Most of these transplantations are
combined with an intestinal graft (Chapter 49).
With long-standing cirrhosis, the development of a constellation of symptoms and signs that represent decompensation
of hepatocellular function or portal hypertension herald a
need for transplantation. These include progressive jaundice,
coagulopathy, protein-calorie malnutrition and growth retardation, impaired cognitive development, encephalopathy,
hypersplenism, variceal hemorrhage, and advanced or refractory ascites. The majority of patients undergoing transplantation in this population are deeply jaundiced because of
secondary or primary biliary cirrhosis from long-standing
intrahepatic and/or extrahepatic biliary obstruction. On physical examination, these patients often have muscle wasting, an
enlarged spleen, a hard palpable liver, abdominal distention
643
644
PART IV
TRANSPLANTATION
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
Total #
2009
2008
2007
2006
2005
572
613
605
577
569
1117 years
610 years
15 years
<1 year
FIGURE 48-1 Distribution of pediatric liver transplantations by age.
Data obtained from www.unos.org.
PEDIATRIC LIVER TRANSPLANT

RECIPIENTS, 2008
Biliary atresia
29%
Acute hepatic
necrosis
12%
Metabolic
disease
9%
Cholestatic liver
disease/cirrhosis
12%
Malignant
neoplasms
8%
Non-cholestatic
cirrhosis
8%
Other
14%
TPN liver
disease
8%
FIGURE 48-2 Indications for pediatric liver transplantation. Data

obtained from www.unos.org, based on 2008 transplantations. TPN, total
parenteral nutrition.
from ascites, and peripheral edema. With decompensation,

long-term survival without liver transplantation is limited,
and referral for transplantation must be made prior to decompensation.3,79
CHOLESTATIC LIVER DISORDERS

The most common indications for liver transplantation are the
cholestatic liver disorders, with the most common being biliary atresia (BA). This group accounts for roughly 40% of
the transplantation performed on children in 2008, and BA
accounts for 70% of this cholestatic group (see Fig. 48-2).2
The management of BA rests on early diagnosis, using surgical
exploration with biopsy as the central confirmatory test
in most cases. The diagnostic workup is detailed in
Chapter 105. Portoenterostomy (PE) is the preferred treatment if diagnosis precedes the development of cirrhosis, usually before 3 months of age, though long-term results from PE
FIGURE 48-3 Decompensated cirrhosis after a Kasai portoenterosotomy. This 6-month-old, 5-kg child presented for liver transplant with the
advanced findings of hepatosplenomegaly, extensive abdominal wall venous collaterals, tense ascites, jaundice, and profound malnutrition.
are optimal if it is done prior to 8 weeks.10 This procedure is

essential for slowing, and, in some cases, arresting, the progression of liver disease to cirrhosis and portal hypertension.
Unfortunately, despite effective biliary drainage, more than
70% of patients will go on to develop decompensated cirrhosis
by the age of 5 years and require transplantation.10,11 However, in many cases, PE allows reasonable growth so that transplantation is forestalled until the child is older and larger.
Primary PE performed late in the course of BA and reexploration for a failing biliary drainage procedure are both usually
unsuccessful and only complicate transplantation outcomes.
Instead, once a PE has failed, patients should be evaluated
for transplantation. Liver transplantation is indicated when
the diagnosis is made beyond 3 months of age, when decompensated cirrhosis is clearly present at any age, or after a PE
has failed. Patients with BA should be managed by a pediatric
hepatologist experienced in transplantation to ensure early
referral for transplantation. This should occur prior to severe
liver decompensation when signs are obvious on physical
examination, especially with an advanced presentation at a
late stage (Fig. 48-3).
Other uncommon etiologies of cholestatic liver injury and
cirrhosis include familial paucity of intrahepatic bile ducts,
which exists in a syndromic (Alagille syndrome) and nonsyndromic form, familial cholestatic syndromes, primary or secondary sclerosing cholangitis, and uncorrectable choledochal
cyst disease, including Caroli disease. The cystic diseases have
a component of uncorrectable extrahepatic obstruction while
the others are the result of malformation or destruction of
intrahepatic bile ducts and/or arterial systems. All these share
in common a variable and unpredictable progression to
advanced fibrosis, cirrhosis, and portal hypertension. These
patients typically present with progressive jaundice at an older
age than patients with BA. Although their management does
not entail a PE, the indications for transplantation in these
patients follow the same rationale as that for BA.
NONCHOLESTATIC CIRRHOSIS
Noncholestatic cirrhosis is an uncommon indication for liver
replacement in children, accounting for less than 10% of all
procedures performed in 2008 (see Fig. 48-2).2 These children usually present later in life than the cholestatic disorders.
CHAPTER 48
Etiologies of cirrhosis and decompensated cirrhosis in these patients include chronic autoimmune hepatitis, neonatal hepatitis,
chronic viral (B or C) hepatitis, and cryptogenic cirrhosis.
ACUTE LIVER FAILURE

Fulminant hepatic failure is usually defined as the onset of encephalopathy within 28 days after the onset of jaundice in a
patient with acute liver failure without evidence of chronic
liver disease. The hallmarks of acute liver failure include profound coagulopathy, acidosis, hypoglycemia, and progressive
hyperbilirubinemia. Acute liver failure patients can develop
acute renal failure, systemic inflammatory response with multiorgan failure syndrome, or cerebral edema progressing to
herniation. Early referral is essential to avoid progression to
a condition that contraindicates transplantation. A variety of
criteria to determine the need for transplantation have been
devised in European centers, where these patients are managed in a highly structured and centralized manner.12 The
most common established etiology in children is viral hepatitis, followed by acetaminophen and other drug toxicities and
Wilson disease. However, in nearly two thirds, an etiology
cannot be identified. Liver transplantation is the only acceptable therapy in patients who meet the criteria of fulminant
hepatic failure, and early referral of all patients with acute liver
failure is essential.5,6
METABOLIC LIVER DISEASE

These disorders have in common an enzyme deficiency or
some other defect in hepatocellular function. This impairment
can result in progressive fibrosis or cirrhosis (e.g., cystic fibrosis, chronic Wilson disease, and neonatal iron storage disease),
with a typical presentation of decompensated cirrhosis. In
other cases, the liver is structurally normal, but harmful
byproducts of metabolism accumulate to cause neurologic injury (e.g., Crigler-Najjar syndrome, ornithine transcarbamylase deficiency, and acute Wilson disease), cardiovascular
disease (e.g., familial hypercholesterolemia), or renal injury
(familial hyperoxaluria). Some disorders are associated
with the development of malignancies (e.g., tyrosinemia),
and transplantation should be considered preemptively.
Transplant evaluation of all patients with known metabolic
disorders of the liver involves a thorough evaluation of extrahepatic function. This will ensure transplantation of only
those patients who can benefit from liver transplantation
and prevents progression of extrahepatic disease. In some patients, simultaneous or sequential or dual organ transplantation may be necessary (e.g., lung, kidney, heart).
TUMORS
The most common liver malignancy in children is hepatoblastoma.13,14 Although sporadic cases have been reported, hepatocellular carcinoma (HCC) is primarily seen in children with
viral hepatitis, tyrosinemia, some of the glycogen storage diseases, or in association with cirrhosis from other causes. Primary liver malignancies in pediatric patients are managed by
surgical resection unless tumor size and/or location preclude
resection. The benefit of neoadjuvant and/or adjuvant chemotherapy and radiation for hepatoblastoma has been well documented.1517 In HCC, multimodal therapy can be used to
LIVER TRANSPLANTATION
645
prevent further cancer progression or perhaps to improve

the outcome of highly selected patients with advanced
HCC. Both may benefit from preoperative transarterial
chemoembolization and/or radiofrequency ablation.18 If the
lesion is unresectable, transplantation can be considered after
excluding extrahepatic disease.15,17 With multimodal therapy
to include liver transplantation, the long-term survival with
hepatoblastoma now exceeds 50%, while outcomes from
HCC are in excess of 70%.15,17,19 The major controversy that
exists is whether transplantation should be attempted for large
hepatoblastomas primarily or as a salvage after recurrence
following resection, and whether focal lung metastases contraindicate transplantation if they can be resected The most
common benign tumor of the liver is hemangioendothelioma,
and although the vast majority regress with growth and medical therapy, occasionally, progression of heart failure or mass
effect warrants transplantation.13,19
MISCELLANEOUS CONDITIONS
These conditions include diagnoses such as Budd-Chiari syndrome, trauma, biliary cirrhosis secondary to intestinal failure,
and long-term total parenteral nutrition (TPN) use. The latter
is detailed in Chapter 49.
Organ Allocation and

Pretransplant Care
------------------------------------------------------------------------------------------------------------------------------------------------
Patients who have evidence of decompensated cirrhosis are

candidates for liver transplantation. However, the small size
of the pediatric patient combined with a nationwide shortage
of organs relative to wait-listed patients makes achieving transplantation in a timely fashion problematic. In 2008, there were
613 pediatric liver transplantations performed. In that same
year, there were 773 pediatric-aged cadaveric liver donors.
The problem with a discrepancy between donors and recipients is primarily a problem in the less-than-6-year age category,
where there were 423 liver recipients but only 274 donors.2
Another problem is, of course, timing: When a pediatric donor
is available, there is not necessarily a size-matched pediatric
recipient available. This creates a relative shortage of organs
that necessitates a system to allocate these scarce organs to
those who will derive the most benefit. Since 2002, the Pediatric End-Stage Liver Disease (PELD) score was implemented
to allocate organs based on this sickest first paradigm.9 The
PELD score consists of five variables: international normalized
ratio (INR), total bilirubin, serum albumin, growth retardation
(2 standard deviations below the median height or weight for
age), and young age (<1 or 1 to 2 years). Status I-A is used to
designate patients with fulminant hepatic failure, primary graft
nonfunction after transplantation, early hepatic artery thrombosis, and miscellaneous acute conditions. Unlike adults, pediatric patients with decompensated cirrhosis requiring
intensive care unit (ICU) stay for accepted reasons, and occasional exceptions can be listed as status I-B. This is because
their mortality is high despite an often minimal change in
PELD score with a decompensation requiring ICU stay. In an
effort to ameliorate the shortage of potential organs, the livers
of all donors 18 years of age and younger are preferentially allocated to pediatric recipients before being offered to adults.
646
PART IV
TRANSPLANTATION
Also, in 2006, organ allocation policy was changed so that

rather than allocating livers from donors less than age 18 years
locally, they are allocated at a regional level to children to increase the probability of use of the organ in pediatric patients
and to facilitate liver splitting. These changes in organ allocation combined with the widespread use of split-liver transplantation (see later) has markedly reduced waiting times and
positively impacted wait list mortality.3,20
Donor Procurement and

Hepatobiliary Anatomy
------------------------------------------------------------------------------------------------------------------------------------------------
HEPATOBILIARY ANATOMY
The performance of a donor hepatectomy or transplant operation requires a thorough understanding of foregut anatomy.
In addition, a very detailed understanding of this anatomy is
essential to the field of segmental liver transplantation and has
impacted hepatobiliary surgery. The blood supply to the liver
is based on a highly variable arterial and portal system. Venous
drainage is through the right, mid-, and left hepatic veins that
join the inferior vena cava, which traverses the dorsal surface
of the liver (the retrohepatic cava). The liver is composed of
the major right and left lobes that are separated by external
landmarks and further subdivided into the right anterior
and posterior sectors and left medial and lateral sectors. This
nomenclature is still used to describe major anatomic liver
resections. However, through the elegant anatomic techniques
of the pioneering surgical anatomist Claude Couinaud,
hepatic anatomy was found to be much more intricate
(Fig. 48-4). The Couinaud nomenclature describes nine
hepatic segments based on portal vein branching in relationship to the transverse plane (a cross-sectional plane located
at midpoints of the hepatic veins) and the longitudinal
planes of the individual hepatic veins (see Fig. 48-4). Each
of the segments is supplied by an independent portal and
hepatic arterial branch and drained by an independent

biliary radicle. The biliary tree is second only to the arterial
system in its variability. The hepatic venous drainage is
intersegmental.
Donor Operation
------------------------------------------------------------------------------------------------------------------------------------------------
The use of organs from cadaveric donors in pediatric liver

transplantation involves selecting an appropriate quality and
size-matched donor, organizing an experienced transplantation harvest team, and performing a precise technical operation that recognizes arterial anatomical variants and allows
for multiorgan procurement. The advent of segmental liver
transplantation has expanded the acceptable donor age to approximately 40 years. Preharvest donor management should
focus on maintenance of hemodynamic stability with adequate but not excessive volume loading, the minimization
of vasopressors, optimization of oxygenation without excessive use of positive end-expiratory pressure (PEEP), and correction of hypernatremia that results from diabetes insipidus.
In the stable donor, once these goals are achieved, the procurement operation can be performed. In the properly selected unstable donor, unnecessary delays are to be avoided, because
expedient hypothermic perfusion and cold storage only help
minimize the ongoing organ ischemia.
The donor operation begins with midline laparotomy and
median sternotomy for wide exposure (Fig. 48-5). The abdominal great vessels are exposed by a medial visceral rotation
of the right colon and small intestine, and the aorta and inferior mesenteric vein are cannulated. The liver quality is
assessed, and the biliary tree is flushed via the gallbladder.
After full systemic heparinization, the supraceliac aorta is
cross-clamped, and the intrapericardial inferior vena cava is
incised to exsanguinate the donor. Then cold-organ perfusion
is begun through the previously placed cannulas, and the abdominal cavity is immersed in ice to attempt achieving a liver
core temperature of 4 C. University of Wisconsin (UW) solution has been used as the standard solution in the United
States since 1987 when it was developed by Belzer and Southhard. This solution extended the limit of preservation to as
long as 12 to 18 hours, after which the incidence of primary
2
7
8
1
4
FIGURE 48-4 Segmental liver anatomy. The division of the liver into independently vascularized and drained segments is based on the parallel
bifurcation of the portal vein and hepatic artery.
FIGURE 48-5 Cadaveric organ procurement. A wide exposure with median sternotomy extending to midline laparotomy is made for multiorgan
procurement. This harvest resulted in the procurement of six organ grafts
from a single recipient, benefiting six different recipients.
CHAPTER 48
graft failure increases substantially. However, the acceptable

preservation time depends on numerous donor and recipient
variables and should still be minimized when possible. This
is particularly the case with reduced-size or split-liver transplantation. The UW solution is a hyperkalemic, hyperosmolar
solution that prevents cell swelling, maintains stable transmembrane electrical gradients upon reperfusion, by preventing efflux of intracellular potassium during storage, and
contains a variety of oxygen free radical scavengers. Many
centers are now using a histidine-tryptophan-ketoglutarate
solution because of its lower potassium content and
viscosity.21 Once procured, the harvest team typically transports the liver graft to the transplantation center and prepares
it for engraftment for a separate recipient team.
Segmental Liver
Transplantation: Living Donor,
Reduced Size, and Split
------------------------------------------------------------------------------------------------------------------------------------------------
The shortage of pediatric organs coupled with a significant

wait-list death rate has driven the development of alternative
organ sources. Three alternatives to use of a whole organ graft
are available to these patients: living donor, reduced-size and
split-liver grafts.7,20,22,23 Living donor transplantation was developed as an alternative to scarce whole organ grafts and typically uses a segment 2 and 3 (left-lateral sector, LLS) graft.
Because of the small but real risk of safety in a healthy donor,
reduced-size transplantation was simultaneously developed as
an alternative and involves resecting the LLS graft prior to or
after cold-organ perfusion and discarding the remaining liver.
Obviously, this benefits the pediatric recipient but wastes an
organ that could be used by an adult recipient. Splitting the
whole organ into a right trisegment and LLS graft to use in
an adult and pediatric recipient, respectively, was first
reported by Pichlmayr in Hanover in 1988. This can be done
either prior to cold organ perfusion (in-situ technique)
(Fig. 48-6) or after cold-organ perfusion and removal of the
liver from the donor (ex-vivo technique).22 This provides a
LLS: 2,3
RTS: 1,48
FIGURE 48-6 Cadaveric in-situ split-liver procedure. The liver is separated just to the left of the umbilical fissure into a right trisegment (RTS)
graft and a left-lateral segment (LLS) graft.
647
suitable graft for the pediatric population without worsening

the already severe organ shortage in the adult population. Although the initial results were discouraging, with increased
experience, the survival rates of patients and grafts are nearly
equal to whole organ and living donor transplantation, though
the risk of vascular and biliary complications is somewhat
higher.27,2325 At the University of California in Los Angeles
(UCLA), we are only performing living donor transplantations
when a whole or split graft is not available in a timely fashion
or for special indications. There has been a marked reduction
in transplant wait time since the routine use of segmental
grafts with this strategy.3,20,23
Liver Transplant Operation

------------------------------------------------------------------------------------------------------------------------------------------------
The performance of the whole organ cadaveric liver transplant

procedure has changed little during the last 2 decades.
Although there are tremendous individual and institutional
differences in the subtleties and how certain techniques are
used or not used, the basic steps in the procedure remain
the same. What follows is a description of how the procedure
is generally performed at UCLA today and has been applied in
more than 3000 cases.26 The procedure can be roughly divided into four major phases, each with its own anatomic
and physiologic challenges: hepatectomy phase, anhepatic
phase with engraftment, reperfusion with arterialization,
and biliary reconstruction. Perhaps the most challenging step
during liver transplantation is the hepatectomy. Coagulopathy,
portal hypertension, and poor hepatic and renal function create a surgical environment wherein continuous bleeding is
possible. During this phase, the anesthesiologist plays a key
role in maintaining volume by rapid transfusion, correcting
coagulopathy and fibrinolysis, and maintaining body temperature. The goal of this phase is to devascularize the liver by
ligating and dividing the hepatic artery and portal vein as well
as to mobilize the suprahepatic and infrahepatic vena cava to
enable removal. These goals are achieved while leaving in the
recipient adequate lengths of each vessel for later implantation
of the donor graft. In the majority of pediatric liver transplant
operations, the retrohepatic vena cava is retained as the liver is
dissected off the vena cava by dividing the tributaries from the
right and caudate lobes, and often only partial occlusion of the
vena cava is necessary. Meticulous but expedient surgical technique is essential during the hepatectomy to ensure optimal
patient outcome. During the anhepatic phase, the anesthesiologist must support certain aspects of hepatic function to prevent or treat acidosis, hypothermia, coagulopathy, and
occasionally fibrinolysis. In addition, they must ensure adequate circulating volume and maintain hemodynamic stability.
In children, venovenous bypass is rarely used.
While the patient is anhepatic, the liver graft is taken out of
hypothermic storage and engrafted. This begins with the
suprahepatic caval, followed by the infrahepatic caval and
the portal anastomoses. If the retrohepatic cava was retained,
the piggyback technique is used, in which the suprahepatic
cava of the graft is sewn to the cloacae created from the confluence of the recipient hepatic veins, and the donor infrahepatic cava is ligated (Fig. 48-7). Prior to reperfusion, the liver
is flushed with a cold colloid and albumin solution through
the donor portal vein to lessen the potential reperfusionassociated complications.
648
PART IV
TRANSPLANTATION
Suprahepatic caval
anastomosis
Portal venous
anastomosis
Hepatic arterial
anastomosis
Roux-en-Y
choledochojejunostomy
FIGURE 48-7 Whole organ engraftment. Both the standard orthotopic and piggyback techniques are depicted.
Reperfusion is then undertaken in a controlled manner.

Communication between the surgical and anesthesia teams
is essential to allow the anesthesiologist time to institute preparative and preventive measures. Reperfusion is undertaken
by first removing the suprahepatic vena cava clamp, then the
infrahepatic vena cava clamp and, lastly, the portal venous
clamp. As blood is reintroduced into the liver allograft and
allowed to drain into the right atria, many serious and lifethreatening complications can develop. The major challenges
encountered by the anesthesiologist at this point are lifethreatening hyperkalemia, acidosis, arrhythmias, and hemodynamic instability with or without surgical or coagulopathic
bleeding. Factors that contribute are the return of cold, acidotic, and hyperkalemic blood directly into the right atrium.
It is at this point that maintenance of physiologic stability by
the surgeon and anesthesiologist in the preceding phases, the
preoperative state of the recipient, and the intrinsic quality of
the graft converge to determine early graft function as well as
the course of the remainder of the operation. Without a doubt,
this is one of the most hazardous portions of the liver transplant process.
The hepatic arterial anastomosis is then performed. In general, arterial inflow is obtained from one of the branches of the
celiac trunk. However, in some instances, inflow from these
vessels is not adequate, thus necessitating the use of aortic
conduits. A conduit can be placed either on the supraceliac
or infrarenal aorta. In some cases, when the arteries are of very
small caliber (<3 mm), the arterial anastomosis is performed
prior to reperfusion. The biliary tree is then reconstructed by
choledochocholedochostomy or by Roux-en-Y choledochojejunostomy, the latter being more common in small children
and used exclusively in partial liver grafts because of the size
of the donor duct (Fig. 48-8). After ensuring sufficient hemostasis, drains are placed and the abdominal cavity is closed.
The patient is then transferred directly to the ICU. Segmental
transplantation, using split, reduced-size, or living donor
grafts, involves variations in the manner in which the anastomoses are performed, but the general steps are the same (see
Fig. 48-8).
IVC
PHA
SA
A
CA
IVC
PV
FIGURE 48-8 Left-lateral segment engraftment. A, aorta; CA, celiac artery; IVC, inferior vena cava; PHA, proper hepatic artery; PV, portal vein;
SA, splenic artery. (Reprinted with permission from Goss J, Yerziz H, Shackelton H, et al: In situ splitting of the cadaveric liver for transplantation.
Transplantation 1997;64:871-877.)
Post-transplant Care
------------------------------------------------------------------------------------------------------------------------------------------------
EARLY POSTOPERATIVE CARE

The early and long-term postoperative care of the liver transplant recipient is almost as important as the performance of
the operation in ensuring optimal outcomes. The immediate
postoperative care is aimed at assessing graft function, providing supportive care for the recipient, and early detection of
complications. Graft function can be assessed in many ways.
Physiologic and clinical assessment can be done almost immediately with a warm, arousable, hemostatic, and hemodynamically stable patient whose liver is making golden-brown bile
(in the infrequent case where a biliary drainage tube is placed),
CHAPTER 48
characteristics that are the hallmarks of a functional graft.

Graft function is then confirmed biochemically by evidence
of synthetic and metabolic function (e.g., correcting prothrombin time, reversal of acidosis). The degree of preservation injury, as measured by liver enzyme levels, does not
linearly correlate with graft function, but grafts with severe injury are more likely to exhibit delayed function or nonfunction. Failure of a graft without vascular compromise
(primary nonfunction [PNF]) is treated by retransplantation
in almost all cases, with outcome directly related to the time
to retransplantation. The incidence of PNF in pediatric patients is 5% to 10%.7,8,23,24,27
Technical Complications
------------------------------------------------------------------------------------------------------------------------------------------------
Technical complications can be divided into vascular, biliary,

and general surgical complications. In the early postoperative
period, infectious and general surgical complications of liver
transplantation today are similar to those that occur after
any major abdominal operation. However, the incidence of
fungal infection is higher, and the incidence of bowel perforation in pediatric recipients is as high as 19% in some series.
Also, early exploration or computed tomography (CT) imaging should be considered when sepsis is suspected and no
other etiology can be found.
VASCULAR COMPLICATIONS
Major vascular complications include hepatic artery thrombosis (HAT), portal vein thrombosis, and caval thrombosis or
stenosis. Intravenous low-dose unfractionated heparin with
or without low-molecular-weight dextran is routinely used
for prophylaxis against vascular thromboses. Duplex ultrasonography and computerized tomographic or conventional
angiography are accepted means of diagnosis. HAT is the
most common complication, and its incidence varies from
5% to 18% depending on patient age and type of
graft.7,8,23,24,27,28 Early vascular complications are usually
technical in nature, while immunologic and infectious
(e.g., cytomegalovirus [CMV]) etiologies have been ascribed
to those occurring months after transplantation. HAT occurring in the first week is commonly associated with graft nonfunction and biliary necrosis or leak, while those occurring
later do not necessarily affect graft function immediately,
but usually produce biliary complications. These include
intrahepatic biliary abscesses, biliary anastomotic stricture,
and sclerosing cholangitis with sepsis, all of which lead to
significant morbidity. If diagnosed early, some patients can
be managed by thrombectomy and surgical revision. However, most early HATs require urgent retransplantation. Late
HATs with preserved graft function can be managed by radiologic interventional techniques and retransplanted remote
from initial transplantation. Thrombosis of the portal vein
occurs in 2% to 4% of pediatric liver transplantations and
is usually associated with loss of the graft. Prompt retransplantation is required for patient salvage. Late portal vein
thrombosis usually presents with recurrent variceal bleeding
or ascites and can be managed medically, endoscopically, or
surgically with shunting or retransplantation. Vena caval or
hepatic vein thrombosis or stenosis occurs in 3% to 6% of
649
pediatric patients and presents with variceal bleeding and/

or ascites and is usually best managed with balloon dilation
in interventional radiology 7,8,23,24,27,28
BILIARY COMPLICATIONS
Biliary complications that are not associated with HAToccur in
3% to 20% of patients depending on the type of graft and
whether a choledochojejunostomy was used. These usually
result from technical factors, but occasionally warm ischemia
or immunologic and infectious factors can be implicated (e.g.,
CMV). Diagnosis is by cholangiography and treatment can
be by endoscopic or radiologic intervention or by surgical
revision.7,2325,27
Immunosuppressive Therapy
and Rejection
------------------------------------------------------------------------------------------------------------------------------------------------
Immunosuppression for liver transplantation in the modern

era rests on a class of drugs known as calcineurin inhibitors
(CNI), the prototype being cyclosporine (Table 48-1). Cyclosporine, especially its microemulsion formulation, which allows better bioavailability and more consistent therapeutic
levels, revolutionized organ transplantation by reducing the
incidence of rejection in all solid organs. The secondgeneration CNI, tacrolimus, was first used clinically in
1990. The greater potency of tacrolimus allowed for a further
reduction in the early incidence of rejection following liver
transplantation, while also allowing the earlier weaning of steroid therapy. Also, the incidence of chronic rejection has significantly decreased with the use of tacrolimus, which can also
effectively treat episodes of acute rejection. Currently, most
liver transplant centers use a tacrolimus-based regimen combined with steroid therapy with or without adjunctive agents.
Cyclosporine and tacrolimus share certain acute and longterm side effects while having some that are unique to the
agent. The most important of these is nephrotoxicity, which
occurs in an acute variety from vasoconstriction of the afferent
renal arterioles and is reversible, as well as a more chronic
variety marked by tubular atrophy, interstitial fibrosis, and
glomerulosclerosis. The latter is variably reversible depending
on the degree of disease. To minimize acute toxicity and to
allow lower early CNI levels, especially with pretransplant renal
insufficiency, a purine antimetabolite, mycophenolate mofetil,
is sometimes used as an adjunctive agent (see Table 48-1).
The newest class of immunosuppressants is the inhibitors of
mammalian target of rapamycin, the prototype being sirolimus.
This agent has been used sparingly in pediatric liver transplantation, and only preliminary data exist. Although this drug
has no nephrotoxicity, it has other long-term sequelae, such
as hypercholesterolemia. No perfect immunosuppression
(i.e., one with minimal side effects) has been developed yet.
Acute rejection (AR) is common in pediatric liver transplantation, with the peak incidence being within the first
6 months, during which 30% to 50% of patients experience
at least one episode.7,8,23,27 It is less common after the first
post-transplant year, occurring in 10% or less of patients.
AR is suspected with elevated aspartate or alanine transaminase levels or by elevated alkaline phosphatase levels and
gamma-glutamyl transferase levels. AR is an alloantigen
specific, T-cellmediated inflammatory process that targets
650
PART IV
TRANSPLANTATION
TABLE 48-1
Modern Immunosuppressants Used in Liver Transplantation
Name
Mechanism of Action
Principal Use
Common Toxicities
Calcineurin
inhibitors (CNI)
cyclosporine
tacrolimus
Exact and complete mechanism

unknown; inhibits IL-2 and other
cytokine gene transcription, thus
preventing T-helper cell expansion
Induction and maintenance of

immunosuppression long term; tacrolimus
is the only agent approved for
monotherapy, while cyclosporine must
generally be used with another agent
long term
Glucocorticoids
methylprednisolone
prednisone
Diffuse action on immune system by its

anti-inflammatory properties,
especially inhibition of IL-1
Induction of immunosuppression and

maintenance; may be weaned off long term
in some patients
Mycophenolate
mofetil
Purine antimetabolite, semiselective

for salvage pathway present primarily
used in lymphocytes
Inhibits cell cycle progression in
stimulated cells, thus preventing clonal
expansion of stimulated B and T cells
Used as an adjunctive agent to reduce the

dose of CNI or steroid
Shared: nephrotoxicity,
hypertension, hyperglycemia,
neurotoxicity (seizures, myoclonus,
essential tremors)
Cyclosporine: hirsutism, gingival
hyperplasia, more diabetes
Tacrolimus: diarrhea, anorexia, more
neurotoxicity, more hypertension
Hyperlipidemia, osteopenia,
hypertension, diabetes, impaired
wound healing, growth retardation,
Cushingoid features, striae, acne
Myelosuppression, diarrhea,
anorexia, nausea, vomiting, GI
mucosal ulceration
Hyperlipidemia, impaired wound
healing, pneumonitis, oral ulceration
Mammalian target
of rapamycin
inhibitors
rapamycin
OKT3 monoclonal
antibody
Antilymphocyte
globulin
(Thymoglobulin)
IL-2 receptor
antagonists
basiliximab
daclizumab
Clonal deletion of (CD3) T cells

Exact and complete mechanism
unknown, but produces central and
peripheral deletion of lymphoid cells
Competitive inhibition of IL-2
receptors
Unclear, use in pediatric patients

preliminary; may be useful in minimizing
CNI dose when toxicity exists or in
refractory AR or chronic allograft rejection
Severe or refractory acute allograft
rejection
Severe or refractory acute allograft
rejection
Induction of immunosuppression as an
adjunct to CNI; used to minimize other
immunosuppression (CNI, steroids)
SIRS and other infusional reactions, increased risk of viral infections and PTLD
Increased risk of viral infections and
PTLD, lower incidence of infusional
reactions than OKT3,
thrombocytopenia
Increased risk of viral infections,
possible PTLD, rare infusional
reactions
AR, acute rejection; GI, gastrointestinal; IL-1, IL-2, interleukin-1, interleukin-2; PTLD, post-transplant lymphoproliferative disorder; SIRS, systemic inflammatory
response syndrome.
vascular endothelium and biliary epithelium but not hepatocytes. This is based on the greater expression of donor human
leukocyte antigens on the former cell types. The histologic
hallmark of AR is a mixed cell inflammatory infiltrate (polymorphonuclear cells, lymphocytes, and eosinophils) in the
portal triad, with evidence of inflammation of the endothelium and/or biliary epithelial injury. Rejection can be graded
as mild, moderate, and severe depending on the proportion
of involved portal triads, the degree of infiltrate and injury,
and the presence of central vein endothelial inflammation,
which is a sign of severe AR. Treatment of AR is centered
on a high-dose methylprednisone bolus, but cases unresponsive to this may require use of antibody therapy (OKT3,
antithymocyte globulin [ATG], see Table 48-1). Mild AR can
often be treated by simply increasing the tacrolimus level,
though steroid bolus should be considered if there is not a
prompt response. AR does not influence long-term graft survival in adults or children, unless it occurs in multiple or steroid refractory episodes, or if it occurs beyond the first year
post-transplant.4,7,29 AR accounts for less than 3% of overall
patient and graft loss. However, treatment for AR is an important risk factor for the development of cytomegalovirus and
Epstein-Barr viral infections in children. The latter is a risk factor for the development of post-transplant lymphoproliferative disorder. Therefore a balance between adequate
immunosuppression to prevent AR and over-immunosuppression to avoid toxicity is necessary. Currently, long-term
morbidity from immunosuppressive drug therapy is the major
challenge facing long-term survival and quality of life in the pediatric population.
Chronic rejection is a common cause of late graft loss in

children, because disease recurrence is uncommon. It is not
felt to be entirely alloantigen driven, and may be due to a number of factors that share final common pathway of graft injury.
Its hallmark is the intrahepatic loss of bile ducts and has been
termed vanishing bile duct syndrome in the past because of
this histologic finding on biopsy. This diagnosis is suspected if
progressive jaundice and rising levels of alkaline phosphatase
occur. Currently, there is no prophylactic or therapeutic agent
for chronic rejection, though progression of graft fibrosis may
be forestalled by sirolimus, based on animal and preliminary
clinical data.30 Although maintenance immune suppression is
often enhanced, the only definitive treatment when decompensated graft failure occurs is retransplantation.
Infectious Complications
------------------------------------------------------------------------------------------------------------------------------------------------
Post-transplant infections are the most common cause of morbidity and mortality after liver transplantation (Table 48-2).
The highest incidence of bacterial and fungal infections is in
the first month after transplantation. Fungal infections occurring months to years after transplantation are unusual and are
more commonly the atypical or endemic organisms, such as
Cryptococcus spp., Mucormycosis spp., Blastomycosis spp., or
Coccidiomycosis spp. Viral infections are the most common infections after the early post-transplant period. Cytomegalovirus (CMV) and Epstein-Barr virus (EBV) infections account for
the vast majority of opportunistic viral infections. Overall
CHAPTER 48
651
TABLE 48-2
Infectious Complications after Liver Transplantation
Organism
Presentation
Diagnosis
Antimicrobials
Cytomegalovirus
(CMV)
Infection results from reactivation of virus;

blood transfusion; infected transplanted
organ
Mild viral, flulike syndrome
Invasive tissue infection (retinitis,
pneumonitis, myocarditis, enterocolitis,
hepatitis, CNS)
Spectrum: infectious mononucleosis to
lymphoproliferative disease to lymphoma
to EBV-associated soft tissue tumors
Occurs with EBV and immunosuppression,
10%-15% infant liver transplantation
GI tract, neck, thorax, CNS
Quantitative CMV-DNA
PCR
pp65 antigen
Tissue cultures
Blood or fluid cultures
Biopsy with
immunostains
Quantitative EBV-DNA
PCR
Blood smear
Biopsy with
immunostains
CT scans of suspected
sites
HSV-1 and HSV-2
antibodies
Biopsy with viral cultures
BAL, lung biopsy
Prophylaxis: IV ganciclovir, oral valganciclovir

Therapy: IV ganciclovir with or without CMV
immunoglobulin
Epstein-Barr virus
(EBV)
Herpes simplex
virus (HSV)
Pneumocystis
Skin lesions, GI tract disseminated

herpesfever, fatigue, abnormal liver
functions, hepatitis, pneumonia
Atypical pneumonia, can progress to
life-threatening pneumonitis
Candida
Local mucous membrane, invasive tissue

infection, fungemia
Aspergillus
Entry via upper or lower respiratory tract

with metastatic spread (CNS, intraabdominal, solid-organ)
Gram negative: enterobacteria, E. coli,
Pseudomonas
Gram positive: Enterococcus, Staphylococcus
Bacteria
Blood, fluid, and tissue

cultures, fundoscopic
exam
Blood, fluid, and tissue
cultures, BAL, CT scans
Blood, fluid and tissue
cultures, BAL, CT scans,
surgical exploration
Prophylaxis: IV ganciclovir, oral valganciclovir

Therapy: Acyclovir, reduction or withdrawal of
immunosuppression; possible use of systemic
chemotherapy for lymphoproliferative disorders
or lymphoma
Acyclovir
Prophylaxis: low-dose oral sulfamethoxazole/

trimethoprim, Dapsone, or pentamidine
Therapy: high-dose IV sulfamethoxazole/
trimethoprim
Prophylaxis: fluconazole, possibly lipid formulation
of amphotericin B in very-high-risk patients
Therapy: fluconazole (for sensitive candidal species)
or lipid formulation of amphotericin B, caspofungin,
or voriconazole (insensitive Candida or Aspergillus)
Varies
BAL, bronchoalveolar lavage; CNS, central nervous system; CT, computed tomography; GI, gastrointestinal; IV, intravenous; PCR, polymerase chain reaction.
reduction and more selective immunosuppression and prophylaxis with ganciclovir have reduced the incidence and
morbidity of these infections. The other agents responsible
for infectious morbidity, their presentation, diagnosis, and
treatment are included in Table 48-2. Of particular importance in children is the prophylaxis and effective treatment
of EBV. This is associated with the development of numerous
malignant consequences. The most common of these is a diffuse proliferation of lymphoid tissue known as post-transplant
lymphoproliferative disorder (PTLD). PTLD can present as a
mononucleosis-like syndrome with diffuse lymphadenopathy
or as lymphoma involving any organ. A variety of other tumors
are also associated with EBV infections.29 The general therapy
for PTLD is reduction or elimination of immunosuppression,
and, occasionally, surgical intervention and/or chemotherapy.
The complete discussion of these disorders is beyond the
scope of this chapter but is extensively covered elsewhere.29
Outcome and the Future

------------------------------------------------------------------------------------------------------------------------------------------------
Numerous factors are known to impact patient and graft

survival in this population of liver transplant patients.*
Overall, survival has improved with 1-year and 5-year
patient survival exceeding 90% and 80%, respectively, in patients less than 18 years of age.3,4 Age, nutritional status, urgency of transplantation, the indication for transplantation,
* References 3, 4, 7, 8, 23, 24, 27, 31.
and presence of renal dysfunction are all major factors that

determine the outcome in any individual patient. Although
early data suggested that patients with BA have worse outcomes because of their often-malnourished state, young age,
and previous surgical intervention, more recent data suggest
that this difference is not significant.11,31 Patients with metabolic disease do exceedingly well, because they often are
older, do not have liver failure and its sequelae, and have
not previously undergone abdominal operation. Finally,
transplantation for malignancy in children is associated
with survival that is substantially less than that of transplantation for other indications but much better than
the natural history of the disease. Numerous large series
exist in the literature detailing the improvement in outcome with experience.4,7,8
Although outcomes have improved, many issues still remain to be resolved. The first and foremost is the organ
shortage. The number of listed patients is increasing, while
the number of suitable donors (even with segmental liver
transplantation and improved organ allocation policies)
has plateaued. Strategies aimed at expanding the donor pool
and allocating organs to those patients that not only have the
greatest survival benefit compared with pretransplant survival, but also the greatest chance of optimal post-transplant
outcome are essential. National policies aimed at effectively
identifying donors for splitting and development of local,
regional, and national sharing of split grafts still await
refinement. Finally, the development of gene therapy or
optimization of hepatocyte transplantation as alternatives
652
PART IV
TRANSPLANTATION
to transplantation for metabolic diseases may alleviate some

of the organ shortage.
Another important challenge for the liver transplant community is the perfection of immunosuppression. Currently,
all immunosuppressants have long-term side effects that result
in impaired growth and development, infectious morbidity, malignancies, and numerous medical complications such as renal
failure. The development of drug therapy that minimizes or
eliminates these is essential. Furthermore, a better understand-
ing of the immunology of peripheral T-cell tolerance and

chronic rejection is essential. Although the last decade saw improvements in many technical and immunosuppressive aspects
of liver transplantation, improvements in survival and quality of
life in the next decade will rest firmly on a better understanding
of our immune system on a cellular and molecular level.
expertconsult.com.
CHAPTER 49
Pediatric Intestinal
Transplantation
Yann Revillon and Christophe Chardot
Intestinal failure (IF) is characterized by the inability of the

gastrointestinal tract to provide sufficient digestion and absorption capacities to cover the nutritional requirements
for maintenance in adults and for growth in children.1
The first-line treatment for IF is parenteral nutrition (PN).
In patients with life-threatening complications of PN, intestinal transplantation (IT)isolated or combined with the liver
and/or other organsprovides children with a second chance
for survival. Since the early days of IT in the 1980s (treatment
with cyclosporine-based immunosuppressive regimens), significant progress has been made in the medical and surgical
management of children requiring IT, with short-term results
(1-year patient survival) now approaching those of liver
transplantation.
Indications for Intestinal

Transplantation
------------------------------------------------------------------------------------------------------------------------------------------------
The causes of intestinal failure in children can be divided into

five groups2 (Fig. 49-1): (1) short bowel, mainly resulting from
gastroschisis, midgut volvulus, necrotizing enterocolitis, and
intestinal atresia; (2) motility disorders: long-segment Hirschsprung disease and chronic intestinal pseudo-obstruction;
(3) epithelial disorders with intractable diarrhea, such as

microvillous inclusion disease and tufting enteropathy;
(4) children with a failed intestinal transplant; and (5) miscellaneous, including tumors.
Parenteral nutrition, including home PN, is the first-line treatment for children with intestinal failure and allows satisfactory
growth and acceptable (although not normal) quality of life in
most patients.1,3 However, life-threatening complications of
PN may occur, primarily line sepsis, loss of venous access resulting from thrombosis, and liver disease leading to cirrhosis.
In such cases, intestinal transplantation may be the only lifesaving alternative. Depending on the underlying disease and the
complications of PN, transplantation of additional organs may
be required: liver in patients with cirrhosis, stomach and duodenopancreas in patients with extended motility disorders, and
kidney(s) in patients with renal failure.4
The management of children with intestinal failure requires
a multidisciplinary approach and is a continuous process that
may last the whole life of the child. In most patients, the disease
starts in the neonatal period and requires initial surgery. The
possibility that the child may need other operations in the
future should be considered at each surgical intervention.
Adequate parenteral nutrition and prevention of line infections
have paramount importance for the long-term prognosis of
children with intestinal failure.3,5 If long-term dependence
on PN is expected, early contact with a team specializing in
the management of children with intestinal failure is recommended before the onset of PN-related complications, to
optimize the overall management of the child: adaptation of
long-term PN and prevention of its complications, education
of parents about home PN, and anticipation and preparation
of further steps of the medicosurgical management, which
may include nontransplant surgery or transplantation. This
early contact with the intestinal failure team is recommended
for every child whose requirements for PN are anticipated to
be more than 50% at 3 months after initiating PN.6 In a retrospective study, including 302 children followed for home PN in
our center between 1980 and 1999, the median duration of
home PN was 1.3 years. By January 2000, 54% had been
weaned from PN, 26% were still receiving PN, 16% had died,
and 4% had undergone intestinal transplantation. Patient
survival rates at 5, 10, and 15 years were 89%, 81%, and
72%, respectively. Nine percent of children with primary digestive disease died versus 38% of children with nonprimary
digestive disease.3 In a multicenter prospective European
study, including 688 adults and 166 children on home PN,
the candidacy rate for transplantation was estimated at 16%
in adults and 34% in children, and it varied greatly among
home PN centers.7,8 The 5-year survival rate on home PN
was 87% in noncandidates for transplantation, 73% in
candidates with home PN failure, 84% in those with high risk
of death attributable to the underlying disease, 100% in those
with IF with high morbidity or low acceptance of PN, and 54%
in ITrecipients (P < 0.001). Nontransplant surgery may have a
place in the rehabilitation program of such children, especially
bowel lengthening procedures in those with a short bowel and
limited hepatic or vascular complications.9,10
The appropriate timing of intestinal transplantation sometimes is not easy to determine, because it is difficult to predict
the ability of the native intestine to adapt, as well as the course
of PN-related complications. On the one hand, intestinal
adaptation may allow digestive autonomy in some patients with
653
654
PART IV
TRANSPLANTATION
Short bowel 63%

Motility disorders 17%
Epithelial disorders 9%
Miscellaneous 3%
Retransplantations 8%
FIGURE 49-1 Indications for intestinal transplantation. From the Intestine Transplant Registry 2003 report2: 606 grafts (223 isolated bowel,
306 liver and intestine, 77 multivisceral) in 563 children (age 18 years).
a short bowel,9 and PN-associated liver disease may regress after

optimization of PN, especially regarding the quality of lipid intake.1113 On the other hand, transplantation results are better
in children who are in good general condition, as opposed to
children with end-stage disease.2 However, early referral to a
specialized center for intestinal failure does not necessarily
result in finding an indication for transplantation. In a series
of 118 patients referred to our center for transplantation assessment, 10 could be weaned off PN, 12 patients were unsuitable for
transplantation, 65 patients were listed for transplantation,
and 31 remained potential candidates.
The two following examples of children treated by our
team illustrate how the same condition (short bowel
syndrome) may require very different therapies, based on
the specifics of the patient.
Yasmine was born in 1971. In 1981, she presented with a
midgut volvulus and complete necrosis of the small bowel
and right colon. She underwent a duodenocolic anastomosis,
and she has been on PN since then. Because of loss of venous
access, an arteriovenous fistula was created in 1983. Today,
she eats normally and receives home PN 5 nights per week.
She has three bowel movements per day. Her general condition is good. She sometimes complains of anal burns or
abdominal distension. She has an ovarian cyst and renal
stones. She works in business, was married in 2002, and gave
birth to two children. She enjoys dancing, skiing, and diving.
She does not wish to receive an intestinal transplant.
Virginie was born in 1988, and underwent an extensive intestinal resection at birth after midgut volvulus. Because of
complications of PN, she underwent intestinal transplantation,
in 1989, with cyclosporine-based immunosuppression. Today,
she eats normally and is off PN, with normal intestinal biopsies.
Her weight and height are normal. She has mild intellectual
retardation and does not work. Her renal function is moderately impaired. She is the worlds longest survivor (22 years)
with a functional intestinal graft. However, it is still impossible
to predict whether this situation will last for a normal lifespan.
Assessment and Preparation

for Intestinal Transplantation
------------------------------------------------------------------------------------------------------------------------------------------------
Potential candidates for IT usually have a complex medical

history and may have undergone several prior operations.
A detailed workup is needed to precisely evaluate (1) the level
of IF and its potential reversibility; (2) the history of PN and

central-line complicationsnumber of catheters, number of
episodes of line sepsis, and bacteria involved (antibiotic
resistance profiles); (3) thrombotic complications and current
cartography of patent vascular access; (4) intestinal failure
associated liver disease (IFALD)liver fibrosis or cirrhosis,
jaundice, ascites, portal hypertension (esophageal and/or
gastric and/or peristomal varices, thrombocytopenia), liver
insufficiency; (5) surgical status of the abdomenprevious
operations, length and function of remaining bowel, stomas;
(6) function of other organs, especially heart, lungs (pulmonary shunts or pulmonary hypertension), and kidneys; (7)
neurologic development and potential neurologic impairment; (8) serologic status and immunizations; (9) immunologic status (anti-HLA antibodies); and (10) sociofamilial
and psychological assessment and ability of the family to manage the child before and after transplantation. The indications
for transplantation as well as all of these issues and alternate
therapies are discussed in a multidisciplinary meeting.
Whatever the proposal for treatment, the child will require
careful follow-up, with further reassessments, because the
indication for transplantation and the type of graft needed
may change with time.
The general condition of the child has a strong impact on
the results of transplant surgery; therefore careful preparation
is required, focusing especially on (1) optimization of PN to
improve nutritional status and reduce its toxicity; (2) prevention and treatment of infections (optimization of central
venous-line management) and immunizations; (3) treatment
of the complications of liver disease, primarily ascites and
portal hypertension (sclerosis or ligation of esophageal
varices, transjugular intrahepatic portosystemic shunt). Education of the child and family about transplantation is done
simultaneously.
Transplantation Surgery
------------------------------------------------------------------------------------------------------------------------------------------------
The donor is usually a deceased donor, although a few livingrelated donations have been reported for isolated small bowel
transplants.14 The volume of the graft must correlate with the
abdominal cavity of the recipient; this depends on (1) the
donor to recipient weight ratio, (2) the native organs removed
and the type of graft implanted, and (3) whether the abdominal cavity is small (IT for short bowel) or distended (IT
for intestinal motility disorders with chronic intestinal
distention).
A wide variety of grafts can be implanted, from isolated
small bowel to multivisceral grafts, including stomach, pancreas and duodenum, small bowel, right colon, liver, and kidneys (Fig. 49-2). These grafts can be classified as
1. Isolated intestinal transplantation: small bowel right
colon. This type of graft is generally indicated for IF with
normal motility of the stomach and duodenum, and without significant liver disease. The native stomach, duodenum, pancreas, spleen, and liver are preserved. The
superior mesenteric artery of the graft is connected to
the recipient infrarenal aorta, and the mesentericoportal
axis of the graft is joined to the native infrarenal vena
cava. The proximal jejunum of the graft is connected to
the native jejunum.
CHAPTER 49
B
FIGURE 49-2 A, Multivisceral transplantation: the graft before implantation. The graft includes stomach, pancreas, duodenum, small bowel, right
colon, liver, and two kidneys. B, Multivisceral transplantation: the
graft after implantation and reperfusion.
2. Modified multivisceral transplantation: stomach, pancreas

and duodenum, small bowel right colon. This type of
graft is indicated for IF with impaired motility of the native
stomach and duodenum (i.e., pan-intestinal Hirschsprung
disease and chronic intestinal pseudo-obstruction), without significant liver disease. The upper part of the native
stomach, duodenum, pancreas, spleen, and liver are preserved. The arterial axis of the graft (including celiac trunk
and superior mesenteric artery) is connected to the recipient infrarenal aorta, and the mesentericoportal axis of the
graft is joined to the recipient infrarenal vena cava. The native and transplanted hemistomachs are connected, the
native first portion of duodenum is closed, and the native
jejunum is connected to the transplanted jejunum as a
Roux loop.
3. Combined liver and intestinal transplantation: liver,
pancreas and duodenum, small bowel, right colon. This
type of graft is indicated for IF with normal motility of the
stomach and duodenum, and with significant liver disease.
The native liver is removed, and a portocaval anastomosis
is fashioned between the native portal vein and vena cava.
The native stomach, duodenum, pancreas, and spleen are
PEDIATRIC INTESTINAL TRANSPLANTATION
655
preserved. The arterial axis of the graft (including celiac

trunk and superior mesenteric artery) is connected to
the recipient infrarenal aorta, and the suprahepatic vena
cava of the graft is connected to the native suprahepatic
vena cava in a piggyback fashion. The first portion of
the duodenum of the graft is closed, and the graft jejunum
is connected to the native jejunum as a Roux loop.
4. Multivisceral transplantation: liver, stomach, duodenum
and pancreas, small bowel, right colon, kidneys. This
type of graft is indicated for (1) IF with impaired motility of
the stomach and duodenum (pan-intestinal Hirschsprung
disease and chronic intestinal pseudo-obstruction), with
significant liver disease; (2) when en-bloc ablation of native
organs (liver, pancreas, duodenum, and intestine) is
needed,4 either due to previous surgeries and portal hypertension, making selective dissection of native abdominal
organs impossible, or (rarely) because of a tumor. All
abdominal organs anterior to the aorta and vena cava are
removed. The arterial axis of the graft (including celiac
trunk and superior mesenteric artery) is connected to
the recipient aorta, and the suprahepatic vena cava of
the graft is connected to the native suprahepatic vena cava
in a piggyback fashion. The native lower esophagus is
connected to the transplanted stomach.
In all cases, a distal ileostomy is performed to provide easy
access to the graft for intestinal biopsies. When the right colon
is transplanted, its distal end is either anastomosed to the
remaining native rectum (patients with short gut or mucosal
diseases) or a temporary distal colostomy is created (Hirschsprung disease). Cholecystectomy and gastrostomy (for
continuous enteral feeding) are generally performed if not
done previously.
Abdominal wall closure is an issue after intestinal transplantation, because of the size discrepancy between the graft
and the abdominal cavity of the recipient and post-reperfusion
edema of the graft. Reduction of the liver and/or intestinal
graft is possible.15,16 Staged abdominal closure, using a temporary Silastic sheet and a vacuum dressing, avoids abdominal
compartment syndrome. The final abdominal closure is
usually possible after 5 to 7 days (Fig. 49-3).
Postoperative Care
------------------------------------------------------------------------------------------------------------------------------------------------
The intestinal transit generally resumes quickly after surgery,

and enteral feeding is progressively introduced 2 to 7 days after the operation. Transit time is accelerated because of the
grafts denervation, and if dysfunction of the graft (mainly rejection or infection) has been ruled out, antimotility agents,
such as loperamide or codeine, can be used to slow down
the intestinal motility. In an uncomplicated postoperative
course, full enteral feeding can be achieved 1 month after
the operation.
Various surgical complications may occur (obstructions,
peritonitis, fistulas, and pancreatitis), and may be difficult to
detect under steroid therapy. Vascular monitoring of the graft
relies on observation of the color of the stoma and, if the graft
includes the liver, repeated ultrasonography (US) of the liver.
The intestine is highly immunogenic, and IT requires highlevel immunosuppression. As understanding of the mechanisms of rejection progresses and new immunosuppressive
656
PART IV
TRANSPLANTATION
drugs become available, immunosuppressive protocols

evolve. The current standard immunosuppressive regimen
is a combination of tacrolimus, steroids, and basiliximab
or daclizumab (antiIL-2 receptor antibodies). Monitoring
of intestinal rejection is based on stoma output, protein
concentration in the stools, and repeated intestinal biopsies
through the stoma. Other markers, such as stool calprotectin
or serum citrulline, have also been used. In case of
biopsy-proven rejection, first-line treatment relies on highdose steroid pulses. Second-line treatments are available but
expose the child to complications of overimmunosuppression, primarily opportunistic infections, and post-transplant
lymphoproliferative disease (PTLD). In case of uncontrolled
rejection, removal of the transplanted intestine may be
needed. Recently, humoral rejection has increasingly been
studied: Donor-specific anti-HLA antibodies are monitored,
and high levels can be treated by high-dose intravenous
(IV) immunoglobulins, rituximab (anti-CD20 antibody),
and plasmapheresis.
Because of the high level of immunosuppression, opportunistic infections are a constant threat after IT. Various infectious agents can be involved, the most common being
cytomegalovirus (CMV) and Epstein Barr virus (EBV). CMV
can cause severe graft enteritis and trigger rejection. EBV
can trigger lymphoproliferation. Monitoring of the viral loads
is currently determined by polymerase chain reaction (PCR),
which guides the prophylactic or curative treatments.
Post-transplant lymphoproliferative disease is nowadays
detected at earlier stages. First-line treatment relies on reduction of immunosuppression and rituximab.
Drug toxicity is an issue after IT, because these children
receive many drugs, some of them at high doses. Impairment
of renal function, hypertension, and seizures are the most
common side effects, which are usually reversible after dose
reduction or a switch to alternate therapies.
Progressively, all treatments are decreased or withdrawn.
Stoma closure can be considered when the child has been
stable, without rejection, under maintenance immunosuppression for several months.
Results of Intestinal
Transplantation
------------------------------------------------------------------------------------------------------------------------------------------------
C
FIGURE 49-3 A, Intestinal transplantation, end of operation. The edema
of the graft after reperfusion prevents primary abdominal closure. A Silastic
silo is performed. B, Staged abdominal closure. The Silastic silo is covered
with a vacuum dressing. The silo is progressively tightened over the next
days at the bedside, as edema progressively resolves and graft reintegrates
the abdomen. C, Final abdominal closure. After 5 to 7 days, the edema of
the graft has diminished, and final abdominal closure can be achieved; the
musculoaponeurotic layer can be closed either completely or with a
wound prosthesis (for instance, a GORE-TEX sheet).
Short-term results of intestinal transplantation have improved

with increasing experience.2 In the United States, current
1-year patient and graft survival is 89% and 79%, respectively,
for isolated bowel recipients, and 72% and 69%, respectively,
for liver-intestine recipients. However, medium-term results
remain unsatisfactory; by 10 years, patient and graft survival
falls to 46% and 29%, respectively, for isolated bowel recipients, and 42% and 39%, respectively, for liver-intestine recipients.17,18 According to the International Intestine Transplant
Registry (1985 to 2003 data: 989 grafts in 923 patients),2
causes of death were sepsis (46.0%), multiorgan failure
(2.5%), graft thrombosis (3.2%), graft rejection (11.2%),
post-transplant lymphomas (6.2%), respiratory causes
(6.6%), technical reasons (6.2%), and other causes (17.3%).
Our team in Paris performed 97 transplants in 90 children,
between November 1994 and April 2011, using tacrolimusbased immunosuppression: isolated bowel in 55; stomach,
pancreas, duodenum, and bowel (modified multivisceral)
in 1; combined liver and bowel in 39; liver, stomach, pancreas,
CHAPTER 49
PEDIATRIC INTESTINAL TRANSPLANTATION
657
Patient survival
1
.8
.6
.4
.2
0
0
10 12
14 16 18
Years after transplantation

90 patients
1-year
5-year 10-year 15-year
Patient survival
73.0%
61.8% 48.6%
Standard error
4.5%
Number of patients reaching considered follow-up
66
5.2%
43
43.7%
6%
6.3%
21
duodenum, and bowel in 1; liver, stomach, pancreas, duodenum, bowel, and two kidneys in 1 (see Fig. 49-2). In 63 of
97 transplants (65%), the graft included the right colon.
One-year, 5-year, 10-year, and 15-year patient survival rates
are 73.0%, 61.8%, 48.6%, and 43.7%, respectively, and
1-year, 5-year, 10-year, and 15-year graft survival rates are
59.5%, 45.0%, 33.6%, and 31.2%, respectively (Figs. 49-4
and 49-5). Early mortality is higher, but long-term graft survival is better after combined liver and intestine transplantation compared with isolated intestinal transplantation. This
is probably due to the protective effect of the liver against
intestinal rejection.19,20
In a study of 31 children treated by our group, and who
are alive with their graft 2 to 18 years after transplantation,21
FIGURE 49-4 Patient survival after intestinal transplantation in the tacrolimus era. Paris series from
November 1994 to March 2011: 90 patients.
all were weaned from PN after transplantation, and 26 of 31

(84%) remained PN-free at last follow-up. Enteral nutrition
was still required for 14 of 31 (45%) patients 2 years
after transplantation. All children had high dietary energy
intakes. The degree of steatorrhoea was fairly constant, with
fat and energy absorption rates of 84% to 89%. After transplantation, two thirds of children had normal growth,
whereas in one third, growth remained delayed, concomitant to a delayed puberty. Endoscopy and histology analyses were normal in asymptomatic patients. Five intestinal
grafts (16%) were removed 2.5 to 8 years after transplantation for acute or chronic rejection. Late complications also include impairment of renal function and
malignancies.
1
Graft survival
.8
.6
.4
.2
0
0
10 12
14 16 18
Years after transplantation

97 grafts
1-year
Graft survival
59.5% 45%
Standard error
Number of grafts reaching considered follow-up
5%
53
5-year 10-year 15-year
5.3%
31
33.6%
31.2%
5.7%
5.8%
14
FIGURE 49-5 Graft survival after intestinal transplantation in the tacrolimus era. Paris series from November
1994 to July 2010: 97 grafts.
658
PART IV
TRANSPLANTATION
Results of intestinal transplantation have improved in the

recent decades,2,18 because of better preparation of patients
and timing of transplantation, progress in surgical techniques,
availability of new immunosuppressive drugs and improved
immunosuppressive regimens, and better monitoring and
treatments of postoperative complications. The scarcity of
grafts remains an important issue, and patients still succumb
while waiting for a graft. With expected improvements in the
outcomes (especially in the long term), intestinal transplantation may move from a lifesaving procedure to an improving
quality-of-life procedure,22 which may also have economic
advantages compared with PN.
Conclusion
------------------------------------------------------------------------------------------------------------------------------------------------
Home PN remains the first-line treatment of intestinal

failure. Intestinal transplantation and its technical variants
are indicated only in case of life-threatening complications
of PN. Intestinal failure requires a multidisciplinary approach
in specialized centers. Early assessment of the child in such a
center, before the onset of complications of PN, is recommended. This does not mean early transplantation, but
adequate planning of medicosurgical strategies to provide
the child with the best chances of survival and an optimal
quality of life.
Acknowledgments
The authors, Yann Revillon and Christophe Chardot, who are part of the surgical team, wish to thank the following pediatric multidisciplinary team
members*:
1. Current team members treating intestinal failure and performing
transplantations:
Surgery: Sabine Irtan, Sabine Sarnacki, and Yves Aigrain.
Gastroenterology, hepatology, and nutrition: Florence Lacaille, Virginie
Colomb, Cecile Talbotec, Franck Ruemmele, Muriel Girard, Dominique Debray, Jean-Pierre Hugot, and Olivier Goulet.
Pathology: Nicole Brousse, Virginie Verkarre, Danie`le Canioni, Julie
Bruneau, and Jean-Christophe Fournet.
Intensive care: Fabrice Lesage, Laurent Dupic, Jean Bergounioux, Olivier Bustaret, Sandrine Jean, and Philippe Hubert.
Radiology: Karen Lambot, Sophie Emond, Laureline Berteloot, and
Francis Brunelle.
Anesthesiology: Nade`ge Salvi, Nathalie Bourdeau, and Caroline Telion.
Research laboratory: Nadine Cerf-Bensoussan.
2. Former members of the team: Claude Ricour, Jean-Pierre Cezard, Dominique Jan, Jean-Luc Michel, Frederique Sauvat, Patrick Jouvet, and
Francis Jaubert.

expertconsult.com.
*All members are affiliated with Hopital Necker-Enfants Malades, Paris,

France, except Jean-Pierre Hugot and Jean-Pierre Cezard, who are affiliated
with Hopital Robert Debre, Paris, France.
Historical Notes
------------------------------------------------------------------------------------------------------------------------------------------------
CHAPTER 50
Heart
Transplantation
Stephanie M. P. Fuller and Thomas L. Spray
Thoracic organ transplantation has been successfully performed in pediatric patients since the mid-1980s and now
serves as an important option in the treatment of both congenital and end-stage heart and lung disease in children. Approximately 400 pediatric heart transplantations are
performed annually in the United States, or roughly 16%
of all pediatric solid organ transplantations.1 Despite the
clinical success of heart and lung transplantation in children,
limited donor availability has prevented broader application
of this therapy. Infants awaiting heart transplantation face the
highest wait-list mortality among all children and adults
listed for a heart transplantation in the United States, with
one in four infants dying before a donor heart can be identified.2 Complications, such as acute and chronic rejection,
graft coronary artery disease (CAD), and bronchiolitis obliterans, as well as the infectious and neoplastic complications
of current methods of immunosuppression, threaten cardiac
transplant longevity. This chapter focuses on the clinical
aspects of heart transplantation in infants and children,
including indications, preoperative evaluation, operative
techniques, postoperative management, complications, and
outcomes.
Kantrowitz and colleagues2a performed the first pediatric

heart transplant in 1967 when they transplanted the heart
of an infant with anencephaly into a 3-week-old infant with
tricuspid atresia. The next year, Cooley2b transplanted the
heart and lungs of a newborn with anencephaly into a
3-month-old with an atrioventricular septal defect and pulmonary hypertension. Although neither of the infants survived for
more than a few hours because of allograft rejection, these pioneering procedures emphasized the technical feasibility of thoracic organ transplantation in children. It was only in 1980 with
the introduction of cyclosporine as an immunosuppressive
agent that meaningful clinical success became possible. In November 1985, Bailey performed the first successful cardiac
transplantation on a 4-day-old neonate with hypoplastic left
heart syndrome (HLHS) at Loma Linda.3,4 During the last 2 decades, outcomes have been improved by technical advances,
better immunosuppression, including reduced steroid use
and the advent of induction therapy, a decreased incidence of
rejection, increased attention to viral prophylaxis, and aggressive treatment of post-transplant lymphoma and other posttransplant complications.
Indications
------------------------------------------------------------------------------------------------------------------------------------------------
As published by the Registry for the International Society for

Heart and Lung Transplantation in the Thirteenth Official
Pediatric Report in 2010, the number of pediatric heart transplantations has remained relatively constant during the last
10 years (Fig. 50-1).5 The most common indications for cardiac
transplantation in the pediatric population remain congenital
cardiac disease and cardiomyopathy, as demonstrated in
Figures 50-2 and 50-3. Congenital heart disease is seen more
commonly in infants, whereas cardiomyopathy is more
prevalent in older children. As expected, the incidence of
retransplantation increases with increasing patient age.
When examining the congenital heart disease population,
the most common anomaly treated by transplantation is hypoplastic left heart syndrome (HLHS), a group of defects characterized by aortic or mitral atresia/stenosis with a diminutive
left ventricle. Initial poor results with a staged palliative
approach to HLHS led some centers to consider orthotopic
heart transplantation as the primary treatment of this anomaly.
Long transplantation waiting lists have led other institutions
to advocate a stage I palliation (Norwood or Sano procedure)
to help stabilize the patient and then list the patient for transplantation.6 However, with improvement in early survival
from the Norwood procedure followed by a Fontan repair,
the majority of cardiac centers have abandoned primary transplantation as initial therapy for HLHS. Transplantation is an
option now reserved for patients with unusually high risk,
including aortic atresia with a diminutive ascending aorta
and severe tricuspid or atrioventricular valve regurgitation.7
Other forms of congenital heart disease that have been treated by cardiac transplantation during infancy include an
unbalanced atrioventricular canal, single ventricle, the Ebstein
659
Number of transplants
660
PART IV
500
450
400
350
300
250
200
150
100
50
0
TRANSPLANTATION
1117 years
110 years
<1 year
82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08
19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20
NOTE: This figure includes only the heart transplants that are reported
to the ISHLT Transplant Registry. As such, this should not be
construed as evidence that the number of hearts transplanted
worldwide has increased and/or decreased in recent years.
anomaly, L-transposition of the great arteries, and pulmonary

atresia with an intact ventricular septum (Table 50-1).8,9 Even
the most complex forms of congenital heart disease, such as
heterotaxy syndromes with anomalies of systemic and venous
drainage, are amenable to cardiac transplantation with suitable reconstruction.10 Other pediatric candidates include infants with congenital heart disease who have undergone
previous corrective or palliative procedures, yet who exhibit
residual or progressive cardiac dysfunction manifested by left
ventricular failure that ultimately requires transplantation.
Postoperative cardiac dysfunction is often related to atrioventricular or semilunar valvar insufficiency that eventually results in dilated cardiomyopathy. In some cases, ventricular
function may be preserved, but the indication for transplantation is for hemodynamic compromise secondary to anatomic
abnormalities not amenable to surgical intervention, intractable arrhythmias, or complications that arise following the
Fontan operation, such as protein-losing enteropathy. Of note,
FIGURE 50-1 Age distribution of heart recipients by year of

transplantation. (From Kirk K, Edwards LB, Kucheryavaya AY, et
al: The Registry of the International Society for Heart and Lung
Transplantation: Thirteenth official pediatric heart transplantation
report2010. J Heart Lung Transplant 2010;29:1119-1128.)
multiple previous palliative procedures do not preclude successful transplantation.11 In the current era, despite the need
for repeat sternotomy, the tendency toward diffuse coagulopathy and potentially prolonged ischemic times, patients undergoing transplantation for congenital heart disease experience
no actuarial difference in survival compared with those
patients undergoing transplantation for cardiomyopathy
who undergo first-time sternotomy and dissection of mediastinal structures.8,12
Cardiomyopathy is the other most common indication for
heart transplantation in infancy and childhood. Most pediatric
heart transplantations outside infancy are performed for dilated, idiopathic cardiomyopathy. Other causes of cardiomyopathy include viral, familial, and hypertrophic. Despite the
diverse causes of cardiomyopathy, several variables have been
associated with poor outcome, including a very high left
ventricular end-diastolic pressure, a left ventricular ejection
fraction less than 20%, ventricular arrhythmia, and a family
33%
Myopathy
17%
Congenital
0%
1%
Other
2%
80%
3%
ReTX
19881995
63%
1/19966/2009
% of cases
100
75
50
Myopathy
Congenital
25
0
86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09
19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20
FIGURE 50-2 Infant heart recipient diagnosis according to year of transplantation. ReTx, retransplant. (From Kirk K, Edwards LB, Kucheryavaya AY, et al:
The Registry of the International Society for Heart and Lung Transplantation: Thirteenth official pediatric heart transplantation report2010. J Heart Lung
Transplant 2010;29:1119-1128.)
CHAPTER 50
HEART TRANSPLANTATION
661
Myopathy
25%
27%
Congenital
68%
Other
65%
3%
3%
8%
ReTX
2%
1/19966/2009
19881995
% of cases
100
Myopathy
75
Congenital
50
25
0
86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09
19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20
FIGURE 50-3 Diagnosis of heart recipients aged 11 to 17 years according to year of transplantation. ReTx, retransplant. (From Kirk K, Edwards LB, Kucheryavaya AY, et al: The Registry of the International Society for Heart and Lung Transplantation: Thirteenth official pediatric heart transplantation report
2010. J Heart Lung Transplant 2010;29:1119-1128.)
history of cardiomyopathy.13,14 Cardiomyopathy attributable

to inflammation or arrhythmia tends to have a more favorable
outcome, and these patients should be supported as long as
possible before transplantation to allow for the possibility of
spontaneous recovery. Other less common indications for cardiac transplantation are doxorubicin-induced cardiotoxicity
from chemotherapy for malignancy and obstructive cardiac
tumors, such as fibromas and rhabdomyomas that are not
amenable to surgical resection.
Regardless of the diagnosis, there are several clinical indications for heart transplantation in children, many of which
have been borrowed from the adult population. These include
the need for ongoing intravenous inotropic support or mechanical circulatory support. Transplantation is indicated in
those patients who experience a progressive deterioration of
ventricular function despite optimal medical care or those
with life-threatening arrhythmias unresponsive to medical
treatment, ablation, or automatic implantable defibrillator.
Progressive pulmonary hypertension secondary to systemic
TABLE 50-1
Distribution of Anatomic Diagnoses in Children Older Than
6 Months of Age and Adults Undergoing Transplantation for
Congenital Heart Disease in the Pediatric Heart Transplant
Study and Cardiac Transplant Research Databases
Diagnosis
Single ventricle
Dextrotransposition of the great arteries
Right ventricular outflow tract lesions
Ventricular/atrial septal defects
Left ventricular outflow tract lesions
Levotransposition of the great arteries
Complete atrioventricular canal
Other
Total
% of Patients
176
58
49
38
38
39
37
53
488
36
12
10
8
8
8
8
11
100
Modified from Chen JM, Davies RR, Mital SR, et al: Trends and outcomes in
transplantation in complex congenital heart disease: 1984-2004. Ann
Thorac Surg 2004;78:1352-1361.
ventricular failure is an indication for cardiac transplantation,

whereas pulmonary hypertension may otherwise be an indication for lung transplantation. Growth failure secondary to severe heart failure and unacceptably poor quality of life are
considered indications as well as certain high-risk conditions
following the Fontan procedure, such as protein-losing enteropathy and plastic bronchitis.15
Preoperative Evaluation
------------------------------------------------------------------------------------------------------------------------------------------------
The pretransplant evaluation is a multidisciplinary screening

process that serves as the key to successful organ transplantation (Table 50-2). Potential recipients go through a thorough
TABLE 50-2
Pretransplant Recipient Evaluation
Cardiac
catheterization
Exercise testing
Assessment of endorgan function
HLA sensitization
Psychosocial
evaluation
Financial evaluation
Blood work
Assess pulmonary vascular resistance and

reactivity
Delineate complex anatomy
Obtain mVo2
Pulmonary function tests to evaluate lung
function
Liver and kidney function
Panel reactive antibody
Social work assessment of family dynamic
Psychiatric evaluation
Insurance evaluation
Complete blood count with differential
chemistry
Lipid profile
Thyroid function
Hepatic function test
Blood type
Urinalysis
Viral testing for cytomegalovirus, EpsteinBarr virus, herpesvirus, HIV, varicella,
toxoplasmosis, hepatitis virus, tuberculosis
HIV, human immunodeficiency virus; HLA, human leukocyte antigen; mVo2

myocardial oxygen consumption.
662
PART IV
TRANSPLANTATION
physical and psychosocial evaluation with careful examination of the cardiac, pulmonary, neurologic, renal, infectious,
and socioeconomic systems. The presence of an adequate family support system is of paramount importance to survival
postoperatively. Parents must demonstrate the ability and
resources to comply with the complex medical regimens
required and to cope with the potential for long or frequent
hospitalizations even years after transplantation. As part of
this multidisciplinary evaluation, patients undergo screening
laboratory tests, including a viral serology panel (e.g., human
immunodeficiency virus [HIV], cytomegalovirus [CMV], human Epstein-Barr virus [EBV], hepatitis).
Cardiac evaluation is performed mainly by echocardiography and cardiac catheterization in which the anatomy of the
systemic and pulmonary venous connections of the heart
and lungs are precisely identified. Important hemodynamic
data, including systemic cardiac output and pulmonary vascular resistance (PVR), both indexed to the patients body area,
are obtained at cardiac catheterization and used to screen candidates. These numbers become significant, because the major
contraindication to transplantation is fixed pulmonary hypertension unresponsive to pulmonary vasodilators. Patients with
elevated PVR (>4 to 6 Wood units) are tested with pulmonary
vasodilators, including sodium nitroprusside, oxygen (Fio2
100%), and inhaled nitric oxide, to establish whether the pulmonary vascular bed is reactive. In general, the presence of a
fixed PVR in excess of 6 to 8 Wood units is a contraindication
to orthotopic heart transplantation, because the donor heart is
unable to tolerate right-sided dilation caused by high pulmonary vascular resistance. Patients who demonstrate improvement with vasodilators may undergo transplantation with a
survival rate comparable to that in patients with normal
resistance.11 Although patients with fixed pulmonary hypertension have successfully undergone transplantation, they
have a much higher mortality rate, usually because of postoperative right ventricular failure. Other contraindications
to cardiac transplantation include multiple noncardiac congenital anomalies, active malignancy, infection, severe metabolic disease (i.e., diabetes mellitus), multiple organ failure,
multiple congenital anomalies, and the lack of an adequate
family support system, in addition to socioeconomic factors
that lead to noncompliance with drug regimen and followup care (Table 50-3).
TABLE 50-3
Potential Contraindications to Cardiac Transplantation
General
Specific
Presence of any noncardiac condition that significantly

shortens life expectancy
Active infection
Active ulcer disease
Active neoplasm
Morbid obesity (BMI > 32)
Renal insufficiency with creatinine greater than 2 times
normal
Hepatic dysfunction with elevated transaminases or
cirrhosis
Elevated, nonreactive pulmonary vascular resistance
Recent pulmonary embolic event with infarction
Recreational drug use
Recurrent medical noncompliance
BMI, body mass index.
Children suffering from cardiomyopathy and manifesting

symptoms of chronic congestive heart failure that limit activity
or uncontrollable arrhythmias are often referred for transplantation, particularly if they are unresponsive to medications.
The timing for transplantation, especially for those children
with hypertrophic cardiomyopathy, is less clear because some
patients may improve with medication and conservative therapy. As previously stated, the mortality for idiopathic dilated
cardiomyopathy in children is highest in the first year after
diagnosis and is mainly determined by the degree of left
ventricular failure.
Children listed for heart transplantation should be closely
monitored until their transplantation, either as outpatients, if
their condition permits, or while hospitalized. Good nutritional status should be maintained, and supplementation,
such as tube feedings or total parenteral nutrition, is used
as needed. A close watch for infectious complications is important, and any subtle indications of infection should be
thoroughly investigated. Major infections require patients to
have their transplantation status put on hold until they are
treated adequately. Anticongestive therapy should be optimized with digoxin, diuretics, and afterload reduction with
captopril or other angiotensin-converting enzyme inhibitors.
If heart failure worsens, hospitalization may be required for
inotropic support with dobutamine or phosphodiesterase inhibitors such as milrinone. Long-term therapy may require the
placement of an intravenous access device such as a Broviac
catheter.
The use of mechanical support as a bridge to cardiac transplantation in critically ill children had been limited mostly to
those with postcardiotomy ventricular failure. In general, the
results have been poor, although several studies show survival
rates ranging from 45% to 73% when extracorporeal membrane oxygenation (ECMO) is used as a bridge to cardiac
transplantation.16,17 ECMO is restricted to short-term use.
Additional limitations are the inability to ambulate and
undergo effective physical therapy while on ECMO, as well
as the damage to circulating red blood cells and platelets requiring persistent transfusion. In the current era, children
have excellent survival with the use of long-term ventricular
assist devices (VADS) as a bridge to transplantation. Although
adult systems can be used in adolescent patients, the Berlin
Heart VAD (Berlin Heart AG, Berlin, Germany) is a pulsatile,
paracorporeal VAD that is suitable in neonates and infants for
both single and biventricular support. The North American
experience from 2000 to February 2007 details approximately
80 patients supported for more than 200 days, with the smallest patient being 3.0 kg. Overall, approximately 55% of the
patients have undergone transplantation, 13% were weaned,
and 25% died during device support.18
A neonate referred for cardiac transplantation requires several other unique considerations. Infants with complex congenital heart disease, such as HLHS, are commonly
confined to a neonatal intensive care unit and are usually
maintained on a continuous infusion of prostaglandin E1 to
prevent closure of the ductus arteriosus if there is ductaldependent physiology. Implantation of expansile stents in
the ductus may allow for discontinuation of prostaglandin
therapy while waiting. Initial palliative procedures, such
as the Norwood procedure for HLHS or a Blalock-Taussig
shunt for lesions with ductal-dependent pulmonary blood
flow, can be performed in the face of a prolonged wait for a
CHAPTER 50
donor. Balloon atrial septostomy, with or without stenting to

improve mixing of saturated and desaturated blood and to decompress the left atrium, can be helpful if there is a restrictive
patent foramen ovale. Other important issues are the maintenance of adequate nutritional support, avoidance of renal and
metabolic complications, and prompt and thorough treatment
of any infectious complications, especially line sepsis, in these
fragile infants. Common neonatal problems, such as seizures,
necrotizing enterocolitis, and intraventricular hemorrhage are
also seen. At the minimum, 10% to 20% of infants die while
awaiting a donor heart.
In all cases, important consideration is given toward
pretransplantation recipient human leukocyte antigen
(HLA) sensitization. Circulating antidonor antibodies may
result in either cellular or humoral rejection culminating in
early graft failure.19 The presence of HLA antibodies is
reported as a panel reactive antibody (PRA), and a panel percentage of greater than 10% is considered elevated. Patients
prone to developing anti-HLA antibodies include those
who have received blood and platelet transfusions during
prior surgeries, postgravid adolescent girls, children who
have undergone implantation of cryopreserved tissue valves
or allograft conduits, and patients with previous organ
transplants. Strategies for reduction in PRA include the use
of intravenous immunoglobulin as well as agents that may
inhibit antibody production by B cells. Candidates who are
high risk are managed with pretransplantation plasmapheresis
that is continued postoperatively, resulting in good short-term
outcomes. However, long-term outcomes are unknown.
The United Network for Organ Sharing determines organ
allocation and, in 2002, revised their classification for pediatric patients awaiting heart transplantation. Status 1A applies
to patients requiring ventilatory or mechanical circulatory
support (i.e., left ventricular assist device, ECMO, or a balloon
pump) or multiple- or high-dose inotropes, infants younger
than 6 months with pulmonary pressure greater than 50%
of systemic levels, or any patient with a life expectancy of less
than 14 days without a heart transplantation. Status 1B applies
to patients requiring single-dose inotropic support or infants
younger than 6 months who have significant failure to thrive
(less than the 5th percentile for weight or height or loss of 1.5
standard deviations [SD] of expected growth). All other
patients with less acuity are classified as status 2. A patients
status may change depending on changes in clinical condition, or the patient may be placed on hold (status 7) because
of an infectious, malignant, or other complication and then
reactivated.
Donor Evaluation and Organ

Procurement
------------------------------------------------------------------------------------------------------------------------------------------------
The criteria for an ideal organ donor are as follows: meets requirements for brain death, consent from next of kin, ABO
compatibility in older children, weight compatibility (1 to
3 times that of the recipient), normal echocardiogram, age
younger than 35 years, and normal heart by visual inspection
at the time of harvest. A history of cardiopulmonary resuscitation is not an absolute contraindication to cardiac donation
for pediatric recipients. All potential donors are evaluated
carefully for the cause of death, including the presence of
663
chest trauma, need for cardiopulmonary resuscitation, and

cardiac function before death. For neonates, most donors have
suffered sudden infant death syndrome or birth asphyxia,
whereas older donors are victims of violence and car
accidents.
The shortage of suitable organ donors, especially for
neonatal recipients, has led to many attempts at expanding
the donor pool. Hearts from donors with moderately impaired
ventricular function by echocardiography (left ventricular
shortening fraction greater than 25% without major wall motion abnormalities) have been successfully transplanted into
infant recipients.10 Donor-to-recipient weight ratios of up to
4:1 have been used in infants. Tamisier and colleagues demonstrated that the higher the PVR, the larger the donor heart
needed for successful transplantation and that hearts with
PVR values thought to be in excess of normal can also be
used.20 Although ideal donor ischemia time is from 2 to
4 hours, ischemic times have been successfully extended
beyond 9 hours. Deviations from the ideal donor criteria
should be individualized, and even though the use of a
marginal donor for a dying infant maintained on ECMO
may be justified, use of the same heart for a child who is stable
as an outpatient might not.
ABO-incompatible transplantation has been introduced as
a method to decrease recipient waiting time and associated
waiting list mortality.19 Because neonates do not have the ability to produce antibodies to T-cell antigens, including major
blood group antigens, ABO incompatibility becomes a negligible complication. ABO-incompatible transplantation has
been infrequently used in the United States, and the age at
which it is no longer feasible is still not clearly defined. Despite
ABO-incompatible listing, it has not yielded lower wait-list
mortality under the current UNOS allocation algorithm.21
Good donor management is a vital part of successful organ
transplantation. The main goals are maintenance of normothermia, euvolemia, and adequate tissue perfusion and prevention of infection. Often, donors with poor cardiac
function on initial evaluation will respond to volume loading
and low-dose inotropic support with a significant improvement in function after heart retrieval, usually as part of a multiorgan retrieval procedure. Similar to recipients, all donors are
screened for agents that might cause serious infection in an
immunocompromised host, such as CMV, EBV, HIV, hepatitis,
and Toxoplasma. The presence of antibodies is not a contraindication to transplantation but helps guide post-transplant
therapy.
The four major goals in procurement of a donor heart are to
(1) work effectively with the other teams to ensure the optimal
condition of each recovered organ, (2) evaluate the hemodynamic status of the patient and the gross function of the heart
by inspection, (3) use an effective cardioplegia and venting
procedure that maximizes preservation of the heart, and (4)
expertly remove the heart and adjoining vascular connections
to ensure optimal anatomy for implantation. Procurement is
performed through a median sternotomy. Donor blood is
obtained for viral titers and retrospective HLA typing. The initial dissection involves separating the aorta from the main pulmonary artery to allow cross-clamping. Careful inspection of
the heart is performed, and the patient is systemically heparinized. Procurement commences when the aorta is crossclamped. Cardioplegia solution is infused through the aortic
root, and the heart is vented through the right atrial
664
PART IV
TRANSPLANTATION
appendage or superior or inferior vena cava for the right side

and through the superior pulmonary vein or left atrial appendage for the left side. The superior vena cava is dissected
free of its pericardial attachments up to the innominate vein,
and the azygous vein is ligated and divided. The pericardial
reflections around the right superior pulmonary vein and
the inferior vena cava are sharply divided. The cardiectomy
begins with inferior vena cava transection at the pericardial reflection. The main pulmonary artery is divided and then the
posterior pericardial attachments and the superior vena cava.
Last, the aorta is transected at the level of the innominate artery or more distally if the aorta is needed for the recipient.
The donor heart is immersed in cold (4 C), sterile saline
and then triple-bagged in a sterile manner for transport. In
general, the cold ischemia time should be limited to a maximum of 4 to 5 hours.
Recipient Preparation and

Techniques of Implantation
------------------------------------------------------------------------------------------------------------------------------------------------
The standard technique for orthotopic heart transplantation

was first described by Lower and Shumway in 1960 and consists of biatrial anastomoses, thus avoiding individual caval
and pulmonary vein connections (Fig. 50-4).22 Currently,
however, the majority of cardiac transplant centers now use
the bicaval technique, because it preserves atrial morphology
and kinesis and is simpler when reconstruction after previous
congenital heart repair is necessary. Once adequate hemodynamic monitoring is in place and the recipient is properly
anesthetized, a median sternotomy is performed and the heart
is suspended in a pericardial cradle. If previous sternotomies
have been performed, appropriate precautions should be
taken, including exposing the groins in the sterile field for
access for femoral bypass. Once in the chest, the main pulmonary artery is dissected off the aorta past the bifurcation, and
the pericardial reflection is mobilized off the aortic arch. Normally, aortic and bicaval cannulation is used.
In the case of a neonatal recipient with HLHS, the aortic
arch vessels are mobilized proximally and controlled with
snares, and the descending thoracic aorta is dissected to a level
2 to 3 cm below the insertion of the ductus arteriosus. The
right and left pulmonary arteries are mobilized and controlled
with snares in preparation for cardiopulmonary bypass. After
heparinization, the main pulmonary artery is cannulated for
arterial inflow, and a single venous cannula is placed in the
right atrium, because circulatory arrest will be used. Immediately on instituting cardiopulmonary bypass, the pulmonary
arteries are snared tight and the body perfused though a patent
ductus arteriosus. The recipient is cooled to 18 C for circulatory arrest.
Once the donor organ is available in the operating room and
the patient has been adequately cooled, circulatory arrest is
established, the arch vessels are snared tightly, and the patient
is exsanguinated into the venous reservoir. The aorta is divided
just above the valve and incised longitudinally along the lesser
curve of the aortic arch to a level 1 to 2 cm below the ductal
insertion site on the descending aorta. The ductus is ligated
next to the pulmonary artery and divided, and then the main
pulmonary artery is transected just below the bifurcation.
The right atrial incision is started superiorly at the base of
the appendage. This incision is then carried down into the coronary sinus and across the atrial septum into the left atrium.
The superior aspect of the right atrial incision is next carried
across the septum to open the roof of the left atrium. The lateral
wall of he left atrium is incised above the left pulmonary veins
with the left atrial appendage included with the specimen.
The donor organ is prepared on the back table in cold saline solution. The right atrium is incised from the inferior vena
FIGURE 50-4 Standard heart transplantation using biatrial anastomosis. A, A recipient ventricular mass has been removed, and the left atrial anastomosis
has been started. B, Final appearance after all anastomoses are completed.
CHAPTER 50
cava laterally to the base of the appendage; the area of the

sinoatrial node is avoided if atrial anastomoses rather than
caval anastomoses are to be performed. The pulmonary vein
confluence is excised off the back of the left atrium, leaving
an opening comparable in size to the recipient left atrial cuff.
The pulmonary artery is transected just below the bifurcation
to provide a wide anastomosis. The aorta is trimmed, depending on the level required in the recipient. Care must be taken
to check for and adequately close a patent foramen ovale,
which is frequently present, especially in infant hearts. Failure
to do so may result in significant postoperative right-to-left
shunting in the face of pulmonary hypertension.
The implantation is begun by forming an anastomosis between the lateral wall of the left atrium from the level of the left
atrial appendage inferiorly. A left ventricular vent is placed
through the right superior pulmonary vein, and the left atrial
anastomosis is completed by reconstructing the intra-atrial
septum. The arch of the aorta is then reconstructed. The right
atrial anastomosis is begun at the inferior vena cava orifice and
then taken superiorly along the intra-atrial septum. The ascending aorta is then cannulated by a new purse-string suture,
air is evacuated, and cardiopulmonary bypass is resumed. The
snares are released from the head vessels and warming is commenced. The pulmonary anastomosis is then performed in an
end-to-end fashion. If time permits, this step may be done
during circulatory arrest in a drier field. After adequate warming, the patient is weaned from cardiopulmonary bypass and
the cannulas removed (Fig. 50-5).22 Right atrial, left atrial,
and, occasionally, pulmonary artery pressure catheters are
placed before discontinuing bypass and brought out through
the skin below the incision.
In older children with cardiomyopathy or infants without
aortic arch abnormalities, the recipient procedure is similar to
that performed in adults. The ascending aorta is mobilized to
the pericardial reflection and used for arterial cannulation.
The child is cooled to 28 C to 34 C, because the implantation is performed under aortic cross-clamp rather than circulatory arrest. After the left atrial anastomosis has been
completed, the right atrial connection can be sewn either directly or by using a bicaval technique if a previous cavopulmonary connection has been performed. This may decrease
the incidence of tricuspid regurgitation in certain patients.
The aortic anastomosis is then completed in an end-to-end
fashion in the midascending aorta. The pulmonary artery
anastomosis may or may not be performed during aortic
cross-clamp, depending on how long the implant procedure
takes.
Numerous other variations of the implantation procedure
can be used, depending on the recipient anatomy. Modifications accounting for a persistent left superior vena cava, previous cavopulmonary shunt or Fontan procedure, corrected
transposition of the great arteries, and situs inversus totalis
have been described.
Postoperative Management
------------------------------------------------------------------------------------------------------------------------------------------------
The recipient is returned from the operating room to an isolation room in the intensive care unit. Mechanical ventilation
is required initially but is weaned as rapidly as possible. Antibiotics are continued until all monitoring lines and chest
tubes have been removed.
665
Some level of inotropic support is required in virtually all

heart transplant recipients. Isoproterenol is often an ideal
choice because of its pulmonary vasodilatory effects, as well
as its inotropic and chronotropic effects, because many patients have a slower than optimal heart rate initially. This transient sinus node dysfunction is rarely permanent. Dobutamine
and dopamine, especially at renal doses, are also frequently
used to augment ventricular contractility. Epinephrine and
norepinephrine are usually reserved for poor graft function.
Sodium nitroprusside infusion or phosphodiesterase inhibitors are used for afterload reduction in the early postoperative
period. Right ventricular dysfunction secondary to pulmonary
hypertension may respond to phosphodiesterase inhibitors,
which are used for afterload reduction in the early postoperative period. Right ventricular dysfunction secondary to pulmonary hypertension may respond to phosphodiesterase
inhibitors such as milrinone. Inhaled nitric oxide has been
shown to be an effective selective pulmonary vasodilator with
few systemic side effects and is useful in cardiac transplant recipients with pulmonary hypertension.
Transplant Immunosuppression
------------------------------------------------------------------------------------------------------------------------------------------------
A combination of immunosuppressive agents is used for the

prevention and treatment of rejection. Standard triple-drug
immunosuppression therapy consisting of prednisone, cyclosporine, and azathioprine has been successfully used
in pediatric cardiac transplant recipients and remains the
most common regimen.23 In 2009, more than 70% of transplant recipients received induction immunotherapy. The
induction and maintenance doses of medications used for
immunosuppression at the Childrens Hospital of Philadelphia are listed in Table 50-4. Because of the adverse effects
of corticosteroids, withdrawal from prednisone is usually
attempted 6 months after transplantation.24 Up to 80% of
recipients may be successfully weaned from steroids; only
a quarter of these patients have an episode of rejection in
the first 6 months.25
Most patients are maintained on an immunosuppressive
regimen that is a combination of calcineurin inhibitor and
cell-cycle inhibitor. Tacrolimus (formerly called FK-506)
has been shown to be an effective immunosuppressive agent
in children, and its use has increased over the last 5 years,
with approximately 66% of all pediatric cardiac transplant
patients receiving it for maintenance immunosuppression
1 year after transplantation in the place of cyclosporine.
Overall, patients taking tacrolimus appear to have a lower incidence of rejection. Side effects of azathioprine therapy,
such as bone marrow depression, have precipitated the use
of mycophenolate mofetil (MMF) in its place. It is estimated
that approximately 66% of patients use MMF as a cell-cycle
inhibitor. It is well tolerated with few side effects and has
been shown in large clinical trials to have benefits in survival
and treated rejection episodes.26
An increasing number of centers use induction immunosuppression in pediatric cardiac recipients, with nearly 70%
of patients now receiving a polyclonal antiT-cell preparation, OKT3 (a murine monoclonal CD3 antibody), or
an interleukin-2 receptor antibody immediately after
transplantation. However, there have been no significant
666
PART IV
TRANSPLANTATION
C
FIGURE 50-5 Technique for transplantation in hypoplastic left heart syndrome (with the use of bicaval anastomosis). A, Recipient anatomy before cardiectomy. B, Appearance of the recipient after cardiectomy. Note that the aortic incision must be extended into the descending aorta beyond the level of
the arterial duct. C, Final appearance after all anastomoses are completed.
differences in the average number of rejection episodes in patients treated for rejection regardless of the type of induction
used. In addition, there is no significant difference in survival
between the induction groups or between use of induction
versus no induction. Induction therapy does not increase
the risk of CMV disease or post-transplant lymphoproliferative disease.

Infectious prophylaxis includes oral nystatin for fungal
prophylaxis and oral trimethoprim-sulfamethoxazole 3 times
per week. Pentamidine inhalation treatment is an effective
CHAPTER 50
667
TABLE 50-4
Heart Transplantation Immunosuppression Regimen at the Childrens Hospital of Philadelphia
Drug
Dosage
Rabbit antithymocyte globulin

(ATG)
Azathioprine/mycophenolate
mofetil (MMF)
1.5 mg/kg IV given in operating room before transplantation for sensitized patients and once daily for 5 days;
titrated to CD3 count
2 mg/kg IV given in the operating room before transplantation
Then 2 mg/kg IV given once daily for 5 days (neonates), 7 days (infants), 9 days (adolescents)
Change to MMF 600 mg/m2 IV given twice daily
Change to MMF orally once intestinal function resumes
0.05 mg/kg every 12 hours orally
0.02 mg/kg/hr IV infusion beginning in the operating room before transplantation
Then 0.02 mg/kg/hr IV infusion for 24 hours
Change to ATG on postoperative day 3 and give 1.5 mg/kg IV once daily for 3 days (neonates), 5 days (infants),
or 7 days (adolescents)
Change back to cyclosporine orally once ATG course completed
Dosing should be carefully adjusted to maintain levels of 125-150 mg in neonates, 175-200 mg in children, 250 mg
in 6- to 12-year-olds, and 250-300 mg in adolescents
15 mg/kg in operating room before transplantation
3 mg/kg IV twice daily for 3 doses
0.5 mg/kg twice daily for sensitized patients followed by oral prednisone taper
Tacrolimus
Cyclosporine
Solumedrol
alternative to trimethoprim-sulfamethoxazole for Pneumocystis

carinii prophylaxis if bone marrow suppression is a problem.
Routine CMV prophylaxis is used in cardiac transplant recipients at our institution.
Early Complications
------------------------------------------------------------------------------------------------------------------------------------------------
Acute rejection and infection are the most common early complications after cardiac transplantation. Nearly 60% to 75% of
patients have at least one episode of rejection, and it should be
expected that about a third will have an episode in the first
3 months and 50% within the first year after transplantation.27
Some studies suggest that infants may be less prone to rejection than older children. Rejection surveillance is based on
clinical evaluation, echocardiography, and endomyocardial
biopsy. Clinical assessment includes observation of changes
in a patients activity or appetite. Atrial or ventricular ectopy,
including tachycardia, is suspicious for rejection and mandates evaluation. Echocardiography is particularly useful
in neonates, in whom biopsy is technically difficult and
carries significant risk because of patient size. Echocardiographic evaluation is typically performed weekly for the first
month and then monthly for the first year after transplantation. Echocardiography-guided transjugular endomyocardial biopsy has been shown to be an effective means of
monitoring pediatric transplant recipients for rejection and
remains the gold standard for detection of rejection.28 An
aggressive approach, consisting of routine endomyocardial
biopsy weekly for the first month after transplantation, every
second week for the second month, and then once monthly
for the remainder of the first year, has been adopted at the
Childrens Hospital of Philadelphia for rejection surveillance.
Subsequent biopsies are obtained twice annually or whenever
rejection is clinically suspected. Most biopsies are performed
on an outpatient basis. The international grading system for
cardiac transplant rejection is shown in Table 50-5.
Episodes of acute rejection are usually treated with a 3-day
course of intravenous methylprednisolone (10 mg/kg). OKT3
and antithymocyte globulin are reserved for an incomplete response or rejection refractory to steroids. Response is confirmed by follow-up biopsy 1 to 2 weeks after treatment.
TABLE 50-5
International Society of Heart and Lung Transplantation
Grading System for Evaluation of Cellular Rejection
2005 Classification
0
No acute rejection
1R
Interstitial and/or perivascular infiltrate with up to 1 focus of
myocyte damage
2R
Two or more foci of infiltrate with associated myocyte damage
3R
Diffuse infiltrate with multifocal myocyte damage, edema,
hemorrhage, vasculitis
1990 Classification
0
No acute rejection
1A
Focal, mild acute rejection
1B
Diffuse, mild acute rejection
2
Focal, moderate acute rejection
3A
Multifocal moderate rejection
3B
Diffuse, borderline severe rejection
4
Severe acute rejection
Modified from Billingham ME, Cary NRB, Hammond ME, et al: A working
formulation for the standardization of nomenclature in the diagnosis of
heart and lung rejection: Heart Rejection Study Group. J Heart Lung
Transplant 1990;9:587-593; Stewart S, Winters GL, Fishbein MC, et al:
Revision of the 1990 working formulation for the standardization of
nomenclature in the diagnosis of heart rejection. J Heart Lung Transplant
2005;24:1710-1720.
Although infectious complications are common in cardiac

transplant recipients, infection-related deaths do not appear
to be. Bacterial infections are most frequent in the early
post-transplant period, but can occur late after transplantation
and usually respond to proper antibiotic therapy. Of viral infections, CMV appears to be the most common and is treated
with intravenous ganciclovir. Viral respiratory infections usually occur at a frequency similar to that in normal children and
appear to be well tolerated by the recipient.
Aside from rejection and infection, the immediate postoperative complications after heart transplantation are hypertension, seizures, renal dysfunction, and diabetes. Nearly 10% of
infant heart transplant recipients require perioperative peritoneal dialysis. Among neonates, 10% to 15% require phenobarbital therapy for postoperative seizures.
668
PART IV
TRANSPLANTATION
Late Complications
------------------------------------------------------------------------------------------------------------------------------------------------
The primary late complications in pediatric cardiac transplant

recipients are chronic rejection, post-transplant lymphoproliferative disease (PTLD), and transplantation CAD. Rejection
may account for up to 40% of deaths after cardiac transplantation. Lymphoproliferative disease is associated with
EBV infection and is currently treated by a reduction in immunosuppressants, acyclovir, and chemotherapy.29
The onset of transplant CAD has a prevalence of 10% to 15%
and may be suggested by symptoms of congestive heart failure in
recipients. Echocardiograms are performed routinely during
follow-up visits of heart transplant recipients, and worsening ventricular function is a sign of graft CAD. A new onset of arrhythmias
after transplantation, especially ventricular arrhythmias, may also
be an indication of underlying CAD. Additionally, CAD may be
found on routine follow-up catheterization or intracoronary ultrasonography, without any previous suggestion of disease. A number of causes have been implicated in the development of graft
CAD, including chronic cellular rejection, hyperlipidemia, vascular rejection, and CMV infection. Unlike adult cardiac transplant
recipients, CAD appears to develop in pediatric patients relatively
early after transplantation, with one series demonstrating an incidence of 35% by 2 years after transplantation. A review of 815 pediatric transplant patients found nearly 8% to have significant
CAD by angiogram or autopsy findings. The mean time after transplantation to diagnosis was 2.2 years, with one patient having significant CAD 2 months after transplantation. Only 20% of patients
in whom graft CAD was diagnosed were still alive, and most of the
deaths were sudden or unexpected. Retransplantation appears to
be the only viable option for these patients, although the results in
general are not encouraging, with 1- and 3-year survival rates of
71% and 47%, respectively, and CAD developing in the second
grafts in 20% of retransplantation patients. However, the Loma
Linda group has reported a significantly better retransplantation
experience in infants who were first transplanted when younger
than 6 months. In this group, a 10-year actuarial survival rate
of 91% was observed after retransplantation. Potential medical
treatment targeted at cholesterol and lipid-lowering therapies
are currently under investigation.
Results
------------------------------------------------------------------------------------------------------------------------------------------------
The largest group of infant cardiac transplant recipients

reported in the literature is from Loma Linda, where 233
heart transplantations in infants younger than 6 months
100
<1 year (N = 2,049)

1117 years (N = 3,027)
90
have been performed. Nearly 65% were for HLHS, and the
rest were for other complex congenital anomalies (29%) or
cardiomyopathy or tumor (8%). The operative (30-day) survival rate was 89%, with the primary causes of mortality being primary graft failure, technical problems, pneumonia, or
acute rejection. The overall 1-year survival rate was 84%,
with a 5- and 10-year actuarial survival rate of 73% and
68%, respectively. In addition, patients undergoing transplantation when younger than 30 days had a significantly
better outcome than did older infants, with an actuarial
survival rate of 80% and 77% at 5 and 10 years, respectively,
potentially related to improved immune tolerance in the
younger subgroup.
Stanford University reported its series of 72 patients younger than 18 years who have undergone heart transplantation
since 1977. Only 25% were younger than 1 year (mean of
9 years), and nearly two thirds had cardiomyopathy unrelated
to congenital heart disease. The operative survival rate was
87.5%, with deaths mainly caused by pulmonary hypertension/right ventricular failure and acute rejection. There were
20 late deaths, 24% were due to rejection, and 17% were
due to graft CAD. Actuarial survival rates at 1, 5, and 10 years
were 75%, 60%, and 50%, respectively.
At St. Louis Childrens Hospital, 45 heart transplants were
performed from 1983 to 1993, more than half in infants with
HLHS. The infant group had a survival rate (92%) similar to that
of the Loma Linda series, whereas the pediatric group (older
than 1 year) had an 80% early survival rate. Morales and colleagues published results of their experience spanning more
2 decades at Texas Childrens Hospital and reported no change
in mortality in survivors after the first post-transplant year.30
Results from the Registry of the International Society for
Heart and Lung Transplantation reveal a perioperative mortality rate higher for infants than for older children (Fig. 50-6).
Despite the much greater early mortality, however, the half-life
of 18.3 years is longer than that of the childhood or adolescent
survivors. For the childhood age group of 1 to 10 years, the
half-life was 17.5 years versus 11.3 years for the adolescent
age group, thus conferring the younger patients a significant
survival advantage. If those patients who died within the first
year after transplant were excluded, the median conditional
survival was 21.4 years for those who underwent transplantation in the first year of life, 19.3 years for those aged between 1 and 10 years, and 15.2 years for older children
(Fig. 50-7). Survival has been improving in relation to the
era of transplantation, with the median survival increased
110 years (N = 2,929)

Overall (N = 8,005)
Survival (%)
80
70
60
0<1 vs. 110: P <0.0001;
0<1 vs. 1117: P = 0.3284;
110 vs. 1117: P = 0.0003.
50
40
30
20
Half-life <1: 18.3 years; 110: 15.5 years; 1117: 11.3 years
10
0
0
9 10 11 12 13 14 15 16 17 18 19 20
Years
FIGURE 50-6 Survival analysis for transplantations performed January 1982 to June 2008. (From Kirk K, Edwards LB,
Kucheryavaya AY, et al: The Registry of the International Society
for Heart and Lung Transplantation: Thirteenth official pediatric
heart transplantation report2010. J Heart Lung Transplant
2010;29:1119-1128.)
CHAPTER 50
669
100
Half-life: <1: 21.4; 110: 19.3 years; 1117: 15.2 years
90
Survival (%)
80
70
0<1 vs. 110: P = 0.0138;
0<1 vs. 1117: P <0.0001;
110 vs. 1117: P <0.0001.
60
50
FIGURE 50-7 Survival analysis for transplants performed

January 1982 to June 2008 and surviving to 1 year after transplantation. (From Kirk K, Edwards LB, Kucheryavaya AY, et al:
The Registry of the International Society for Heart and Lung
Transplantation: Thirteenth official pediatric heart transplantation report2010. J Heart Lung Transplant 2010;29:11191128.)
40
<1 year (N = 1,422)

1117 years (N = 2,399)
110 years (N = 2,272)

Overall (N = 6,093)
30
0
9 10 11 12 13 14 15 16 17 18 19 20
Years
acute rejection, lymphoma, graft failure, and infection. Retransplantations now account for 5% of all transplantation operations. Survival for retransplantation is decreased when the
intertransplantation interval was less than 3 years and is relative
to indication for primary transplantation (Fig. 50-10).31,32
Aside from survival, it has been demonstrated that transplanted hearts in children appear to grow normally, and the
left ventricle increases muscle mass to maintain the normal left
ventricular mass-to-volume ratio with time. Exercise testing in
older children has shown peak heart rate and oxygen
consumption to be consistently two thirds of that predicted
in heart transplant recipients. Somatic growth appears to be
normal in infants after heart transplantation, and neurologic
development is generally preserved, although some neuro-
from 9.5 years for the period 1982 to 1989, to 11.7 years for
the period 1990 to 1994, to 14.3 years for the period 1995 to
1999 (Fig. 50-8). Averaged over 15 years, an infant recipient
would have an approximate 2% per year risk of mortality,
whereas for older children, it remains approximately 4%,
again indicating a longer-term survival advantage for younger
cardiac transplant recipients.
The most predictive risk factors for 1-year mortality in the
pediatric population remain congenital heart disease, donor
age, pulmonary artery systolic pressure greater than 35 mm
Hg, and the need for mechanical ventilation and hospitalization
while awaiting transplantation. Among the most significant risk
factors for 5-year mortality are dialysis, congenital heart disease,
and female gender (Fig. 50-9). Causes of death include CAD,
100
19821989 (N = 846)
19901994 (N = 1,803)
19951999 (N = 1,846)
20006/2008 (N = 3,510)
Survival (%)
80
FIGURE 50-8 Survival analysis by era for transplantations

performed January 1982 to June 2008. (From Kirk K, Edwards
LB, Kucheryavaya AY, et al: The Registry of the International
Society for Heart and Lung Transplantation: Thirteenth official pediatric heart transplantation report2010. J Heart
Lung Transplant 2010;29:1119-1128.)
Half-life 19821989: 9.5 years; 19901994: 11.7 years; 1995

1999: 14.3; 20006/2008: n.c.
60
40
20
All P values significant at P = 0.01
0
0
FIGURE 50-9 Relative incidence of leading causes of death

for deaths occurring January 1998 to June 2009. CAV, coronary
artery vasculopathy; CMV, cytomegalovirus. (From Kirk K,
Edwards LB, Kucheryavaya AY, et al: The Registry of the International Society for Heart and Lung Transplantation: Thirteenth official pediatric heart transplantation report2010.
J Heart Lung Transplant 2010;29:1119-1128.)
Percentage of deaths
40
9 10 11 12 13 14 15 16 17 18 19 20
Years
CAV
Acute rejection
Primary failure
Graft failure
Infection
(non-CMV)
30
20
10
0
030 days
(N = 226)
31 days1
year
(N = 266)
>1 year3
years
(N = 173)
>3 years5
years
(N = 173)
>5 years10
years
(N = 306)
>10 years
(N = 190)
670
PART IV
TRANSPLANTATION
<1 year (N = 48)

5+ years (N = 145)
1<3 years (N = 34)

Primary TX (N = 5,417)
100
Comparison of survival for

retransplant groups: P = 0.0019
90
Survival (%)
3<5 years (N = 48)
80
70
60
50
40
0
Time (years) since most recent transplant

Only patients who were younger than 18 years old at the time of retransplant
are included.
logic abnormalities may be seen in up to 20% of neonatal recipients on long-term follow-up.
Conclusion
------------------------------------------------------------------------------------------------------------------------------------------------
Despite further improvements in surgical technique, immunosuppression, perioperative management, and rejection surveillance, long-term results of pediatric heart transplantation
have shown little change, with a 15-year survival rate of approximately 50%. Chronic rejection, graft CAD, and the
FIGURE 50-10 Survival rates for retransplantations, stratified by intertransplantation interval, for retransplantations
performed January 1994 to June 2008. (From Kirk K, Edwards LB, Kucheryavaya AY, et al: The Registry of the International Society for Heart and Lung Transplantation:
Thirteenth official pediatric heart transplantation report
2010. J Heart Lung Transplant 2010;29:1119-1128.)
long-term effects of steroids on growth continue to cloud

the development of cardiac transplantation as the primary
treatment of complex congenital heart disease. However, for
many children with end-stage cardiomyopathy and structural
heart disease not amenable to corrective surgery, transplantation is the only option. Future areas of research include the use
of xenografts, ABO incompatibility, permanent mechanical
support, and widening the bridge to transplantation with
smaller and more adaptable assist devices.
expertconsult.com.
Organ Allocation
------------------------------------------------------------------------------------------------------------------------------------------------
In 2005, a new system to allocate lungs to recipients was

established across the United States. Previously, lungs were allocated based on the amount of time the recipient had accrued
on the waiting list (first come, first served). In an attempt to
make distribution of lungs more equitable, a lung allocation
score (LAS) was devised and implemented.4 This score attempts to prioritize organs to patients on the list most in need
of the organ (the sickest) as well as to those most likely to
do well post-transplantation. This score is used in children
12 years and older. For those younger than 12 years, the
old system still remains in effect at present.
Indications
------------------------------------------------------------------------------------------------------------------------------------------------
CHAPTER 51
Pediatric Lung
Transplantation
Sanjiv K. Gandhi, Albert Faro,
and Charles B. Huddleston
Isolated lung transplantation is applicable to any child with

life-threatening and progressive disability because of pulmonary parenchymal or vascular disease. In general, this treatment modality is indicated for increasing the duration of
life and for improvement in the quality of life. The current
long-term survival after lung transplantation is approximately
50% at 5 years. Thus the selection of patients for transplantation and the timing of the procedure are critically important.
One would like to be able to predict when a child would be
within 2 years of dying without any form of medical treatment.
Obviously, this may be very difficult. The major diagnostic
groups for pediatric lung transplantation are cystic fibrosis
(CF), interstitial lung disease with pulmonary fibrosis, primary pulmonary hypertension, pulmonary hypertension
associated with congenital heart disease, retransplantation,
and a miscellaneous category (Table 51-1). Chronic obstructive lung disease, the most common indication for transplantation in adults, is remarkably absent from this list.3
CYSTIC FIBROSIS
The first reported attempt at lung transplantation occurred
in 1963 and was performed by Dr. James Hardy at the University of Mississippi Medical Center.1 The patient did not
survive the hospitalization, dying 18 days after the transplant. There were a number of additional attempts at this
during the next few years, with most failures related to poor
healing of the airway anastomosis. Approximately 20 years
after Dr. Hardys ill-fated effort, the first truly successful lung
transplant was performed in Toronto, Canada, by a team led
by Dr. Joel Cooper. This patient had a single-lung transplant
for pulmonary fibrosis and survived for more than 6 years,
ultimately dying of renal failure.2 During the subsequent
years, and particularly in the late 1990s, pediatric lung
transplantation has emerged as a viable treatment option
for children with end-stage pulmonary parenchymal and
vascular diseases. However, the number of children undergoing transplantations throughout the world since 1989 remains relatively small, representing only 4% of all lung
transplantations performed.3 In this chapter, pediatric lung
transplantation is described as an isolated procedure, and
heart-lung transplantation is not included.
This disease, the most common lethal hereditary disease in

North America, comprises the largest diagnostic group of children younger than age 18 years who undergo lung transplantation. Although the median survival now exceeds 38 years, one
eighth of the deaths from CF in 2008 still occurred in the pediatric age group. Without question, the most common cause of
death is respiratory related. About 250 to 300 transplantations are performed annually in the United States for CF, and,
although this number is growing slowly each year, donor
availability still remains a major limiting factor.5
As with other diagnostic groups, timing of transplantation
in the course of a chronic disease is a crucial issue. Kerem, in
the early 1990s, demonstrated that a 1-second forced expiratory volume (FEV1) less than 30% of predicted, a PaO2 less
than 55 mm Hg, and/or a PCO2 greater than 50 mm Hg were
associated with a survival beyond 2 years of less than 50%.6
The impact of these factors is magnified in the pediatric age
group, particularly in girls. However, more recent studies
on the natural history of CF patients, once they have severely
compromised lung function, show that an isolated measure of
the FEV1 alone may not be sufficiently predictive. The rate of
671
672
PART IV
TRANSPLANTATION
TABLE 51-1
Indications for Lung Transplantation in Children
Cystic fibrosis
Pulmonary fibrosis
Pulmonary vascular disease
Primary pulmonary hypertension
Eisenmenger syndrome
Bronchiolitis obliterans
Retransplantation
Other
decline in the FEV1 may be a more accurate determinant of

survival.7,8 Other factors that may serve as relative indicators
in deciding to proceed toward transplantation include the
need for continuous supplemental oxygen, increased frequency of hospitalizations, and diminished weight for height
(below the 80th percentile).6 More recently, Liou and colleagues validated a formula that included microbiological
data, body mass index measures, and presence of diabetes,
in addition to lung function and gender, as important determinants in prognosticating 5-year survival.9 The presence of
antibiotic-resistant organisms in the sputum is a relative
contraindication to lung transplantation. The synergistic
effectiveness of antibiotic combinations is not useful in
treating pulmonary exacerbations and is no longer readily
available. Chronic infection with Burkholderia cenocepacia,
pretransplantation, is associated with a particularly poor
post-transplantation prognosis and therefore is of particular
concern.10,11 Portal hypertension with hepatic cirrhosis
occurs in 5% to 10% of patients with CF. These children are
at risk for variceal bleeding as well as derangements of synthetic function. In general, if the synthetic function is preserved, decompression of the portal venous system with
percutaneous procedures will lower the risk to a level
satisfactory for lung transplantation.12 However, when there
is also synthetic dysfunction of the liver, combined liver-lung
transplantation may be the only appropriate option.13,14 Diabetes mellitus is generally not considered a contraindication to
transplantation unless there is evidence of vasculopathy, bearing in mind that control of serum glucose will be more difficult
after transplantation.15 As many as 10% to 15% of CF lung
transplantation candidates will have had prior thoracotomies
for either pneumothorax or pulmonary resection. Most centers do not consider this a contraindication to transplantation,
although the resultant adhesions from these prior operations
do increase the difficulty and the risk of bleeding.16 Mechanical ventilation or presence of a tracheostomy are not in and of
themselves contraindications to transplantation, but the overall medical condition of patients requiring this level of support
must be carefully considered.17
PULMONARY VASCULAR DISEASE

This rather broad classification of patients includes those with
primary pulmonary hypertension (PPH) and those with pulmonary hypertension associated with congenital heart disease
(PH/CHD). The latter category includes patients with Eisenmenger syndrome but is not limited to this. These patients
die of either progressive right-sided heart failure, arrhythmias,
or a lethal episode of hemoptysis. It is difficult to predict when
a patient might have a fatal arrhythmia or episode of hemoptysis. However, most patients with pulmonary vascular disease
will die of progressive right-sided heart failure over a protracted period of time.18 In the past several years, a number
of somewhat selective pulmonary vasodilators have become
available for use in these patients. These include intravenous
prostacyclin,19 prostacyclin analogues iloprost (inhaled)20
and betaprost (oral),21 and bosentan,22 an endothelin receptor
antagonist. These drugs have enabled patients to delay the
need for transplantation for years. In fact, the number of
patients undergoing transplantation for pulmonary vascular
disease has significantly dropped in recent years. The timing
of transplantation for patients with pulmonary vascular disease is influenced significantly by the response to medical
therapy and the underlying cause of the pulmonary vascular
disease. Although primary pulmonary hypertension and
Eisenmenger syndrome result in identical histologic changes
in the pulmonary vascular bed, the latter of these two is
associated with a much more favorable long-term prognosis.
A retrospective analysis by Hopkins of 100 adults with severe
pulmonary hypertension resulting from either Eisenmenger
syndrome or PPH revealed that, in the former group, actuarial
survival without transplantation was 97% at 1 year, 89% at
2 years, and 77% at 3 years. In contrast, survival was 77%,
69%, and 35% during the same respective time intervals in
the PPH cohort.23 It is presumed that the intracardiac defect
allows the right ventricle to decompress via the defect when
the afterload in the pulmonary vascular bed becomes prohibitively high. On the basis of this and other observations, atrial
septostomy performed in the cardiac catheterization suite has
been demonstrated to provide clinical benefit in patients with
PPH.24 Results from a multicenter study of patients with PPH
performed before the advent of long-term intravenous prostacyclin therapy demonstrated a median survival from time of
diagnosis of 2.8 years. In that study, a formula was developed
incorporating hemodynamic variables to assist in predicting
the 2-year mortality,18 and it was recommended that patients
should be listed when this figure is less than or equal to 50%.
Studies regarding natural history in adults have been applied
to children, but it is unclear whether this disease behaves the
same in a younger population. Clabby and co-workers
reviewed 50 patients from many centers to provide a means
of estimating survival in children with PPH.25 There was a direct correlation of mortality with the product of the mean right
atrial pressure and the pulmonary vascular resistance.25 With
progress in the medical therapy of PPH to identify selective
pulmonary vasodilators as well as the underlying mechanisms
of this disease, these formulas predicting survival may be obsolete. The durability of medical therapy is unclear. How this
therapy might be applied to secondary pulmonary hypertension, such as Eisenmenger syndrome, is speculative.
The two main issues in considering patients with Eisenmenger syndrome or PH/CHD for lung transplantation are the timing of listing and the complexity of the cardiac lesion to be
repaired. As noted earlier, it is clear that, once the diagnosis
is made, these patients can live much longer than those with
PPH.23 The mode of death in these patients is by progressive
heart failure, pulmonary hemorrhage, stroke, or sudden death,
presumably due to arrhythmias.25 Patients should be listed
when symptoms develop, when there has been a single pulmonary hemorrhage, or perhaps arbitrarily when they reach their
late 30s. Most patients with PH/CHD have an atrial septal defect, ventricular septal defect, or patent ductus arteriosus. All of
these require relatively simple cardiac repairs. However, there
CHAPTER 51
are patients with unrepaired atrioventricular canal defects,

transposition of the great arteries, and truncus arteriosus who
would require more complex procedures. An alternative for
these patients would be heart-lung transplantation. The likelihood of obtaining a donor heart-lung block for anyone more
than 40 kg is low because of the distribution policy for thoracic
donor organs. In addition, the long-term survival after heartlung transplantation is particularly poor (approximately 40%
at 5 years post-transplantation).3 These two issues must be factored into the decision as to whether one should perform the
higher-risk procedure of lung transplantation in combination
with repair of a complex cardiac lesion or heart-lung transplantation. Some patients with congenital heart disease who have
undergone repair may not experience the expected decline in
pulmonary vascular resistance after appropriate correction. Occasionally the repair has been performed relatively late in life,
but there are children who have undergone timely repair and
still present later with severe pulmonary hypertension. It is
not clear how to classify these patients. In general, this is a less
uniform group than either the patients with PPH or those with
Eisenmenger syndrome. They seem to follow a clinical course
similar to that seen in patients with PPH and should be treated
in a similar fashion.23
Another diagnostic group with pulmonary vascular disease
are patients with an inadequate pulmonary vascular bed.
Examples of this include pulmonary atresia, ventricular septal
defect and multiple aortopulmonary collaterals, and congenital
diaphragmatic hernia, where there is primarily a general deficiency of pulmonary parenchyma. In the former group, complete correction (repair of the ventricular septal defect
combined with reconstruction of the right ventricular outflow
tract with a conduit to the unifocalized aortopulmonary
collaterals) represents a high-risk but viable option for the
majority of these patients. However, when the anatomy of the
aortopulmonary collaterals is not amenable to unifocalization
or when unifocalization has not produced satisfactory growth
of the pulmonary vascular tree, the result is progressive cyanosis or progressive pulmonary hypertension or both. Lung transplantation with repair of the residual cardiac defect may be
the only feasible option for survival. Children with congenital
diaphragmatic hernias, despite having undergone a successful
hernia repair, may still be left with inadequate pulmonary
parenchyma and vascular bed to handle the full cardiac output.
The resultant severe pulmonary hypertension is the usual cause
of death in these infants and is an indication for transplantation.
The problem here is that these infants often will require extracorporeal membrane oxygenation (ECMO) support during the
perioperative period. This reduces the time that patients such as
this can wait for a donor offer once listed for lung transplantation. It is possible that a single-lung transplant on the affected
side would be sufficient in this circumstance. In this scenario,
once the patient has grown, it may be possible to remove the
transplanted lung altogether, leaving the patient with
a presumably normal contralateral lung to maintain normal
respiratory function. In reality, those patients with insufficient
pulmonary reserve will have to be identified very early in the
course for lung transplantation to be a realistic option. The
mortality is quite high even when donor organs are identified.
In all the previous situations, isolated lung transplantation
is appropriate only when left ventricular function is normal.
Poor left ventricular function will result in elevated left ventricular end-diastolic pressure post-transplantation, which
PEDIATRIC LUNG TRANSPLANTATION
673
will add significantly to problems with pulmonary edema

and early graft failure. Right ventricular function is frequently
poor, particularly in the patient group with PPH. That should
not be a deterrent to isolated lung transplantation, because the
right ventricular function always returns to normal within a
relatively short period of time.26
Although a prior thoracotomy is generally not a contraindication to lung transplantation in patients with pulmonary parenchymal disease, this is not true for those with pulmonary
vascular disease, especially when secondary to congenital heart
disease and associated with cyanosis. The adhesions that develop
after a thoracotomy for palliation of cyanotic congenital heart
disease are extremely vascular. Intercostal and internal mammary arteries will form direct connections through the pleura
into the parenchyma of the lung in a compensatory attempt to
enhance pulmonary blood flow. The bleeding that occurs during
the recipient pneumonectomy portion of the transplantation
procedure is often horrendous and life threatening.
PULMONARY FIBROSIS
These patients account for 5% to 10% of pediatric patients
undergoing lung transplantation.3 Placed in this category
are those patients with usual interstitial fibrosis, radiationinduced fibrosis, bronchopulmonary dysplasia, and pulmonary fibrosis secondary to chronic aspiration. The progression
of these disease processes is quite variable. Generally, patients
should be listed when normal activities are markedly limited
and minor viral illnesses lead to significant deterioration. Most
patients will be oxygen dependent and may well have evidence of coexistent pulmonary hypertension. For those in
whom aspiration is the underlying problem, the source of
the aspiration must be eliminated.
The prognosis of children with idiopathic pulmonary fibrosis is not altogether clear. This may be because there is not a
usual interstitial pulmonary fibrosis disease in children; the
underlying causes are frequently unique and unusual. Decisions regarding listing for transplantation are somewhat difficult because of this. Pulmonary fibrosis presenting during
infancy was once believed to have a poor prognosis; however,
some studies have demonstrated improved survival with high
doses of corticosteroid therapy.27 The prognosis for adults
with total lung capacity less than 60% predicted is still poor;
nearly all are dead within 2 years.28 It is difficult to translate
this information into the pediatric experience. Pulmonary
hypertension frequently accompanies this disease as it progresses. These patients should be evaluated and listed for
transplantation when they become symptomatic. If there is
a favorable response to corticosteroids, they can be followed
with standard (age > 5 years) or infant (length < 90 centimeters) pulmonary function tests. One problem with managing this disease is that patients with progression of their
disease tend to remain on relatively high doses of corticosteroids and come to transplantation in a rather cushingoid state.
This should not exclude them from transplantation.
BRONCHIOLITIS OBLITERANS
AND RETRANSPLANTATION
Bronchiolitis obliterans is not a specific disease but rather
a histologic description characterized by the obstruction
and destruction of the distal airways. It may occur as a
674
PART IV
TRANSPLANTATION
consequence of any severe lung injury, including viral

pneumonia, graft-versus-host disease after bone marrow
transplantation, autoimmune diseases, chemical injury,
Stevens-Johnson syndrome, and others. Of course, it is a
relatively common late complication of lung transplantation
(see later). The underlying etiology is unknown. Primary
bronchiolitis obliterans (not related to prior lung transplantation) is a perfectly legitimate indication for lung transplantation: It is a slowly progressive disease in virtually all
cases with no known effective treatment. When this disorder
occurs as a consequence of an isolated lung injury, transplantation is a fairly straightforward decision process. However,
those patients with prior bone marrow transplantations (usually for leukemia) offer special considerations.29 Although
the standard definition of cure is remission of the malignancy for more than 5 years, most of these patients present
within 2 or 3 years of treatment. Another problem is the
deranged immune competency seen after bone marrow transplantation and how the immunosuppressant agents used after
lung transplantation might further affect this. We have found
that these patients have less acute rejection than most other
lung transplantation recipients but may be more prone to
opportunistic infections.29 However, the number of patients
transplanted in this setting is low. For patients who acquire
bronchiolitis obliterans through other immunologic injuries,
such as autoimmune disorders, there are concerns about
the likelihood of recurrence in transplanted lungs. One would
have to ascertain that the primary process has completely
abated before transplantation.
Retransplantation for acute graft failure after transplantation has an extremely poor prognosis.30 Retransplantation
for bronchiolitis obliterans is a controversial issue. Bronchiolitis obliterans accounts for the majority of deaths occurring
more than 90 days post-transplantation.3 This figure is borne
out in our pediatric series.31 Although early mortality after
retransplantation is higher than for first-time lung transplantations, those who do survive this early phase have long-term
survival similar to the non-redo transplantations.30 Risk
factors for poor early outcome include nonambulatory status,
short period of time since the first transplantation, transplantation at a center with limited experience, and dependence on
mechanical ventilation. We have further noted that a low glomerular filtration rate is an independent risk factor. Because
patients continue to die on the waiting list, one could argue
that no patient should ever be retransplanted because this
might deprive an otherwise lower-risk patient from receiving
organs in a timely fashion. At present this issue is unresolved.
One can only advise use of proper judgment in selecting only
the best candidates when the issue of retransplantation arises.
MISCELLANEOUS
A variety of diagnoses fall into this group. Congenitally based
pulmonary parenchymal diseases constitute one of the more
interesting broad categories. Typically, these full-term newborns present with severe respiratory distress and no obvious
cause, such as meconium aspiration, sepsis, or persistent fetal
circulation. The diagnoses falling into this category include
surfactant protein B deficiency, other forms of pulmonary
alveolar proteinosis, alveolar-capillary dysplasia, pulmonary
dysmaturity, congenital interstitial pneumonitis, and others.
These infants usually have severe respiratory failure and
require a high level of ventilatory support. Often extracorporeal membrane oxygenation has been or is currently being
used. An open-lung biopsy is often necessary to either make
the diagnosis or to exclude other diagnoses. Surfactant protein
(SP) B or C deficiency and the ABCA 3 mutation can now be
diagnosed by looking for the specific genetic mutation in
peripheral blood or cheek swabs and assaying tracheal effluent
for the presence of this surfactant protein.32 All children will
survive less than 3 months even with aggressive therapy.
Abnormalities in SP-C and ABCA3 can have more varied presentations. Additionally, because the surfactant proteins are
expressed only in the lungs, extrapulmonary organ dysfunction is rare.33 Until other therapies become available, lung
transplantation is the only viable therapeutic option. In general, the waiting time for an organ offer is relatively short in
infants. Therefore one might realistically believe that an infant
with a 3-month life expectancy could undergo transplantation
and survive. When an infant is on ECMO, every effort should
be made to wean from it, using whatever means possible,
including a high-frequency oscillating ventilator and/or nitric
oxide. Although ECMO is not an absolute contraindication to
transplantation, one should be very cautious in this setting
because of the relatively high incidence of other organ
dysfunction.
Contraindications
------------------------------------------------------------------------------------------------------------------------------------------------
Contraindications to transplantation in children are also based

on experience obtained in adults (Table 51-2). Absolute contraindications include systemic disease with major extrapulmonary manifestations or severe dysfunction of other organ
systems. Thus widespread malignancy, collagen vascular disease, human immunodeficiency virus infection, and severe
neuromuscular disease are absolute contraindications. The
acceptable degree of renal insufficiency is open to some interpretation. Given the nephrotoxicity of cyclosporine and
tacrolimus, the drugs that form the basis of nearly all immunosuppressant regimens, a serum creatinine value greater than
2.0 mg/dL and a probable need for post-transplantation
dialysis are clinical parameters that would mitigate strongly
against proceeding with transplantation. A glomerular
TABLE 51-2
Contraindications to Lung Transplantation
Absolute
Malignancy
Human immunodeficiency virus infection
Multisystem organ failure
Left ventricular dysfunction
Active collagen vascular disease
Severe neuromuscular disease
Relative
Renal insufficiency
Liver function impairment
Malnutrition
Resistant organisms in the sputum
Poorly controlled diabetes mellitus
Osteopenia
Prior thoracotomies in the presence of pulmonary vascular disease
Prior pneumonectomy with mediastinal shift
Extreme prematurity
Inadequate psychosocial support system
Poor compliance
CHAPTER 51
filtration rate less than 50 mL/min has been associated with a

poor outcome in some patients. Significantly deranged hepatic
synthetic function precludes transplantation unless concomitant liver transplantation is also being undertaken. More
complex issues include severe malnutrition, poorly controlled
diabetes mellitus, osteopenia, vertebral compression fractures,
and the need for mechanical ventilation. None of these factors
in and of themselves serves as an absolute contraindication.
Nonetheless, all such concerning aspects of the clinical
presentation must be evaluated and carefully considered in
the scope of the patients overall state of health to assess the
likelihood for successful recovery after transplantation.
Chronic administration of corticosteroids before transplantation is considered to be undesirable, and, when possible, one
should reduce the total daily dose or change to an everyother-day dosage schedule. Previously, corticosteroids were
believed to have a significant negative impact on airway healing, particularly in the case of double-lung transplantation
with a tracheal anastomosis. Bilateral sequential lung transplantation with bronchial anastomoses has obviated this
problem to a large degree. Severe psychiatric disorder in either
the patient or, in the case of a young child, the care provider, is
a strong relative contraindication. Finally, a history of poor
compliance with either a medical regimen or in keeping
follow-up appointments is considered by most to be a strong
relative contraindication to transplantation. Graft failure due
to lack of proper care not only results in death to the recipient
involved, but also results in either a delayed or denied
transplantation for a more appropriate candidate.34
SPECIAL CIRCUMSTANCES
Some infants born extremely prematurely survive the early
days of their lives only to develop severe bronchopulmonary
dysplasia with respiratory failure within the first year of life.
The incidence of significant cerebral injury in this group is
high; approximately 50% of those surviving have some
disability.35 We can only assume that the incidence is higher
in those with severe residual lung disease requiring transplantation. It is often difficult to assess the neurologic status in
these infants because of their small size and often the need
for sedation and neuromuscular paralysis for maintenance
of satisfactory ventilation. It is probably unwise to submit
an infant born at less than 28 weeks estimated gestational
age to lung transplantation, unless there has been an opportunity for an accurate neurologic examination. Imaging studies
may offer some reassurance but are inconclusive. Another unusual situation that arises where lung transplantation may be
considered appropriate is the child with severe acute respiratory distress syndrome. Those children still in the acute phase
of this illness often have other organ dysfunction, and their
condition is too unstable for them to wait the obligatory time
once listed for transplantation, given the current organ allocation system for children less than 12 years of age. Those who
survive the early phase of acute respiratory distress syndrome
and are left with fibrotic lungs and stable ventilatory requirements should be evaluated. Finally, occasionally a patient with
a history of prior pneumonectomy will be referred for lung
transplantation. After pneumonectomy in children, the mediastinum shifts to the affected side. This distorts the hilar structures to the point that bilateral or single lung transplantation is
virtually impossible. When possible, a patient undergoing
675
pneumonectomy who might require lung transplantation in

the future should have a prosthetic spacer placed in that side
of the chest to maintain normal mediastinal geometry.
Donor Evaluation and Organ

Procurement
------------------------------------------------------------------------------------------------------------------------------------------------
Donor availability remains a major limitation to the applicability of transplantation for end-stage lung disease. Donors must
be matched by ABO blood type compatibility and within a reasonable size range of the recipient. Height is used as the most
accurate correlate to lung size. Height that falls within 15% to
20% of the recipient height is probably suitable. Extending
this range upward is certainly feasible, because it is not difficult to reduce the size of the lungs by trimming off the edge or
even using only the lower lobes. However, extending the lower
limit should be done with great caution, because the transplanted lungs may not fill the chest and may be more prone
to pulmonary edema. Donors are excluded in the presence
of positive HIV serology, active hepatitis, history of asthma, tuberculosis, or other significant pulmonary disease. A history of
limited cigarette smoking is probably acceptable if other parameters of the evaluation fall within the guidelines. In general, the upper limit of donor age is approximately 55 years.
The chest radiograph should be free of infiltrates, and the arterial oxygen tension should be more than 300 mm Hg on an
inspired oxygen fraction of 1.0 with an appropriate tidal volume and 5 cm H2O positive end-expiratory pressure. Mild
pulmonary contusions and subsegmental atelectasis would
not necessarily exclude a donor as long as these criteria are
met. Flexible bronchoscopy should be performed to examine
the airways for erythema suggestive of aspiration of gastric
contents. In addition, this provides an opportunity to assess
the nature and quantity of pulmonary secretions. The presence of purulent secretions that do not clear well with suctioning should exclude the donor even if the chest radiograph is
clear and the oxygenation is adequate.
The surgical part of the procurement process is performed
through a median sternotomy. Both pleural spaces are opened
widely to allow visual inspection of the lungs and also the topical application of cold saline and slush. The trachea is dissected
out between the superior vena cava and aorta. It may be helpful
to develop the interatrial groove, also, to allow a more accurate
division of the left atrial tissue that must be shared with the
cardiac donor team in most situations. The principles of the procurement process beyond this are (1) anticoagulation with
high-dose (300 units/kg) heparin; (2) bolus injection of prostaglandin E1 (50 to 70 mcg/kg) directly into the main pulmonary
artery; (3) decompressing the right side of the heart by incising
the inferior vena cava; (4) decompressing the left side of the
heart by amputating the left atrial appendage; (5) high-volume
(50 mL/kg), low-pressure flush of cold (4 C) pulmonary preservation solution of choice; (6) topical application of cold saline
and slush to the lungs; and (7) continued ventilation of the
lungs with low volumes and low pressures using an FiO2 of
0.4. When all the preservation solution has been administered,
the lungs are excised en bloc. The trachea is divided while the
lungs are held in gentle inflation (pressure of 20 cm H2O)
with the FiO2 at 0.4. The lungs are then extracted, placed in a
bag containing the preservation solution used for the flush,
and then placed in cold storage for transport.
676
PART IV
TRANSPLANTATION
Much research has been devoted to finding the ideal preservation solution to extend potential ischemic times and avoid
reperfusion injury.36 A full discussion of this complex topic
goes beyond the scope of this chapter. The most commonly
used preservation solutions at this time are modified EuroCollins solution, University of Wisconsin solution, Perfadex,
and Celsior. None of these is clearly superior to the others,
and all work reasonably well. However, none reliably allows
for preservation times greater than 8 hours, and none
completely avoids reperfusion injury.
Technique of Transplantation
------------------------------------------------------------------------------------------------------------------------------------------------
The surgical technique used for children is like that for adults,
except that virtually all children will require cardiopulmonary
bypass, whereas that is not always necessary in adults.
Transplantation without cardiopulmonary bypass would
require single-lung ventilation during the procedure. Maintaining single-lung ventilation in these small children is
extremely difficult, because the airways are too small to accommodate double-lumen endobronchial tubes. Bilateral
lung transplantation is performed for nearly all children
because of concerns over the growth potential of the transplanted lungs. Trans-sternal bilateral anterior thoracotomy
incision (the so-called clamshell incision) through the fourth
intercostal space provides excellent exposure of the heart and
hilar regions. Though absorbable suture theoretically provides
the greatest potential for growth, some surgeons use nonabsorbable suture material for the anastomoses.37 We recommend a simple end-to-end rather than a telescoping
anastomosis for the airway because of the high incidence of
stenosis in the latter.38,39 If the patient requires concomitant
repair of an intracardiac lesion (e.g., with Eisenmenger
syndrome), that is best performed after the recipient pneumonectomies and before implanting the donor lungs. Many of
these patients have significant aortopulmonary collaterals
resulting in significant pulmonary venous return to the heart
while on cardiopulmonary bypass. After the recipient lungs
have been removed, the absence of pulmonary venous return
to the heart from bronchial arteries and other collateral vessels
will allow for a bloodless operative field for the intracardiac
repair. The subsequent period during which allograft implantation is performed provides sufficient time for cardiac
reperfusion before weaning from cardiopulmonary bypass.
Living donor lobar transplantation and the use of cadaveric
lobes, has become less commonplace as an alternative to
standard cadaveric whole lung transplantation, since implementation of the LAS.40 Although the upper lobes have been
used, lower lobes seem better suited anatomically, with each
lobe serving as an entire lung. When lobes come from a living
donor, there is less bronchial and vascular tissue with which to
work and thus longer cuffs of the bronchus, pulmonary artery,
and pulmonary vein of the recipient will facilitate the procedure. A technique has been devised whereby a single left
lung can be partitioned such that the upper lobe is used on
the right and the lower lobe on the left.41 The circumstances
under which one might use this technique would be quite
unusuala single left lung from a large donor being made
available to a desperately ill child. Nonetheless, it is another
attempt at solving the ongoing problem of inadequate donor
organ supply.
Immunosuppression
------------------------------------------------------------------------------------------------------------------------------------------------
Although the precise protocols differ from one center to

another, most use the so-called triple-drug immunosuppression approach (Table 51-3). Combinations of these immunosuppressant drugs allow for a better overall effect with a
relatively less toxic dose of any one agent. Drug regimens generally include cyclosporine or tacrolimus in combination with
azathioprine or mycophenolate mofetil (MMF) and prednisone. Most pediatric lung transplantation centers now use
tacrolimus because of the cosmetic advantages it offers versus
cyclosporine and therefore perhaps improving adherence,
especially among adolescent recipients. The use of induction
cytolytic therapy using antithymocyte globulin is somewhat
controversial because of the potential of infectious complications associated with their use. Basiliximab, a specific monoclonal antibody to interleukin 2, is an alternative to cytolytic
agents to induce tolerance.42 Rather than being cytolytic,
these drugs work by blocking a critical pathway in the activation of lymphocytes involved in cellular rejection. The low
infection rate using these monoclonal antibodies has stimulated the reemergence of induction therapy early after lung
transplantation.43 The initial target trough cyclosporine blood
level is 300 to 400 ng/mL by whole blood monoclonal assay.
When tacrolimus is used, that target trough level is 10 to
15 ng/mL. The initial corticosteroid dose is 0.5 mg/kg daily
of prednisone or methylprednisolone. MMF is given at a dose
of 600 mg/m2 twice daily while azathioprine is given in a
dose of 2.5 to 3.0 mg/kg daily. Acute rejection is treated with
3 consecutive days of intravenous methylprednisolone at a
dose of 10 mg/kg/day. Rejection refractory to methylprednisolone is treated with antithymocyte globulin for 7 to 10 days.
Recurrent (greater than three) bouts of acute rejection may
also prompt a change of the baseline immunosuppression.
Although the corticosteroid dose is gradually tapered with
TABLE 51-3
Immunosuppressant Agents
Class of Drug
Side Effects
Interleukin-2 Synthesis Inhibitors

Cyclosporine
Hypertension, seizures, nephrotoxicity, hirsutism,
gingival hyperplasia
Tacrolimus
Hyperglycemia, seizures, nephrotoxic
Lymphocyte Proliferation Inhibitors
Azathioprine
Leukopenia, nausea
Mycophenolate
Leukopenia, nausea, diarrhea, elevated liver
mofetil
enzymes
Sirolimus
Hypertriglyceridemia, delayed wound healing
Corticosteroids
Hypertension, hyperglycemia, cushingoid
appearance
Induction Agents
Antithymocyte
Fever, chills, leukopenia, cytomegalovirus
globulin
infections, post-transplantation
lymphoproliferative disorder
OKT3
Fever, chills, cytomegalovirus infections, posttransplantation lymphoproliferative disorder
Daclizumab,
Nausea, diarrhea
basiliximab
CHAPTER 51
time, we do not believe it is appropriate to stop this drug

altogether. The side effects of immunosuppressive drugs in children are similar to those seen in adults. Sirolimus (rapamycin)
is chemically similar to tacrolimus but inhibits the proliferative
response of lymphocytes to interleukin-2.44 It does not share
the nephrotoxic potential of tacrolimus. It is currently reserved
for situations of failure of other immunosuppressant drugs.
Some caution should be exercised in using sirolimus as initial
immunosuppression early after transplantation because there
has been evidence of impaired wound and airway healing
resulting in serious complications.45
All patients receive prophylaxis against pneumocystis
jiroveci pneumonia with either sulfamethoxazole-trimethoprim
orally 3 times per week or when sulfa allergy or intolerance is
present one may consider monthly treatment with aerosolized
pentamidine or daily therapy with atovaquone. Prophylaxis
against mucocutaneous Candida infections is also used.
Surveillance
------------------------------------------------------------------------------------------------------------------------------------------------
Surveillance after transplantation is based on periodic spirometry and bronchoscopy with biopsies and bronchoalveolar
lavage. Before discharge from the hospital, patients are
provided with a home spirometer and are asked to perform
spirometry at least once daily. A decrease in FEV1 of greater
than 10% from baseline is considered an indication for evaluation. All patients, regardless of size, undergo regularly
scheduled surveillance bronchoscopy to diagnose lower
respiratory infections, subclinical graft rejection, and airway
anastomotic complications. Virtually all episodes of suspected
rejection should be confirmed with transbronchial biopsies.
The main challenge occurs in small infants in whom a miniforceps is used through either the 2.8-mm or the 3.5-mm
pediatric flexible fiberoptic bronchoscope. However, obtaining an adequate specimen with these forceps can be challenging. Recently a 4.0 mm bronchoscope with a 2.2-mm suction
channel was introduced into clinical practice, thus allowing
the use of adult-sized forceps for many young children. At
our institution, bronchoscopy with biopsy is performed at
7 to 10 days and at 1, 2, 3, 6, 9, 12, and 18 months after transplantation as a surveillance procedure. Worsening pulmonary
function, infiltrates on a chest radiograph, or deterioration in
clinical status, such as fever or an oxygen requirement, also
prompt bronchoscopy and biopsy. Bronchoalveolar lavage is
performed at these procedures for quantitative bacterial,
routine viral, and fungal cultures.
Complications
------------------------------------------------------------------------------------------------------------------------------------------------
AIRWAY ANASTOMOTIC COMPLICATIONS

Anastomotic complications can involve either the airway or
the vascular anastomoses. Airway dehiscence was the major
source of postoperative morbidity and mortality in the early
days of lung transplantation when tracheal anastomoses were
performed. Not until this problem was solved by using an
omental wrap for the airway anastomosis could clinical lung
677
transplantation progress.46 Currently, dehiscence is rare in

spite of the fact that most surgeons do not use the omental
wrap any longer, but rather approximate donor and recipient
peribronchial tissue over the anastomosis. Dehiscence of the
airway may be either partial or total. Partial dehiscence can
usually be treated expectantly but puts the airway at increased
risk of late stenosis.47 Complete dehiscence requires emergent
therapy and is generally a lethal complication. Although reanastomosis should be attempted when possible, it is associated with a high rate of failure, and transplantation
pneumonectomy is required. Smaller airway size in children
prompted concerns about whether the incidence of bronchial
anastomotic stenosis would be higher and also whether the
anastomoses would grow. Current evidence suggests that
the airways at the anastomoses grow and that the incidence
of bronchial stenosis is not affected by age or size at the time
of transplantation.48,49 Bronchial stenosis is usually treated
with dilatation initially with either progressively larger rigid
bronchoscopes or with an angioplasty balloon. Balloon dilatation of a stricture may be preferable, because it is less likely
than a rigid bronchoscope to injure the distal airway. Repeat
bronchoscopy 10 to 14 days after initial dilatation of a bronchial stenosis is necessary to judge the overall effectiveness
and to assess the likelihood of recurrence. Depending on
the severity of the initial stricture or the rapidity with which
it recurs, one might consider placing a stent. There are two
basic types of stents applicable to this situation: Silastic and
wire mesh. In general, wire mesh stents are easier to insert
but much more difficult to remove, and Silastic stents are
harder to place and easier to remove. Alternatives to stent
placement include sleeve resection (of the bronchus or upper
lobe) or retransplantation. Resection has been performed with
good results in adults but would be a very difficult procedure
in children.50 Retransplantation should be reserved for situations in which the stricture extends beyond the bronchial
bifurcation on either side and cannot be managed with either
endobronchial techniques or local resection.
VASCULAR ANASTOMOTIC COMPLICATIONS

Problems with either the arterial or venous anastomoses are
rare. In most instances, a stenosis in either of these is secondary to excessive length on the donor pulmonary artery or left
atrial cuff or torsion of either of these structures when performing the anastomosis. Stenosis in one of the pulmonary
artery anastomoses may or may not be manifest by right ventricular hypertension. Because pulmonary artery catheters
are not often placed in children, one should check the right
ventricular pressure by direct puncture once off cardiopulmonary bypass. If elevated, the pressure distal to each anastomosis should be checked also by direct puncture. Unilateral mild
to moderate pulmonary arterial anastomotic stenosis may not
result in significant elevation of right ventricular pressure.
A perfusion lung scan is routinely performed within 24 hours
of the transplantation to screen for technical problems with
the vascular anastomoses. Any significant discrepancy between right- and left-sided perfusion should be immediately
evaluated with either direct visualization in the operating
room or angiography. Stenosis in either or both pulmonary
venous anastomoses is manifest by pulmonary hypertension,
profuse pink frothy sputum, and diffuse infiltrates on a chest
radiograph. These findings may also be present with a severe
678
PART IV
TRANSPLANTATION
reperfusion injury or diffuse alveolar damage. However, the

pulmonary capillary wedge pressure is generally normal in
the latter two instances and elevated with a stenosis in the
pulmonary venous anastomosis. Transesophageal echocardiography is particularly helpful in the diagnosis of pulmonary
venous anastomotic problems. Confirmation of the diagnosis
usually requires direct measurement of the pulmonary venous
and left atrial pressures, particularly in small children. Early
correction is mandatory.
BLEEDING
A number of factors place these patients at increased risk for
bleeding after transplantation. Nearly all transplantations in
children require prolonged cardiopulmonary bypass for
recipient pneumonectomies and implantation of donor
organs. Additionally, many of these patients have undergone
prior thoracotomies or sternotomies. Patients with cyanotic
heart disease and a prior thoracotomy have the greatest risk
of serious bleeding, as mentioned earlier.
PHRENIC NERVE INJURY

This complication occurs in about 20% of lung transplantations and is secondary to trauma because of stretch while
retracting to expose the hilar regions; it is more common on
the right side.51 Recovery of diaphragmatic function within
6 months of transplantation is the general rule. The reason
for the right side being injured more commonly probably
relates to the proximity of the nerve to the pulmonary artery
and the superior vena cava on that side. The superior vena
cava (and thus the phrenic nerve) must be retracted to expose
the proximal right pulmonary artery. Prior thoracotomy puts
the nerve at greater risk for injury, because it may be obscured
by adhesions.
HOARSENESS
Vocal cord paralysis caused by recurrent laryngeal nerve injury
has an incidence of approximately 10%. This diagnosis is
made at the time of flexible fiberoptic bronchoscopy with
direct examination of the cords. In most cases, anatomic asymmetry improves without directed therapy within 6 months of
transplantation. The left vocal cord is nearly always the one
involved, and the injury presumably occurs as a result of
dissection of the left pulmonary artery in the region of the
ligamentum arteriosum.
GASTROINTESTINAL COMPLICATIONS
Many centers now routinely assess for the presence of gastroesophageal reflux because of its potential association with the
development of bronchiolitis obliterans (BO).52 Since instituting routine 24-hour pH probe monitoring at the 2-month
post-transplantation evaluation, we have found that almost
70% of our patients have evidence of acid reflux. The etiology
of this high incidence of gastroesophageal reflux is not clear
but may be due to injury to the vagus nerves bilaterally in
the process of performing the recipient pneumonectomies.
Decreased intestinal motility is also a common problem in
all age groups. Patients with CF are at risk for distal intestinal
obstruction syndrome. This can be avoided by aggressively
treating with osmotic cathartics after transplantation. Gastrografin enemas may be necessary if there is no response to oral
cathartics.
ATRIAL FLUTTER
Atrial arrhythmias are relatively common with significant
episodes of atrial flutter occurring in 10% of pediatric lung
transplantation recipients. Many require long-term treatment.53 Investigation into this entity using a model of lung
transplantation has shown that the suture lines for the left
atrial anastomoses provide sufficient substrate for the maintenance of atrial flutter when initiated by programmed extrastimulus.54
GRAFT COMPLICATIONS
Reperfusion injury manifesting as graft failure with diffuse
infiltrates on chest radiography, frothy sputum, and poor
oxygenation is the most common graft complication early after
lung transplantation, occurring in 20% to 30% of transplantation recipients.55 It is the most common cause of death
within the first 30 days after transplantation.3 The underlying
cause is probably multifactorial, with both donor and recipient conditions contributing to this problem. The best preventive measures include careful evaluation and procurement of
the donor organs as well as having a recipient free of active
infection or other acute problems. A well-conducted transplantation procedure is also of utmost importance. The treatment of reperfusion injury is mostly supportive, although
nitric oxide56 and prostaglandin E157 may be of some primary
benefit.
Rejection is a common occurrence after lung transplantation, perhaps more so than in other solid organ transplantations (Fig. 51-1). The lung has a much larger
endothelial surface than other organs. Because the major histocompatibility antigen expression on endothelial surfaces is
the primary signal for local immune recognition, the lung
would seem to be the least easily camouflaged organ in the
body. In addition, the lung graft comes with its own parenchymal bronchial lymphocytes and macrophages. Gradually,
these are replaced by the recipient lymphocytes and macrophages. This rather intense immunologic activity adds to
the risk of rejection. Acute graft rejection early after transplantation presents in such a nonspecific fashion that each suspected episode should be documented with histologic
evidence obtained by either transbronchial biopsy or openlung biopsy. The great majority of episodes of acute rejection
occur in the first 6 months after transplantation. Although
the incidence of acute rejection in all children is about the
same as that seen in adults, it appears that infants have a
much lower incidence.58,59 The precise reason for this is
unclear but may have to do with the relative immaturity of
the immunologic system in infants.
Antibody-mediated rejection (AMR) is now recognized as a
serious and relatively common complication of lung transplantation. With the advent of better detection assays, one
can now quantitate the amount of circulating donor-specific
antibody in the recipient. However, it is also becoming
increasingly clear that nonhuman leukocyte antigen (HLA)
antibodies may also be responsible and that, in fact, autoantibodies to previously sequestered antigens may play a vital role
CHAPTER 51
679
FIGURE 51-1 Acute rejection. Multiple lymphocytes are present in a perivascular position involving many blood vessels, which can be seen better on
higher power. This was interpreted as grade A2 acute rejection.
in the development of bronchiolitis obliterans. Increasing

experience with C4d immunostaining of biopsy specimens
and C4d and C3d immunofluorescence permit histologic
confirmation of AMR. The exact frequency of this complication in lung transplantation recipients is not yet known.
AMR has shown itself to be fairly refractory to therapy.
We use a protocol that includes plasmapheresis, intravenous bortezomib, intravenous immunoglobulin (IVIG), and
rituximab.
Bronchiolitis obliterans is viewed by most clinicians to be a
manifestation of chronic rejection and occurs in nearly 50% of
all long-term survivors.60 The precise cause is unknown,
although donor ischemic time, episodes of early acute rejection, and history of lymphocytic bronchitis have been identified as risk factors.61,62 Bronchiolitis obliterans presents as a
significant fall in FEV1 without other obvious cause. The chest
radiograph is generally clear, and the computed tomographic
examination of the chest usually demonstrates evidence of air
trapping with mosaicism and occasionally bronchiectasis.
Ventilation/perfusion lung scanning may demonstrate air trapping with xenon retention. Bronchoscopy with transbronchial
biopsy and bronchoalveolar lavage should be done as part of
the evaluation to rule out other potential causes, such as acute
rejection or infection, and to assess the degree of active lymphocytic infiltration of the airways. The histologic picture of
bronchiolitis obliterans is one of dense scarring of the membranous and respiratory bronchioles (Fig. 51-2). It may be inferred by the absence of identifiable bronchioles on biopsy
material. Diagnosis of bronchiolitis obliterans by histologic
examination of transbronchial biopsy material may be very
difficult, however, and many do not consider it necessary to
establish the diagnosis. A staging system has been established
based on the degree of decline of FEV1 or forced expiratory
flow (FEF)25-75 from the peak value: post-transplantation
stage 0p 10% (FEV1) or 25% (FEF25-75), stage 1 20%
to 35%, stage 2 35% to 50% decline, and stage 3 more
than 50% decline.63 The usual treatment for bronchiolitis
obliterans in the United States is to augment immunosuppression, usually beginning with antithymocyte globulin daily for
7 to 10 days; the clinical response has been variable. A change
in the maintenance immunosuppression may also be appropriate. Antiproliferative agents may provide a more effective
approach, but that has yet to be proved. Results from small
studies suggest that azithromycin may be effective in stabilizing and perhaps even improving lung function.64,65
Researchers at Duke University demonstrated that fundoplication in patients with BO and gastroesophageal reflux may
potentially improve BO grade if performed early.52 Total lymphoid irradiation and photopheresis are other modalities that
have been proposed.66 Patients not responding to these
measures may be suitable candidates for retransplantation.
As mentioned earlier, this is a somewhat controversial topic,
because there is a shortage of donor organs, and the results
with retransplantation overall are not quite as good as with
first-time transplantations. However, if the candidates are
ambulatory, not ventilator dependent, and at an experienced
lung transplantation center, the survival results are not
significantly different from first-time transplantations.30
Post-transplantation lymphoproliferative disease (PTLD)
occurs in 10% to 19% of pediatric patients undergoing lung
transplantation. PTLD occurs more frequently in association
with a primary Epstein-Barr virus (EBV) infection.67 Children
may be somewhat more prone to this complication, because
they are frequently seronegative for EBV infection at the time
of transplantation and are therefore likely to acquire a primary
EBV infection during their post-transplantation life. Reduction in immunosuppression is the mainstay of early therapy,
although this may be insufficient and not uncommonly
leads to the subsequent development of bronchiolitis obliterans. Rituximab, an anti-CD20 monoclonal antibody, has
been used effectively in the treatment of PTLD.60 Other
treatment modalities include conventional chemotherapy,68
irradiation, and infusion of human leukocyte antigen
matched T lymphocytes.69
680
PART IV
TRANSPLANTATION
FIGURE 51-2 Histologic slide taken from the lung of a patient undergoing retransplantation for bronchiolitis obliterans. Small airways are obliterated by
fibrous tissue.
INFECTION
Although infection is generally common after any solid organ
transplantation, lung transplantation recipients are at greater
risk. Donors are all on mechanical ventilation, resulting in
colonization of the airway with bacteria from an intensive care
unit. The lung is the only solid organ constantly in contact
with the nonsterile outside world. An endotracheal tube necessary early after the transplantation bypasses some of the
natural defenses available to the respiratory tract. Obligate
denervation of the lung that occurs with transplantation results in the cough reflex being markedly diminished or absent
altogether. These and numerous other factors demand that the
caregivers maintain constant vigilance in the diagnosis and
treatment of respiratory infections and also emphasize to
the recipient the importance of pulmonary toilet.
All potential candidates are screened for the presence of
organisms in the airway and evidence of previous infections.
Evidence of prior viral infections is evaluated by serologic
testing for antibodies to cytomegalovirus; herpes simplex
virus; varicella; EBV; hepatitis A, B, and C; and human immunodeficiency virus. Viral serologic screening is less informative in young infants whose immunoglobulin pool reflects
passively transferred maternal antibodies. The initial antimicrobial therapy given in the early post-transplantation period is directed in part by the results of pretransplantation
studies. Ganciclovir is given at a dose of 5 mg/kg/day for
6 weeks for any positive donor or recipient serology for
cytomegalovirus. If patients have evidence of present or past
Aspergillus infection, antifungal therapy with either intravenous anidulafungin or voriconazole followed by oral voriconazole is used, depending on the clinical situation.
A number of viral respiratory infections are quite common
in pediatric patients. Adenovirus and parainfluenza viruses
are particularly bothersome in children. As for cytomegalovirus,
primary disease is generally more likely to be severe than reactivation disease.2 As mentioned earlier, primary infection with
EBV is an important risk factor for the development of PTLD.
Fungal infections are uncommon but potentially devastating. Nystatin oral suspension is employed to reduce the risk of
infection from Candida species. Virtually all infections caused
by Candida species can be successfully treated with oral or
intravenous triazole antifungal agents. Invasive Aspergillus infections, however, are much more difficult to treat and may
result in widespread dissemination if appropriate antifungal
therapy is delayed.
Bacterial infections are common after lung transplantation.
Bacterial lower respiratory tract infections, which include both
purulent bronchitis and pneumonia, occur in most patients at
some point after transplantation. Patients with CF are more
likely to experience this complication, with the organism usually the same as that colonizing the airway before transplantation. Prophylaxis against lower respiratory tract infections in CF
lung transplantation recipients may be accomplished by administering aerosolized antibiotics (tobramycin or colistin) just as
one might for end-stage CF.
OTHER COMPLICATIONS
Hypertension is a common problem after transplantation and
is presumably due to treatment with the calcineurin inhibitors
cyclosporine and tacrolimus, as well as prednisone. Renal insufficiency occurs with increasing time after transplantation
and is also related to treatment with cyclosporine and tacrolimus. Diabetes mellitus occurs in approximately 15% of
patients after transplantation, primarily in patients with CF.2
Tacrolimus predictably increases the likelihood for the development of hyperglycemia.
Survival
------------------------------------------------------------------------------------------------------------------------------------------------
The 3- and 5-year actuarial survival for children undergoing

lung transplantation is approximately 54% and 45%, respectively, according to the International Society for Heart and
CHAPTER 51
681
100
1 year vs. 111 years: P = 0.3124
<1 year vs. 1217 years: P = 0.8387
111 years vs. 1217 years: P = 0.0395
Survival (%)
75
<1 year (N = 84)

111 years (N = 334)
1217 years (N = 756)
50
N at risk = 17
N at risk = 10
25
N at risk = 20
HALF-LIFE
<1 year: 6.4 years
111 years: 6.0 years
1217 years: 4.3 years
0
0
10
Years
Lung Transplantation registry (Fig. 51-3).3 Acute graft failure

accounts for the majority of deaths in the first 30 days. Infection is the cause of death in approximately 50% of those dying
in the first year beyond the transplantation hospitalization.
Bronchiolitis obliterans is the cause of death in 50% of patients
beyond 1 year after transplantation and is clearly the major
impediment to long-term survival.30
Pulmonary Function and Growth

------------------------------------------------------------------------------------------------------------------------------------------------
It is unclear whether transplanted lungs grow in terms of number and size of alveoli, and experimental data are inconclusive.70,71 Measurement of lung growth is fraught with a
number of complicating factors. One cannot use pulmonary
function tests and lung volume size as measured by either
chest radiograph or computed tomography, because there are
a number of elements that affect these studies that would not
accurately reflect the number or size of alveoli. The impact of
lung growth is particularly critical in small infants, because their
transplanted lungs will have to grow substantially over the rest
of their lives to handle the physiologic load presented to them.
Those children in our series too young to undergo standard
pulmonary function testing underwent infant pulmonary function tests that provide a measurement of functional residual
capacity, a reasonable surrogate for lung volume. The average
functional residual capacity per centimeter in height at
3 months after transplantation was 2.3 mL/cm and remained
between 2.1 and 2.8 mL/cm through 15 months after transplantation. During this time, substantial somatic growth
occurred in these infants.72 Thus in the absence of central or
peripheral airway obstruction, these data suggest that lung
growth appropriate for size is occurring. However, we do not
11
12
FIGURE 51-3 Kaplan-Meier survival curve for pediatric lung

transplantation. (From Aurora P, Edwards LB, Kucheryavaya
AY, et al: The Registry of the International Society for Heart
and Lung Transplantation: Thirteenth Official Pediatric Lung
and Heart/Lung Transplantation Report2010. J Heart Lung
Transplant 2010;29:1129-1141.)
know whether this represents an increase in the number of

alveoli and/or an increase in the size of existing alveoli.
Future Considerations
------------------------------------------------------------------------------------------------------------------------------------------------
Factors that limit the success of lung transplantation in children are similar to those in adults: donor shortage, balance
of immunosuppression and prevention of infection, and
development of bronchiolitis obliterans. Xenotransplantation
may eventually offer another solution, but realistically, this
is many years from application. Transplantation across ABO
blood groups, now commonplace in infant cardiac transplantation, is another possibility in small children, though the
overall impact of this would be minor. Newer immunosuppressive agents aimed at more specific areas of the immune
response involved with organ recognition are necessary. Bronchiolitis obliterans remains the Achilles heel of long-term
survival after lung transplantation. Although still not
completely characterized as to its precise cause, most investigators ascribe this development to airway injury leading to
chronic rejection. To that end, clinical and basic research
aimed at understanding the vectors of injury and disease
progression in bronchiolitis obliterans are of paramount importance to the field of lung transplantation. Because the
airway as the site of injury is accessible for assessment and
therapy, bronchiolitis obliterans may provide a model system
whereby chronic rejection, which also affects long-term
success in heart, kidney, and liver transplantation, can be
understood and overcome.
expertconsult.com.
CHAPTER 52
Surgical
Implications
Associated with
Pediatric Bone
Marrow
Transplantation
care, the switch to PBSC has led to mortality rates of less

than 5% for autologous transplantation in many studies.
One of the major advances of PBSC compared with autologous
marrow is more rapid engraftment of the recipient. Faster
hematopoietic recovery results in an abbreviated period of
neutropenia, thrombocytopenia, and anemia, resulting in
lower rates of infection and hemorrhage, less risk of transfusions, and earlier discharge. Harvest and storage of a patients
own hematopoietic stem cells (HSC), followed by reinfusion
after high-dose chemotherapy (HDC), is commonly referred
to as autologous HSCT or stem cell rescue. HDC is generally
administered beyond the tolerance of the patients marrow
(myeloablative), meaning no recovery is possible without
stored HSC.3
The transplantation process is divided into five phases:
(1) conditioning, (2) stem cell infusion, (3) neutropenia,
(4) engraftment, and (5) postengraftment phase. The conditioning phase involves intensive chemotherapy with or without total body irradiation to eliminate the disease. This period
lasts between 7 and 10 days. The stem cell processing and
infusion time varies based upon the size of the patient and
source of the stem cells. The neutropenic phase lasts 2 to
4 weeks and is an interval when the patient is extremely
vulnerable to infections because of the lack of an effective
immune system. Wound healing is impaired, and empiric
antibiotics are generally administered to minimize infectious
complications. Mucus membranes are rapidly dividing tissues, and therefore susceptible to ulceration and the risk for
nosocomial infections is high. Total parenteral nutrition is
widely used in children during this phase.
The engraftment phase takes several weeks as transplanted
cells incorporate into recipient tissues. The development of
graft versus host disease (GVHD) and viral infections are
the greatest clinical obstacles of this phase. The postengraftment phase lasts months to years as the gradual development of tolerance, weaning of immunosuppression, and
immune reconstitution occur.4 Pediatric surgeons are consulted frequently for access or associated complications in
all phases of the transplantation process.
Thomas E. Hamilton and Robert C. Shamberger
Stem Cell Harvest and Vascular

Access
Fifty-three years after the seminal report of hematopoietic

stem cell transplantation (HSCT) in children and adolescents
by Thomas and colleagues,1 pediatric surgeons have retained
an important role in successful implementation of this therapy.
The expansion of HSCT to a variety of malignant and nonmalignant diseases has grown based on successful use of multiple
stem cell sources and innovative conditioning regimens.2
A fundamental understanding of these processes will assist pediatric surgeons when called upon for access or complications
associated with patients undergoing HSCT.
Stem cells can be derived primarily from bone marrow,
umbilical cord blood, or peripheral blood stem cells (PBSC).
Along with advances in infection prophylaxis and supportive
Multiple ports of access are required for peripheral stem cell

harvest. Unlike adults, where large-bore antecubital catheters
can be used for both harvest and infusion, pediatric patients
generally require indwelling central venous catheters. The extracorporeal separation of blood components from patients or
donors has spawned an entire field termed apheresis. As technologies of separation evolve, the one constant is the need for
adequate access to withdraw and return blood components.
Standard Broviac catheters collapse with the negative pressure
required for apheresis (approximately 1 to 2 mL/kg/minute).
Specially designed apheresis catheters allow faster flow rates
because of larger-diameter stiffer walls and shorter catheter
length. Most apheresis catheters are dual lumen offset with
multiple ports to avoid mixing processed and unprocessed
blood. Children weighing more than 10 kg will accommodate 8-Fr or larger (MedComp, Harleysville, Pa.) catheters.
When children weigh less than 10 kg, temporary femoral or
------------------------------------------------------------------------------------------------------------------------------------------------
683
684
PART IV
TRANSPLANTATION
subclavian catheters may be used.3 Because of the larger size

and stiffness of the apheresis catheters and dilators, fluoroscopic guidance and respect for tissue is paramount. As one
of my surgical mentors always said, placing the line in is
never as interesting to the surgeon as taking out the tumor,
but it is just as important to the care of the child.
Complications of Immune
System Ablation and
Immunosuppression
TABLE 52-1
Neutropenic Enterocolitis: Criteria for Appropriate Surgical
Intervention
1. Persistent gastrointestinal bleeding after resolution of neutropenia
and thrombocytopenia and correction of clotting abnormalities
2. Evidence of free intraperitoneal perforation
3. Clinical deterioration requiring support with vasopressors or large
volumes of fluid, suggesting uncontrolled sepsis
4. Development of symptoms of an intra-abdominal process, in the
absence of neutropenia, which would normally require surgery
------------------------------------------------------------------------------------------------------------------------------------------------
INTESTINAL COMPLICATIONS
Abdominal pain and diarrhea are common post-HSCT. A substantial component of the initial inflammatory cascade is
thought to occur in the gastrointestinal tract, and patients with
higher volumes of diarrhea at the time of the preparative regimen have an increased risk of acute GVHD.5 Barker and
associates6 performed a retrospective study of 132 consecutive
pediatric HSCT patients, and diarrhea occurred in 67% of
patients. Common etiologic agents included GVHD (27%),
viral (6%), Clostridium difficile (8%), and unknown (28%).
When stool cultures are negative, endoscopy is considered
to differentiate infectious etiologies from GVHD. Gastric antral
biopsies and small bowel biopsies may be preferred, because
duodenal hematomas have been reported by Ramakrishna and
Treem.7 They recommended avoiding the duodenum if possible and maintaining platelet counts greater than 55,000/mm3
for 48 hours postbiopsy when a duodenal biopsy is necessary.
A prospective multicenter study of pediatric bone marrow
transplantation (BMT) patients who underwent 1120 small
bowel biopsies did not report hematoma as a complication.8,9
Silbermintz and co-workers10 reported successful identification of small bowel graft versus host disease by capsule endoscopy in a child with refractory hemorrhage, when upper and
lower endoscopies were nondiagnostic. Identification of small
intestinal cytomegalovirus (CMV) disease has also been
reported by capsule endoscopy.11 The future role for capsule
endoscopy is increasing as the intestinal complications after
HSCT predominate in the small intestine.12 Accurate visualization may help guide the need for more intensive immunosuppressive therapy or to avoid immunosuppressive therapy.
Neutropenic enterocolitis (typhlitis) is characterized by
necrotizing inflammation of the colon in a severely immunocompromised patient (Table 52-1). Clinically, fever, abdominal pain, tenderness, and neutropenia are present
(Fig. 52-1). The incidence is low in children after HSCT.
Barkers retrospective review of 132 consecutive pediatric
HSCT patients reported an incidence of 3.5%.6 Early experience with neutropenic enterocolitis was marked by controversy regarding the timing of surgical intervention, and
mortality exceeded 50%.1316 Delineation of the criteria for
surgical intervention reserved for clearly identified surgical
complications has contributed to a substantial decrease in
mortality and morbidity.15,17
Multiple contemporary series now report excellent outcomes using a strategy of bowel rest, prompt institution of
appropriate intravenous fluid resuscitation, broad-spectrum
antibiotics and antifungal therapy, nutritional support with
total parenteral nutrition (TPN), and the use of granulocyte
FIGURE 52-1 Typhlitis in a teenager with acute myelocytic leukemia

who had undergone bone marrow transplantation (BMT). Note the thickened, onionskin appearance of the cecal wall and the pinpoint lumen.
This patient underwent right hemicolectomy with ileostomy and mucous
fistula; 1 year later, still in clinical remission from her leukemia, her
condition was successfully reversed. Most cases of typhlitis are handled
nonoperatively.
colony-stimulating factor (GCSF).1821 In 2002, Otaibi reported a series of 142 HSCT transplantations performed in
Alberta Childrens Hospital. Ninety-seven patients developed
abdominal pain, and only five developed radiographically
proven typhlitis. No patients required surgical intervention.20
Mullassery, from The Royal Liverpool Childrens Hospital reported a 5-year retrospective series in 2009 in which 18 of
596 patients had radiographically confirmed typhlitis and
three required surgical intervention. One child, each, had
extensive colonic necrosis, perforated gastric ulcer, and perforated appendix. A single mortality was also reported from
fulminant gram-negative sepsis without intervention.19
HEPATOBILIARY COMPLICATIONS
Abnormal liver function studies are commonly identified in
HSCT patients. An extensive 40-year review of hepatobiliary complications in HSCT has recently been published by
McDonald.22 Liver complications have become far less frequent as the understanding of how to prevent and treat severe
hepatobiliary problems has emerged. Surgeons are frequently
consulted to discern whether abnormal liver function studies
are secondary to obstruction or parenchymal dysfunction.
Biliary obstruction occurs secondary to calculous disease.
CHAPTER 52
SURGICAL IMPLICATIONS ASSOCIATED WITH PEDIATRIC BONE MARROW TRANSPLANTATION
Safford and colleagues23 reported a series of 575 patients, of

which 235 received ultrasonography of the abdomen for pain,
jaundice, sepsis, or metastases. Cholelithiasis was identified in
20 cases (8.5%). The overall incidence of cholelithiasis
reported in the study was far greater than the 0.13% to
0.21%. incidence in children.24 When the reason for HSCT
was considered in Saffords series, 27% of patients who had
HSCT for bone marrow failure versus 7.4% for neoplasia
developed cholelithiasis (P < 0.01). This suggests a role for
hemolysis. Despite the high incidence of cholelithiasis,
85% of children did not require surgical intervention.
Nine (45%) died from primary disease, five (25%) showed
sonographic resolution, and three (15%) had nonoperative
follow-up for persistent cholelithiasis. Surgical interventions included one cholecystostomy, one open cholecystectomy, and one laparoscopic cholecystectomy in three
(15%) patients who developed acute cholecystitis, with a
mean time to operative intervention of 1.9 years without
complication. There was no morbidity or mortality associated with conservative management of cholelithiasis in any
child.23
Hepatic veno-occlusive disease (VOD), also known as sinusoidal obstruction syndrome (SOS) is a serious and frequent
complication after HSCT. It is clinically heralded by a triad
of (1) painful hepatomegaly, (2) hyperbilirubinemia, and
(3) unexplained fluid retention. Milder cases resolve spontaneously, but severe cases may become rapidly fatal. Prognosis
varies with extent of injury and the development of multiorgan system failure.25 The pathogenesis involves sinusoidal
endothelial cell and hepatocyte damage from high-dose alkylating chemotherapeutic agents.22,25 Plasminogen activator
inhibitor (PAI-1) serum levels have both diagnostic and prognostic value as a marker for VOD.26 The intensity of the conditioning regimen and type of transplantation are important
determinants for risk of VOD. Pediatric patients younger than
6.5 years at transplantation appear to be at increased risk of
hepatic VOD.27 Defibrotide (DF) is a polydisperse mixture
of single-stranded oligonucleotide with local antithrombotic,
anti-ischemic, and anti-inflammatory activity that has clinical
efficacy in severe VOD. DF appears to modulate endothelial
cell injury without enhancing systemic bleeding and protects
the hepatic sinusoids without compromising the antitumor
cytotoxic effects of therapy.28 Ho and associates25 reviewed
hepatic VOD and multiple clinical trials with DF, where
30% to 60% complete remission rates are reported, even in
patients with severe VOD and multiorgan system failure. Corbacioglu and colleagues29 reported that 34 (76%) patients
with severe VOD post-HSCT, treated with DF, achieved a complete response, and in multivariate analysis, early intervention
(1 day vs. 5.5 days in nonresponders) was the only significant
factor. Richardson and coworkers30 recently reported a phase
II multicenter randomized trial in adult and pediatric patients
that established the safety and efficacy of defibrotide. Early stabilization or decreased bilirubin was associated with better
response and day 100 post-HSCT survival, and decreased
plasminogen activator inhibitor type 1 (PAI-1) during treatment was associated with better outcome; changes were similar in both treatment arms. A dosage of 25 mg/kg/day was
selected for ongoing phase III trials of the treatment of VOD.
Focal nodular hyperplasia following pediatric HSCT has
been identified in 17 of 137 patients prospectively studied,
with a median delay of 6.4 years from a series reported by
685
Sudour and associates.31 The authors postulate an iatrogenic

vascular origin, because 16 patients received myeloablative
preconditioning, and only three had evidence of SOS. No
complication or malignant transformation was reported; clinical and diagnostic imaging follow-up is recommended.
HEMORRHAGIC CYSTITIS
Hemorrhagic cystitis (HC) occurs in 10% to 20% of pediatric
HSCT patients. Decker and colleagues32 have recently
reviewed the pediatric experience. HC is characterized by diffuse vesical bleeding which ranges from microscopic hematuria to gross hemorrhage with clot formation and urinary
obstruction requiring instrumentation for evacuation. With
severe HC, prolonged hospitalization with significant morbidity may occur. High-dose chemotherapy and immunosuppression that accompany HSCT make the pediatric patient
particularly susceptible. Investigations point to a multifactorial pathophysiology of HC. Damage to the transitional epithelium by radiation, chemotherapy, and infectious agents have
been postulated. BK virus is now a known pathogen with
increasing evidence for a major role in HC. High-dose cyclophosphamide and bisulfan are well studied alkylating agents
used in conditioning protocols for HSCT that are known to
cause HC. Three main strategies for HC prophylaxis include
mesna, hyperhydration with forced diuresis, and continuous
bladder irrigation (CBI). Three-way catheter drainage is often
difficult in pediatric populations and may require a suprapubic tube. Ultrasonography may underestimate the clot burden,
and cystoscopy provides visualization with the opportunity
for clot evacuation and fulguration of the bladder epithelium.
Escalation to more intensive therapies, such as instillation of
drugs into the bladder (intravesical therapy), carry increased
risk. Many agents, including aluminum potassium sulfate,
prostaglandins, and E-aminocaproic acid and cidofovir have
been used for intravesical therapy, but none are well studied
in a pediatric population. HC is a self-limiting condition
once engraftment and immune reconstitution occur, effective
defense barriers of the bladder mucosa are reconstituted,
and viral replication is controlled. More intensive therapies
should be undertaken under the auspices of a multidisciplinary team.33
PULMONARY COMPLICATIONS
Opportunistic infections are common in HSCT patients.
Pneumonia is the most common infectious complication but
must be distinguished from noninfectious causes, such as
bronchiolitis obliterans, diffuse alveolar hemorrhage (DAH),
and a constellation of noninfectious fever accompanied by
either skin rash, pulmonary infiltrates, or diarrheatermed
engraftment syndrome.34 Early recognition and treatment of
post-HSCT pneumonia favorably impacts survival.35,36 A
recent article from Shannon and co-workers37 from M.D.
Anderson Cancer Center has examined the utility of early versus late fiberoptic bronchoscopy (FOB) with bronchoalveolar
lavage (BAL) post-HSCT in 501 consecutive adult patients.
Five hundred and ninety-eight fiberoptic bronchoscopies
(FOB) with bronchoalveolar lavage (BAL) were performed
for the evaluation of pulmonary infiltrates. The overall diagnostic yield was 55%. The diagnostic yield was 2.5-fold
higher when the FOB was performed within the first 4 days
686
PART IV
TRANSPLANTATION
and highest (75%) when performed within 24 hours of clinical presentation. The rates of adjustment in antimicrobial
therapy were not different with early versus late treatment
(51%); however, late FOB-guided antibiotic adjustments were
associated with 30-day pulmonary-associated deaths that
were threefold higher (6% vs. 18%, P 0.035). The authors
conclude early referral for FOB may yield a higher diagnostic
yield and favorably impact survival in adult patients.
The utility of lung biopsy in pediatric patients post-HSCT is
controversial. Shorter and colleagues38 reported a 10-year experience of 126 HSCT patients from Childrens Hospital of
Philadelphia from 1976 to 1986. Twenty-one patients had open
lung biopsies; 14 showed no causative organisms. One patient
had CMV, and three patients had Pneumocystis carinii. Thirteen
patients died because of continued deterioration postbiopsy.
Hayes-Jordan and associates39 reported a retrospective series
of 528 patients post-HSCT from St. Judes Childrens Research
Center from 1991 to 1998. Eighty-three patients developed
pulmonary infiltrate within 6 months; 43 (52%) had BAL
and 19 (23%) had open lung biopsies, 6 (7%) underwent needle biopsy, and 5 (7%) underwent transbronchial biopsy.
Histology identified infections in 6 (30%), bronchiolitis obliterans organizing pneumonia (BOOP) in 5 (26%), interstitial
pneumonia in 4 (21%), gangliosidosis in 1, and lymphocytic
infiltrate in 1. Despite changing the clinical plan, based on histology in 17 of 19 (90%) patients, improvement in outcome
was only seen in 8 (47%). Postoperative morbidity at 30 days
was 47%, including prolonged intubation (7 patients), pneumothorax (2 patients), and pleural effusion (1 patient).
Thirty-day survival was 63.2%, and no patient with multiorgan
system failure, ventilator dependence, or postoperative complication survived postopen lung biopsy. Careful patient selection and consideration of less-invasive modalities should be
strongly considered in these extremely high-risk patients. Minimally invasive surgical techniques have been applied both
diagnostically and therapeutically in childhood cancer;40 however, the decrease in pulmonary compliance and increase in
cardiac afterload is often prohibitive for thoracoscopic techniques in the post-HSCT patient population.
Invasive pulmonary aspergillosis (IPA) is a common infection in the HSCT population. A potentially lethal complication
of HSCT is pulmonary hemorrhage secondary to the angioinvasive nature of this agent. IPA is one specific opportunistic
infection where surgical therapy remains beneficial. Gow and
colleagues41 reported on 43 patients with invasive pulmonary
aspergillosis, spanning 9 years, from St. Judes Childrens
Cancer Research Hospital. Eighteen patients had surgical
intervention, (16 thoracotomies [89%] and 2 thoracoscopies).
Fourteen had one operation; 4 patients had two. Surgical
resection of the affected parenchyma significantly improved
survival (P < 0.001). The four survivors had disease amenable
to wedge resection, the longest interval at the time of report
being 43.5 months. When feasible, a surgical approach should
be strongly considered, because, left untreated, invasive
pulmonary aspergillosis is almost always fatal.
even more challenging. As opposed to the otherwise healthy

child, who typically has a solitary organism after a traumatic
event, immunocompromised patients may have enteric translocation of gram-negative organisms. Extreme pain (often out
of proportion to physical findings), fever, and tachycardia
are the hallmarks of soft tissue infection. Johnston and
colleagues42 presented a retrospective series over an 11-year
period, where seven neutropenic patients with deep soft
tissue infections were identified. The median number of days
postinitiation of chemotherapy was 14, pain was present in
all patients, and 86% had fever and tachycardia. The pathogenic organism was from the gastrointestinal tract in four
of seven patients. Five patients survived and were treated
with urgent surgical debridement, intravenous antibiotics,
GCSF, and hyperbaric oxygen. Butterworth and associates43
reported an 11-year retrospective series of 19 patients with
necrotizing soft tissue infections in healthy and immunocompromised children. An interesting finding in this series
was that the immunocompromised patients were less likely
to have severe tenderness and more likely to have polymicrobial perineal/buttock infections. When diagnostic uncertainty exists, judicious use of magnetic resonance imaging
(MRI) may prove beneficial if tolerated by the patients clinical
status.
Post-transplantation Malignancies
As survival increases post-HSCT, attention has become
directed toward late effects, including post-tranplant malignancies (PTMs). New malignancies post-HSCT fall into
three broad categories: post-tranplant lymphoproliferative
disorders (PTLD), hematologic malignancies (primarily
treatment-related myelodysplastic syndrome and acute myeloid leukemia [MDS/AML]) and solid tumors. Baker and
co-workers,44 from the University of Minnesota, reported
147 PTMs in 137 pediatric and adult patients of 3,372 patients
poststem cell transplantation. The majority of PTMs were
PTLD (44), with MDS/AML in 36 patients. Sixty-two solid
tumors were reported in 57 patients. A significant finding
was the risk of solid tumor malignancy continues to increase
with each successive year of follow-up. The cumulative incidence of developing a solid tumor did not plateau and was
3.8% (95% confidence interval [CI], 2.2 to 5.4) at 20 years
post-HSCT. The most common solid tumors reported were
11 basal cell and 8 squamous cell carcinomas of the skin,
4 breast, 5 carcinoma in situ, 8 melanoma, and 4 soft tissue
sarcomas. Age greater than or equal to 20 years at the time
of HSCT was the only significant predictor for development
of a solid tumor (relative risk [RR] 2.0; 95% CI, 1.1 to 3.5;
P 0.03).44 Forty-two percent of patients died from their
post-transplantation solid tumor. Future efforts at long-term
surveillance are warranted.
Conclusion
------------------------------------------------------------------------------------------------------------------------------------------------
SOFT TISSUE INFECTIONS

Necrotizing soft tissue infections are rapidly progressive and
carry a significant mortality and morbidity without prompt
surgical intervention. Neutropenia makes the clinical picture
HSCT has evolved into a well-established clinical modality

with increasing utility for multiple malignant and nonmalignant disease processes in children. Pediatric general surgeons
are integral members of the multidisciplinary team required
for HSCT because of the multitude of organ systems involved.
CHAPTER 52
SURGICAL IMPLICATIONS ASSOCIATED WITH PEDIATRIC BONE MARROW TRANSPLANTATION
Pulmonary, gastrointestinal, hepatic, genitourinary, and soft

tissues are all subject to surgical complications. Vascular
access will continue to be required for medications, apheresis, and transfusions. Pediatric surgeons must have knowledge of the HSCT process and the surgical complications
during all phases of transplantation, including long-term
687
survivors, in order to provide expert advice for treatment

and to contribute to the understanding of this rapidly expanding field.
expertconsult.com.
The first two sections discuss two of the more common

skeletal anomalies treated by craniofacial surgeons: craniosynostosis and jaw anomalies that require orthognathic surgical
correction. The final section deals with facial asymmetry
and hypoplasia disorders that are rare and complex to treat.
The Craniosynostoses
------------------------------------------------------------------------------------------------------------------------------------------------
CHAPTER 53
Craniofacial
Anomalies
Jason J. Hall and H. Peter Lorenz
In 1967, Dr. Paul Tessier presented his modifications of Gilles

previously little known and underutilized techniques to correct congenital or post-traumatic bony anomalies to the International Society of Plastic Surgeons in Rome. Over the next
30 years, his teachings formed the basis of a new subspecialty
of plastic surgerycraniofacial surgery. Tessier developed
precisely placed osteotomies, autologous nonvascularized
bone and soft tissue grafts, and intracranial approaches to
the facial skeleton. His techniques gave children and adults
with previously unfixable craniofacial anomalies a chance
for a more normal appearance. Currently, most major medical
centers have a multidisciplinary team dedicated to the care of
this special subset of patients. The craniofacial anomaly
teams span many different specialties, which include plastic
surgery, pediatric otolaryngology, pediatric neurosurgery,
speech pathology, audiology, oral surgery, dentistry and
orthodontics, social work, pediatrics, and genetics.
Fortunately, many of the anomalies discussed in this chapter are relatively rare, but for surgeons in a pediatric tertiary
care center, these patients are seen as daily occurrences. This
chapter is not meant as a definitive tome on the diagnosis
and treatment of these disorders, but is meant to paint a
picture of the spectrum of craniofacial anomalies with broad
brushstrokes.
Craniosynostosis refers to the premature closure of one or

more cranial sutures and encompasses a wide range of anatomic derangements. Isolated, premature fusion of a single suture causes a predictable cranial deformity that can typically
be recognized without the need for radiologic imaging by
an experienced practitioner. However, complex craniosynostosis, consisting of multiple suture fusions, can be difficult
to diagnose without radiologic imaging. Despite the wide
range of clinical presentation, treatment for craniosynostosis
cases is similar to any other surgical problem in children: accurate diagnosis, appropriate treatment planning, proper timing of surgery with respect to current and future growth, and
surgical technique that gives a predictable correction and
minimizes adverse long-term sequelae.
ETIOLOGY AND PATHOLOGIC ANATOMY

The infant skull undergoes a period of rapid expansion during
the first year of life, which is driven by brain growth. Bone
growth at patent cranial sutures causes calvarial expansion
in a distinct morphologic pattern. Typically, suture fusion occurs from anterior to posterior and lateral to medial.1 The
metopic suture, which normally closes by 8 to 9 months of
age, is the only suture to close completely during infancy;
the remaining sutures do not completely fuse until adulthood.
Most patients presenting with craniosynostosis have prenatal
onset of suture fusion. However, in severe cases of syndromic
craniosynostosis, progressive, multiple suture fusion can occur over the first 3 to 4 years of life.2 In this situation, the calvarial deformity is typically not detected at birth except when
quite severe. After a few months of rapid growth, the deformity typically becomes more apparent.
In 1851, Virchow published his landmark paper that laid
the foundation for our understanding of craniofacial deformities associated with craniosynostosis. He described the growth
pattern of the skull as being restricted in a plane perpendicular
to the fused suture while being amplified in a plane parallel to
the direction of the suture. This compensatory growth pattern
causes predictable deformities in patients with single suture
synostosis. The growth constriction, however, is not confined
to the cranial vault, but also affects the cranial base to varying
degrees. Moss proposed that the cranial base pathology was
the inciting event, and that the calvarial fusion occurred
as a secondary phenomenon.3 However, cranial base anomalies are not corrected by calvarial vault surgery and are now
not thought to be the inciting event.
Recent research has focused on the role of the dura and its
influence on suture patency in the growing skull. A number of
dural-related cytokines, such as heparin-binding factor, fibroblast growth factors (FGFs), bone morphogenic proteins
(BMPs), transforming growth factor(s)-b (TGF(s)-b), and transcription factors Msx2 and TWIST, have a role in the regulation
and coordination of suture patency.4 A mutation of the TWIST
691
692
PART V
HEAD AND NECK
gene on chromosome 7p21 has been linked with SaethreChotzen syndrome.5,4 FGF receptor mutations causing constitutive activation of the receptor occur in many of the human
craniosynostosis syndromes, including Apert, Crouzon,
Muenke, and Pfeiffer syndromes.4 Interestingly, one mechanism of bony fusion across the suture occurring with the
FGF-R mutation is the loss of Noggin expression in the involved
suture mesenchyme. Noggin is a BMP-inhibitor that prevents
bony fusion in the mesenchyme. When Noggin is not present,
bone forms across the mesenchyme, and the suture fuses.6
Most incidences of craniosynostosis are the result of sporadic
genetic anomalies. Yet, a number of both autosomal dominant
and autosomal recessive syndromes, whose most striking phenotype is the pattern of craniosynostosis, are known. Patients
with a family history of craniosynostosis should thus be referred to a dedicated craniofacial team and be evaluated by
a skilled geneticist, as new genetic mutations linked to the
craniosynostosis syndromes are being discovered frequently.
The treatment for craniosynostosis is surgical calvarial vault
remodeling, which is performed to avoid future adverse
sequelae. Chief among these is the avoidance of intracranial
hypertension, which has been linked to brain damage, optic
nerve compression, and cognitive impairment.7 Early surgical
correction (between 3 to 6 months of age) has the advantages of prevention of elevated intracranial pressure and its
attendant consequences, improved reossification of calvarial
bone defects, and the need for a less extensive surgical correction. Correction at a more advanced age (6 to 9 months) is
reported to have more stable long-term results and lower rates
of reoperation. These factors are taken into account by the
craniofacial surgeon and pediatric neurosurgeon during the
treatment planning process.
FIGURE 53-1 A lateral view of a patient with scaphocephaly resulting

from sagittal synostosis. Note the frontal bossing, elongation along the
anteroposterior (AP) axis, and prominent occiput, all of which are characteristic of this condition.
Common Patterns of Single

Suture Craniosynostosis
------------------------------------------------------------------------------------------------------------------------------------------------
The most common form of craniosynostosis is sagittal synostosis, with an incidence of approximately 2 per 10,000 live
births. In concordance with Virchows law, premature fusion
of the sagittal suture leads to compensatory growth in the anteroposterior dimension, resulting in scaphocephaly (Fig. 53-1).
Unilateral coronal synostosis is less common, with an incidence of approximately 0.9 per 10,000 live births. The growth
pattern in unicoronal synostosis is more complex, albeit leading to a stereotypical calvarial phenotype. Ipsilateral to the
fused coronal suture, the supraorbital rim is flattened and
recessed, and the forehead is flattened. As calvarial growth occurs, the contralateral forehead becomes bossed and the nasal
bridge starts to twist, producing a C-shaped facial deformity
(Fig. 53-2).
Premature fusion of the metopic suture results in constriction of growth in an axial plane centered on the caudal forehead.
A palpable bony ridge is present in the midline of the forehead
(described as a keel-shaped forehead or trigonocephaly). The
forehead is narrow and pointed. The medial orbital rims are
consequently closer to the midline, giving the appearance of
hypotelorism. Bilateral lateral brow recession and temporal
hollowing also occurs and exaggerates this appearance.
The least common form of single suture craniosynostosis
affects the lambdoid suture. Common findings are posterior
FIGURE 53-2 A child with left unicoronal synostosis. Characteristic findings include ipsilateral fronto-orbital retrusion, prominent contralateral
forehead bossing, and nasal root deviation toward the affected side.
The C-shaped facial deformity is notable here.
plagiocephaly, ipsilateral mastoid bossing, and both anterior

and inferior ipsilateral ear displacement. A posterior skull base
cant and contralateral forehead bossing occur in more advanced cases.
Multiple suture craniosynostoses are rare, and may present
with a variety of skull-shaped deformities.
Syndromic Craniosynostosis
------------------------------------------------------------------------------------------------------------------------------------------------
Children born with craniofacial dysostosis syndromes require

multiple staged procedures to correct their bony deformities
during their childhood years and extending into early adulthood. These patients have a higher relapse rate because
CHAPTER 53
multiple affected midface and skull base sutures are usually

present. The syndromic craniosynostoses carry a greater risk
of intracranial hypertension, as well.
Many of the craniofacial dysostosis syndromes are inherited in an autosomal dominant pattern and are the result of
mutations in the FGF-receptor genes, which alter the signaling
pathway. Although penetrance is complete in most syndromes, the expression of the mutation is highly variable.
Examples of these include Crouzon, Apert, and Pfeiffer
syndromes, the three most common craniofacial dysostosis
syndromes, all of which have mutations in the FGF-receptor
genes.
Crouzon syndrome is the most common syndromic craniofacial dysostosis, with an estimated incidence of 1 in 25,000
live births.8 Crouzon syndrome is caused by a mutation in
the FGFR2 gene, causing increased receptor activation.9 The
syndrome is characterized by the triad of bicoronal craniosynostosis, exorbitism, and midface retrusion (Fig. 53-3).
Exorbitism is the feature of this syndrome, which results in
prominent globes. As such, ocular protection is a key feature
the clinician must keep in mind when assessing these children. Exposure keratitis and conjunctivitis may prompt earlier
surgical intervention to avoid permanent visual impairment.
Apert syndrome has an incidence of 1 per 160,000 live
births and is also caused by an FGFR2 mutation.10 Unlike
Crouzon syndrome, the genetic defect in Apert syndrome is
a missense mutation, and the vast majority of Apert patients
are the result of a sporadic mutation. The triad of bicoronal
craniosynostosis, midface hypoplasia, and syndactyly of the
hands and feet characterize Apert syndrome. The craniofacial
findings include bicoronal suture fusion, a large anterior fontanelle, midface retrusion, and varying degrees of exorbitism.
FIGURE 53-3 Brachycephaly, exorbitism, and midface retrusion, which

are common findings in both Apert and Crouzon syndromes.
CRANIOFACIAL ANOMALIES
693
Neonatal respiratory distress from a narrowed nasal vault,

sometimes severe enough to warrant tracheostomy, may occasionally occur.11 A specialist in congenital hand anomalies
typically addresses complex syndactyly of the hands and feet.
Unlike Crouzon syndrome, mental retardation is common,
and affects nearly 50% of children with Apert syndrome.
Other craniofacial dysostosis syndromes include Pfeiffer
syndrome, Saethre-Chotzen syndrome, and Carpenter syndrome. These are much less common than either Crouzon
or Apert syndrome and share varying degrees of craniosynostosis, midface retrusion, digital anomalies, and mental retardation. Management priorities in these children are similar
to that of children with either of the previously mentioned
syndromes.
DIAGNOSIS
A thorough history and physical examination is the cornerstone of diagnosing synostosis of the calvarial sutures. Single
suture fusion results in the characteristic patterns of calvarial
morphology described previously and are readily diagnosed
by physical examination.
Measurement of head circumference with plotting on a
standard growth curve gives an indication of head growth relative to the childs body. Physical examination should accurately define asymmetries in the infant skull. Head shape
should be assessed from anterior, posterior, and top-down
views to identify areas of relative bossing or recession. Ear position in both anterior-posterior and craniocaudal planes
should be examined. Lateral examination of the forehead
and face will identify forehead bossing and the position of
the midface and orbits. Facial examination is especially important in identifying children with the craniofacial
syndromes that have midface retrusion as part of their phenotype. Stigmata of intracranial hypertension should be investigated. For infants, these include a history of irritability,
burrowing behavior, or repetitive head slapping.
Since the advent in the early 1990s of the American
Academy of Pediatrics Back to Sleep campaign, designed
to decrease the incidence of sudden infant death syndrome
(SIDS), a sharp increase in the incidence of positional plagiocephaly has occurred. Positional plagiocephaly is the deformation of the calvarium despite the presence of widely patent
cranial sutures. This condition may be difficult to differentiate
from unilateral coronal or lambdoid craniosynostosis, which
both result in forms of plagiocephaly. Usually, an experienced
craniofacial surgeon can make the correct diagnosis based on
physical examination findings alone. However, sometimes
computed tomography (CT) imaging is needed. The correct
diagnosis is critical, however, because children with positional
plagiocephaly are treated with changes in sleeping position or,
in more severe cases, a custom orthotic molding helmet.
Children who are suspected of having craniosynostosis
should be referred to a craniofacial team for evaluation. This
evaluation includes detailed cranial measurements and a thorough physical examination. CT scanning is rarely needed for
diagnosis, but is obtained by many craniofacial surgeons prior
to calvarial vault remodeling to assess for intracranial abnormalities. Given the additional cost, risks of sedation, and potential harmful effects of ionizing radiation on the growing
child, radiologic imaging should be undertaken on a caseby-case basis as determined by the craniofacial surgeon.12
694
PART V
HEAD AND NECK
In addition to findings of suture fusion, stigmata of intracranial hypertension, such as a moth-eaten appearance of the
calvarium on CT images or a copper-beaten appearance
on plain radiographs will be seen. Papilledema may be seen
on fundoscopic examination and is an indication for urgent
cranial vault expansion.
TREATMENT
The mainstay of treatment for premature fusion of cranial sutures is surgical cranial vault remodeling. The two goals of surgery are to release the involved suture through resection and
reconstruction of the cranial vault to a more normal shape.
Surgery is a combined procedure between a craniofacial plastic surgeon and a pediatric neurosurgeon. In general, the pediatric neurosurgeon performs the initial craniotomy, and the
craniofacial surgeon performs the bony reshaping. However,
the procedure is mainly bone surgery and not brain surgery. Cranial vault remodeling is accomplished by removing
the abnormally shaped calvarial bones and recontouring
them. The plasticity of the infant calvarium is utilized as the
bones are bent and shaped to a more anatomic contour. Barrel
stave osteotomies are commonly performed to expand the cranial vault. Wedge osteotomies are done to reduce the vault in
areas of bony excess or bossing. Except in cases of isolated sagittal or lambdoid synostosis, the orbits are deformed, which
necessitates advancement of the supraorbital rim in addition
to reshaping of the forehead and anterior cranial vault. Bones
are fixed into their new position with resorbable hardware
consisting of polyglactic/polyglycolic acid plates and screws.
This type of hardware undergoes degradation over the course
of a year, and is not prone to intracranial migration, which can
occur with traditional titanium hardware. After initial reconstruction in infants, bony defects remain after surgery. The
unique osteogenic potential of the dura and overlying periosteum in infants results in primary bone formation and complete healing of these large bone defects. Bone defects that
remain after 2 years of age typically require secondary bone
grafting. Endoscopic techniques have been described for correction of both sagittal and coronal synostosis.13,14 Although
resection of the involved suture is performed, these techniques rely on a long period of postoperative molding helmet
therapy to achieve final head shape. Despite their reported
benefit of reduced blood loss and transfusion requirements,
endoscopic techniques have not gained wide acceptance, primarily due to their poor head shape outcomes.
Timing of surgery is somewhat controversial among craniofacial surgeons. Some surgeons advocate early correction at
3 to 6 months of age, believing that the rapidly growing brain
will assist in remodeling the skull if the fused suture is released
and the calvarium reshaped. This theoretically reduces the
amount of correction that needs to be performed in the
operating room. Delaying surgery until 9 to 12 months of
age allows the infant skull growth to begin to plateau prior
to surgery. Although this reduces the amount of intrinsic bone
shape normalization resulting from brain growth, the thicker
calvarial bone may provide a more stable skeletal correction
with less relapse. In reality, a large window exists when the
surgery can be performed with acceptable risks and outcomes.
Children who present late for corrective surgery present
unique challenges to the craniofacial surgeon. Bone in these
children is typically thicker and more difficult to contour with
bone-bending forceps and simple barrel stave osteotomies.

The asymmetric skull base is more developed and resistant
to normalization, which increases the chances of a permanent
deformity. Also, the diploic space between the inner and outer
table has begun to form, subjecting the child to increased
blood loss during surgical exposure and bony resection.
The entire reconstructive procedure is more extensive, as even
subcentimeter bone defects must be grafted in order to heal.
Extracranial bone grafts (rib and/or iliac) are needed when
the diploic space has not yet formed (before 5 years of age).
In children with syndromic craniosynostosis and midface
retrusion, further reconstruction is frequently needed. A midface advancement consisting of either a Le Fort III or a monobloc osteotomy is performed between the ages of 5 and
8 years, unless the need for ocular protection forces earlier correction. The choice between procedures depends on forehead
projectiona subcranial Le Fort III is performed if forehead
projection is adequate, whereas a monobloc frontofacial advancement is needed if forehead retrusion has occurred. At
the age of skeletal maturity (typically between the ages of
14 and 16), a Le Fort I maxillary osteotomy, with or without
mandibular osteotomies, is usually needed to formally correct
any malocclusion that commonly exists in these patients.
Orthognathic Surgery
------------------------------------------------------------------------------------------------------------------------------------------------
The term orthognathic is derived from the Greek orthos,

meaning to straighten, and gnathos meaning jaw. Orthognathic surgery is a discipline that crosses the boundaries of
many different specialties, principally oral and maxillofacial
surgery and craniofacial surgery (a distinct subspecialty of
plastic surgery). Orthognathic surgery repositions the dentofacial skeleton to correct either congenital or acquired malocclusions and restore a harmonious balance to the underlying
bony facial form. These movements can involve both the maxilla and the mandible, either in segments or in conjunction
with varying portions of the craniofacial skeleton. Certain procedures mandate the assistance of a pediatric neurosurgeon,
because they require intracranial osteotomies to fully mobilize
the bony segments for repositioning. An orthodontist who is
familiar with the necessary dental movements is also needed to
prepare a patient for surgery. The orthodontist also finetunes the dental occlusal relationships postoperatively, which
is imperative for a successful outcome.
Although secondary trauma can lead to orthognathic surgery, the field predominantly addresses congenital and developmental deformities leading to malocclusion that cannot be
corrected orthodontically. Nearly 25% of all patients who have
undergone correction of a cleft lip and palate will develop severe maxillary hypoplasia that warrants surgical correction.
Whether this is due to an intrinsic growth disturbance or restriction of growth of the midface due to postcleft surgical
scarring is a hotly debated topic among craniofacial surgeons.
The age at which orthognathic procedures are performed is
an important consideration in planning subsequent surgeries.
Any skeletal facial advancement procedure performed prior to
the age of bony maturity carries a significant rate of relapse as
the remainder of the face grows; merely advancing the hypoplastic segments will not unlock their growth potential. For
this reason, definitive orthognathic procedures are carried out
CHAPTER 53
in mid- to late adolescence, once the majority of facial skeletal

growth is complete. Thus orthognathic surgery is usually done
between 14 and 16 years for girls and 15 and 17 years for boys.
Epiphyseal closure on anteroposterior hand radiographs is a
good indicator of overall skeletal maturity.
Preoperative evaluation of patients with congenital dentofacial deformities usually occurs in the setting of a cleft/
craniofacial team visit. An orthodontist is present for discussion of treatment options and will be intimately involved in
both the preoperative and postoperative care of these patients.
The patients occlusion is determined according to the Angle
classification, with class 1 being a normal occlusal relationship, class 2 as the typical overbite, and class 3 being an
underbite. The facial profile and aesthetics are analyzed with
special attention paid to the soft tissue bony landmarks. A lateral cephalometric radiograph, which is a standardized view
with the head aligned in the neutral position, is used for treatment planning. The positions of the maxilla and mandible are
determined with respect to the cranial base. Dental models
are cast and mounted on an articulator in their anatomic relationship. Model surgery is performed wherein the models are
precisely moved to place the teeth in proper planned postoperative occlusion. An acrylic splint is made from the models
after each jaw is repositioned, which will be wired into place
in the operating room to assure the bony segments are in
proper position. This splint can be a final interdental splint
for wear postoperatively.
In 1901, Rene Le Fort, a French surgeon and anatomist, carried out a series of rather grotesque experiments and defined a
classification system for facial fractures that bears his name.15,16
These fault lines of the facial bony skeleton were then subsequently adapted for use in elective orthognathic surgery and
have become the mainstay of therapy to correct midfacial skeletal anomalies. The most common is the Le Fort I osteotomy,
which repositions the tooth-bearing segment of the maxilla in
either the anteroposterior or craniocaudal planes, or both. This
osteotomy can be modified to divide the maxilla into two or
three tooth-bearing segments, which can be moved independently to provide a functional, Angle class 1 occlusion and correct both sagittal and transverse maxillary deficiency. The Le
Fort I osteotomy is used to correct cleft lip and palate-related
maxillary hypoplasia. It is also used to correct vertical maxillary
excess or rotational abnormalities of the maxilla.
Another commonly used corrective procedure in craniofacial surgery is the bilateral sagittal split osteotomy (BSSO). In
this technically demanding procedure, the mandible is split
in the sagittal plane from midramus to proximal body, sparing
the inferior alveolar nerve on both sides. The BSSO movements can be adjusted to correct mandibular asymmetry, level
the occlusal plane, or correct overall facial disharmony caused
by overgrowth or undergrowth of the mandible. The BSSO is
often combined with the Le Fort I osteotomy to correct larger
mandibularmaxillary discrepancies. Genioplasty refers to reduction, advancement, or lateral movements of the bony chin
point. It is generally considered a purely aesthetic procedure
that can be added to enhance the facial skeletal balance.
Although not usually considered an orthognathic procedure, but rather a craniofacial procedure, the Le Fort III
osteotomy is performed to reposition the midface (maxilla
and zygomas) relative to the cranial base. In this procedure,
osteotomies are created across the ascending portion of the zygoma, the zygomatic arch, orbital floor and both medial and
695
lateral walls, frontonasal junction, and pterygomaxillary junction. The nasal septum is divided in the coronal plane. This
allows for complete separation of the facial skeleton from
the cranial base and subsequent advancement of the Le Fort
III segment to correct midface hypoplasia. The Le Fort III advancement is used to correct severe malocclusion or upper airway obstruction from overall midface growth restriction. The
majority of patients who undergo a Le Fort III osteotomy have
a named syndrome (Crouzon, Apert, and Pfeiffer being the
most common), and have their first midface advancement at
age 5 or 6 years. Because the growth of the midface is not complete until mid- to late adolescence, a second advancement is
frequently necessary. When midface retrusion is accompanied
by growth restriction of the supraorbital rim and forehead, a
monobloc frontofacial advancement may be necessary. This is an
intracranial and extracranial procedure and requires both a
pediatric neurosurgeon and a craniofacial surgeon for intracranial access. A monobloc advancement is similar to a Le Fort
III osteotomy, except that the supraorbital rim and forehead
are advanced along with the midface. This procedure has
fallen out of favor because of infectious complications arising
from difficulty obtaining adequate separation of the intracranial space and nasal cavities.17 Thus the monobloc is now
usually performed as a staged procedure (fronto-orbital advancement, followed by a Le Fort III osteotomy a few months
later) or is performed with the use of distraction osteogenesis
to allow the soft tissues to grow along with bony skeletal
expansion, which minimizes the risk of intracranial infectious
complications.
Distraction osteogenesis was pioneered by Ilizarov (a Russian
orthopedist) in 1958, and adapted for use in the mandible by
McCarthy in 1992.18,19 It is now commonly used for a number
of applications in craniofacial surgery. The principle behind
distraction osteogenesis is that gradual expansion of the bony
skeletal gap left by a surgically placed osteotomy will allow
lengthening of the bone by gradual osteoblastic activity and
ingrowth of new bone. This results in lengthening of the skeleton in the direction of the vector of expansion. Distraction is
useful in situations that would require extensive bone grafting
to obtain adequate bone length, or in cases in which opposing
soft tissue forces would result in skeletal relapse if the bone
was rapidly advanced and grafted. Clinically, maxillary distraction is most commonly applied when a large discrepancy
exists between the maxilla and mandible as a result of a complete cleft lip and palate and resultant maxillary hypoplasia.
Scarring from the numerous previous procedures makes maxillary advancement alone prone to relapse; by applying the
principle of distraction, the bone and soft tissue are gradually
expanded, and the cleft maxilla can be advanced into a normal
occlusal relationship (Angle class 1) with minimal chances of
relapse resulting from soft tissue resistance. Distraction is also
useful to correct severe mandibular hypoplasia accompanying
conditions such as hemifacial microsomia or Pierre Robin sequence, which will be discussed in subsequent sections of this
chapter.
Craniofacial Clefts
------------------------------------------------------------------------------------------------------------------------------------------------
Dr. Paul Tessier, the father of the field of craniofacial surgery,

also developed a classification system for craniofacial clefts
that is arguably the most widely used of those available today.
696
PART V
HEAD AND NECK
14 131211
1 23
10
4 9
8
0 1
2
3
30
14 13 12
0 1 2 11 10
3 4
9
8
FIGURE 53-5 A child with a typical Tessier number 7 soft tissue cleft.
56
6
0
12 3
30
FIGURE 53-4 Tessiers pre-computed tomography classification system
of rare craniofacial clefts. The lower image represents the cleft location of
the bony skeleton, while the upper illustrates the cutaneous manifestations of the various bony clefts.
The Tessier system is based on specific anatomic derangements that fall along embryonic lines of fusion within the face
(although these were not known at the time they were initially
described) (Fig. 53-4). Tessiers classification system is notated
0 to 14, with clefts 0 to 7 describing facial clefts and clefts 8 to
14 describing cranial vault clefts. Each cleft has unique soft
tissue and bone lines of clefting. Also, the facial and cranial
clefts coincide such that the sum of the two components is
14 (i.e., 0 and 14 clefts coincide, as do 3/11 and 5/9, etc.).
A thorough search along the meridian of the cleft will usually
elucidate subtle (or not-so-subtle) findings.
The majority of craniofacial clefts are rare and will be seen
infrequently during an individual surgeons career. As such,
the remainder of this section of the chapter will be spent dealing with those more common Tessier clefts.
10% of affected children will have symmetric, bilateral

involvement.21
The derivatives of the first and second branchial arch are
abnormal, albeit to varying degrees. Macrostomia is a common
finding, with the commissure of the lip being displaced laterally toward the affected side, resulting in an enlarged oral
opening (Fig. 53-5). Preauricular ear tags, embryologic remnants of the developing ear, are present anterior to the external
auditory canal and contain small stalks of cartilage remnants
(Fig. 53-6). The external ear can likewise be affected and can
range from mild hypoplasia to near complete absence. A conductive hearing loss on the ipsilateral side is commonplace.
The mandible is commonly affected, and the defect of the ascending ramus and condyle may range from mild hypoplasia
to complete absence (Fig. 53-7). The facial nerve may be affected to varying degrees, as well, which in turn contributes
CLEFT NUMBER 7
A number 7 cleft can be protean in physical manifestation and
is known by a number of different names. Hemifacial microsomia, oculoauriculovertebral syndrome (OAV), first and second branchial arch syndrome, and otomandibular dysostosis
syndrome all refer to the physical findings associated with a
number 7 cleft. The number 7 cleft is thought to have an incidence of between 1 in 3000 and 1 in 564220 live births.
A number 7 cleft is usually unilateral, but approximately
FIGURE 53-6 Multiple accessory external ear tags as seen in Tessier 7

clefts.
CHAPTER 53
697
of a free costochondral rib graft, which is performed around 7

years of age. Ear reconstruction with costal cartilage is performed
at the same age, when necessary.
TREACHER COLLINS SYNDROME
FIGURE 53-7 Skeletal findings in hemifacial microsomia. Note the

absence of the ascending ramus and condyle of the mandible.
to weakness and underdevelopment of the muscles of facial

expression that they supply.
Given that the manifestations of the number 7 cleft vary, so
do the treatment options. In general, more severely affected
children require earlier and more significant reconstruction.
Cutaneous manifestationsmacrostomia and branchial arch
remnantsare usually treated with excision and local flap reconstruction during infancy. Underlying bony anomalies are
treated based on severity. Children with mild hypoplasia of
the ramus and condyle are followed throughout growth, and
will usually develop an occlusal cant (Fig. 53-8) as they age.
These children can be treated with orthognathic surgery after cessation of skeletal growth to reposition the hypoplastic mandible
and maxilla with the combination of a Le Fort 1 osteotomy, bilateral sagittal split osteotomies, and an osseous genioplasty to align
the chin point. Those with more severely hypoplastic rami
will undergo distraction osteogenesis of the ramus during childhood, but will usually need formal orthognathic correction during late adolescence.22 Children who are born with complete
absence of an ascending ramus need an extensive reconstruction
to achieve normal function of their mandible and a normal occlusal relationship. This is typically accomplished through the use
FIGURE 53-8 A child with hemifacial microsomia and an occlusal cant.

Note the upward slant of the mandible caused by right-sided ramus
hypoplasia.
First described in 1847,23 Treacher Collins syndrome (or

mandibulofacial dysostosis) is a relatively common syndrome
made up of Tessier clefts 6, 7, and 8. Its incidence is estimated
at 1 in 10,000 live births. Treacher Collins is caused by a mutation on chromosome 5q31.3-33.3 (the TCOF1 gene) and is
an autosomal dominant disorder with variable penetrance.
Treacher Collins patients manifest bilateral anomalies of the
eyelids, zygomas, maxilla, mandible, and ears (Fig. 53-9).
There is typically an antimongoloid slant to the palpebral fissures, with lower lid notching present. Eyelashes are often absent on the medial two thirds of the lower eyelid. The zygomas
are severely hypoplastic or absent. The maxilla is hypoplastic
with a shortened vertical height, which can impede or block
nasal airflow. Shortening of both the length and height of
the mandible results in narrowing of the posterior pharyngeal
airway and can contribute to upper airway obstruction in these
children, which in the past necessitated placement of a tracheostomy. Within the past 20 years, however, distraction osteogenesis has been applied to these patients to relieve early upper
airway obstruction and avoid the need for long-term tracheostomy dependence. Severe microtia is often accompanied by
atresia of the middle and inner ear; placement of boneanchored hearing aids (BAHA) is commonly necessary to allow
for improved hearing and speech development. As is the case
with either isolated or syndromic microtia, reconstruction of
the external ears is usually undertaken at age 7 or 8 with autologous costal cartilage grafts. At around this same time, onlay
bone grafting of the maxilla and zygoma is performed to add
projection and give a more normal midfacial profile. Correction of the eyelid notching with local tissue transfers is also performed at this age should it not be needed earlier because
of corneal protection issues. Definitive orthognathic surgical
correction of the associated deformities of the occlusal plane
is performed in late adolescence.
FIGURE 53-9 Typical findings in a patient with Treacher Collins syndrome.

Down-slanting palpebral fissures, malar hypoplasia, and microtia are common hallmarks of this syndrome. This child lacks true colobomas of the lower
eyelid but has loss of lower eyelid support and excess scleral show.
698
PART V
HEAD AND NECK
Acknowledgments
SELECTED READINGS
Special thanks to Henry K. Kawamoto, Jr., DDS, MD, for Figures 53-4 and
53-7.
Bentz ML, Bauer BS, Zucker RM, eds. Principles and Practice of Pediatric Plastic Surgery. St Louis: Quality Medical Publishing; 2007.
Mathes SJ, ed. Plastic Surgery. Pediatric Plastic Surgery. Vol 4. Philadelphia:
Saunders; 2005.
Posnick JC. Craniofacial and Maxillofacial Surgery in Children and Young
Adults. Philadelphia: Saunders; 2000.
Thaller SR, Bradley JP, Garri JI, eds. Craniofacial Surgery. New York: Informa
Healthcare; 2008.

expertconsult.com.
workup and counseling is mandatory, as is heightened suspicion for other physical and physiologic anomalies.
Etiology
------------------------------------------------------------------------------------------------------------------------------------------------
CHAPTER 54
Understanding and
Caring for Children
with Cleft Lip
and Palate
James Y. Liau, John A. van Aalst, and
A. Michael Sadove
Epidemiology
------------------------------------------------------------------------------------------------------------------------------------------------
The incidence of orofacial clefting varies among racial backgrounds. Because of the close association between cleft lip
and palate, the presence of cleft lip is often described as being
with or without cleft palate. Worldwide prevalence of cleft lip
and palate is 1 per 700 live births.1 People of African descent
have the lowest incidence of cleft lip with/without cleft palate
at 0.5 per 1000 live births, followed by whites (1 per 1000 live
births), and Asians (1.3 per 1000 per live births). Overall, an
isolated cleft lip makes up approximately 21% of all patients
with cleft lip and palate. Unilateral clefts are roughly 9 times
more prevalent than bilateral cleft lips, and males are more
affected than females. The U.S. incidence of cleft palate alone
ranges from 0.3 to 0.5 per 1000 live births.2 A child with a
cleft lip with or without cleft palate has an approximately
30% chance of having an associated syndrome; interestingly,
a child with an isolated cleft palate has a 50% incidence of
an associated syndrome.2,3 Because of this association, genetic
Genetic and environmental factors have both been associated

with clefting. If other family members are affected by a cleft,
offspring have an increased chance of being affected. For
example, if a family already has a child with a cleft, or one parent has a cleft, the chance that the next child will have a cleft is
4%; with two affected children, the chance that a third child
will have a cleft increases to 9%. The probability increases
to 17% in a family if both a child and parent have a cleft.4 This
increase in frequency is seen in spontaneous clefting not associated with syndromes. There are some syndromes in which
clefting can be passed down in an autosomal dominant fashion, such as van der Woude syndrome, where presence of the
cleft palate is an autosomal dominant trait.
Environmental causes of clefting include the use of anticonvulsants, including phenytoin, which is associated with
10-fold increase in clefting; maternal smoking increases the
incidence twofold. Other environmental influences, such as
alcohol use, the use of retinoic acid, and dietary causes,
including zinc and folate deficiencies, can cause syndromes
with clefting; however, these are not directly linked to isolated
cleft lip/palate. People with certain genotypes are more susceptible to certain environmental exposures; hence, women with
less efficient methyl tetrahydrofolate reductase enzymes are
more prone to clefting in the face of folic acid deficiency.512
Embryology
------------------------------------------------------------------------------------------------------------------------------------------------
Basic understanding of midface and palatal embryologic development helps elucidate the pathoanatomy of cleft lip with or
without palate. Orofacial clefting occurs when there is failure
of fusion of maxillofacial structures migrating from lateral to
medial during the initial 4 to 10 weeks of embryonic development. A key anteriorposterior embryologic and anatomic landmark in understanding clefts is the incisive foramen. The
structures that form anterior to the foramen ultimately develop
into the nose, lip, and alveolus; embryologically, these structures
form first and are designated as the primary palate. The structures
that form posterior to the foramen become the hard palate and
soft palate, and are referred to as the secondary palate, because
they fuse secondarily. Clefting of the lip (primary palate) occurs
when the nasomedial and nasolateral prominences of the frontonasal prominence do not meet with the maxillary prominence
(Fig. 54-1, A). Clefting of the palate (secondary palate) occurs
when the lateral palatine shelves do not elevate and fuse at the
midline to each other or to the primary palate (Fig. 54-1, B).
Because the development of this craniofacial area is complex, deformities can occur at multiple points along the embryologic time
line, resulting in a full spectrum or combination of anomalies.
Anatomy
------------------------------------------------------------------------------------------------------------------------------------------------
UNILATERAL CLEFT LIP

A cleft lip affects the anatomy of the lip, philtrum, nose, as
well as the alveolus, depending on the severity of the defect.
Microform clefts are the mildest form of clefting, involving
699
700
PART V
HEAD AND NECK
Primary palate
Nasolateral
Nasomedial
Incisor foramen
Secondary plate
Maxillary prominence
B Lateral palatine shelves
FIGURE 54-1 A, The nasolateral and nasomedial prominence fuse to make the lip and nose. Lack of fusion between the maxillary nasal prominence with
the nasomedial prominence will yield a cleft lip. B, The lateral palatine shelves fuse in the midline, thus forming the secondary palate. The primary palate
fuses posteriorly with the palatine shelves to form a complete palate.
the vermillion and white roll, and can be quite subtle

(Fig. 54-2, A). Incomplete clefts are associated with absence
of skin, mucosa, and orbicularis muscle but have a retained
webbing of skin across the floor of the nasal aperture, referred
to as a Simonart band (Fig. 54-2, B). A complete unilateral
cleft involves the alveolus, and can be associated with a cleft
of the palate (Fig. 54-2, C).
The most obvious deformity involving a cleft lip is discontinuity of the lip itself. However further analysis of the defect
demonstrates deviation of the nasal septum and columella, as
well as widening or flattening of the cleft-side nasal cartilage.

Other defects include the lack of nasal floor and discontinuity
of the lip, muscle, and the alveolar ridge. Successful surgical
correction of the lip must address all of these structures.
BILATERAL CLEFT LIP

Bilateral clefts have a two-sided discontinuity of the lip, with
a central portion, the premaxilla, which is discontinuous
from either of the lateral segments. The premaxilla contains
C
FIGURE 54-2 A, Microform cleft can be seen with mild notching of the lips vermilion along with the minimal distortion of the nose. B, Incomplete cleft
has some webbing across the cleft with some retention of the lateral nose; however, the skin across the cleft is devoid of orbicularis oris muscle and is
functionless. This skin is also called a Simonart band. C, Complete cleft lip has lateralization of the lip and lateral nasal element. Notice the deviation of the
nasal columella and philtrum away from the cleft, and the flattened nose on the cleft side. This picture also demonstrates the clefting of the palate, which
allows an unobstructed view into the nasal airway.
CHAPTER 54
UNDERSTANDING AND CARING FOR CHILDREN WITH CLEFT LIP AND PALATE
701
CLEFT PALATE
B
FIGURE 54-3 A, Bilateral cleft lip with a Simonart band transversing both
clefts. B, Complete bilateral cleft lip. The midline protuberance is called
the premaxilla and is much more protruding than an incomplete bilateral
cleft; tethering of the Simonart bands help in keeping the premaxilla in a
more anatomic position. Absence, or shortening of the columella, widening of the alar bases, and anterior projection of the premaxilla are all trademarks of bilateral cleft lips.
elements of skin, mucosa and bone, which can be asymmetrically deviated to one side or the other, and can be anteriorly positioned, depending on the severity of the
deformity. Bilateral clefts can be incomplete with Simonart
bands (Fig. 54-3, A), or complete with defects that proceed
through the alveolar ridges (Fig. 54-3, B). A central feature
of the bilateral cleft lip deformity is depression of the nasal
tip, a shortened columella, and widely splayed alae.
Clefts of the palate can exist with clefts of the lip or may be
present alone. Anatomically, the hard palate begins immediately posterior to the incisive foramen, with embryologic
fusion of the palate from anterior to posterior. Hence, an isolated cleft of the soft palate may exist; however, an isolated
cleft of the hard palate cannot. A complete cleft of the secondary palate includes both the hard and soft palate, extending
anteriorly from the incisive foramen to the uvula, and this
can be bilateral as well (Fig. 54-4). The primary function of
the palate is to separate the oral cavity from the nasal cavity.
This function is lost in the presence of a cleft. The function
of the soft palate is primarily speech related and dependent
on five paired muscles, the two most important of which
are the levator veli palatini and the tensor veli palatini. Ordinarily, these muscles form a transverse sling enabling the palate to rise and move posteriorly to close the oropharynx from
the nasopharynx. In a cleft palate, these muscles abnormally
insert onto the posterior shelf of the hard palate, and as a consequence, the palate is deficient in its ability to seal off the oropharynx from the nasopharynx.
A submucous cleft is the most minor expression of the
clefting spectrum. The soft palate mucosa is actually intact,
but split posteriorly, resulting in a bifid uvula; there is a midline lucency in the soft palate, referred to as a zona pellucidum, which is a muscle diastasis, and a notch at the
midline, posterior edge of the hard palate.13 As in a full cleft,
the levator and tensor veli palatine abnormally insert onto the
posterior hard palate, preventing the soft palate from moving
appropriately during speech, potentially leading to nasal
sounding speech. The incidence of submucous clefts is
roughly 1 in 1,200 to 2,000; however, this is likely an underestimation, because many patients may not seek treatment or
even know of their submucous cleft unless there is functional
speech deficit.14
Treatment Protocols
------------------------------------------------------------------------------------------------------------------------------------------------
The timing for cleft lip repair in the United States is generally
between 3 to 6 months of age. Depending on the severity of the
deformity, various forms of presurgical orthopedics can be
used to prepare the child for lip surgery. In general, the goals
of these techniques are to improve the alignment of the alveolar segments, decrease the size of the soft tissue cleft, and to
FIGURE 54-4 The figure on the left depicts a cleft of the secondary palate only; there is an intact hard palate. The middle figure depicts a complete
unilateral cleft. The figure on the right depicts a complete bilateral cleft.
702
PART V
HEAD AND NECK
FIGURE 54-6
lip repair).
B
FIGURE 54-5 A, A unilateral cleft lip prior to nasoalveolar molding (NAM)
has a wide alveolar cleft, slumping of the nasal cartilage on the cleft side,
and lateralization of the alar base of the cleft side. B, Nasoalveolar molding
assists in realigning the alveolar segments, reshaping the slumping nasal
cartilage of the cleft side, and medializing the alar base of the cleft side.
improve the symmetry of the nose. The simplest form is a taping regimen (literally from cheek to cheek) that helps to pull
the two sides of the clefts together, with the goal of narrowing
the cleft and realigning the tissues in an anterior-posterior
dimension. Some centers also use a technique termed nasoalveolar molding (NAM) to address the three major components
of the cleft deformity (Fig. 54-5). NAM addresses the slumping of the nasal alar cartilage, helps realign the alveolar ridges,
and brings the soft tissue of the lips into closer proximity.15,16
CLEFT LIP SURGICAL REPAIR

Unilateral
There are multiple surgical techniques for cleft lip repair.
A recent survey of U.S. cleft surgeons found three predominant surgical techniques for unilateral cleft lip repair: the
Millard rotation-advancement technique (46%), the Millard
rotation-advancement technique with modifications (38%),
and triangular flap techniques (9%).17 Techniques for bilateral
cleft lip repair include variations of techniques introduced by
Millard and by Mulliken.18
In a unilateral cleft lip, there is absence of central tissue of
the lip and philtrum, as well as of the nasal columella, depending on the severity of the cleft. The overall goals of surgery are
restoration of lip continuity, which starts with functional orbicularis oris muscle reapproximation, establishing symmetry
of the lip (especially at the central cupids bow) and nose,
with aesthetic placement of scars in anatomic subunits. The
rotation-advancement repair, as described by Dr. Ralph Millard
Markings for the rotation-advancement lip repair (Millard
in 1976, addresses all of these goals (Fig. 54-6). Because there is

a paucity of tissue medial to the cleft, the downward rotation of
the remaining philtrum helps to provide adequate tissue that
matches the contralateral, noncleft side. Advancement of the lateral tissue reconstructs the affected philtrum, thus providing
lip continuity. Medialization of the base of the nose, as well
as further soft tissue dissection of the nose, results in a symmetric reconstruction. Reapproximation of the orbicularis oris
muscle provides oral competence. Further soft tissue arrangement in the nasal floor allows final closure of the lip and nose.
Surgical details are found in several references.1921 A potential
shortcoming of the rotation-advancement technique is the inability to provide adequate philtral length despite aggressive
downward rotation of this tissue. This may result in the high
point of the cupids bow on the cleft side being located in a
position higher than on the noncleft side.
Some of the more commonly used modifications of the
rotation-advancement technique are the Mohler repair, the
Noordoff vermilion flap, and the triangular advancement flap,
although the details of these techniques are beyond the scope
of this chapter (see the referenced articles for more complete
descriptions).2224
Another technique for unilateral cleft lip repair is the triangular flap technique (also known as the Tennison repair),
introduced by Charles Tennison in 1952. Tennison approached
the lack of central soft tissue of the cleft in a very different fashion than Millard. In this technique lip length is achieved by
designing a triangular flap of the lateral, cleft side, which then
inserts into a cut of the medial, noncleft side, thus providing the
extra tissue for appropriate lip height. The muscle is repaired,
as in the rotation-advancement technique, followed by reconstruction of the floor of the nose (full surgical details are found
in the references).25,26 Shortcomings of the Tennison repair
include placement of the scar in a nonanatomic location, thus
drawing attention to the repair, as well as an overly long lip,
depending on the size of the triangular flaps.
CHAPTER 54
Bilateral Cleft Lip Repair

Bilateral cleft lip repairs are especially challenging because of a
central lack of soft tissue, and the anterior displacement of the
premaxilla, which functionally increases the transverse width
of cleft defect. Many surgeons use presurgical orthopedics to
decrease premaxillary protrusion, thus increasing the columellar length and nasal tip projection. This technique also
decreases the distance of the cleft, potentially making surgical
repair easier for the surgeon (Fig. 54-7). The primary goals of
surgery include lip and nasal symmetry, which is achieved
through the creation of a philtral column, including the
703
cupids bow, reapproximation of the orbicularis oris, repositioning the nasal alar cartilages, lengthening the columella,
and closure of the nasal floor (Fig. 54-8). The absence of
the philtral column and cupids bow is especially problematic
since these structures are difficult to replicate in a repair. Postoperatively, these imperfections can be quite noticeable at
conversational distances.27 Realistically, patients with bilateral
cleft lips will ultimately require revisional surgeries to correct
the secondary stigmata of the repair, which include a shortened columella, blunted nasal tip, widened nasal ala, and a
widened philtrum.
CLEFT PALATE SURGICAL REPAIR
FIGURE 54-7 Nasoalveolar mold (presurgical orthopedics) for bilateral

cleft lip, designed to align the alveolar segments, as well as realign the
premaxilla into a more anatomic position.
Following surgical repair of the lip, repair of the palate is

generally performed between 9 to 12 months of age. The
choice of techniques for palate repair depends on the type
of cleft. Recent surveys show that the most commonly used
techniques are the Bardach two-flap palatoplasty (45%)
(Fig. 54-9) and the Furlow palatoplasty (42%) (Fig. 54-10);
the Veau-Ward-Kilner (VWK) pushback (Fig. 54-11) and
the von Langenbeck (Fig. 54-12) techniques, although less
common, are also used.28 The common denominators for
all of these techniques are repair in three layers (nasal mucosa,
muscle layer of the soft palate, and the oral mucosa), and anatomic repositioning of the soft palate musculature. Bilateral
cleft palate repair is similar in principle in that a three-layer
repair is achieved (Fig. 54-13). See the references for complete
descriptions of these techniques.29
FIGURE 54-8 Schematic of the steps involved with a bilateral cleft lip repair. Re-creation of the columella, dissection of the muscle in the lateral lip
elements, re-creation of the nasal floor, reapproximation of the lateral lip muscle, and insetting of the nasal alar bases with trimming of skin for final closure
are all integral parts of bilateral cleft lip repair.
704
PART V
HEAD AND NECK
FIGURE 54-9 The Bardach two-flap palatal reconstruction consists of elevating the palatal mucosa off the hard palate bone as a flap, elevation of the
nasal mucosa, and closure of these two layers separately for the hard palate. Pedicles for the mucosal flaps come from the greater palatine arteries posteriorly. Closure of the soft palate is a three-layer repair, including a nasal mucosal layer, muscle layer (levator veli palatini and tensor veli palatini
realignment), and oral mucosal layer.
FIGURE 54-10 The double opposing Z-plasty, (Furlow palatoplasty) of the soft palate includes realignment of the levator and tensor veli palatine muscles
in the form of Z-plasty. One flap has oral mucosa only, whereas the contralateral side has oral mucosa and muscle. The nasal layer consists of a separate nasal
mucosal layer, which is on the side with oral mucosa and muscle flap, and the contralateral side has nasal mucosa and muscle. Closure of both layers in a
double opposing Z-plasty assists in elongating the soft palate, which should subsequently improve palatal speech function.
FIGURE 54-11 The Veau-Ward-Kilner repair consists of advancing the oral mucosal flaps posteriorly to allow closure of the hard palate. Muscle realignment of the soft palate assists in palatal function.
Muscle repair is integral to the palates primary function,

namely speech. Any repair that does not address the muscle
will fail in the development of normal speech.30 As previously
noted, the levator veli palatine and the tensor veli palatine are
the two most integral muscles in producing a functional palate. Detaching these muscles from their abnormal insertions to
the remnant shelves of the posterior hard palate, and then
reapproximating them to each other in a transverse palatal
sling, is referred to as an intravelar veloplasty. Failure to perform this step will result in a nonfunctional palate.
Multidisciplinary Care
------------------------------------------------------------------------------------------------------------------------------------------------
Patients with orofacial clefts require multidisciplinary care

that is provided by plastic surgeons, otolaryngologists, dentists,
orthodontists, oral surgeons, geneticists, audiologists, and
speech and language pathologists. This team approach yields
more comprehensive and coordinated care, which benefits
the patient.3133 Longitudinal follow-up in a team environment is mandatory, because many issues, including the ability
CHAPTER 54
705
FIGURE 54-12 The von Langenbeck repair consists of relaxing incisions on the lateral palate with subsequent advancement to midline, allowing a palatal
repair of both mucosa and muscle layers. Benefits include keeping a bipedicled flap; however, the advancement can be limited and inadequate with wider
cleft defects.
FIGURE 54-13 Closure of bilateral cleft palate follows in the same principles. Nasal mucosal layers are dissected off the vomer and palatine shelves and
closed. Dissection of the hard palate oral mucosa allows closure separately, thus providing a two-layer closure. The soft palatal muscles are dissected and
realigned to provide palatal function.
to produce normal speech, and dental eruption, continue to

evolve as the patient grows.
Care of the cleft patient begins prenatally when a screening
ultrasound makes the initial diagnosis. Ideally, counseling
with the family begins during the perinatal period, and
focuses on potential feeding concerns postnatally, the time
line for surgical repair of the clefts, and team-centered care.
In the neonatal period, feeding is the primary concern, especially in children with clefts of the palate. Appropriate weight
gain must be monitored, as well as the familys overall psychological comfort with the childs cleft. A genetic evaluation,
looking for evidence of associated anomalies and syndromes,
and further counseling within a cleft team, prepare the family
for the ongoing care of their child.
Secondary Cleft Management

------------------------------------------------------------------------------------------------------------------------------------------------
Although patients with cleft lip and palate undergo initial

repair of their clefts in the first 12 to 18 months of life, these
patients will ultimately require further surgical interventions.
Nasal and lip revision, if needed, can be pursued at 5 years of
age, which coincides with greater self-awareness of physical

differences, and exposure to an expanding group of peers in
school.
Between the ages of 7 and 9 years, during the period of
mixed dentition (when the adult lateral incisor is ready to
erupt through the area of the alveolar cleft), a bone graft generally harvested from the hip, is required in the alveolar cleft.
The new bone allows eruption of the lateral incisor, and completes the continuity of the maxilla. Both before and after the
bone graft, additional dental and orthodontic work may be
required to align the teeth in normal anatomic position.
At facial skeletal maturity, generally at 15 years of age for
females and 17 to 19 years of age for males, orthognathic surgery, with surgical movement of the maxilla, mandible, or
both, may be required to achieve normal occlusion, overall
facial appearance and profile. During the teenage years, further revisions of the lip and nose may be required to give these
patients the desired aesthetic outcome that prepares them
for adult life.
Velopharyngeal insufficiency (VPI) is the condition in
which the repaired cleft palate is physically incapable of
isolating the nasopharynx from the oropharynx, resulting in
706
PART V
HEAD AND NECK
air escape through the nose during speech. The patient with
VPI has a nasal quality to his or her speech. Usually the most
affected sounds are plosives, /p/ and /b/, in words such a
papa and buggy. In a patient with VPI, the pressure is dissipated through the nose, making these sounds more nasal in
quality: mama and muggy. Depending on the severity of
the condition, VPI can range from being barely audible to rendering speech unintelligible. VPI usually occurs in patients
whose palates are short and scarred or who have an inadequately functioning soft palate muscle sling. The advent of
VPI is usually noticed as children become more verbal,
between 3 and 5 years of age. Another time period during
which VPI may arise is during tonsillar and adenoid regression. As these tissues atrophy, the nasopharyngeal and oropharyngeal spaces enlarge; a marginally functional soft palate may
no longer be able to seal the nasopharynx from the oropharynx, resulting in VPI. Because of these ongoing changes, vigilance for VPI must be maintained throughout a childs growth
and development.
In patients suspected of VPI, evaluation by a speech
therapist is vital in determining whether additional speech
therapy or surgical intervention is required. Nasoendoscopy

and video-fluoroscopy may be required to determine the
degree of soft palate incompetence, which in turn helps
to determine the optimal surgical technique to correct the
childs VPI.
Conclusions
------------------------------------------------------------------------------------------------------------------------------------------------
Cleft lip and palate can be visually and functionally devastating to a child. Multispecialty and interdisciplinary team care is
both ideal and necessary for the care of these children because
of the complexity of the anomalies and the longitudinal nature
of cleft care. Establishing rapport with patients and their
families, as well as among team specialists, can lead to lifechanging differences in patients with clefts, allowing them
to lead normal, productive lives, as well as making the formidable problems of cleft care rewarding to treat.
expertconsult.com.
CHAPTER 55
Otolaryngologic
Disorders
Lisa M. Elden, Ralph F. Wetmore,
and William P. Potsic
This chapter is divided according to anatomic structures: the

ear, the nose, the oral cavity and pharynx, the larynx, and
the neck. In each section we review anatomy, embryology,
and examination, before discussing congenital and acquired
disorders, including infections, trauma, and tumors.
Ear
------------------------------------------------------------------------------------------------------------------------------------------------
ANATOMY
The ear is divided into three anatomic and functional areas:
the external ear, the middle ear, and the inner ear. The external
ear consists of the auricle, external auditory canal, and the
lateral surface of the tympanic membrane. The auricle is a
complex fibroelastic skeleton that is covered by skin and subcutaneous tissue that directs sound into the external ear canal.
The external auditory canal is oval with the long axis in the
superior to inferior direction. In neonates, the external canal is
almost entirely supported by soft, collapsible cartilage. As the
temporal bone grows over several years, the bony portion of
the canal enlarges to comprise the inner one third, leaving the
outer two thirds supported by firm cartilage. Hair and cerumen
glands are present in the outer two thirds of the external canal.
The ear canal is lined by skin that is continuous with the lateral
surface of the tympanic membrane, and it is innervated by
cranial nerves V, VII, IX, and X and by the great auricular nerve.
The tympanic membrane separates the external ear canal
from the middle ear. It has three layers: an outer layer of squamous epithelium (skin); a middle layer of fibrous tissue that is
attached to the malleus, the most lateral middle ear ossicle;
and an inner layer of mucosa that is continuous with the
mucosa lining the middle ear. The fibrous layer is also attached
to a thick fibrous annulus that anchors the tympanic membrane to the temporal bone.
The middle ear is an air-filled space within the temporal
bone of the skull that is lined by ciliated, columnar respiratory
epithelium. The middle ear communicates with the mastoid
air cell system posteriorly and is lined by the same mucosa.
It also communicates with the nasopharynx anteriorly
through the eustachian tube. The mucociliary transport system of the middle ear moves mucus and debris into the nasopharynx, where it is swallowed. Secretory cells are not evenly
distributed throughout the middle ear and mastoid complex
and are more numerous anteriorly near the eustachian tube.
Three ossicles are present in the middle earthe malleus,
incus, and stapesthat transmit sound from the vibrating
tympanic membrane to the stapes footplate. Stapes movement
creates a fluid wave in the inner ear that travels to the round
window membrane and is dissipated by reciprocal motion to
the stapes.
There are two striated muscles in the middle ear. The tensor
tympani muscle lies parallel to the eustachian tube, and its
tendon attaches to the medial surface of the malleus. The stapedius muscle lies along the vertical portion of the facial nerve
in the posterosuperior part of the middle ear. Its tendon
attaches to the head of the stapes. These muscles stiffen the
ossicular chain in the presence of sustained loud noise.
The facial nerve traverses the middle ear with its horizontal
portion lying superior to the stapes. Posterior to the stapes, the
facial nerve turns inferiorly in a vertical fashion to exit the stylomastoid foramen deep to the tip of the mastoid. The chorda
tympani nerve is a branch of the facial nerve that innervates
taste to the anterior two thirds of the tongue. It exits the facial
nerve in the vertical segment and passes under the posterosuperior surface of the tympanic membrane, crossing the middle
ear lateral to the long process of the incus and medial to the
malleus. The facial nerve lies within a protective bony canal
throughout its course in the middle ear. However, the bony
canal may be absent (in the horizontal portion) in as many
as 8% to 30% of patients.1 Cranial nerve IX supplies sensation
to the floor of the middle ear.
The inner ear consists of the cochlea, semicircular canals,
and vestibule. The cochlea is a coiled fluid-filled tube consisting of 2 to 23/4 turns surrounded by dense bone. It
contains the membranes that support the organ of Corti and
has hair cells that detect the fluid wave from vibration of
the stapes footplate. The hair cells create the neural impulses
that are transmitted from the auditory nerve (cranial nerve
VIII) to the brain, providing the sensation of hearing.
The three paired semicircular canals (horizontal, superior,
and inferior) are also fluid-filled tubes surrounded by dense
bone. The semicircular canals each have a hair cellcontaining
structure (the ampulla) that detects motion. The utricle and
saccule of the vestibule also have hair cell structures that
detect acceleration.2
707
708
PART V
HEAD AND NECK
EMBRYOLOGY
The external ear develops during the sixth week of gestation
and is completely developed by the 20th week. Six hillocks
fuse to form the basic units of the pinna. Defects in the fusion
of the hillocks lead to preauricular tags and sinuses. The external auditory canal develops from the first branchial cleft.
A solid epithelial plug forms during the beginning of the third
month of gestation and canalizes in the seventh month to form
the external auditory canal.
The middle ear space develops from the first pharyngeal
pouch. The ossicles develop from the first and second pharyngeal arches. The inner ear arises from neuroectodermal tissue
within the otic placode that forms the otic pit.2
Any combination of anomalies may occur. Abnormalities of
the development of the ear may create anomalies of the pinna,
external auditory canal, middle ear structures, and inner ear.
One of the anomalies that involves the external and middle
ear is aural atresia (absence of the external auditory canal).
Absence of the external canal may occur with a deformed or
normal external ear. The ossicles may be deformed and are
usually fused to each other as well as the bony plate representing the undeveloped tympanic membrane. The facial nerve
may also be altered in its course through the temporal bone.
Reconstruction of the atretic canal, removal of the bony tympanic plate, release of the fused ossicles, and reconstruction of
a new eardrum is a complex surgical procedure that may
improve hearing. Rarely, there is incomplete development of
the inner ear structures. The most common of these is dysplasia of the cochlea, and it may vary in severity. Dysplasia
is associated with sensorineural hearing loss in most cases.3,4
EXAMINATION
The examination of the ear should always start with inspection
of the outer ear and surrounding structures. Deformities of the
outer ear structure may suggest the presence of other anomalies, such as a first branchial cleft sinus. A first branchial cleft
sinus usually presents below the ear lobe near the angle of
the jaw. The sinus tract may connect to the ear canal or, rarely,
the middle ear.
The external auditory canal and tympanic membrane are
best examined with a handheld otoscope that has a bright
fiberoptic light source and a pneumatic bulb attached to its
head. The largest speculum that comfortably fits in the external canal should be used to maximize visualization and minimize pain. A very small speculum may be inserted deeply, but
it might lacerate the ear canal as well as limit visibility of the
tympanic membrane. The otoscope permits visualization of
the ear canal and tympanic membrane. A translucent tympanic membrane will also permit visualization of the contents
of the middle ear.
A healthy middle ear contains air and is ventilated via the
eustachian tube that connects to the nasopharynx. Insufflation
of air into the ear canal via the pneumatic bulb should cause
the tympanic membrane to move if the middle ear is normal
(aerated) and fail to move if it is filled with effusion (mucus or
pus). Cerumen may be encountered in the ear canal that
obstructs the view of the tympanic membrane or fails to allow
insufflation to occur with pneumatic otoscopy. Removal of
cerumen may be performed by using an operating otoscope
head and an ear curette. However, the use of a headlight, such
as the Lumiview (Welch Allyn, Skaneateles, NY) or operating

microscope, permits the use of both hands and superior
visualization. Care should be taken to secure the child to prevent sudden movement, and the ear curette should be used
gently to avoid causing pain and a laceration of the ear canal.
A mechanical test of tympanic compliance (tympanometry)
may also be useful to help determine if the middle ear
is normally aerated (type A, peaked tracing), fluid-filled
(type B, flat tracing), or has negative pressure because it is
poorly ventilated, suggesting eustachian tube dysfunction
(type C, negative pressure tracing). Examination of a child
with an apparent or suspected ear condition often requires
objective assessment of hearing by audiometry. Current technology and expertise makes it possible to test a child at any age.
Behavioral audiometry can usually be accurately performed
for a child who is older than 6 months of age by sound-field
testing. Older children are presented with a tone through
insert earphones and are tested across a range of frequencies
between 250 and 8000 Hz for ear-specific testing. The hearing
thresholds are recorded at each presented frequency, and
this represents the air conduction threshold. The sound has
to traverse the ear canal, tympanic membrane, and middle
ear. The inner ear must respond by creating electrical impulses
that are transmitted to the brain. Normal thresholds are less
than 20 dB for children.
Bone conduction thresholds test the sensorineural component of hearing. A bone oscillator is used to test a range of
frequencies by vibrating the skull, which stimulates the inner
ear directly, bypassing the external and middle ear. Normally,
air conduction thresholds require less energy than bone conduction thresholds. If bone conduction thresholds require less
sound intensity to be heard than air conduction, the child has
a conductive hearing loss. If air conduction and bone conduction thresholds are elevated but the same, the child has
a sensorineural hearing loss. Most sensorineural hearing loss
in children is a result of hair cell dysfunction in the organ of
Corti. Hearing loss may be conductive, sensorineural, or
mixed. Objective electrophysical tests, such as brainstem
auditory-evoked response and sound emission tests that
measure the intrinsic sounds from the inner ear (otoacoustic
emissions), may be used in young infants and children who
cannot participate in behavioral audiometry. All of these tools
are used by pediatric audiologists.5
For purposes of describing hearing loss, a threshold of
20 to 40 dB is considered mild, 40 to 65 dB is moderate,
55 to 70 dB is moderately severe, 70 to 90 dB is severe, and
greater than 90 dB is profound. Four of 1000 children are
born with a hearing loss, and 1 of those children is born with
a severe to profound hearing loss.
Conductive hearing loss may be corrected with otologic
surgery. Hearing aids and frequency modulation (FM) amplification systems may be helpful to children with both conductive and sensorineural hearing loss. Assistance may be
needed through auditory training, speech language therapy,
and education to maximally develop communication skills.
When a child has a sensorineural hearing loss that is too severe
to be helped with hearing aids, a cochlear implant may be
considered.
A cochlear implant is an electrical device that is implanted
under the scalp behind the ear. Its processor converts sound to
electrical impulses. A cable travels through the mastoid and
facial recess to reach the middle ear, and the electrode array
CHAPTER 55
is inserted into the scala tympani of the cochlea through an

opening that is made in the cochlea.
Cochlear implants stimulate the neural elements of the cochlea directly and bypass the hair cells. Because the vast majority
of sensorineural hearing loss in children is due to hair cell dysfunction, nearly all children get sound perception from a cochlear
implant. Rare conditions, such as an absent auditory nerve or
an absent cochlea, preclude the use of a cochlear implant.
A multidisciplinary evaluation by a cochlear implantation
team is required to evaluate a child and determine family
expectations before performing a cochlear implantation.
A temporal bone computed tomographic (CT) scan and/or
magnetic resonance imaging (MRI) is performed to assess
the cochlea and auditory nerves.
Children who are born deaf and are younger than the age of
3 years, as well as children who have already developed communication skills, language, and speech before losing their
hearing, derive the greatest benefit from cochlear implants.
Cochlear implantation is approved for children 12 months
of age or older by the U.S. Food and Drug Administration.
Children with cochlear implants should be vaccinated against
Streptococcus pneumoniae, according to high-risk schedules,
and against Haemophilus influenzae, according to standard
schedules, because the implant wire crosses from the middle
ear into the cochlea, increasing the risk of meningitis if the
child gets otitis media. After a cochlear implant is performed,
considerable auditory oral training is required to maximize a
childs benefit to develop skills of audition, speech, and language. A child who has been deaf and without sound perception for several years is expected to benefit to a lesser degree.6
OTITIS MEDIA WITH EFFUSION

AND INFLAMMATORY DISORDERS
Otitis media with effusion is the most common chronic condition of the ear during childhood. All children are born with
small and horizontally oriented eustachian tubes that may at
times be unable to clear mucus that is secreted in the mastoid
and middle ear normally and when the child has an upper
respiratory tract infection. The excess mucus usually clears
within a few weeks as the upper respiratory tract infection
resolves. Younger children (infants to 3 years of age) and
children with craniofacial anomalies, such as cleft palate
and Down syndrome, are more prone to having persistent
middle ear effusions; there is no medication that is consistently effective in resolving such effusions.
Persistent effusion may cause a conductive hearing loss
in the range of 20 to 40 dB. A middle ear effusion may also
function as a culture medium and predispose children to
recurrent acute otitis media (AOM).
When fluid persists in the middle ear for 3 to 4 months,
causing a hearing loss or is associated with AOM, myringotomy
and tympanostomy tube placement is helpful to resolve the
hearing loss and reduce the frequency and severity of infection.
Myringotomy and placement of a tube is performed under
general anesthesia using an operating microscope. A small
incision is made in any quadrant of the tympanic membrane
except the posterosuperior quadrant, where there would be
risk of injuring the ossicles. The mucus is suctioned from
the ear, and a Silastic tube is placed in the myringotomy to
provide prolonged ventilation of the middle ear. The tube will
OTOLARYNGOLOGIC DISORDERS
709
usually extrude and the tympanostomy will heal in 6 months

to 1 year. When the ear is no longer ventilated by a tube, the
eustachian tube must ventilate the middle ear. If fluid recurs
and persists, a repeat procedure may be needed. Most children
outgrow this problem as their eustachian tube grows. Occasionally, adenoid tissue in the nasopharynx may contribute
to the persistence of middle ear effusion and may also be removed at the time that a tube is placed. Children who have
had multiple sets of tubes are candidates for adenoidectomy.
ACUTE OTITIS MEDIA

Acute otitis media is the most common infection of childhood
except for acute upper respiratory tract infections. It is the
most common bacterial infection for which children seek medical care from their primary care physician. Usual pathogens
causing AOM include Streptococcus pneumoniae, Haemophilus
influenzae, and Moraxella catarrhalis.7
AOM usually causes severe deep ear pain, fever, and a conductive hearing loss in the affected ear. The purulence in the
middle ear is also present in the mastoid air cells because they
are connected.
To prevent the overuse of antibiotics, the American Academy
of Pediatrics (AAP) and the American Academy of Family
Practitioners (AAFP) developed guidelines in 2004 to improve
accuracy of diagnosis of AOM.8 Three components should be
present to diagnose AOM, including history of acute onset of
symptoms within 48 hours of presentation, presence of middle
ear effusion confirmed by pneumatic otoscopy or tympanometry, and signs of middle ear inflammation. The tympanic membrane typically is reddened and bulging, with obliteration of
normal landmarks.8
Once an accurate diagnosis is made, the AAP/AAFP guidelines
offer options for treatment in otherwise healthy children. They
advocate that a period of observation (48 to 72 hours) is justified because AOM spontaneously resolves in many children
(80% of episodes of AOM resolve within 2 to 7 days of symptom
onset). Very young children (less than 6 months old) and those
with Down syndrome, immune disorders, craniofacial anomalies, or chronic medical conditions should not be considered
candidates for observation, because they are at higher risk of
developing complications such as mastoiditis or meningitis.
Table 55-1 describes specific recommendations for treatment in otherwise healthy children aged 6 months to 12 years,
TABLE 55-1
AAFP/AAP Guidelines for Treatment of Acute Otitis Media
in Children Younger Than 12 Years
Child Age
Certain Diagnosis
Uncertain Diagnosis
Younger than
6 months
6 months to 2 years
Antibiotics
Antibiotics
Antibiotics
> 2 years to 12 years
Antibiotic if severe
illness*; observe if
nonsevere illness{
Antibiotic if severe
illness*; observe if
nonsevere illness{
Observe
*Severe illness: fever > 39 C and/or moderate to severe otalgia.

{
Nonsevere illness: fever < 39 C and/or mild otalgia.
AAFP, American Academy of Family Practitioners; AAP, American Academy of
Pediatrics.
710
PART V
HEAD AND NECK
based on age, certainty of diagnosis, and severity of symptoms.

All affected children should be given pain control, and
children who are treated with observation should be followed
up in 48 to 72 hours and treated if they continue to manifest
symptoms.
Recommended first-line antibiotic therapy is higher-dose
amoxicillin (80 mg per kilogram per day in two divided doses
for 5 to 10 days). Higher dose therapy can effectively cover even
intermediate and some highly resistant strains of S. pneumoniae.
Infections caused by this organism are most likely to cause
more serious complications and are least likely to spontaneously resolve. Azithromycin, erythromycin, or clarithromycin
can be used as alternatives in patients with type 1 allergy
(anaphylaxis or hives to amoxicillin). Second-line antibiotics
should be considered in patients who fail to improve after several days of first-line antibiotics and includes higher-strength
amoxicillin-clavulanate (90 mg per kilogram per day in two divided doses for 10 days) or oral cefdinir, cefuroxime, cefpodoxime, clindamycin, and, less commonly, intravenous or
intramuscular ceftriaxone.
Occasionally, AOM does not respond as expected to
standard antibiotic therapy. When this occurs, culture and
sensitivity testing can be obtained by tympanocentesis. After
sterilizing the ear canal with alcohol, a 22-gauge spinal needle
can be placed through the posterior or anterior inferior quadrant of the tympanic membrane and fluid can be aspirated
with a small syringe.
Complications of AOM are uncommon if appropriate antibiotic therapy is used. The conductive hearing loss resolves as
the middle ear effusion clears. However, infection may necrose
the tympanic membrane, causing a spontaneous perforation.
Small perforations usually heal in less than 7 days, but larger
perforations may persist, cause a conductive hearing loss, and
require a tympanoplasty for closure. The ossicular chain may
also be disrupted by necrosis of the long process of the incus
requiring ossicular reconstruction to restore hearing.
Acute coalescent mastoiditis occurs when infection erodes
the bony mastoid cortex and destroys bony septae within the
mastoid. A subperiosteal abscess may also develop over the
mastoid process. There is usually postauricular erythema
and edema over the mastoid area. The auricle is displaced
laterally and forward (Fig. 55-1). Otoscopy reveals forward
displacement of the posterior superior skin of the ear canal.
In addition to antibiotics, treatment should include a widefield myringotomy from the anterior inferior quadrant to the
posterior inferior quadrant, a tympanostomy tube placement
for middle ear drainage, and a postauricular mastoidectomy to
drain the subperiosteal abscess and the mastoid.
Facial nerve paralysis may occur from inflammation of
that portion of the facial nerve that is exposed in the middle
ear during AOM. Treatment with parenteral antibiotics and
ototopical antibiotic drops applied in the ear canal through
a tympanostomy tube almost always results in complete recovery of facial function. A short course of oral steroids may also
be helpful. Facial nerve recovery may take a few weeks to several months.
Intracranial complications of AOM include meningitis, epidural abscess, brain abscess, otitic hydrocephalus, and lateral
sinus thrombosis. Meningitis is the most common intracranial
complication of AOM and may be associated with profound
sensorineural hearing loss and loss of vestibular function.
Treatment of the intracranial complications of AOM is focused
FIGURE 55-1 Acute mastoiditis. Extension of the acute inflammatory

process from the middle ear and mastoid air cell systems to the overlying
soft tissues displaces the auricle in an inferior and lateral direction from the
side of the head. Fluctuance may be palpated over the mastoid cortex, and
a defect in the cortical bone can frequently be appreciated. Surgical drainage with mastoidectomy is required.
on appropriate treatment of the intracranial process, in addition to a wide-field myringotomy and tympanostomy tube
placement in the affected ear.9
OTITIS MEDIA WITH EFFUSION/CHRONIC

OTITIS MEDIA/CHRONIC SUPPURATIVE
OTITIS MEDIA
Otitis media with effusion is a descriptive term that refers to
persistent middle ear effusion that usually is serous or mucoid
in nature. Chronic otitis media is a term used to describe the
effusion if it lasts longer than 3 months. Otitis media with
effusion may occur following an ear infection, but can occur
spontaneously, especially when the nose has been congested.
It may be associated with hearing loss and the child may or
may not be symptomatic with pain, irritability, or poor
balance. Most effusions resolve spontaneously within weeks,
and most children affected are younger than 5 years of age.
In otherwise healthy children, hearing tests or hearing
screens should be performed once the effusion has been present for more than 3 months, and sooner if significant hearing
loss is suspected or if the child is at high risk for developing
significant speech and language delays. The associated hearing
loss usually falls in the mild range (30 dB), but even in normal
children, may contribute to the development of speech and
language delays. Speech and language tests should be considered if hearing loss is documented. Children should be evaluated for surgical treatment with bilateral myringotomy and
tube placement if they have ongoing pain or irritability attributable to the effusion, structural changes to the tympanic
membrane (such as thinning or deep retractions), documented speech and language delays, or those who are at high risk
for complications if observed (Down syndrome, those with
CHAPTER 55
existing speech delay, autism, or neurocognitive delays). Adenoidectomy would be considered as well if the adenoids are
found to be enlarged, especially if the child has symptoms
of heavy snoring, sleep apnea, or chronic nasal congestion.10
Chronic suppurative otitis media occurs when otorrhea
(drainage of pus or mucous) persists for more than 3 months,
either through a perforation of the tympanic membrane or
through a tube in the tympanic membrane. A cholesteatoma
of the middle ear may also be present in patients who have
perforated tympanic membranes. A cholesteatoma is a squamous epithelial-lined cyst that may be congenital or acquired.
Congenital cholesteatomas are caused by epithelial rests that
persist in the middle ear during temporal bone development.
They present behind an intact tympanic membrane and
appear as a white, smooth mass, most often located in the
anterior superior quadrant of the middle ear. They expand
over time and are filled with squamous debris and may erode
the ossicular chain and extend into the mastoid.
Acquired cholesteatomas develop from skin entering the
middle ear after a tympanic membrane perforation or a retraction
pocket from eustachian tube dysfunction and are usually
located in the posterior-superior quadrant of the middle ear
space. Cholesteatomas are usually painless, cause a conductive
hearing loss, and, in acquired cases, often present as otorrhea.
The otorrhea should be treated with ototopical antibiotic
eardrops, but the only treatment of cholesteatomas is complete
surgical excision by tympanomastoid surgery and ossicular
reconstruction.11 The potential complications of cholesteatomas
are the same as those for acute suppurative otitis media (ASOM).
TRAUMA
Objects stuck deeply into the ear canal, such as a cottontipped applicator, may perforate the tympanic membrane.
This usually causes acute pain, bleeding, and a conductive
hearing loss. If the ossicular chain is not disrupted, the vast
711
majority of these perforations will heal spontaneously in about

2 weeks. If the tympanic membrane is perforated and the
middle ear is contaminated with water, topical antibiotics
should be given.
Lacerations of the auricle should be cleaned to prevent
tattooing and repaired by careful approximation of the skin
and soft tissue to restore the contours of the ear. The cartilage itself does not usually need to be sutured. Partially or totally
avulsed tissue should be replaced. If necrosis of tissue occurs,
it can be debrided as needed. In severe injuries of the auricle, oral
antibiotic treatment to cover S. aureus and Pseudomonas species is
helpful to prevent chondritis and loss of the cartilage framework.
Blunt trauma to the ear is commonly seen in wrestlers, in
children with poor neuromuscular tone, or in children with
self-injurious behaviors. Blood or serum collects between
the periosteum and the auricular cartilage. If the cartilage is
fractured, the collection may occur on both sides of the ear.
Evacuation of the collection is required to restore the contours
of the ear, prevent infection, and prevent scarring with formation of a cauliflower ear. Aspiration of the fluid and placement of a mastoid dressing for compression may be tried
but is most often unsuccessful. Incision and drainage provides
for complete evacuation of the blood or serum. Cotton dental
rolls placed on each side of the auricle and held in place with
bolster mattress sutures is the most effective management.
The dental rolls should be left in place for 7 to 10 days while
the patient also continues with a course of oral antibiotics.
No outer dressing is required except in a child with cognitive
impairment, who may pick at the bolsters.11
Blunt head trauma may disrupt the inner ear membranes
causing sensorineural hearing loss and vertigo. No treatment
is required, and the injury and symptoms may resolve spontaneously, but the sensorineural hearing loss may persist.
Severe head trauma may cause fracture of the temporal bone
of the skull. Temporal bone fractures can be classified as
longitudinal, transverse, or mixed (Fig. 55-2) but are often
FIGURE 55-2 A, Longitudinal temporal bone fracture. These fractures run parallel to the petrous pyramid. The otic capsule is generally not affected by the
fracture lines. Balance, hearing, and facial function are generally preserved. B, Transverse temporal bone fracture. These fractures generally extend through
the cochlea and facial canal and result in deafness, vertigo, and facial nerve paralysis of immediate onset. Facial nerve exploration with repair should always
be considered in these cases.
712
PART V
HEAD AND NECK
complex and do not neatly fit into one category or another.

A high-resolution, thin-section CT scan of the temporal bone
will define the extent of the fracture. The middle ear and
mastoid are filled with blood when a fracture is present.
The blood causes a conductive hearing loss that resolves when
the ear clears.
Otoscopic evaluation of a child with a temporal bone
fracture may reveal a laceration of the ear canal and tympanic
membrane. Blood is usually present in the ear canal, and
the tympanic membrane appears to be dark blue because
the middle ear is filled with blood. There is often ecchymosis
of the mastoid area (Battles sign).
It is important during evaluation of a skull and temporal
bone fracture to note and record the function of the facial
nerve if the patient is not unconscious. Facial nerve paralysis
may be immediate or delayed in onset. Delayed facial nerve
paralysis has a good prognosis for spontaneous recovery.
Immediate complete facial paralysis may indicate disruption
of the nerve or compression by bone fragments. Immediate
facial nerve paralysis requires exploration and repair once
the patient is stable and sufficiently recovered from any associated trauma. The facial nerve should be decompressed in the
mastoid, middle ear, and middle cranial fossa. Bone chips
impinging on the nerve should be removed, and the nerve
should be sutured or grafted if needed. All patients with
temporal bone fractures should have an audiogram once
their condition has stabilized. If the fracture disarticulates
the ossicles, a conductive hearing loss will persist after the
blood has cleared from the middle ear and mastoid.
Fractures of the temporal bone may transverse the cochlea
and vestibular apparatus. These fractures usually cause
a severe sensorineural hearing loss and loss of vestibular
function on the affected side. Most children compensate for
vestibular injuries within weeks, but sensorineural hearing
loss is less likely to improve. A concussive injury of the cochlea
may also simultaneously be present in the opposite ear in
severe head trauma.
Temporal bone fractures may permit leakage of cerebrospinal fluid (CSF) into the middle ear and mastoid. CSF may also
drain through the lacerated tympanic membrane, causing CSF
otorrhea. These leaks usually stop spontaneously, but persistent CSF otorrhea may require a lumbar drain to reduce
the pressure and permit healing. Rarely, tympanomastoid
exploration is required to close the leak. Persistent CSF leaks
in the ear are associated with meningitis.
TUMORS
Benign and malignant tumors of the ear are rare. Glomus tympanicum tumors and neuromas of the facial nerve may present
in the middle ear. Also, eosinophilic granuloma and rhabdomyosarcoma may involve the structures of the temporal
bone.12,13
Nose
------------------------------------------------------------------------------------------------------------------------------------------------
ANATOMY
The nose can be divided into three anatomic sections. The
bony vault is the immobile portion of the nose. It consists
of the paired nasal bones, the frontal process of the maxillary
bone, and the nasal process of the frontal bone. The cartilaginous vault is supported by the upper lateral cartilages and the
cartilaginous nasal septum. The nasal lobule is supported
by the lower lateral cartilages and the cartilaginous septum.
The nasal septum is formed by the quadrilateral cartilage
anteriorly. The posterior septum is composed of bone from
the vomer, perpendicular plate of the ethmoid, nasal crest
of the maxillary bone, and palatine bone.
Both the internal and external carotid artery systems supply
blood to the nose. The roof and lateral wall of the internal nasal cavity are supplied by the anterior and posterior ethmoidal
arteries, sphenopalatine artery, and greater palatine artery. The
septum is supplied by the anterior and posterior ethmoidal
arteries, palatine artery, and the superior labial artery. The
convergence of these vessels in the anterior segment of the
nose is referred to as the Kiesselbach plexus or the Little area.
Venous drainage is accomplished mainly by the ophthalmic,
anterior facial, and sphenopalatine veins.
The olfactory bulb is positioned high in the roof of the nasal
cavity and is responsible for the sense of smell. Sensory information is transported by nerves that penetrate the cribriform
plate and traverse cranial nerve I (the olfactory nerve) to
the brain. Smell is also an important component of what is
perceived as taste.
Bony projections, called turbinates, form the lateral nasal
wall and significantly increase the surface area of the nose,
allowing for more efficient humidification and warming
of the air to 36 C. Three turbinates are usually present
(i.e., inferior, middle, and superior). A supreme turbinate,
which is essentially a flap of mucosa, is occasionally present.
The turbinates contribute to the turbulent airflow that creates
approximately 50% of the total airflow resistance to the lungs.
Cleaning of air is accomplished through the nasal hairs
(vibrissae) and the mucosal surface. Anteriorly, the nose is
lined with stratified squamous epithelium, which changes
to respiratory epithelium immediately anterior to the turbinates. Trapped debris is transported in a posterior direction
into the nasopharynx by a mucociliary transport mechanism.
Speech is affected by nasal anatomy and pathologic conditions. Hyponasality from nasal obstruction or hypernasality
from an excessive air leak can affect voice quality and intelligibility of speech.
EMBRYOLOGY
The nose serves as a drainage port for the paranasal sinuses.
The meati are spaces between the lateral aspect of the nasal
turbinates and the medial aspects of the lateral nasal wall.
Each meatus is named for the turbinate that surrounds it.
The maxillary, frontal, and anterior ethmoidal sinuses drain
into the middle meatus. The posterior ethmoidal sinuses drain
into the superior meatus. The sphenoidal sinus drains into an
area known as the sphenoethmoidal recess that is located posterior and superior to the superior turbinate. The nasolacrimal
duct drains into the inferior meatus.
The nasal cavities develop from the nasal pits in the 4-week
embryo. These pits deepen and move medially to form the
nasal cavity. The oronasal membrane that separates the nose
from the mouth resolves in the seventh week to permit
communication between the nose and nasopharynx.
The paranasal sinuses develop from an outpouching of the
lateral nasal walls during the third and fourth months of
CHAPTER 55
development. The maxillary and ethmoidal sinuses are

present at birth. The frontal and sphenoidal sinuses develop
several years after birth. The frontal sinus begins to develop
at 7 years of age but is not fully aerated until adulthood.14
INFLAMMATORY CONDITIONS
Viral rhinosinusitis (the common cold) accounts for the
majority of nose and sinus infections. It is caused by many
strains of viruses and is a self-limited infection. Symptoms
of fever, nasal congestion, headache, and clear rhinorrhea
usually resolve over 5 to 7 days. Treatment is symptomatic.
BACTERIAL RHINOSINUSITIS
Acute bacterial rhinosinusitis may often follow an acute viral
upper respiratory tract infection. The most common bacteria
causing rhinosinusitis are Streptococcus pneumoniae, Haemophilus
influenzae, and Moraxella catarrhalis. Acute rhinosinusitis causes
malaise, headache, and nasal congestion. There may also be
pain localized to the sinus region or pain on palpation over
the maxillary or frontal sinuses. Chronic sinus infection may
persist after the acute phase, and symptoms often last longer than
30 days.
The gold standard for diagnosing sinusitis is CT of the
sinuses, but a thorough history and nasal examination is
usually sufficient to diagnose acute rhinosinusitis. The nasal
cavity can be visualized by using a large speculum on an
otoscopic head. The posterior nasal cavity can be visualized
with either a straight-rod endoscope or a flexible fiberoptic
nasopharyngoscope.
The treatment of rhinosinusitis includes oral antibiotics,
short-term use of topical nasal decongestants (e.g., oxymetazoline), and saline nasal sprays. Topical nasal corticosteroid
sprays may be helpful for the treatment of both acute and
chronic sinusitis.
Chronic sinusitis in a child may be exacerbated by gastroesophageal reflux disease, immunodeficiencies, mucociliary
dysfunction, and, more commonly, upper respiratory allergy.
These predisposing conditions should be managed while
treating the sinus infection. If the signs and symptoms of
chronic sinus infection persist, a sinus CT is required to
evaluate the condition of the sinus mucosa and the drainage
pathways. Endoscopic sinus surgery may be necessary to open
the involved sinuses to provide drainage.
Chronic inflammation of the nasal and sinus mucosa may
lead to nasal and sinus polyp formation that chronically
obstructs the nose and sinuses. Antrochoanal polyps are large
polyps that originate from the walls of the maxillary sinus and
extend through the nasal cavity into the nasopharynx. Nasal
polyps may be removed endoscopically, but a large antrochoanal polyp may require removal through an open maxillary sinus procedure. Nasal polyps in a child should always
prompt an evaluation for cystic fibrosis.
COMPLICATIONS OF SINUSITIS
The sinuses surround the orbit so a common complication
of acute rhinosinusitis in children is orbital cellulitis with
erythema and edema of the eyelids. Chemosis (edema of the
ocular conjunctiva) is usually absent. However, if a periorbital
subperiosteal abscess forms adjacent to an infected sinus,
713
there may be proptosis, chemosis, ophthalmoplegia, and loss

of vision. Infection in the ethmoidal sinuses most commonly
results in this complication. Subperiosteal periorbital abscess
is demonstrated best by sinus CT (with axial and coronal
cuts). Initial treatment should include intravenous antibiotics. Endoscopic or external drainage may be required in
some cases.
Intracranial complications of sinusitis include cerebritis,
meningitis, cavernous sinus thrombosis, as well as epidural,
subdural, and brain abscesses. Treatment of impending
or confirmed intracranial complications requires surgical
drainage of the involved sinus and concurrent treatment of
the intracranial lesion by a neurosurgeon.15
FUNGAL SINUSITIS
Fungal sinusitis may occur in immunocompromised children,
specifically severe diabetics, children undergoing chemotherapy, and bone marrow transplant recipients. The more
common invasive fungi include Mucor and Aspergillus species.
The treatment of fungal sinusitis involves surgical drainage
and intravenous antifungal agents.
However, a chronic form of fungal sinusitis is allergic
fungal sinusitis. The presence of fungi causes inflammatory
cells to proliferate in the sinuses, causing symptoms of
nasal plugging and facial pain, along with discharge or
polyps. These patients usually have other signs of allergy,
such as asthma. The treatment of this condition is corticosteroids and debridement of the involved sinuses. The diagnosis
is made by sinus CT findings and the presence of eosinophils
as well as fungi in the sinus secretions that are removed at the
time of surgery.16
CONGENITAL MALFORMATIONS
Pyriform Aperture Stenosis
Congenital stenosis of the anterior bony aperture causes
partial nasal obstruction that may be severe enough to cause
difficulty feeding, respiratory distress, and failure to thrive.
Anterior rhinoscopy demonstrates a very constricted nasal
opening bilaterally. CT of the nose shows marked narrowing
of the pyriform aperture.
Neonates are obligate nasal breathers, and severe stenosis
must be surgically corrected. Because the stenotic segment
is very anterior and the remainder of the nasal cavity is normal,
removal of the constricting bone with drills is done through
a sublabial approach. The nasal openings are stented with
3.0-mm endotracheal tube stents that are sutured in place
and removed after a few days.
Choanal Atresia
Choanal atresia may be unilateral or bilateral. The obstructing
tissue is usually a bony plate, but a few cases will have only
membranous atresia. Unilateral choanal atresia presents as
chronic unilateral rhinorrhea. There is no significant respiratory distress. Because neonates are obligate nose breathers, bilateral choanal atresia is associated with severe respiratory
distress, difficulty feeding, and failure to thrive. The diagnosis
is suspected if catheters cannot be passed through the nose
and into the pharynx. The obstruction may be visualized with
a narrow flexible nasopharyngoscope after the nasal cavity has
714
PART V
HEAD AND NECK
been suctioned of mucus and the nasal mucosa has been constricted with a nasal decongestant (e.g., oxymetazoline). The
diagnosis is best made with CT of the nasal cavity. CTwill demonstrate the atresia, define the tissue (bony or membranous),
and show the configuration of the entire nasal cavity.
Choanal atresia may be successfully treated by removing the
obstructing tissue transnasally. Curettes, lasers, microdebriders,
bone punches, and drills may all be effective to remove the atresia plate. However, when the bony plate is very thick and there
is an extremely narrow posterior nasal cavity, a transpalatal
repair is more direct. A transpalatal repair provides better access
for more effective removal of the bony plate and posterior
septum (Fig. 55-3). Stents fashioned from endotracheal tubes
are placed and secured with sutures to the septum. They are
removed after several weeks. The stents must be moistened
with saline and suctioned several times daily to prevent mucus
plugging and acute respiratory distress. Transpalatal repair of
choanal atresia has a lower incidence of restenosis.11
Nasal Dermoid
FIGURE 55-3 Choanal atresia. This disorder frequently presents at birth
with respiratory distress.
FIGURE 55-4 Nasal dermoid presenting in the midline as a pit.
Nasal dermoid cysts or sinuses present in the midline of the

nasal dorsum (Fig. 55-4). They usually appear as a round
bump or a pit with hair present in the pit (Fig. 55-5). They
also may become infected. Nasal dermoid sinuses may extend
through the nasal bones into the nasofrontal area and have an
intracranial component. Both CT and MRI may be necessary to
demonstrate the extent of the dermoid. Surgical removal is
required to prevent infection and recurrence. This may be
done between ages 3 and 5 years if prior infection has not
occurred. Dermoids confined to the nose are resected
completely using a midline incision with an ellipse around
the sinus tract. The tract is followed to its termination,
and the nasal bones may need to be separated to reach the
end of the tract.11 If an intracranial component is present, a
combined craniotomy and nasal approach with a neurosurgeon is recommended.
FIGURE 55-5 Nasal dermoid. These lesions typically present on the nasal
dorsum as a single midline pit, often with a hair extruding from the depths
of the pit. The pits may also be found on the columella. The dermoid will
then tract through the septum toward the cranial base.
CHAPTER 55
Nasal Glioma and Encephalocele

A nasal glioma presents as an intranasal mass and may be
confused with a nasal polyp. The mass contains dysplastic
brain tissue and may have an intracranial connection. CT and
MRI are important to define the extent of the glioma and intracranial component as well as to plan the surgical approach.
An encephalocele presents as a soft compressible mass and
may also be confused with a nasal polyp or a nasal dermoid.
Intranasal encephaloceles extend through a defect in the skull
at the cribriform plate. CT and MRI define the extent of the
encephalocele and are necessary to design the surgical approach. Surgical removal often includes a frontal craniotomy.
Nasal encephaloceles may be associated with CSF rhinorrhea
and meningitis.
715
Epistaxis in children usually occurs in Littles area of the

anterior septum and frequently results from digital trauma
(nose picking). The bleeding usually stops with pressure by
squeezing the nasal ala. Infrequently, cauterization of the
vessels under general anesthesia is needed to reduce the
frequency of bleeding. In cases that fail to stop with pressure,
the nose should be packed with absorbable materials (such as
Gelfoam or cellulose) or nonabsorbable gauze. Finally, in
severe cases, embolization or emergent surgery has been used
to control bleeding from the internal maxillary and anterior
ethmoid arteries, which are the primary sources of nose
bleeds. Hematology consultation should be considered in
severe or recurrent cases to evaluate for coagulopathies.
Nasal Foreign Bodies
TRAUMA
Anosmia
Head trauma can lead to temporary or permanent anosmia
(lack of sense of smell). In one large study of head trauma
patients (n 190), 11% reported loss of sense of smell that
persisted after their initial recovery and later was confirmed
by smell tests. Those at higher risk had trauma that led to
intracranial hematoma and/or hemorrhages or injury near
the skull base.17
Nasal Fracture
An infant may be born with the soft nasal bones and the
septum deviated to one side either as a result of a difficult delivery or from persistent intrauterine compression of the nose.
The nasal structures can most often be returned to the midline
with digital manipulation. If the nasal deformity is partially
reduced, the nose usually straightens with growth during
the first 12 to 18 months of age.
Nasal bone and nasal septal fractures in older children
usually occur from a blow to the face during sports. There
is usually a brief period of epistaxis and deviation of the nasal
dorsum to one side. Swelling occurs rapidly, and the degree
of the cosmetic deformity or the need for fracture reduction
may not be easily determined. At the fourth to sixth day after
injury, the edema subsides and the need for reduction can be
determined. Nasal bone radiographs are of little help in making this judgment; so, the need for nasal fracture reduction is
usually based solely on clinical examination. Effective closed
nasal fracture reduction may be done up to 2 weeks after
the injury. Closed reduction under general anesthesia is the
method of choice. Oral antibiotics prevent infection and are
essential if nasal packing is used to support the nasal bone.
Although nasal fracture reduction is not urgent, a septal
hematoma from a fractured septum should be excluded by
the initial physician seeing the child. A septal hematoma that
remains untreated may cause cartilage necrosis and loss of nasal
support, with a resulting saddle-nose deformity. Treatment of a
septal hematoma is with incision and evacuation of the clot. The
mucoperichondrial flap should then be sutured in place by
bolster sutures through the septum. A small rubber band drain
may be required and, if used, should remain in place for 12 to
24 hours, and antibiotics be given for 10 to 14 days to prevent
secondary infection while a drain is in place.
Children may be observed inserting a foreign body into their

nose, or they may inform their parents of the event. Most
children, however, present with a foul-smelling unilateral purulent nasal discharge and deny putting anything into their nose.
Most nasal foreign bodies are painless and do no harm to the
nose but cause a foul nasal discharge. Disc batteries, on the other
hand, cause very rapid alkali burns of the nasal cavity and pain.
Batteries must be removed from the nose quickly because the
chemical burn occurs in minutes to hours. If extensive tissue necrosis occurs, it may cause a nasal stenosis or septal perforation.
Removal of a nasal foreign body is aided by decongesting
the nasal mucosa and using a headlamp to visualize the foreign
body. A variety of forceps or hooks may be used. If the object is
deep in the nose, the removal is best performed under general
anesthesia. The endotracheal tube prevents aspiration of the
object into the tracheobronchial tree if it is pushed back into
the nasopharynx. One must remember that multiple foreign
bodies may be present on one or both sides of the nose.
Nasal Lacerations
Nasal lacerations should be closed with care to match edges
and restore the contours of the nose. Standard wound closure
technique is used. The nasal mucosa does not need to be
sutured unless a large flap is displaced.
NASAL/NASOPHARYNGEAL TUMORS
Rhabdomyosarcoma, lymphoma, squamous cell carcinoma,
and esthesioneuroblastoma may occur in the nose and sinuses
of children. Fortunately, these malignant tumors are very rare
in children. The treatment of children with malignant tumors
of the nose and sinuses usually involves a multidisciplinary,
multimodal approach.
Juvenile nasopharyngeal angiofibroma is a benign tumor of
adolescent males that originates from the lateral wall of the
nose and nasopharynx. The tumor may completely obstruct
the nose and fill the nasopharynx. This type of angiofibroma
may also extend intracranially through the base of the skull.
Patients with these tumors present with nasal obstruction,
recurrent epistaxis, and rhinorrhea.
The tumor may be seen with a flexible fiberoptic nasopharyngoscope or a rod lens telescope after decongesting the nasal
mucosa. It appears as a smooth reddish mass. Biopsy of the
mass should be avoided because of the potential for severe
bleeding. CT and MRI define the extent and location of the
tumor. On imaging, the mass originates in the pterygopalatine
716
PART V
HEAD AND NECK
fossa within the aperture of the pterygoid (vidian) canal. It

causes anterior bowing of the posterior wall of the maxillary
sinus and erosion of the greater wing of the sphenoid as it
grows into the nose and nasopharynx. MR angiography helps
to delineate the blood supply, which may originate from both
the internal and external carotid arteries. Contrast angiography may be reserved for presurgical planning and embolization of the copious blood supply that is often present.
The treatment of juvenile nasopharyngeal angiofibroma is
complete surgical resection after preoperative embolization.
Depending on the material used, the embolization may be
effective for days to weeks. A variety of surgical approaches
may be used, including endoscopic resection of small tumors
using instruments to reduce blood loss, such as suction
cautery or coblation tools. Extensive tumors may require a
combined midfacial and craniotomy approach.18
Some authors have proposed radiation therapy as the
primary treatment of juvenile nasopharyngeal angiofibroma,
but many surgeons are concerned about the long-term
effects of radiation in children, including the induction of
malignant tumors.
Nasopharyngeal carcinoma can occur in adolescents and is
more common in those of Asian or African descent. It arises
from the epithelium of the nasopharynx and histologically
is composed of lymphoepithelial cells of variable stages of
differentiation. Epstein-Barr viral infection has been implicated as a possible cause in some cases, but genetic factors
appear to make some individuals more susceptible to developing this tumor. Most children present with advanced
disease and tend to have undifferentiated subtypes. They usually have a history of unilateral nasal plugging and otalgia or
hearing loss caused by a blocked eustachian tube. They may
also present with metastasis in the posterior triangle lymph
nodes. Treatment consists of radiotherapy and, in some cases,
adjuvant chemotherapy.
Oral Cavity/Pharynx
and the musculus uvulae. Defects in formation of the hard

and/or soft palate result in clefting. The sensory and motor
innervation of the palate is through the trigeminal nerve and
pharyngeal plexus.
The circumvallate papillae divide the tongue into the
anterior two thirds that lies in the oral cavity and the posterior
one third lying in the oropharynx. The innervation and vascular supply to the two major divisions of the tongue reflect
their differences in originthe anterior two thirds of the
tongue being a first branchial arch derivative (trigeminal),
whereas the posterior one third being a combination of third
and fourth arch derivatives (pharyngeal plexus). The hypoglossal nerve supplies motor innervation to the intrinsic
musculature. In addition to the intrinsic tongue musculature,
the action of four extrinsic muscles combine to provide mobility. The genioglossus protrudes and depresses, the hyoglossus
retracts and depresses, the styloglossus retracts, and the palatoglossus elevates. In addition to the circumvallate papillae,
other taste buds on the tongue surface include conical,
filiform, fungiform, and foliate papillae.
The pharynx is a fibromuscular tube that extends from the
skull base to the level of the cricoid cartilage of the larynx and
can be divided into three levels. The nasopharynx extends
from the skull base to the level of the soft palate, the oropharynx extends from the soft palate to the tongue base, and the
hypopharynx extends from the tongue base to the cricoid cartilage. Three muscular constrictors combine to form the muscular portion of the pharynx: superior, middle, and inferior
constrictors. The Passavant ridge is a muscular segment of
the superior constrictor that is involved in velopharyngeal closure. Lower fibers of the inferior constrictor help to form the
upper esophageal sphincter. The motor and sensory innervation of the pharynx is from the glossopharyngeal and vagus
nerves via the pharyngeal plexus.
A collection of lymphoid tissue within the pharynx forms
the Waldeyers ring, which includes the palatine tonsils, the
adenoids (pharyngeal tonsil), and lymphoid follicles lining
the lateral and posterior pharyngeal walls.
------------------------------------------------------------------------------------------------------------------------------------------------
ANATOMY
ACUTE PHARYNGOTONSILLITIS
The boundaries of the oral cavity include the lips anteriorly, the
cheeks laterally, and the palate superiorly. The posterior boundary is a plane that extends from the soft palate to the junction of
the anterior two thirds and posterior one third of the
tongue. The oral cavity is composed of the vestibule, the space
between the lips and cheeks and alveolar ridges, and the oral cavity proper. The vestibule and oral cavity proper are separated by
the alveolar ridge and teeth. The vestibule is divided in the midline by the frenula of the upper and lower lips. The alveolar ridge
is contiguous superiorly with the hard palate. The parotid ducts
(Stensen ducts) enter the vestibule opposite the second maxillary
molars. The submandibular ducts (Wharton ducts) enter the
floor of mouth near the lingual frenulum.
The palate is formed by a fusion of the primary palate
anteriorly and medial growth of the palatal processes that form
the secondary palate. The hard palate divides the nasal and oral
cavities and is formed by the premaxilla and the horizontal
plates of the palatine bones. The soft palate is formed by a muscular aponeurosis of the tensor veli palatini tendon. Five muscles insert into this aponeurosis and include the tensor veli
palatini, levator veli palatini, palatoglossus, palatopharyngeus,
In addition to the acute onset of sore throat, viral pharyngitis

typically presents with fever and malaise. Signs include erythema of the pharynx and cervical lymphadenopathy.
Depending on the viral agent, associated symptoms of nasal
obstruction and rhinorrhea may also be present. Rhinovirus,
coronavirus, parainfluenza virus, respiratory syncytial virus,
adenovirus, and influenza virus are agents responsible for viral
pharyngitis.
Primary herpetic gingivostomatitis, caused by herpes simplex virus types 1 or 2, presents as fever, adenopathy, and vesicles and ulcers on the lips, tongue, buccal mucosa, soft palate,
and pharyngeal mucosa. Herpangina and Coxsackie virus
(hand-foot-and-mouth disease) are viral infections that involve the oropharynx. Epstein-Barr virus (EBV) infection
(infectious mononucleosis) presents as acute pharyngotonsillitis (often with white sloughing debris on the tonsils), fever,
generalized adenopathy, malaise, and splenomegaly. Although
EBV infection is suspected by the appearance of 10% or more
atypical lymphocytes on a complete blood cell count and the
presence of a positive Monospot test, the definitive diagnosis
is confirmed by elevated titers of EBV. A short course of
CHAPTER 55
717
corticosteroids has been proven to reduce the lymphoid hypertrophy that can cause acute airway obstruction.
Group A beta-hemolytic streptococci (GABHS, i.e.,
S. pyogenes) commonly infect the pharynx. In addition to sore
throat, associated symptoms include fever, headache, and
abdominal pain. Associated signs include pharyngeal erythema,
halitosis, tonsillar exudates, and tender lymphadenopathy.
Lack of cough helps differentiate it from other upper respiratory
tract infections. Diagnosis may be confirmed initially with a
rapid streptococcal antigen test. Because rapid antigen testing
is more sensitive than formal plating on blood agar, a negative
test does not need confirmation, but positive rapid streptococcal
tests should be confirmed with formal plating. Other bacterial
pathogens that cause acute pharyngitis include Haemophilus
influenzae and groups C and G beta-hemolytic streptococci.
Occasionally, concurrent infection with penicillin-resistant
Staphylococcus aureus may interfere with treatment of a GABHS
infection.19 Although many cases of GABHS infections respond
to treatment with penicillin Vor amoxicillin, emerging resistance
to oropharyngeal pathogens mandates treatment of recalcitrant
cases with an antibiotic having known effectiveness against
beta-lactamaseproducing organisms. In cases in which a lack
of compliance is suspected, intramuscular benzathine penicillin
or ceftriaxone may be used.
Acute pharyngitis may also be associated with acute bacterial infections of the nose, nasopharynx, and sinuses. These
infections may be caused by a variety of viral and bacterial pathogens; in addition to a sore throat, symptoms include fever,
mucopurulent nasal drainage, nasal obstruction, and facial pain.
tonsilliths. Signs include erythema of the tonsils, cryptic

debris, and chronically enlarged cervical lymphadenopathy.
A variety of viral and bacterial agents can be blamed for
chronic infection of the pharynx. Cultures may or may not
be positive in these patients because surface cultures may
be negative while core tissue is positive. Antibiotic therapy
directed at oral anaerobes or S. aureus may be helpful in resistant cases. Children with infections unresponsive to medical
management are candidates for tonsillectomy.
Periodic fever, aphthous ulcers, pharyngitis, and cervical
adenitis (PFAPA) is a syndrome that occurs most commonly
in young children (mean age 39 months). The cause is unknown. It is characterized by recurrences of fevers that usually
last 3 to 7 days, along with aphthous stomatitis, pharyngitis,
cervical adenitis, and headache. The recurrences occur in
cycles of every 1 to 2 months, and the child is well between
episodes. Throat cultures are negative. Antibiotics are not
effective in treating this condition, but steroids (prednisone
1 mg per kilogram in a single dose) have been shown to reduce
the duration of fever in individual episodes (from 4 days to
1 day in one study); however, sometimes steroids may also reduce the duration of intervals between infections. Most children have spontaneous resolution of these fevers over several
years (mean time to resolution is 32 months). Tonsillectomy
has been shown to be effective in significantly reducing the
duration of this syndrome and frequency of episodes.21,22
RECURRENT PHARYNGOTONSILLITIS
Injuries to the oropharynx and palate are relatively common in

children, usually occurring when a child runs with a toy or
stick in his mouth. Most result in mucosal lacerations that
spontaneously heal, but larger lacerations may require sedation or anesthesia to repair. Use of prophylactic antibiotics
is reserved for larger wounds. Although rare, blunt (and less
often penetrating) injuries can occur when the object strikes
the jugular vein or the carotid artery that can result in immediate neurovascular injury and poor neurologic outcomes.
However, more subtle injuries to the intima of the carotid
can lead to pseudoaneurysms that may later develop emboli.
These emboli can cause brain infarcts with severe neurologic
sequelae over the following several days. Ideally, if a vascular
injury has been identified, then aspirin, or, less often, anticoagulant therapy could be used to prevent these emboli from
forming. Unfortunately, no specific clinical factors (including
size or location of wound) have been shown to correlate with
the presence of a subtle vascular injury. Computed tomography angiography (CTA) has been used to rule out a significant
vascular injury, but benefit from CTA remains controversial,
because only 2.8% of studies are positive.23
Recurrent infection of the pharynx may be either viral or

bacterial. GABHS are the most worrisome bacterial organisms,
because recurrent infection may lead to complications such as
scarlet fever, acute rheumatic fever, septic arthritis, and acute
glomerulonephritis. In addition to a history of multiple positive cultures for S. pyogenes, elevated antistreptolysin-O (ASO)
titers may identify patients with chronic infection who are
at risk for developing complications. Some asymptomatic
children may be chronic carriers of GABHS, and elevated
ASO titers may not be a reliable indicator for distinguishing
between an active infection and the carrier state.
Treatment of recurrent streptococcal infection or the child
who is a carrier should include a trial course of an antibiotic
shown to reduce carriage (e.g., clindamycin, vancomycin, or
rifampin). Children with recurrent pharyngotonsillitis unresponsive to medical therapy or those who suffer a complication should be considered for surgical management.
Whereas treatment of each child should be individualized,
suggested guidelines for surgical candidates include seven infections in 1 year, five or more infections per year for 2 years,
or three or more infections per year for 3 years.20 Other factors
to be considered in using a surgical option include severity of
infection, response to antibiotic therapy, loss of time from
school, and need for hospitalization.
CHRONIC PHARYNGOTONSILLITIS
The pharynx and, specifically, the tonsils may be the target of
chronic infection. Affected children complain of chronic
throat pain, halitosis, and production of white particles or
ORAL TRAUMA
PERITONSILLAR CELLULITIS/ABSCESS
Localized extension of tonsillar infection may result in peritonsillar cellulitis. The same pathogens that cause acute
pharyngotonsillitis are responsible for peritonsillar cellulitis.
In addition to a severe sore throat, symptoms and signs include drooling, trismus, muffled voice, ipsilateral referred
otalgia, and tender lymphadenopathy. The affected tonsil is
usually displaced in a medial and inferior position. Peritonsillar cellulitis may progress to frank abscess formation (quinsy).
718
PART V
HEAD AND NECK
FIGURE 55-6 A, Retropharyngeal abscess. Computed tomography of the cervical area demonstrates fluid loculated in the retropharyngeal space.
The abscess is typically unilateral and frequently extends into the medial aspect of the peripharyngeal space. In the absence of associated complications,
drainage can be done intraorally (arrow). B, Lateral neck abscess on the left side (arrow).
Early cases of peritonsillar cellulitis may respond to oral

antibiotics, such as the penicillins, cephalosporins, erythromycins, or clindamycin. Unresponsive cases of cellulitis
or abscess should be treated with intravenous antibiotics.
In children with suspected abscess formation, a variety of
surgical drainage procedures can be performed. Needle aspiration or incision and drainage have been shown to be equally
effective.24 In persistent cases or in those children who will
require general anesthesia for drainage, consideration should
be given to performing a tonsillectomy (quinsy tonsillectomy).
RETROPHARYNGEAL/PARAPHARYNGEAL
SPACE INFECTIONS
Signs and symptoms of deep neck space (retropharyngeal/
parapharyngeal) infections that involve the pharynx typically
are as fever, drooling, irritability, decreased oral intake, torticollis, and/or trismus. Often there is a history of a preceding
viral illness. Stridor or symptoms of upper airway obstruction
may be seen in half of patients.25 A neck mass or enlarged cervical nodes may be present, depending on the location of the
infection. Usual pathogens include coagulase-positive staphylococci and GABHS. Anaerobic bacteria have been found in as
many as 50% of cases.25 Complications of deep neck space infections include airway obstruction, bacteremia, rupture of
the abscess into the pharynx with aspiration, mediastinal extension of infection, jugular thrombosis, and carotid artery
rupture.
In suspected cases, the diagnosis of a retropharyngeal/
parapharyngeal space infection is confirmed with either
contrast mediumenhanced CT or MRI. Widening of the retropharynx on a lateral neck radiograph suggests a retropharyngeal infection. Although ultrasonography can detect the
presence of an abscess cavity, CT or MRI are most helpful in
demonstrating the extent of infection and the location of
surrounding structures of importance, specifically the great
vessels. Contrast mediumenhanced CT is particularly useful
in distinguishing a phlegmon (cellulitis) from cases of frank
suppuration. Demonstration of a hypodense region with

surrounding rim enhancement has been shown to correlate
with an abscess in 92% of cases (Fig. 55-6).
The initial management of a deep neck infection should begin with intravenous antibiotics, including clindamycin, cefazolin, beta-lactamase penicillins, or a combination thereof.
Sixty-seven percent of children with these infections (including those presenting with cellulitis or early abscess) require
eventual drainage. Surgical drainage should be reserved for
those children who present with airway symptoms along with
obvious abscess and for those who fail to show clinical improvement or progress to frank abscess formation on CT after
48 to 72 hours of intravenous (IV) antibiotics. The usual approach to surgical drainage is intraoral, if the abscess points
medial to the great vessels, or extraoral, if the infection points
lateral to the great vessels.
Complications of deep neck infections should be treated
aggressively. Mediastinal spread requires prompt surgical
drainage in most cases. An infected jugular thrombosis
(Lemierre syndrome) can be a source of metastatic spread of
infection as septic emboli. Signs and symptoms include spiking chills and fever (picket-fence fevers) and a neck mass
despite appropriate antibiotic therapy. Anticoagulation or excision of the infected thrombus may be required to eradicate
the infection.
SLEEP-DISORDERED BREATHING
In the past decade, the impact of sleep-disordered breathing
(SDB) on the health of children has been well described,
beginning with the report of normative sleep data by Marcus
and colleagues.26 Children appear to have briefer but more
frequent episodes of partial (hypopnea) and complete (apnea)
obstruction. Because an apnea of less than 10 seconds may
represent several missed breaths in a child, an apnea of any
duration is abnormal. In most cases the site of obstruction
during sleep is in the pharynx. In contrast to adults with this
disorder, in whom the pharyngeal impingement is due to
CHAPTER 55
adipose tissue surrounding the pharyngeal musculature, the

major cause of airway obstruction in children results from
adenotonsillar hypertrophy.
The apnea index (AI) represents the number of apneas in
an hour, with a normal value being less than 1 in children.
Because most children have an increased frequency of partial
obstructions compared with adults, a measure of hypopneas
may be more significant. A hypopnea is variably described
as a reduction in airflow or respiratory effort or oxygen
desaturation or combination thereof. The apnea/hypopnea
index (AHI) is a measure of both apneas and hypopneas in
an hour and may be a better reflection of SDB in children.
An AHI greater than 5 is abnormal in adults, whereas, an
AHI greater than 1.0 to 1.5 is abnormal in children. The upper
airway resistance syndrome represents obstructed breathing
with normal respiratory indices but with sleep fragmentation and electroencephalographic arousals that indicate
disordered sleep.
The major group at risk for SDB includes children with
adenotonsillar hypertrophy secondary to lymphoid hyperplasia (Figs. 55-7 and 55-8). Whereas the age of affected children
ranges from 2 years through adolescence, the prevalence mirrors the age of greatest lymphoid hyperplasia, 2 to 6 years, the
age the tonsils and adenoids are largest in size. Other at-risk
groups include syndromic children with Down syndrome
who also have relative macroglossia and tend to have larger
tonsils and adenoids, children with craniofacial disorders,
and patients with cleft palate or storage diseases (Hunter
and Hurler syndromes). Adverse effects of obstructive sleep
apnea on children include poor school performance, failure
to thrive, facial and dental maldevelopment, and, rarely, severe
cardiac impairment, including systemic hypertension, cardiac
arrhythmias, and cor pulmonale with heart failure.
Daytime symptoms include noisy mouth-breathing, nasal
obstruction and congestion, hyponasal speech, and dyspnea
FIGURE 55-7 Tonsillar hypertrophy. Tonsillar hypertrophy is rated on a

scale of 1 to 4. Grade 1 tonsils are hypertrophic, grade 2 tonsils extend
slightly beyond the tonsillar pillars, grade 3 tonsils extend in a medial
direction beyond the anterior tonsillar pillars, and grade 4 tonsils touch
in the midline.
719
FIGURE 55-8 Adenoid hypertrophy. Hypertrophy of the adenoids may

cause the nasopharynx to be obstructed with tissue. Smaller amounts of
tissue are also able to obstruct nasal respiration by growing into the
posterior choana as shown in this photograph.
on exertion. In contrast to adults, hypersomnolence is uncommon in children because of the lower incidence of gas exchange
abnormalities, specifically hypercarbia. Children may complain of headaches, seem irritable, and perform poorly in
school. Nighttime symptoms are more obvious and include
snoring, gasping, and choking respirations, apnea, coughing,
and a variety of other behaviors, including sleepwalking, sleeptalking, rocking, head banging, and bruxism. Enuresis may
appear in children with airway obstruction and then resolve
after surgical treatment. In addition to enlarged tonsils, signs
include the presence of a posterior pharyngeal flap in cleft
palate patients, a craniofacial disorder, adenoid facies, and,
rarely, evidence of right-sided heart failure.
The diagnosis of SDB is suggested by history and physical
examination. Confirmation of obstruction and apnea may be
made with overnight pulse oximetry and video or audio
monitoring of sleep. The gold standard in the diagnosis of
obstructive sleep apnea remains formal polysomnography,
including measures of nasal and oral airflow, transcutaneous
oxygen and carbon dioxide, chest wall movements, electrocardiography, extraocular muscle movements, electroencephalography, leg movements, and gastric pH monitoring in
selected cases. Depending on the suspected site of obstruction, adjuvant studies, such as a lateral neck radiograph,
MRI of the head and neck, and flexible upper airway
endoscopy, might be helpful.
The nonsurgical management of SDB consists of weight
loss in obese patients and treatment of underlying allergies
and gastroesophageal reflux. Nasal and dental appliances
to maintain airway patency that may be useful in adults are
usually poorly tolerated in children. Nasal continuous positive
airway pressure, the mainstay of treatment in adults, is tolerated in many children and should be considered as a treatment
option, especially in patients in whom other therapies have
been exhausted or proven ineffective.
720
PART V
HEAD AND NECK
The initial surgical treatment for most children with SDB remains a tonsillectomy and adenoidectomy, a therapy that is usually curative. In patients with documented sleep apnea or a sleep
disorder, both procedures should be used even if the tonsils appear small. Tonsillectomy and adenoidectomy techniques that
have been standard for decades have been supplanted in some
institutions by new technology, including use of coblation, harmonic scalpel, and the microdebrider. Efficacy of these newer
techniques versus established methods remains unproven.
Complications after tonsillectomy and adenoidectomy usually consist of respiratory compromise and acute or delayed
bleeding. Since the advent of modern pediatric anesthesia,
respiratory complications, such as aspiration with resultant
pneumonia and lung abscess, are rare. Humidification, intraoperative corticosteroids, and antibiotics have all been shown
to improve the postoperative course after tonsil and adenoid
surgery. Young children are most vulnerable to complications,
and, in most institutions, children younger than 3 to 4 years
of age are observed overnight for signs of dehydration and
respiratory compromise.
Adjuvant surgery in the management of SDB includes
craniofacial repair or posterior flap revision surgery in appropriate patients. Midface, mandibular, and hyoid advancement
have proved useful in selected patients, along with nasal
surgery such as septoplasty, partial inferior turbinectomy, or
nasal polypectomy. Tracheostomy remains the treatment of last
resort in patients who fail to respond to other forms of therapy.
ANKYLOGLOSSIA
Ankyloglossia or tongue-tie is a common congenital disorder
involving the lingual frenulum (Fig. 55-9). Neonates with
diminished tongue mobility resulting from a foreshortened
frenulum may have problems in sucking and feeding. Because
the frenulum is thin and relatively avascular in neonates and
young infants, it can often be incised as an office procedure.
In older children the greatest effect of ankyloglossia is on
speech and it can lead to dental caries because it may be difficult to clean the lower teeth. Because the tip of the tongue
FIGURE 55-9 Ankyloglossia. Abnormal development of the lingual

frenulum that limits extension of the tongue tip beyond the mandibular incisors frequently causes articulation disorders and should be
corrected.
curls under on protrusion and has limited lateral and superior

movement, speech articulation may be affected. Surgical treatment in these patients may require a short general anesthetic
because the frenulum is thicker and more vascular, requiring
surgical correction that includes simple division either with
or without a Z-plasty repair.
MACROGLOSSIA
Macroglossia is uncommon. Generalized macroglossia, as seen
in association with Beckwith-Wiedemann syndrome, with
glycogen storage diseases (Hunter and Hurler syndromes)
or hypothyroidism, is rare. Relative macroglossia can be seen
normally on occasion but is most common in Down syndrome. The most serious complication of this condition is
airway obstruction. In infants, macroglossia should be distinguished from focal enlargement of the tongue seen in patients
with a lymphatic malformation or hemangioma. Glossoptosis,
posterior displacement of a normal-sized tongue, is seen in association with cleft palate and micrognathia in infants afflicted
with the Pierre Robin sequence. The airway symptoms in most
of these infants usually improve over the first year or two of
life; so, supportive care is most often recommended (including oral airways and upright positioning with feeding). Infants
with severe airway obstruction secondary to an enlarged or
displaced tongue may require tongue reduction or a temporary tongue-to-lower lip adhesion suture, respectively. Tracheostomy is reserved for the worst cases. Macroglossia in older
children that affects cosmesis, interferes with speech, or
causes drooling may be treated with a variety of other tongue
reduction techniques.
BENIGN LESIONS
Epulis is a congenital granular cell tumor that typically
presents as a soft, pink submucosal mass on the anterior
alveolar ridge of the maxilla (Fig. 55-10). Females are more
FIGURE 55-10 Congenital epulis. The congenital epulis is an unusual benign lesion that frequently arises from the anterior maxillary alveolar ridge.
Airway and feeding difficulties may develop secondary to large lesions.
Surgical excision is required.
CHAPTER 55
FIGURE 55-11 A ranula is a pseudocyst caused by obstruction of a sublingual gland. It generally presents as a unilateral, painless swelling in the
floor of the mouth.
commonly affected, and symptoms are usually confined to

feeding problems. Surgical excision is curative.
Ranula is a pseudocyst located in the floor of the mouth
that may occur congenitally or result from intraoral trauma
(Fig. 55-11). Large ranulas may extend through the mylohyoid musculature and present in the neck as a plunging
ranula. Treatment of ranulas is by excision or marsupialization of the pseudocyst, often in conjunction with excision
of the sublingual gland. Mucoceles are also pseudocysts of
minor salivary gland origin and frequently rupture spontaneously. Recurrent or symptomatic mucoceles respond to
surgical excision.
Hemangioma is a proliferative endothelial lesion found
commonly in the head and neck. Their growth characteristics
include enlargement during the first year of life, followed by
spontaneous resolution. Surgical excision or treatment with
corticosteroids may be necessary in lesions that cause ulceration and bleeding, airway obstruction, cardiovascular
compromise, or platelet-trapping coagulopathy (KasabachMerritt syndrome). Longer-term systemic treatment with propranolol has recently been found to effectively reduce the size
of symptomatic hemangiomas and may work by promoting
vasoconstriction and downregulation of certain growth factors.27 Vascular malformations, including venous, arterial,
or arteriovenous malformations, rarely occur in the oral cavity
and pharynx and necessitate intervention only if they cause
pain, bleeding, ulceration, or heart failure. Management of
complicated cases is by surgical excision or sclerotherapy
for low-flow lesions (venous) and angiographic embolization
for high-flow lesions. Lymphatic malformation, formerly
known as lymphangioma or cystic hygroma, is congenital
and usually presents before 2 years of age. Histologically, lymphatic malformations consist of multiple dilated lymphatic
channels or may contain either capillary or venous elements
(venolymphatic malformations). Lymphatic malformations
have been characterized as microcystic, macrocystic, or mixed
based on their histologic patterns. Lymphatic malformations
721
can occur anywhere in the neck and may cause extensive

cosmetic deformity and functional problems in cases with involvement of the tongue, floor of mouth, mandible, or larynx.
Deep and macrocystic disease may be controlled with aspiration and sclerotherapy performed by interventional radiologists, whereas treatment of microcystic or more superficial
disease usually is surgical. Surgical resection of lymphatic malformations may be fraught with difficulty because they lack a
capsule and are infiltrative. During surgical excision, care
should be taken to avoid damaging nearby vital structures,
and debulking is an acceptable option to total radical excision
in many cases. Postoperative suction drains can be helpful in
preventing the recurrence of lymphatic drainage under skin
flaps. Coblation therapy and carbon dioxide laser therapy
have been used in superficial lymphatic malformations of
the tongue.
Foregut cysts are true cysts, lined with respiratory epithelium, that present in the floor of mouth and should be distinguished from dermoid cysts, lined with stratified squamous
epithelium and skin appendages, which may also be found
in this location. A thyroglossal duct cyst may rarely present
in the base of the tongue. Likewise, aberrant thyroid tissue,
lingual thyroid, presents as a purple mass in the tongue base.
Thyroid tissue in this location is usually hypofunctioning,
and affected children require thyroid supplementation. Other
aberrant rests of tissue, choristomas, consist of gastric, enteric,
or neural tissue of normal histology in an abnormal location.
Second branchial cleft derivatives will rarely present as a
cystic mass near the superior pole of the tonsil. Their extent
and associated tracts can be demonstrated on MRI. A Tornwaldt cyst is a blind pouch in the nasopharynx that represents
a persistence of an embryonic connection between the primitive notochord and the pharynx. Other benign nasopharyngeal masses include nasopharyngeal teratomas, dermoid
lesions (hairy polyp), and nasopharyngeal encephaloceles.
Most of these lesions are best evaluated by CT and/or MRI
to determine their extent and the presence of an intracranial
connection. Surgical excision is curative in most cases.
Squamous papillomas are benign slow-growing lesions
typically found on the soft palate, uvula, and tonsillar pillars
and are the result of infection with serotypes 6 and 11 of
the human papillomavirus (HPV). Because of concern that
these lesions could spread to the larynx or trachea, complete
surgical excision is usually recommended. Pleomorphic adenoma (mixed tumor) is a benign neoplasm of minor salivary
glands with a predilection for the palate, although it may
also be found in the lip and buccal mucosa. Treatment is with
surgical excision.
MALIGNANT LESIONS
Rhabdomyosarcoma, the most frequent soft tissue malignancy
of childhood, typically occurs in the 2- to 6-year-old group
and is derived from embryonic skeletal muscle.28,29 In the oral
cavity and oropharynx, it presents as a rapidly growing mass
in the tongue, palate, and uvula or cheek. These tumors metastasize early to local lymph nodes, lung, and bone. Surgical
therapy is limited to biopsy, excision of small lesions, or surgical salvage of treatment failures. The usual therapy includes
a combination of chemotherapy and radiation therapy.
Lymphoma of the oral cavity and oropharynx typically involves the lymphoid tissue of the Waldeyer ring and presents
722
PART V
HEAD AND NECK
as a mass of the tonsil or in the nasopharynx.30 The diagnosis

may be suspected by evidence of involved adenopathy in the
neck but is confirmed by surgical biopsy. Treatment is with a
combination of chemotherapy and radiation therapy.
Other rare malignant neoplasms of the oral cavity and
pharynx include malignant salivary gland tumors (mucoepidermoid carcinoma) and epidermoid or squamous cell carcinoma. This latter tumor has been reported in organ transplant
patients and adolescents who use snuff or chewing tobacco.31
Treatment is usually surgical depending on the site and extent
of involvement.
Larynx
------------------------------------------------------------------------------------------------------------------------------------------------
ANATOMY
With the exception of the hyoid bone, the major structural
framework of the larynx consists of cartilage and soft tissue.
The hyoid bone lies superior to the larynx and is attached
to it by the thyrohyoid membrane and strap muscles. The
hyoid bone is derived from the second and third branchial
arches. The cartilaginous structures of the larynx are composed of hyaline cartilage, with the exception of the epiglottis,
which is composed of elastic cartilage. The cartilaginous structures of the larynx develop from the fourth, fifth, and sixth
branchial arches. There are nine laryngeal cartilages, three that
are single (thyroid, cricoid, and epiglottis) and six that are
paired (arytenoid, cuneiform, and corniculate). The thyroid
cartilage consists of two quadrilateral cartilages that form
the anterior framework of the larynx. The cricoid cartilage
is the only completely cartilaginous structure in the airway
and provides posterior stability and a base of support for
the cricoarytenoid and cricothyroid joints.
The cricothyroid muscles are paired extrinsic laryngeal
muscles that serve to tilt the larynx down and forward, tensing
the vocal folds. Paired intrinsic musclesthe thyroarytenoid,
thyroepiglottic, and aryepiglottic musclesact as a sphincter
to close the larynx. The vocalis muscle comprises the internal
fibers of the thyroarytenoid muscle and attaches to the vocal
ligament. Action of this muscle serves to regulate the pitch of
the vocal ligament. The other set of paired muscles includes
the posterior cricoarytenoid, lateral cricoarytenoid, and
interarytenoid muscles. The posterior cricoarytenoid muscles
serve to abduct the vocal folds, whereas the cricoarytenoid
and interarytenoid muscles adduct the vocal folds.
The quadrangular membrane is a connective tissue
covering of the superior larynx that ends in a free margin along
the vestibular ligament of the false cord. The conus elasticus is
a membrane of elastic tissue that extends superiorly from
the cricoid cartilage to form the paired vocal ligaments, the
supporting structures of the vocal folds.
The blood supply of the larynx arises from the superior and
inferior laryngeal arteries. The former is a branch of the
superior thyroid artery, whereas the latter is a branch from
the thyrocervical trunk. The intrinsic muscles of the larynx
are innervated by the recurrent laryngeal nerve, which also
supplies sensory branches to the inferior larynx. The superior
laryngeal nerve has two branches: The external branch innervates the cricothyroid muscle, while the internal branch
supplies sensation to the superior larynx.
The larynx has multiple functions within the upper airway.

During respiration, it regulates airflow by opening during
inspiration. The posterior cricoarytenoid muscle contracts
with each inspiration to abduct the cords just before activation
of the diaphragm. The protective function of the larynx
produces two reflexes: cough and closure. Cough is important
to expel mucus and foreign objects. The closure reflex
serves to prevent aspiration of foreign matter. In addition to
closure, the larynx elevates during swallowing. Both closure
and elevation occur simultaneously along with relaxation of
the cricopharyngeus muscle during the swallow of a bolus. Finally, the larynx plays an important role in speech production
by generating sound. Vibration of the mucosa covering the
vocalis structures produces sound whose pitch and register
is altered by changes in tension, length, and mass of the underlying vocalis muscle and ligament.
The larynx of an infant sits much higher than that of an
adult. The cricoid is located at the level of C4, whereas the
tip of the epiglottis is at C1. The close approximation of
the epiglottis to the soft palate makes the infant an obligate
nose breather. By 2 years of age, the larynx has descended
to the level of C5 and reaches the adult level of C6 to C7
by puberty. The glottis of the newborn is 7 mm in the anteroposterior dimension and 4 mm in the lateral dimension. The
narrowest area of the infant airway, the subglottis, is approximately 4 mm in diameter.
UPPER AIRWAY ASSESSMENT

Symptoms of acute airway obstruction include dyspnea,
cough, vocal changes, dysphagia, and sore throat. Dyspnea
and rapid or labored breathing are indications of inadequate
ventilation and may be triggered by changes in PCO2 and PO2.
A stimulus anywhere in the airway may produce cough. It is
difficult to localize the site of the stimulus from the quality of
the cough. Changes in the childs vocal character, such as
hoarseness or a muffled or weak cry, may help in localizing
the area of obstruction. Dysphagia for solids and/or liquids
is often associated with airway obstruction. Depending
on the cause of airway obstruction, affected patients may
complain of sore throat.
The childs overall appearance is the first sign to be assessed
in airway obstruction, because airway status often dictates
how quickly further evaluation and intervention need to be
performed. The level of consciousness should be determined,
because the unconscious or obtunded patient may need immediate airway management. Along with cyanosis in a patient
without cyanotic heart disease, the presence of anxiety, restlessness, and diaphoresis are all ominous signs of impending
airway compromise. Other symptoms of airway obstruction
include tachypnea and substernal retractions. The child with
airway obstruction is often tachycardic. The presence of bradycardia is a late indicator of severe hypoxia. The presence of a
muffled cry often suggests obstruction at the level of the pharynx, whereas a barking cough is associated with laryngeal
inflammation and edema. Stertor is a snorting sound whose
origin is often in the pharynx. Stridor is noise produced by turbulent airflow in the laryngeal or tracheal airway. Inspiratory
stridor suggests turbulence at or above the glottis. Expiratory
stridor results from turbulent airflow in the distal trachea or
bronchi. Biphasic stridor suggests a tracheal or subglottic
source. A barking or croupy cough usually occurs when the
CHAPTER 55
subglottic trachea is involved. The degree and loudness of the

sound is not always indicative of the severity of obstruction,
because stridor can become softer just before complete
obstruction. Other important signs of airway obstruction
include drooling and use of accessory respiratory muscles.
In addition to determination of the childs physical status,
assessment of the degree of airway obstruction should include
an evaluation of the ventilatory status. Pulse oximetry provides
an immediate record of arterial oxygenation, while transcutaneous monitoring of carbon dioxide is a good indicator of
ventilation. The lateral neck radiograph remains the best study
for the initial evaluation of a child with airway obstruction,
because it demonstrates the anatomy from the tip of the nose
to the thoracic inlet. It can demonstrate findings of retropharyngeal or subglottic swelling from edema or infection and
identify free air in the soft tissue spaces. The anteroposterior
view of the neck is also helpful, specifically in defining areas of
narrowing, such as a steeple sign associated with subglottic
edema. A chest radiograph is also important in the initial
assessment to identify foreign bodies or other conditions such
as unilateral emphysema, atelectasis, or pneumonia that may
account for the childs respiratory compromise. If time
permits, a barium swallow or airway fluoroscopy may provide
additional information.
Additional airway evaluation may include a brief flexible
endoscopic examination. The nose is first sprayed with a
combination of 2% lidocaine and oxymetazoline, and the
child is gently restrained. The airway can be examined from
the nares to the glottis. Attempts to pass a flexible scope
through the glottis in a child with airway obstruction should
be avoided. Likewise, flexible endoscopy should be avoided in
a child with supraglottitis because of the possibility of precipitating complete obstruction. Children with suspected
airway pathology distal to the glottis or those in whom the
possibility that flexible endoscopy could compromise the
airway should undergo any airway examination in the operating room where rigid endoscopes and other airway equipment
is immediately available to secure the airway if necessary.
Nonsurgical intervention in the child with acute airway
obstruction may begin with just observation alone in a high
surveillance unit. Humidified oxygen administered by face
mask will improve PO2 and clearance of secretions. Racemic
epinephrine administered by nebulizer acts to reduce mucosal
edema and is useful in conditions such as laryngotracheobronchitis (infectious croup). Because its length of action lasts
30 to 60 minutes, treated patients should be observed for signs
of rebound for 4 to 6 hours after administration. Corticosteroids have been shown to have value in the management of
postintubation croup, adenotonsillar hypertrophy that results
from EBV infection, allergic edema, and spasmodic and viral
croup. Corticosteroids and propranolol have been used successfully in infants to treat subglottic hemangiomas.32,33
Other adjuvant therapies include antibiotics and inhalation
of helium/oxygen mixture (heliox). Although viral agents are
often responsible for inflammation in the larynx and trachea,
bacterial superinfection is also common. Because of the prevalence of penicillin-resistant organisms, broad-spectrum
antibiotics, including a higher-generation cephalosporin,
penicillinase-resistant penicillin, or beta-lactamase penicillin,
are useful in preventing or eradicating infection. Heliox is a
mixture of gas in which helium is used to replace nitrogen.
The advantage of the helium-oxygen mixture is that its low
723
density reduces air turbulence and gas resistance, allowing improved delivery of oxygen in patients with airway obstruction.
Nonsurgical airway management may include use of nasal
or oral airways, endotracheal intubation, and, rarely, transtracheal ventilation. Nasal airways of rubber or other synthetic
material can be easily inserted into the nose of most children
after adequate lubrication with a water-soluble lubricant.
Their best use is in cases where the pharynx is the site of obstruction. Oral airways are not as readily tolerated by children
and only serve as a brief solution to an airway problem. During
the 1970s, endotracheal intubation with polyvinyl chloride
tubes revolutionized the management of supraglottitis, and
even today intubation remains the mainstay of initial airway
therapy in most children with severe airway obstruction.
The size of the endotracheal tube used correlates with the
age of the child. The subglottis, the narrowest part of the infant
airway, typically admits a 3.5- or 4.0-mm inner-diameter tube.
The tube used in children older than 1 year can be roughly
estimated by using the following formula: tube size (age
in years/4) 4. Once the airway has been established, the tube
should be carefully secured and the child appropriately sedated and/or restrained, if necessary, to avoid accidental
self-extubation. Another method of airway management
should be considered in children with an unstable cervical
spine or in whom oral or neck trauma makes visualization
difficult. Transtracheal ventilation, insertion of a 16-gauge
needle through the cricothyroid membrane for the delivery
of oxygen, should be reserved for emergencies and used only
until a more stable airway can be obtained.
The surgical management of the child with acute airway
obstruction should begin with endoscopy. The larynx can
be visualized with one of a variety of pediatric laryngoscopes
and the airway secured with a rigid pediatric ventilating
bronchoscope of appropriate size. Once the airway is secured,
a more stable form of airway management can be used. Rarely,
in a child with acute airway obstruction, an airway cannot be
established, and a cricothyrotomy may need to be performed.
As in adults, this procedure avoids some of the risks of bleeding and pneumothorax inherent in a formal emergency tracheostomy. A small endotracheal or tracheostomy tube can be
inserted through the incision in the cricothyroid membrane,
but conversion should be made to a more stable airway as
soon as possible. Tracheostomy remains the preferred airway
in cases of acute obstruction in which a translaryngeal
approach is unsuccessful or must be avoided. The emergent
tracheostomy should be avoided if at all possible to lessen
complications of bleeding, pneumothorax, pneumomediastinum, subcutaneous emphysema, or damage to surrounding
structures. The incidence of these complications can be
reduced by careful attention to surgical technique, good lighting, and adequate assistance.
CONGENITAL LARYNGEAL ANOMALIES

Laryngomalacia is the most common cause of newborn stridor
and is caused by prolapse of the supraglottic structures (arytenoid cartilages, aryepiglottic folds) during inspiration
(Fig. 55-12). Symptoms typically appear at birth or soon
thereafter and include high-pitched inspiratory stridor, feeding difficulties, and, rarely, apnea or signs of severe airway obstruction. Gastroesophageal reflux disease (GERD) is common
in children with laryngomalacia and tends to worsen the
724
PART V
HEAD AND NECK
FIGURE 55-12 Laryngomalacia. This disorder classically presents as an

omega-shaped epiglottis. The arytenoid mucosa is redundant, and the
aryepiglottic folds are foreshortened. The result is a hooding of tissue over
the glottic inlet that leads to airway obstruction on inspiration.
airway symptoms, because it creates swelling of the posterior

cricoid region of the larynx. The diagnosis of laryngomalacia is
confirmed by flexible endoscopy of the larynx, and other airway pathology can be excluded with lateral neck, chest, and
airway fluoroscopy. Barium swallow radiography is helpful
to identify the presence of GERD. In most cases, laryngomalacia is self-limited and resolves by 18 months of age. Changes
in positioning and feeding, treatment of reflux, and, in some
neonates, use of monitoring may be necessary. In severe cases,
surgical intervention with either a supraglottoplasty (surgical
division with or without partial resection of the aryepiglottic
folds) or a tracheostomy may be necessary.
Tracheobronchomalacia is defined as collapse of the tracheobronchial airway. It may be congenital or acquired (from
long-standing intubation and infection) and may be segmental
or involve the entire tracheobronchial tree. Depending on the
extent and location, symptoms include low-pitched biphasic
or expiratory stridor and signs of respiratory compromise.
The diagnosis is usually made with endoscopy, although fluoroscopy of the airway may often demonstrate it. Treatment
ranges from observation in most cases to airway management
with a tracheostomy tube and positive-pressure ventilation
in severe cases. Tracheomalacia may be localized, especially
when associated with esophageal atresia, and aortopexy is occasionally the treatment of choice if due to extensive compression from vessels (see Chapter 69). Tracheal stents have also
been used for more extensive tracheomalacia.
Vocal fold paralysis is the second most common congenital
laryngeal anomaly (after laryngomalacia) and may be unilateral or bilateral. Congenital vocal fold paralysis may be caused
by neurologic abnormalities (hydrocephalus, Arnold-Chiari
malformation), birth trauma, or, rarely, in association with
neoplasms of the larynx or neck. Acquired vocal fold paralysis
may result from trauma or from neoplasms of the chest or
neck, or it may be iatrogenic, typically after surgery of the
neck, esophagus, or arch of the aorta. The risk of vocal cord
paralysis is higher in premature babies who have surgery
before they reach normal birth weights. Neonates with bilateral involvement typically present with high-pitched inspiratory or biphasic stridor but a good cry. Respiratory
compromise and feeding difficulties may accompany the stridor because the vocal cords cannot abduct and the resultant
airway is narrow. However, compensatory extralaryngeal muscles can help adduct the cords to produce a strong voice. In
infants with unilateral involvement, the airway may be adequate because the affected vocal cord remains partly lateralized at rest. Unlike the case of bilateral vocal cord
paralysis, the extralaryngeal muscles cannot cause the cord
to adduct upon vocalization or during a swallow. As a result,
these infants are at increased risk of aspiration and often have
breathy, weak voices. The diagnosis of unilateral or bilateral
vocal fold paralysis is confirmed with flexible or rigid endoscopy. Additional studies in the evaluation of patients with vocal fold paralysis include lateral neck and chest radiography,
barium swallow, and CT or MRI of the head and neck. Most
children with unilateral involvement can be observed. As they
grow, they may be candidates for vocal cord medialization procedures, whereby, agents such as Gelfoam or Teflon are
injected lateral to the cord to improve vocalization. Another
treatment option is ansa cervicalistorecurrent laryngeal
nerve anastomosis to reinnervate the affected cord. This increases the tone, bulk, and tension of the cord, but does
not restore normal mobility.34 Infants with bilateral vocal fold
paralysis often require a tracheostomy. In addition, infants
with associated feeding difficulties may need a gastrostomy.
In older children (4 or 5 years of age) with bilateral vocal cord
paralysis, a more permanent solution, such as a cordotomy or
arytenoidectomy, can be considered to improve the glottic airway and to allow for decannulation of the tracheotomy tube.
Congenital subglottic stenosis is the third most common
congenital laryngeal anomaly and is defined as a neonatal
larynx in a term baby without a history of prior instrumentation or intubation who fails to admit a 3.5-mm endotracheal
tube (Fig. 55-13). The underlying abnormality is a cricoid
cartilage that is either small or deformed. Children with Down
syndrome are at higher risk for this condition. Infants with
congenital subglottic stenosis present with inspiratory
FIGURE 55-13 Subglottic stenosis. Congenital and acquired stenosis

create airway obstruction, depending on the severity and type of stenosis.
Various forms of reconstruction are available (see Chapter 65).
CHAPTER 55
725
INFLAMMATORY DISEASE
OF THE UPPER AIRWAY
FIGURE 55-14 Subglottic hemangiomas typically arise from the posterior lateral aspect of the larynx. Small lesions may be managed conservatively, whereas lesions with aggressive growth patterns that do not
respond to propranolol or steroids require tracheotomy to bypass the laryngeal obstruction.
or biphasic stridor, barking cough, and other symptoms of

airway obstruction. The diagnosis is often suggested by narrowing of the subglottis on a lateral neck radiograph and
confirmed by endoscopy. Treatment depends on the severity
of symptoms and ranges from observation to laryngeal reconstruction to tracheostomy.
A child with a subglottic hemangioma presents with the
onset of progressive stridor during the first few months of life
(Fig. 55-14). Hemangiomas are proliferative endothelial
lesions that can form in the submucosa of the posterior
and lateral subglottis. Occasionally, they may involve the
subglottis in a circumferential pattern. Associated cutaneous
hemangiomas may be found in approximately 50% of
patients, but only 1% of patients with cutaneous lesions have
airway lesions. Symptoms are dependent on the amount
of airway compromise and include biphasic stridor, barking
cough, difficulty feeding, and other symptoms and signs
of airway obstruction. The diagnosis may be suggested on
a lateral neck radiograph but is confirmed with endoscopy.
Nonsurgical management of infants with a subglottic hemangioma includes observation or treatment with systemic
corticosteroids or propranolol. Surgical therapy includes
laser excision, open excision through a laryngofissure, or a
tracheostomy.
A laryngocele is an air-filled dilatation of the saccule of
the larynx that communicates with the laryngeal airway.
It may present internally into the posterior superior false
cord region or externally through the thyrohyoid membrane.
A saccular cyst is fluid filled and protrudes between the
true and false vocal folds. The diagnosis of this lesion is
confirmed endoscopically, and CT of the larynx is helpful
in assessing its extent and if it is fluid or air filled. Treatment
is with endoscopic marsupialization or excision through a
laryngofissure.
Laryngotracheobronchitis (viral croup) is an inflammation of

the subglottic airway caused by a variety of parainfluenza and
influenza viral agents. The infection may involve the entire
glottis and extend into the trachea and bronchi. Affected children fall typically into the 1- to 3-year-old group; males are
more commonly affected than females. Symptoms and signs
of viral croup include biphasic stridor, barking cough, and
hoarseness, often in association with a prodromal viral upper
respiratory tract infection. The diagnosis of croup is made
clinically, but endoscopic examination may help to exclude
other pathologic processes. Care should be taken not to instrument the subglottis, causing more swelling and inflammation and precipitating acute obstruction. Lateral neck
radiography demonstrates subglottic narrowing, whereas
anteroposterior neck films show a steeple sign, the result
of subglottic edema. Treatment of viral croup is typically supportive with humidification. Treatment with nebulized racemic epinephrine in the emergency department or hospital
setting often relieves symptoms; however, rebound of signs
may occur several hours later, and children should be monitored accordingly. A meta-analysis of randomized controlled
trials has shown treatment with glucocorticoids is effective
in improving symptoms within 6 hours, for up to 12 hours,
with significant improvement in croup scores, shorter hospital
stays, and less use of epinephrine.35 Severely affected children
may require intubation for respiratory failure (less than 5% of
affected patients). A smaller than normal tube should be chosen to avoid edema and scarring. In rare cases, a tracheostomy
may be required if the inflammation fails to resolve.
A child younger than 1 year of age with recurrent bouts of
croup should be suspected of having either congenital
subglottic stenosis or a hemangioma. Spasmodic croup is
the recurrence of crouplike symptoms in a child who is otherwise well. Fever is rarely present, and the attacks frequently
occur at night. Gastroesophageal reflux disease has been suggested as a possible inciting process. Treatment of spasmodic
croup is usually observant, although corticosteroids or antireflux medications may prove beneficial.
Supraglottitis (epiglottitis) is an infectious disease that involves the supraglottic larynx. In children, the most common
pathogen is Haemophilus influenzae type B (HIB), followed by
S. pneumoniae and S. aureus. The incidence of supraglottitis
in children has diminished markedly since the introduction
of the conjugated HIB vaccine in the early 1990s.36 However,
HIB-related supraglottitis continues to occur in children who
have been vaccinated, with a reported 2% vaccine failure rate.
Alternatively, S. pneumoniae, S. aureus, and viruses are more
likely to cause supraglottitis in adolescents and adults.
Children who develop supraglottitis are somewhat older than
those seen with croup in the 2- to 6-year-old group. Symptoms
and signs have a rapid onset, progress quickly to frank airway
obstruction, and include stridor, dysphagia, fever, muffled voice,
and signs of systemic toxicity. Affected children frequently sit
and assume the sniffing position in an attempt to maximize their
airway. Intraoral or endoscopic examination should be avoided
in suspected patients because of concern for precipitating complete obstruction. Lateral neck radiography demonstrates a classic
thumbprint sign of the epiglottis but should only be obtained
if facilities are present in close proximity to secure the airway.
726
PART V
HEAD AND NECK
Prompt airway management is essential in children with

supraglottitis. In severe cases, the childs airway should be
secured in either the emergency department or operating
room with team members, including a pediatrician, anesthesiologist, critical care physician, otolaryngologist, or pediatric
surgeon or others familiar with the pediatric airway. After
inducing the child with general anesthesia, the airway should
be intubated. Examination of the supraglottis may be made,
and cultures of the larynx and blood are obtained. Equipment
to perform a tracheostomy should be readily available. The
child should remain intubated for 24 to 72 hours and should
be supported with intravenous fluids and antibiotics that treat
antibiotic-resistant H. influenzae, S. pneumoniae, and S. aureus
(third-generation cephalosporins or ampicillin-sulbactam).
Bacterial tracheitis (membranous croup) often occurs as
a complication of another infection, such as measles, varicella,
or other viral agents. The most common organisms include
S. aureus, GABHS, M. catarrhalis, or H. influenzae. It can occur
in any age child and present with stridor, barking cough, and
low-grade fever. Symptoms and signs then progress to include
high fever and increasing obstruction and toxicity. The diagnosis may be suspected by diffuse narrowing of the tracheal
air shadow on chest radiograph but is confirmed by endoscopic examination in the operating room. Purulent debris
and crusts can be removed at this time. Cultures of secretions
and crusts may be helpful in guiding intravenous antibiotic
therapy that should be aimed initially at the usual pathogens.
The airway should be secured with an endotracheal tube or,
rarely, a tracheostomy. Repeat endoscopic examination of
the airway may be warranted to continue debridement and
to determine the feasibility of extubation.
CHRONIC AIRWAY OBSTRUCTION

The chronic management of subglottic stenosis and other
prolonged airway disorders is discussed in Chapter 65.
BENIGN LARYNGEAL NEOPLASMS

Recurrent respiratory papillomatosis (RRP) is the most common benign neoplasm of the larynx in children. Squamous
papillomas involve the larynx and, occasionally, the trachea
and lower respiratory tract as exophytic lesions. Because of
its recurrent nature, RRP causes morbidity and, rarely, mortality resulting from malignant degeneration. Patients may be
almost any age, but the disease is more aggressive in children.
Human papillomavirus (HPV) subtypes 6, 11, 16, and 18 have
all been identified within papilloma tissue. The first two
subtypes have been associated with genital warts, whereas
the latter two have been associated with cervical and laryngeal
cancers. The exact mechanism of HPV infection in the larynx
remains unknown. In most cases, transmission of virus to the
child is thought to occur via vaginal birth in a mother with
cervical HPV infection or warts. However, children can still
get RRP even when born by cesarean section.
Children afflicted with RRP present initially with hoarseness but may also have symptoms and signs of airway obstruction, including stridor. Lateral neck radiography may suggest
laryngeal involvement, but the diagnosis is confirmed by
direct laryngoscopy and biopsy (Fig. 55-15). In addition to
the trachea and bronchi, squamous papillomas may also be
found in the oral cavity.
FIGURE 55-15 Recurrent respiratory papillomatosis. Severe papillomatosis may completely obstruct the larynx. Papillomas are characterized
by malignant degeneration and aggressive growth patterns.
Surgical excision is the mainstay of therapy in patients with

RRP. In the past, papillomas were excised using the carbon
dioxide laser. More recently, the laryngeal microdebrider has
become the preferred method of excision in many centers.
In aggressive cases with swift recurrence and accompanying
airway obstruction, tracheostomy may be necessary for airway
management, although tracheostomy has been implicated in
the spread of disease to the trachea and lower respiratory tract.
Medical adjuvant therapy that has been used with mixed
results includes interferon, photodynamic therapy with dihematoporphyrin ether, indole-3-carbinol, or antiviral agents
such as cidofovir.
Other benign laryngeal neoplasms are rare and include
connective tissue tumors such as chondromas or fibromas,
neurogenic tumors such as neurofibromas, or granular cell
tumors and other cell types such as hamartomas or fibrous histiocytomas. Malignant tumors of the larynx are also rare and
include squamous cell carcinoma and a variety of epithelial
and connective tissue malignancies, such as spindle cell carcinoma, rhabdomyosarcoma, mucoepidermoid carcinoma, and
chondrosarcoma. Metastatic tumors and lymphoma may also
rarely involve the larynx in children. Diagnosis is suspected
by the sudden appearance of stridor, hoarseness, and airway
obstruction and confirmed by biopsy. Treatment is dependent
on cell type and may include surgical excision, radiation
therapy, and/or chemotherapy.
Neck
------------------------------------------------------------------------------------------------------------------------------------------------
ANATOMY
The surgical anatomy and embryology of the neck is discussed
in Chapter 59.
CLINICAL EVALUATION
The initial examination of a disease or disorder of the neck
begins with a thorough history. A detailed history can often
serve to focus the differential diagnosis of a neck disorder.
The age of the child is an important first consideration.
The appearance of a neck mass in an infant often suggests a
CHAPTER 55
congenital disorder, whereas the sudden appearance of a mass

in an adolescent might suggest a malignant process. Inflammatory diseases of the neck may occur in any age group but
typically mirror the incidence of upper respiratory tract infections in children. Growth and temporal relationships are
often important clues to a diagnosis. Neck masses that grow
rapidly suggest either an inflammatory or malignant process,
whereas slow-growing masses are typically benign. A history
of systemic infection elsewhere in the body or recent travel or
exposure to farm animals often points to an infectious origin.
A history of trauma to the neck may explain the sudden
appearance of a neck mass. Likewise, changes in the size of
a neck mass with eating may suggest a salivary gland origin.
Vascular lesions enlarge with straining or crying. Finally, systemic symptoms of fever, weight loss, night sweats, or fatigue
in association with the sudden development of a neck mass
may indicate a malignant process.
The physical examination of a child with a neck mass
should begin with a comprehensive examination of the entire
head and neck. Because the vascular, neural, and lymphatic
patterns of the head drain into the neck, the source of neck
disorders may be found in the head. Depending on the differential diagnosis, a physical examination of the entire body,
including an assessment of lymph nodes in the groin and
axillae and the presence of an enlarged spleen or liver, is essential. Palpable lymph nodes in the neck of children are a
common finding, but lymph nodes larger than 2 cm fall outside the range of normal hyperplastic nodes and should be
either monitored or investigated. The sudden appearance of
large nodes in either the posterior cervical or supraclavicular
regions may suggest a malignancy, especially if unilateral.37
The consistency of a neck mass is also important in narrowing
the differential diagnosis. Hard masses tend to be associated
with either infection or malignancy. Fixation of a neck mass
to skin or nearby structures is also suggestive of a malignancy.
Cysts or abscesses tend to have a characteristic feel on palpation and are usually ballotable, and the overlying skin may be
inflamed if infected. Depending on the differential diagnosis
after a history and physical examination, radiologic studies
may be useful. A lateral neck radiograph may demonstrate
an abnormality of the nasopharynx, retropharynx, or cervical
spine. Likewise, a chest radiograph may identify a malignancy,
sarcoidosis, or tuberculosis. Infection or a neoplastic process
in the sinuses may appear on a sinus series. CT and MRI are
useful in the evaluation of a neck mass. Demonstration of
hypodensity on CT suggests an inflammatory or necrotic
process. Ring enhancement of a hypodense region on a contrast CT scan is indicative of an abscess. MRI is excellent for
distinguishing fine detail within soft tissue and in the evaluation of vascular lesions of the neck. Finally, ultrasonography is
helpful in distinguishing solid and cystic masses and may be
the only imaging modality required in the assessment of neck
masses. Use of ultrasonography preoperatively in patients
with a thyroglossal duct cyst is also a simple and economic
way to assess the presence of normal thyroid tissue when it
is not easily felt. Ultrasonography should be used in the assessment of any thyroid mass, while thyroid scanning is
now thought to be of limited value in the pediatric age group.
Selected laboratory studies may be helpful in the evaluation
of a child with a neck disorder. A complete blood cell count
with differential may identify patients with either a malignancy or systemic infection. Serologic testing for EBV or
727
cytomegalovirus infection, toxoplasmosis, or cat-scratch disease may be diagnostic. Thyroid function testing is essential
in any child with a suspected thyroid disorder. Finally, collection of urine for catecholamine metabolites (vanillylmandelic
acid) may assist in the diagnosis of neuroblastoma.
If the diagnosis remains in doubt at this point, incisional or
excisional biopsy may be indicated. Biopsy provides material
for pathologic examination, culture, and other more sophisticated testing if necessary. Fine-needle aspiration of a neck
mass in children for suspected malignancy is not as reliable
as in adults.
CONGENITAL TRACTS AND CYSTS

Congenital sinuses and cysts are discussed in Chapter 59.
INFLAMMATORY AND INFECTIOUS MASSES

Viral adenitis is the most common infectious disorder to involve the neck in children. Enlarged or hyperplastic lymph
nodes are frequently the result of viral upper respiratory tract
illnesses. Common pathogens include rhinovirus, adenovirus,
and enterovirus, but measles, mumps, rubella, varicella, EBV,
and cytomegalovirus may also cause lymphadenopathy. The
diagnosis is often suspected by other findings in the history
or physical examination and can be confirmed by serologic
testing. Acute human immunodeficiency virus infection may
present, as do other viral syndromes, with fever, headache,
malaise, gastrointestinal symptoms, and a neck mass.
The usual source of bacterial cervical adenitis is the pharynx. Causative organisms are often streptococcal or staphylococcal species. Patients present with systemic symptoms of
fever and malaise in addition to a neck mass that is diffusely
swollen, erythematous, and tender. In contrast to viral adenitis, which is frequently bilateral, bacterial infections of the
neck are usually unilateral. CT with contrast medium enhancement may be helpful in the evaluation of large infectious
neck masses that may contain an abscess cavity (Fig. 55-6, B),
although ultrasound examination can provide similar information without radiation. Needle aspiration of suspected
infectious masses may provide material for culture and
decompress the mass.
Most children with bacterial cervical adenitis respond to
oral antibiotics chosen to cover group A streptococci and
S. aureus, but those who fail to improve require IV antibiotics.
The initial choice of antibiotic is important. A recent study has
shown a predominance of S. aureus (63%) compared with
Streptococcus group A isolates (22%) obtained from those abscesses requiring surgical drainage. Of those with S. aureus
infections, 27% were methicillin-resistant Staphylococcus
aureus (MRSA), and all of these were sensitive to clindamycin
and trimethoprim-sulfamethoxazole. Of the methicillinsensitive Staphylococcus aureus (MSSA) isolates; 100%, 86%,
and 82% were sensitive to trimethoprim-sulfamethoxazole,
clindamycin, and ciprofloxacin, respectively.38
Cat-scratch disease is caused by Bartonella henselae infection. The clinical picture includes the sudden appearance of
unilateral lymphadenopathy after a scratch from a cat. Fever
and malaise may be accompanying symptoms in many cases.
Serologic testing for antibodies to Bartonella is diagnostic.
Cat-scratch disease is usually self-limited, although some
728
PART V
HEAD AND NECK
benefit has been described with the use of erythromycins and

other antibiotics.39
In the past, most mycobacterial infections have been
caused by atypical organisms, such as Mycobacterium aviumintracellulare, M. scrofulaceum, M. bovis, or M. kansasii. These
organisms are commonly found in the environment in dirt,
dust, water, and sometimes in food. In the past decade or
so, mycobacterial tuberculosis has made a resurgence as
the pathogen responsible for a neck infection. A chest radiograph should be obtained if M. tuberculosis is suspected.
M. tuberculosis is usually associated with abnormal chest
radiograph and the presence of a positive tuberculous skin
test. Tuberculosis should be treated with appropriate antituberculous chemotherapy.
Children with nontubercular (NTM) or atypical mycobacterial infections have weakly positive or negative skin tests and
present with a typical indolent course consisting of slowly
growing, nontender nodes in the preauricular, intraparotid,
submandibular, or posterior triangle regions that do not respond to antibiotics. Systemic symptoms are rare. After several
days to weeks, the skin overlying the node typically assumes a
violet color, and the area may become fluctuant and tender to
palpation. The diagnosis is mainly clinical, because the organism will often take several weeks to grow in culture, and acidfast bacilli are not always demonstrated. The treatment is
surgical and consists of excision of the involved node(s).
Combination therapy using clarithromycin and rifabutin
may be effective but requires a prolonged course; it is generally
reserved for recurrences or nodes that are not safely accessible
by surgical approach.
Rarely, the neck may be involved with infections such as
tularemia, brucellosis, actinomycosis, plague, histoplasmosis,
or toxoplasmosis. Inflammatory disorders that may affect the
neck include Kawasaki syndrome, sarcoidosis, sinus histiocytosis (Rosai-Dorfman disease), Kikuchi-Fujimoto disease, and
PFAPA syndrome (periodic recurrent fever).
MALIGNANT NEOPLASMS
Thyroid malignancies are discussed in Chapter 58, and
malignant lymphadenopathies in Chapters 38 and 57.
Neurofibromatosis is a benign disorder that in some forms
(plexiform) may infiltrate surrounding tissues. For this reason,
CT and/or MRI are vital in the preoperative evaluation of these
lesions. When the tumors are multiple and extensive, surgical
resection is reserved for symptomatic lesions, because complete excision is usually impossible without compromising
neurovascular structures. Neuroblastoma is a malignancy that
develops from neural crest cells and may present as a solitary
tumor or as lymphadenopathy. Clinical staging determines the
mode of therapy that includes surgery, chemotherapy, and
radiation therapy.
Rhabdomyosarcoma rarely presents as a primary tumor in
the neck, more often being found as a primary tumor in the
orbit, temporal bone, or nasopharynx. The diagnosis is made
by biopsy, and patients are staged according to involvement.
Treatment includes surgery, chemotherapy, and radiation
therapy.
Malignancies of almost any type and location in the body
can metastasize to the neck. The most common are thyroid
malignancies. In adolescents, carcinomas, especially those
arising in the nasopharynx, may spread to the neck lymphatics.
expertconsult.com.
process is similar for all of the major salivary gland embryogenesis.2 During early gestation, the parotid ductules begin
to grow around the facial nerve and its branches. This is of
great clinical and surgical significance, because the facial nerve
may be compressed or invaded by parotid gland lesions, or its
branches may be injured during parotid gland surgery.3
Anatomy and Physiology

------------------------------------------------------------------------------------------------------------------------------------------------
CHAPTER 56
Salivary Glands
Douglas Sidell and Nina L. Shapiro
Salivary gland disorders are rare in children. They often present as a painful or, less commonly, a painless swelling in
the affected gland. Disease processes may be of infectious,
inflammatory, systemic, autoimmune, congenital, neoplastic,
or traumatic origin.1 Treatment is guided by the medical or
surgical nature of the specific disease process.
Classification
------------------------------------------------------------------------------------------------------------------------------------------------
Salivary glands may be divided into major and minor categories. The former category includes the parotid, submandibular, and sublingual glands, all of which are paired structures
with their own well-defined anatomy, including blood supply
and ductal drainage. Their function is augmented and facilitated by the minor salivary glands, which include the
mucus-secreting tissues in the buccal mucosa, palate, mucosal
surfaces of the lips, and floor of the mouth.
Embryology
------------------------------------------------------------------------------------------------------------------------------------------------
In the sixth week of gestation, solid epithelial buds of ectoderm from the developing mouth invaginate into the surrounding mesenchyme. A groove from this invagination
develops into a tunnel, which subsequently forms branches
of salivary ductal tissue. The mesenchymal tissue forms
the capsule and connective tissue of the salivary glands. This
The parotid gland is located in the space between the external

auditory canal and the mandible. Its main duct (Stensen duct)
crosses the masseter muscle and opens in the buccal mucosa
at the level of the second maxillary molar. The deep lobe of
the parotid gland lies medial to the facial nerve branches
and the mandible. Deep lobe parotid gland masses may extend
to the parapharyngeal space and present as intraoral growths.
The parotid gland is the only salivary gland containing lymph
nodes, which may become apparent during certain pathologic processes, such as atypical mycobacterial adenitis (see
Chapter 57). Accessory parotid tissue is present in some children and in approximately 20% of adults. It can occur superficial to the masseter and is often mistaken for a neoplasm.4
The submandibular gland is located in the submandibular triangle of the neck. The main submandibular duct (Wharton
duct) exits the gland at a right angle and enters the mouth just
lateral to the midline lingual frenulum. The sublingual gland is
located at the lateral aspect of the floor of the mouth.1
The salivary glands serve to lubricate the mouth for
hygiene, speech, and deglutition; to moisten food for taste
and mastication; and to initiate early starch digestion with
a-amylase.1 These processes may be initiated by various
stimuli, including cerebral, visual, olfactory, or gustatory.
Pathology
------------------------------------------------------------------------------------------------------------------------------------------------
The majority of salivary masses in children are congenital

vascular lesions, with hemangiomas seen in 50% to 60% of
salivary gland masses and lymphatic malformations in approximately 25%.5 Acquired lesions are of inflammatory, infectious,
autoimmune, traumatic, or neoplastic origin. Salivary gland
swelling is characteristic of nearly all glandular pathologic processes, and may be accompanied by pain, tenderness, or abnormal ductal discharge.4 Advanced stages of disease may lead to
cranial nerve involvement with resultant paresis or paralysis.
Diagnosis
------------------------------------------------------------------------------------------------------------------------------------------------
HISTORY
A careful history should focus on the duration of the lesion, its
bilateral or unilateral presentation, and whether there is any
symptom fluctuation associated with eating. A complete
medical history is essential, because the salivary glands may
be involved in several systemic conditions.
PHYSICAL EXAMINATION
The physical examination should include careful inspection of
the overlying skin, both local and distant, to evaluate for any
cutaneous hemangiomas, as well as of the intraoral mucosa to
729
730
PART V
HEAD AND NECK
evaluate for intraoral extension of the mass. Longitudinal duct

massage will assess for duct obstruction or purulent material
in the saliva. Benign salivary lesions tend to be mobile, soft,
and spongy, whereas malignant and infectious lesions are
more often fixed and firm on palpation.
DIAGNOSTIC IMAGING
Plain radiographs of the salivary glands are helpful in detecting
salivary duct calculi or diffuse glandular calcification.1 Sialography is useful in identifying strictures, sialectasis, calculi, or saccular dilatation (Fig. 56-1).6 High-resolution ultrasonography
is a useful, noninvasive technique in the diagnosis of sialectasis
and salivary gland calculi.7 The addition of color-flow Doppler
imaging can provide accurate information regarding the consistency of the lesion and its vascular pattern (Fig. 56-2).8
Computed Tomography
Computed tomography (CT) is an excellent diagnostic modality for assessing both the pathology and anatomy of the
salivary glands. It can aid in distinguishing intrinsic or extrinsic lesions. Use of an intravenous contrast agent can help
detect an abscess or delineate the vascularity of congenital
and acquired vascular lesions.1 These features help in both
medical and surgical planning.9,10
Magnetic Resonance Imaging
Magnetic resonance imaging (MRI) provides the best soft tissue detail of the salivary glands, and it is the only imaging
technique that can delineate the facial nerve anatomy within
the parotid glands. Signal intensity variations (T1- and T2weighted images) provide additional valuable information
regarding the nature of the mass.11,12
FIGURE 56-2 Doppler ultrasound study shows vascular pooling in a

parotid hemangioma.
BIOPSY
Fine-needle aspiration (FNA) biopsy is an excellent tool in the
diagnostic evaluation of salivary gland masses.13,14 The overall
diagnostic accuracy is 84%, with a sensitivity and specificity
approaching 92% for parotid lesions.1517 Obtaining an adequate needle biopsy specimen may preclude the necessity for
surgical therapy or aid in surgical planning. Understandably,
the accuracy and dependability of FNA rely heavily on the
expertise of the cytopathologist and may vary based on institution or clinical setting.13 For deeper salivary gland tumors,
fine-needle aspiration may be performed under image guidance. Open excisional biopsy is the definitive tool for investigation and may be curative. If the size and location of the
lesion are favorable, the entire tumor may be resected intact
with a clear surrounding cuff of normal tissue. The diagnosis
of Sjogren syndrome may be obtained by incisional biopsy of
the minor salivary glands of the labial mucosa, or, alternatively, of the parotid gland.18
SIALENDOSCOPY
FIGURE 56-1 Sialogram shows saccular sialolithiasis of the parotid gland.
Sialendoscopy involves semirigid endoscopy and microinstrumentation to evaluate and treat certain disorders of the parotid
and submandibular glands. Although relatively new, this technique is increasing in popularity and has been demonstrated to
effectively classify and treat ductal lesions, such as stricture and
calculi. Sialendoscopy has been reported to produce a greater
sensitivity in detecting salivary calculi than conventional radiography, MRI, or ultrasonography. Duct marsupialization and
intraductal calculi retrieval have been demonstrated, allowing
for the early treatment of some lesions without the requirement
for open surgery.19,20
CHAPTER 56
Inflammatory Disease
SALIVARY GLANDS
731
CHRONIC SIALADENITIS
------------------------------------------------------------------------------------------------------------------------------------------------
VIRAL SIALADENITIS
Acute inflammation of the salivary glands may be viral in up to
85% of cases, and the majority of viral sialadenitis involves the
parotid glands. Viral infections are characterized by a benign
self-limiting course over 2 to 3 weeks. Antipyretics, analgesics,
and anti-inflammatory agents may be given for relief of symptoms. Causative organisms include coxsackievirus A and
echovirus. Before the nearly universal implementation of the
mumps vaccine in 1967, mumps virus (paramyxovirus) was
the most common cause of acute parotid inflammation in
children.2124 Other potential causes include cytomegalovirus
(CMV), which is most commonly seen as a component of
disseminated CMV infection in infants and young children,25
and Epstein-Barr virus (EBV), which in healthy children is
associated with infectious mononucleosis and in chronically
ill children may be associated with human immunodeficiency
virus (HIV) infection.26,27
BACTERIAL SUPPURATIVE SIALADENITIS

Acute suppurative sialadenitis most often presents as rapidly developing pain, swelling, and occasional ductal discharge, with
associated fever and poor oral intake. It is primarily seen in
the parotid glands and less commonly in the submandibular
or sublingual glands. The causative organisms are usually
Staphylococcus aureus and Streptococcus viridans.28 Acute sialadenitis often occurs in dehydrated patients because of decrease in
salivary flow and dry oral mucosa.29 Most cases will respond to
antistaphylococcal antibiotics, with careful attention to hydration, oral hygiene with mouthwashes, warm local compresses,
and sialogogues, such as sour lemon drop candies, to stimulate
salivary flow. Rarely, despite treatment, the infected tissue will
coalesce to form an abscess. Treatment of a salivary gland abscess includes intravenous antibiotics and surgical drainage.30
If an abscess develops in the parotid gland, fascial incisions parallel to the course of the facial nerve are made to drain the
abscess. If the facial nerve is paretic preoperatively, abscess
drainage will usually facilitate resolution of nerve function.1
Chronic sialadenitis is the most common cause of inflammatory salivary gland disease in children and may lead to
structural changes in the gland and acinar destruction
(Fig. 56-3). There are obstructive and nonobstructive causes
of this condition. Obstruction is caused by ductal stenosis,
which may be congenital, caused by a stone, or result from
chewing or biting the ductal opening. In such cases, the duct
should be probed and stented for continuous drainage.
Nonobstructive chronic sialadenitis may occur in conjunction
with metabolic disorders, such as Sjogren syndrome, or
chronic granulomatous disease, such as sarcoidosis, tuberculosis, or atypical mycobacterial disease.
The treatment of obstructive sialadenitis is initially conservative, with warm compresses and anti-inflammatory medications. Ductal dilatation or marsupialization may be necessary
for recalcitrant disease. Gland excision is rarely required.
Sialolithiasis (salivary gland or duct calculi) is rare in children and occurs in the submandibular gland in 80% of cases.
When the stone is located at the distal salivary duct, it may be
excised by a simple incision at the ductal orifice. Temporary
stent placement may be necessary. Rarely, a large calculus will
be located in the proximal salivary duct or salivary gland
parenchyma and may require complete gland excision with
the stone-containing duct.
Cystic Disease
------------------------------------------------------------------------------------------------------------------------------------------------
Cystic disease may be acquired, congenital, or traumatic. Congenital cystic disease may occur in the salivary glands, but it is
not of salivary gland origin. Work type I and type II first branchial cleft cysts may present as parotid gland masses, and
depending on the orientation of the tract, may have accompanying otorrhea.31,32 Congenital lymphatic malformations may
also present in the parotid, submandibular, or sublingual
glands. Large, bilateral intraparotid lymphoepithelial cysts
are characteristic of HIV infection.1 Small mucous retention
cysts may present in the minor salivary glands of the labial
or buccal mucosa; these cysts usually result from single or
FIGURE 56-3 A, Sialogram of patient with history of recurrent parotid swelling. Note normal ductal system with early diffuse punctate sialectasis.
B, Parotid gland swelling between acute attacks of inflammation.
732
PART V
HEAD AND NECK
FIGURE 56-4 Floor of mouth ranula with posterosuperior lingual elevation.
repeated local trauma to the minor salivary glands and may

lead to recurrent local mucosal swellings. If they do not
resolve spontaneously, they will require complete excision.
Local drainage or marsupialization will result in recurrence.
RANULA
A ranula is a mucus extravasation cyst of the sublingual gland.
Initial presentation is a bluish, cystic mass at the floor of mouth,
which may lead to lingual elevation or difficulty with deglutition.
They may extend to the neck through the mylohyoid (plunging
ranula) (Fig. 56-4). Surgical management is controversial and
ranges from simple transoral marsupialization to combination
transoral-transcervical approaches.33,34 Despite controversy,
recurrence rates as high as 67% have been described with marsupialization alone. Recent evidence, derived from the largest
review to date, suggests that the excision of the ipsilateral sublingual gland produces the lowest incidence of recurrence.35
During sublingual gland excision, care must be taken to avoid
Wharton duct injury, and it can be avoided by placing a lacrimal
probe in the duct intraoperatively. The lingual nerve must also be
meticulously dissected just deep to the sublingual gland.
Neoplasms
------------------------------------------------------------------------------------------------------------------------------------------------
Salivary gland neoplasms are extremely rare in children and

comprise less than 1% of all pediatric neoplasms.5,28,36,37 Less
than 5% of salivary gland neoplasms occur in patients younger
than 16 years of age.38,39 However, when present, a pediatric
salivary tumor must be assessed to rule out malignancy.4042
In the pediatric population, greater than 90% of salivary
neoplasms occur in the parotid gland.43 Caution should be
exercised when evaluating adolescents, because imaging characteristics change over time as the gland is replaced with fat.
Occasionally, this can cause benign disease to be mistaken for
an infiltrative tumor on CT.4
BENIGN NEOPLASMS
AND MALFORMATIONS
Benign neoplasms account for 60% of salivary tumors in children and are most commonly vascular in origin.5 Vascular
lesions include hemangiomas and lymphatic malformations,
which are both congenital in origin (Fig. 56-5).
FIGURE 56-5 Vascular malformation of the parotid gland, showing large,

irregular vascular spaces. (Hematoxylin-eosin stain, 50.)
Hemangiomas
Hemangiomas are one of the most common salivary (primarily
intraparotid) neoplasms in children, with infantile hemangiomas comprising greater than 90% of all salivary lesions in children less than 1 year of age.4 Hemangiomas usually present in
infancy as a soft, nontender parotid swelling, with or without
associated pigmented cutaneous lesions.44 Diagnosis is usually
confirmed with ultrasonography, which demonstrates a lobulated, hypervascular mass, with arterial and venous signals visible on color-flow Doppler.4,45 MRI may also be useful but is
rarely required. Parotid hemangiomas often resolve spontaneously and do not require treatment. If they are rapidly growing
or are causing functional impairments, such as facial nerve
weakness, external auditory canal obstruction, or cutaneous
breakdown, systemic therapy such as corticosteroids, propranolol, or interferon alfa-2a or alfa-2b are viable options to inhibit
vascular growth and promote involution of the tumor.4649
Lymphatic Malformations
Lymphatic malformations are less common than hemangiomas.
They do not undergo spontaneous involution, are usually present at or soon after birth, and grow with the growth of the
child.44 They are not true salivary lesions, but they are commonly seen in the submandibular and parotid region in infants
and young children.50 Lymphatic malformations are susceptible
to infection, with potential for cellulitis, intralesional bleeding,
abscess formation, or lymphangiomatous extension to the floor
of mouth or trachea with airway compromise. Treatment modalities have been an area of much investigation. Surgical resection
must be complete to obviate recurrence. This is often difficult,
because of the fragility of the tumor lining, its infiltrative nature,
and its proximity to major vessels and branches of the facial
nerve.51,52 In an effort to avoid surgical morbidity, success with
intralesional sclerotherapy has been demonstrated, resulting in
reduction in tumor size and minimal scarring or recurrence.53
Pleomorphic Adenoma
Pleomorphic adenomas (benign mixed tumors) are the most
common nonvascular benign salivary tumors in children
(Fig. 56-6).42,54 They present as firm, rubbery masses, most
often in the parotid gland, with an average age at presentation
of 9.5 years within the pediatric population.54,55 The tumor
CHAPTER 56
SALIVARY GLANDS
733
involving the submandibular or minor salivary glands, concomitant neck dissection and adjuvant radiation therapy is
recommended by many institutions.36,64,65
Acinic Cell Carcinoma
Acinic cell carcinomas present in a similar fashion as mucoepidermoid carcinomas. They tend to be low grade, and treatment
is similar to that of mucoepidermoid carcinoma (Fig. 56-7).
Adenoid Cystic Carcinoma
Adenoid cystic carcinoma is a rare, high-grade salivary gland
tumor. Perineural invasion may result in cranial nerve deficits.
There is a high incidence of regional nodal metastases, as well
as distant metastases to the lungs, liver, and bone. Treatment
includes wide surgical resection, neck dissection, and adjuvant radiation therapy.61
FIGURE 56-6 Pleomorphic adenoma (mixed tumor) of the parotid gland.
Epithelial areas are mixed with myxomatoid and chondroid stroma.
(Hematoxylin-eosin stain, 50.)
presents as a painless, slowly growing mass and is rarely infiltrative.56 They have variable echogenicity on imaging, with increased heterogenicity seen in larger lesions secondary to
necrosis or cystic changes.4 Treatment of superficial lobe
tumors includes superficial parotidectomy with facial nerve
dissection and preservation. Recurrence rates have been
reported to be up to 40%; so, long-term follow-up is recommended.57,58 Simple excisional biopsy should be avoided,
because it is associated with a higher recurrence rate. Rarely,
recurrent pleomorphic adenomas may undergo malignant
degeneration.59
Rhabdomyosarcoma
Rhabdomyosarcoma may present as a parotid mass. Histologic
variants include undifferentiated and embryonal types
(Fig. 56-8). Treatment and outcomes depend on tumor stage
and may include wide local surgical resection with radiation
and chemotherapy.
Monomorphic Adenomas
Monomorphic adenomas are rare in children. Histologically,
they may resemble adenoid cystic carcinoma, a highly aggressive malignant salivary tumor.60 Treatment includes complete
surgical resection and close long-term follow-up.
Papillary Cystadenoma Lymphomatosum
(Warthin Tumor)
These tumors are most commonly seen in men and are often
bilateral parotid lesions. They may rarely present as benign
parotid tumors in children.1 Treatment is similar to that for
pleomorphic adenomas.
FIGURE 56-7 Acinic cell carcinoma of the parotid gland showing

invasive proliferation. (Hematoxylin-eosin stain, 100.)
MALIGNANT NEOPLASMS
Malignant salivary neoplasms are rare in children. When present, they are often low-grade lesions, located most commonly
in the parotid gland, and have a female preponderance.54
Diagnostic evaluation should include CT or MRI and fineneedle aspiration biopsy. Treatment is surgical, with complete
tumor excision with clear margins. Invasive malignancies may
require sacrifice of the facial nerve branches. Postoperative
radiation therapy is recommended for high-grade lesions.61,62
Mucoepidermoid Carcinoma
Mucoepidermoid carcinoma is the most common pediatric
salivary malignancy and is most commonly low grade and
located in the parotid gland. Surgery is usually curative.39,63
For high-grade mucoepidermoid carcinomas, or those
FIGURE 56-8 Rhabdomyosarcoma of the parotid gland showing

spindle cell sarcoma with myogenous differentiation. (Hematoxylin-eosin
stain, 100.)
734
PART V
HEAD AND NECK
Surgical Considerations
------------------------------------------------------------------------------------------------------------------------------------------------
PAROTID GLAND
An S-shaped incision is made, beginning in the preauricular
crease and extending in a curvilinear fashion to the postauricular region, followed by an inferior extension to 2 fingerbreadths below the angle of the mandible (Fig. 56-9). Skin
flaps are elevated, and the greater auricular nerve and posterior facial vein will be identified and may need to be sacrificed
to expose the posterior border of the parotid gland. Blunt dissection along the tragal pointer and mastoid process, following the posterior belly of the digastric muscle, will allow
visualization of the main trunk of the facial nerve as it emerges
from the stylomastoid foramen. Meticulous dissection along
the facial nerve branches in an anterior direction will elevate
Parotid gland sup. lobe

Sup. temporal v. and a.
FACIAL NERVE
Mastoid process
Gr. auricular ns.
Parotid gland deep lobe
Zygoma
Pterygoid fossa
Cartilage of external
auditory canal
Masseter m.
FACIAL NERVE
Mastoid process
FIGURE 56-9 Technique for parotidectomy.
CHAPTER 56
the superficial lobe of the parotid gland. Careful blunt dissection, with use of the bipolar cautery and facial nerve monitor, will maximize excellent surgical results with minimal
morbidity.64,67
SUBMANDIBULAR GLAND
For submandibular gland resection, a horizontal skin incision
is made in a natural skin crease approximately two fingerbreadths inferior to the body of the mandible. The dissection
plane is carried to the investing fascia of the submandibular
gland. Exposure should reveal the mylohyoid muscle anteriorly, the sternocleidomastoid muscle posteriorly, and the
digastric muscle inferiorly. Identification and division of
the anterior facial vein, just deep to this fascia, will facilitate
protection of the facial nerve. Anterior retraction of the
mylohyoid muscle and downward retraction on the submandibular gland will enable identification of the lingual nerve
and Wharton duct. Division of the duct will free the lingual
nerve from the gland and allow complete blunt dissection
of the gland.66
SALIVARY GLANDS
735
Conclusion
------------------------------------------------------------------------------------------------------------------------------------------------
Although salivary gland disorders are rare in childhood,

knowledge of the anatomy of the major salivary glands and
understanding of both systemic and neoplastic physiology is
critical. Neoplasms of the salivary glands are very rare in children and are commonly benign.31,68 Evaluation and management should be tailored to the specific entity. A multitude of
diagnostic tools are available and may include radiologic or
pathologic studies.
Inflammatory and infectious disorders are often treated
medically, whereas neoplastic disorders require surgical intervention. Patients and families must be counseled regarding
potential short-term and long-term complications of facial
nerve injury.
Despite the rigorous demands of parotid and submandibular gland surgery, in experienced hands, with adequate monitoring and meticulous dissection and hemostasis, surgical
results are excellent.67
expertconsult.com.
Anatomy
------------------------------------------------------------------------------------------------------------------------------------------------
The regional lymph node groups of the head and neck are
shown in Figure 57-1. The precise borders for these groups
have been classified by the American Head and Neck Society
and are shown in Figure 57-2.4 Drainage to lymphatic basins
usually follows predictable, anatomic routes, with the nomenclature reflecting the site of the lymph nodes. The face and
oropharynx drain predominantly to the preauricular, submandibular, and submental nodes; the posterior scalp drains to the
occipital nodal group; and the mouth, tongue, tonsils, oropharynx, and nasopharynx drain to superficial and deep
chains of the anterior cervical nodes. Significant lymphatic
collateralization exists.
Differential Diagnosis
------------------------------------------------------------------------------------------------------------------------------------------------
CHAPTER 57
Lymph Node
Disorders
Faisal G. Qureshi and Kurt D. Newman
Lymphadenopathy is defined as an enlargement or a change in

the character of a lymph node. Pathologic lymphadenopathy is
usually a symptom of infectious, noninfectious conditions, or,
in rare cases, malignant disease. Lymphadenopathy, especially
cervical lymphadenopathy, is quite common in childhood,
with a reported prevalence of 28% to 55% in otherwise
normal infants and children.1,2 In addition, children have
palpable nodes in most of the superficial lymphatic basins,
including cervical, axillary, and inguinal regions that are nonpathologic; there is progressive increase in lymphoid mass
from birth until early adolescence. This lymphoid tissue then
normally diminishes throughout puberty.3
Many lymph nodes are palpable in children, and generally,
cervical nodes less than 2 cm, axillary nodes less than 1 cm,
and inguinal nodes less than 1.5 cm are considered physiologic in young children. Palpable epitrochlear and supraclavicular nodes should, however, be viewed with suspicion and
trigger investigations.
The primary goal of a consulting surgeon is to determine
the need for a tissue diagnosis. A key consideration is to resolve the familys fears of malignancy in an efficient and
cost-effective manner. This chapter focuses primarily on
lymphadenopathy in the cervical region. Some comments are
made about other regions.
Most lymphadenopathy is benign in nature and is generally

associated with a short duration of symptoms. Table 57-1
shows a list of differential diagnoses. Generalized lymphadenopathy is defined as enlargement of more than two noncontiguous lymph node groups.
MALIGNANCY
Malignancy accounts for 11% to 24% of the diagnoses,
depending on the nature of the group reporting their result.
The higher rates are reported in series from oncology practices.5,6 Malignant processes are more common in the age
group of 2 to 12 years old and very rare in the age group of
less than 2 years old. Malignancy as a cause is also more common in children with chronic generalized lymphadenopathy,
nodes greater than 3 cm in diameter, and nodes in the supraclavicular region. Associated symptoms of night sweats,
weight loss, and hepatosplenomegaly also increase the chance
of malignancy. Finally abnormal laboratory and radiologic
evaluation are associated with increased malignancy rates.7
Soldes and colleagues reviewed predictors of malignancy in
children with peripheral lymphadenopathy and determined
that increasing node size, increasing number of sites of adenopathy, and age were associated with an increasing risk of
malignancy (P < 0.05).8 In addition, supraclavicular adenopathy, an abnormal chest radiograph, and fixed nodes were
all significantly associated with malignancy.
The most common malignancies as a cause of lymphadenopathy are Hodgkin and non-Hodgkin lymphomas, leukemia, and metastatic disease.
Evaluation
------------------------------------------------------------------------------------------------------------------------------------------------
A careful history, physical examination, appropriate laboratory evaluation, and targeted imaging will usually help in deciding the need for tissue sampling. Persistent or progressive
new-onset lymphadenopathy of greater than 4 to 6 weeks duration usually triggers a workup by the referring pediatrician.
Indeed, most children with acute lymphadenopathy are rarely
ever evaluated by pediatric surgeons. Most will improve with
antibiotic therapy initiated by their pediatrician or the lymphadenopathy will resolve spontaneously when related to viral
illnesses.
737
738
PART V
HEAD AND NECK
Retroauricular
II
Occipital
III
Parotid
VI
Submandibular
Superficial cervical
Supraclavicular
IV
Submental
Jugulodigastric
Deep cervical
FIGURE 57-1 Regional lymph node groups of the head and neck.
FIGURE 57-2 Lymphatic node levels of the neck. Level I: submental and
submandibular; level II: superior jugular; level III: middle jugular; level IV:
inferior jugular; level V: supraclavicular or posterior; and level VI: central
or anterior.
TABLE 57-1
Differential Diagnosis of Lymphadenopathy in Children
Generalized lymphadenopathy:
infectious
malignant
others
Localized lymphadenopathy:
infectious
malignant
site specific
Localized cervical masses:

non-nodal masses
Viral: CMV, HIV, rubella, varicella, measles, EBV, herpes, hepatitis

Bacterial: typhoid, tuberculosis, mycobacterial, syphilis, LGV, leptospirosis, brucellosis
Protozoal: for instance, toxoplasmosis, leishmaniasis
Fungal: for instance, coccidioidomycosis, Cryptococcus, histoplasmosis
Other: syphilis, Lyme disease
Lymphoma, leukemia, neuroblastoma, thyroid tumor, metastasis (e.g., osteosarcoma, glioblastoma)
Autoimmune disorders: for instance, JRA, SLE, drug reactions, CGD, lymphohistiocytosis, LCH, dermatomyositis
Storage disorders: for instance, Gaucher disease, Niemann-Pick disease
Miscellaneous: Addison disease, Castleman disease, Churg-Strauss syndrome, Kawasaki disease, Kikuchi disease,
lipid storage disease, sarcoidosis
Staphylococcus aureus, group A Streptococcus (e.g., pharyngitis), anaerobes (periodontal disease), acute bacterial
lymphadenitis, cat-scratch disease, tularemia, bubonic plague, diphtheria, chancroid, viral URI, mononucleosis,
tuberculosis/atypical mycobacterium
Lymphoma, leukemia, neuroblastoma, rhabdomyosarcoma, parotid tumor, nasopharyngeal tumor, solid tumor
metastasis
Cervical: Kawasaki disease
Occipital: tinea capitis, pediculosis capitis
Preauricular: cat-scratch disease, chronic eye infections
Supraclavicular: histoplasmosis, coccidioidomycosis
Mediastinal: sarcoidosis, cystic fibrosis, histoplasmosis,
Axillary: local infection, brucellosis, immunization reactions, JRA
Inguinal: syphilis, LGV, diaper rash
Mumps, thyroglossal duct, branchial cleft cyst, sternocleidomastoid tumor, cervical ribs, lymphatic malformation,
hemangiomas, laryngocele, dermoid cyst
CGD, chronic granulomatous disease; CMV, cytomegalovirus; EBV, Epstein-Barr virus; HIV, human immunodeficiency virus; JRA, juvenile rheumatoid arthritis;
LCH, Langerhans cell histiocytosis; LGV, lymphogranuloma venereum; SLE: systemic lupus erythematosus; URI, upper respiratory infection.
CHAPTER 57
Once a child is referred to a surgeon, important historical

questions include duration, progression, location, and associated symptoms, such as pain, fever, weight loss, and night
sweats. Additional clinical information includes recent illnesses, especially upper respiratory tract symptoms, infections, trauma, bites, and dental problems. Drug use and
sexual activity are important questions, especially in adolescents. Recent immunizations, especially bacillus CalmetteGuerin (BCG) should be evaluated. Social history, including
recent travel, animal exposure, and exposure to tuberculosis
and tropical diseases should be sought.
Once a general physical examination is completed, including a search for organomegaly, specific evaluation of the enlarged lymph nodes and other nodal basins should be
performed. The skin and subcutaneous tissue that is drained
by the affected lymph nodes should be evaluated; the characteristics of the lymph node should be noted. Normal nodes are
usually soft, mobile, small, and nontender. Lymphadenopathy
secondary to infections is also usually soft and can be mobile.
However, on occasion, bacterial invasion of lymph nodes can
result in erythema, tenderness, and fluctuance. With time,
infected nodes can become adherent and have no inflammatory signs. Firm, fixed, nontender rubbery nodes can indicate
a neoplastic process in older children.8
A thorough history and physical examination usually help
separate local from generalized processes and help guide further evaluation, including laboratory and radiologic evaluation.
INVESTIGATION
Laboratory Studies
Most patients have had a laboratory evaluation prior to referral
to surgery. These tests usually include a complete blood
count (CBC) with manual differential, sedimentation rate,
and C-reactive protein. However, these are not always helpful
in determining the specific etiology of the disease process. Pancytopenia can be seen in leukemia; lymphocytosis is seen with
mononucleosis, cytomegalovirus (CMV), and toxoplasmosis.
Based on the history and physical examination, more specific tests for Epstein-Barr virus (EBV), CMV, toxoplasmosis,
brucellosis, histoplasmosis, syphilis, bartonellosis, and coccidioidomycosis should be considered. Tests for human immunodeficiency virus (HIV) should also be considered,
based on the history as well as the tuberculin skin test.
Serum lactate dehydrogenase should be assayed when
suspecting leukemia or lymphoma as a byproduct of high cell
turnover.
Radiologic Evaluation
Diagnostic imaging can be used to determine the characteristic
of the lymphadenopathy, identify potential sources of infection, identify mediastinal and abdominal masses, and to help
differentiate enlarged lymph nodes from other pathology.
Chest radiographs, ultrasonography with Doppler, and computer tomography have all been used in the evaluation of
adenopathy.
In children with long-term lymphadenopathy, a two-view
chest radiograph is helpful to rule out mediastinal masses that
may compress the airway with or without significant symptoms. A chest radiograph should be performed prior to any
operative intervention, including biopsies done under general
LYMPH NODE DISORDERS
739
anesthesia. Patients with large mediastinal masses compressing the airway should not undergo general anesthesia,
because this could result in airway collapse (see Chapter 38).9
Ultrasonography (US) is helpful when the nodes are difficult to palpate and to help differentiate nodes from other
structures, such as thyroglossal duct cysts and dermoid cysts
in the neck, and undescended testis and inguinal hernias in
the groin, US may also be helpful in determining the characteristics of the node. Fluctuance and abscess formation will
help guide therapies such as needle aspiration or incision
and drainage.
Attempts have been made to use ultrasonography and
Doppler characteristics to differentiate neoplastic from nonneoplastic etiologies. Reactive lymphadenopathy is associated
with central necrosis, central hyperechogenicity, long to
shortaxis ratio (>2.0), hilar vascularity, and low pulsatility
index.1014 However, these modalities are not sensitive or
specific enough to primarily rule out neoplastic processes.
The decision to delay biopsy diagnosis should not be dependent on US/Doppler findings.
Computed tomography (CT) is useful in patients with mediastinal masses and suspected intra-abdominal malignancies.
Airway compromise may be best evaluated by chest CT. Interventional radiologists sometimes use CT scans to help guide
biopsies from mediastinal masses.
Diagnostic Procedures
The decision to proceed with obtaining tissue from the involved lymph node is made in conjunction with the referring
physician and after appropriate physical, laboratory, and radiologic evaluation as required. Often, the child has been observed for several weeks prior to referral to a surgeon. Small,
soft, mobile nodes should not undergo biopsy, because these
are most likely benign unless they are in the supraclavicular
region. Tissue diagnosis is helpful when lymph nodes persist
or enlarge after adequate antibiotic therapy, when they are associated with signs or symptoms of malignancy, and, finally, if
the diagnosis is questioned.
Most authors recommend waiting at least 4 to 6 weeks before obtaining tissue samples. Earlier biopsy should be considered for nodes in the supraclavicular or epitrochlear region,
nodes greater than 3 cm in diameter, and for children with
a history of malignancy, weight loss, night sweats, fever, or
hepatosplenomegaly. Similarly, physical characteristics of the
lymph node may also indicate earlier biopsy.15,16
Fine-Needle Aspiration Fine-needle aspiration (FNA) has
been used extensively in adults, with practical advantages, including its simplicity, speed in the outpatient setting without
sedation, as well as its cost effectiveness. In addition, the sensitivity and specificity reaches more than 90%.
The use of FNA in children has increased, especially in
countries where tuberculosis is prevalent.1720 Aspirates
should be sent for Gram stain, acid-fast stain, and cultures
for aerobic/anaerobic bacteria, mycobacteria, and fungi.
However, the use of FNA in children has not become universal, because the aspirate usually provides a small sample,
which limits the ability to perform flow cytometry, chromosomal analysis, and electron microscopy. Most pediatric hematologists and pathologists prefer excisional biopsy, because
it allows the assessment of nodal architecture and permits
740
PART V
HEAD AND NECK
the use of special stains. In addition, some children will not

permit FNA without some sedation, which negates a primary
benefit of FNA. Aspirates may also have a higher rate of falsenegative rates in the diagnosis of Hodgkin disease, a common
malignant condition in children. Finally, the risk of seeding
the needle site tract with malignant cells, although small, is
a legitimate concern of physicians and parents alike.21
As can be seen, the pathologic diagnosis varied depending

on the reporting group and the associated referral pattern,
with higher malignant rates documented by oncology
groups7,21a and higher infectious rates reported by authors
in developing countries.5
Excisional Biopsy Excisional biopsy provides enough tissue to perform flow cytometry, chromosomal analysis, electron microscopy, and the use of special stains. Indications
for an excisional biopsy include
1. Lymph nodes that are hard/matted
2. Lymph nodes fixed to surrounding tissue
3. Progressively enlarging nodes without response to antibiotic therapy
4. Presence of abnormally enlarged nodes after 4 to 6 weeks
5. Supraclavicular, epitrochlear lymph nodes
6. Hepatosplenomegaly
7. Mediastinal or hilar masses
8. Laboratory anomalies, especially anemia, leukocytosis,
leucopenia, and thrombocytopenia
9. Symptoms such as fever, weight loss, and night sweats
10. Suspicion of atypical mycobacterial adenitis
11. Diagnostic dilemma
Most excisional biopsies are done under general anesthesia
or sedation and, very rarely, under local anesthesia. The biopsy
should be coordinated with pathology so that the lymph node
can be sent as a fresh specimen. The nodes should not be fixed
in formalin. As discussed earlier, a chest radiograph should be
obtained to rule out a mediastinal mass that may compromise
the airway prior to exposing children to general anesthesia or
sedation.
In a recent review, Oguz et al. reviewed their experience
with 457 children (2 months to 19 years old) with lymphadenopathy who were referred to their oncology group; 346
(75.7%) had benign processes, and 111 (24.3%) had malignant disease. Of these, 134 patients underwent excisional
biopsy for indications highlighted previously. Table 57-2
highlights the findings on excisional biopsy and compares
them to findings by other authors.7
------------------------------------------------------------------------------------------------------------------------------------------------
TABLE 57-2
Excisional Biopsy Results
Excisional Biopsy Results

Malignant
Hodgkin lymphoma
Non-Hodgkin
lymphoma
Nasopharyngeal cancer
Thyroid cancer
Miscellaneous
Benign
Chronic lymphadenitis
Hyperplasia
Tuberculosis
Reactive
Miscellaneous
Oguz
et al,
20067
(n = 134)
Moore
et al,
20035
(n = 1332)
Yaris
et al,
200621a
(n = 38)
79.8%
40.2%
29.1%
11.8%
6%
2.1%
50%
3.7%
2.2%
4.2%
20.1%
5.9%
5.9%
2.9%
2.2%
3.2%
3.9%
88.2%
11.3%
47.8%
25%
4.1%
50%
25%
15.7%
Management of Adenopathy
Darville and colleagues have suggested a helpful algorithm for
the management of cervical lymphadenopathy (Fig. 57-3).22
This algorithm is a useful tool to help surgeons determine their
role in the management of enlarged lymph nodes. As suggested elsewhere in this chapter, most of the medical evaluation and management has usually been performed by the
referring physician; however, it is the surgeons responsibility
to review each case prior to intervention.
SURGICAL MANAGEMENT
Surgical management is usually limited to diagnostic FNA,
excisional biopsy, incision and drainage, and total excision.
Further details are provided under the specific conditions discussed later.
ACUTE LYMPHADENITIS
The most common cause of self-limiting, acute, inflammatory
lymph node is a viral infection.23 Acute bilateral cervical adenopathy is most often caused by a viral respiratory tract infection
(rhinovirus, parainfluenza virus, influenza virus, respiratory
syncytial virus, coronavirus, adenovirus, reovirus) and is usually hyperplastic in nature.24 Viral-associated adenopathy does
not suppurate and usually resolves spontaneously.
Unilateral lymphadenitis is usually caused by streptococcal
or staphylococcal infection in 40% to 80% of the cases.25
These are usually large (>2 cm), solitary, and tender in the
preschool child.26 The submandibular, upper cervical, submental, occipital, and lower cervical nodes are affected in
decreasing order of frequency.27 Suppurative adenitis is
associated with group A streptococcal or penicillin-resistant
staphylococci. Staphylococcus infection leading to lymphadenitis seems to occur more commonly in infants.28 Other less
frequent causal organisms include Hemophilus influenzae type
B, group B streptococci, and anaerobic bacteria. Communityacquired methicillin-resistant Staphylococcus aureus (MRSA) is
now more commonly being isolated from superficial abscesses
and suppurative lymphadenitis in children. Clindamycin is an
appropriate agent to use under these circumstances.25,29
Suppurative lymphadenitis presents with local inflammatory
signs, including unilateral tender adenopathy involving the
submandibular or deep cervical nodes draining the oropharynx. Erythema, fever, malaise, and signs of systemic illness
may occur. The primary infection in the head and neck regions
should be looked for with careful attention to the oropharynx
and middle ear. Appropriate treatment should be started, usually an empirical 5- to 10-day course of an oral b-lactamase
resistant antibiotic. Intravenous antibiotics should be started
if systemic signs are present or in very young infants. A response
should be observable within 72 hours, and failure of therapy
usually necessitates additional diagnostic testing. This is usually
fine-needle aspiration or ultrasonography.
CHAPTER 57
741
Cervical lymphadenitis
Asymptomatic
Symptomatic
Small node (<3 cm)
Large node (>3 cm)
Nonfluctuant
Fluctuant
Culture possible
primary focus
Culture possible
primary focus;
Needle aspirate
with stains and
cultures
Culture possible
primary focus;
Needle aspirate with
stains and cultures;
Blood culture;
Draw acute serum
Empiric antibiotic therapy for 23 weeks
Small node (<3 cm)
Observe
Node size
persists or enlarges
within 46 weeks
Good clinical response

but node remains
large and tender
Cultures negative, no change in size or larger
Continue empiric treatment

CBC, ESR, CXR, PPD skin test
Serology (EBV, CMV, Toxoplasma gondii, VDRL,
ASO titer, tularemia, B. henselae, Brucella, fungi)
Diagnosis in doubt;
Node persists, enlarges,
is fixed to underlying
tissues, or is hard
Ultrasound to look for

fluctuance, abscess
formation
Therapeutic needle
aspiration/incision
and drainage of
abscess
Excisional biopsy
FIGURE 57-3 Evaluation and treatment algorithm. ASO, antistreptolysin titer; CBC, complete blood count; CMV, cytomegalovirus; CXR, chest radiograph;
EBV, Epstein-Barr virus; ESR, erythrocyte sedimentation rate; PPD, purified protein derivative; VDRL, Venereal Disease Research Laboratory. (Reprinted with
permission from Elsevier.22)
Aspirate culture by FNA can guide further organismspecific antibiotic treatment, including clindamycin if MRSA
is encountered. If no fluid is aspirated, sterile saline can be
injected and then aspirated to obtain material for culture.26
In addition, repeated aspiration together with antibiotics is
an effective treatment for fluctuant lymphadenitis.30 As stated
previously, however, FNA may require sedation or anesthesia
in young children.
Ultrasonography may help to differentiate between solid
and cystic masses and can identify fluid that may require operative drainage. Incision and drainage is a more definitive
surgical approach to suppurative fluctuant lymphadenitis.
Gauze packing has been used to prevent early skin closure
and achieve hemostasis; however, the use of minimal incisions, with vessel loops functioning as drains, has been gaining wider acceptance recently.31
PERSISTENT LYMPHADENITIS
Persistent lymphadenitis that does not resolve despite 2 to
4 weeks of appropriate therapy warrants additional diagnostic
workup. Some common causes of persistent lymphadenitis
are discussed in this section.
Atypical Mycobacterial Adenitis
The genus Mycobacterium is characterized on light microscopy
to be bacilli distinguished by their dense lipid capsules. The
lipid capsules resist decolorization by acid alcohol after
staining and thus are termed acid-fast bacilli. In the United
States, 70% to 95% of mycobacterial lymphadenitis cases
are caused by atypical mycobacteria (nontuberculous strains).
The most common agents include M. avium-intracellulare,

M. scrofulaceum, M. fortuitum, and M. chelonei.32 In contrast
to tuberculous adenitis, atypical (or nontuberculous) mycobacterial adenitis is generally considered a local infectious process, without systemic involvement in immunocompetent
hosts. Disseminated disease is more commonly observed in
patients with underlying acquired or congenital immunodeficiency states. Atypical mycobacterial adenitis is not contagious, and the portal of entry in otherwise healthy children
is the oropharynx.33
Atypical mycobacterial adenitis usually occurs in young
children between 1 and 5 years of age. The common clinical
presentation is focal, unilateral involvement of the jugulodigastric, preauricular, or submandibular nodal group. There
is rapid onset of nodal enlargement, and the skin gradually
develops a pink or red hue; with time, the overlying skin becomes thin.26 In contrast to acute suppurative lymphadenitis,
there is no response to first-line antibiotics, and the clinical
course is described as indolent, with the involved nodal group
being minimally tender, firm, and rubbery to palpation, well
circumscribed, and sometimes adherent to underlying structures. Although remarkably nontender, these lesions develop
a draining sinus tract in 10% of patients.34,35 Signs of systemic
illness or inflammation are usually minimal or nonexistent.
Chest radiographs are usually normal.
Differentiating atypical mycobacterial and mycobacterial
tuberculous cervical lymphadenitis can occasionally be challenging, based purely on epidemiologic and clinical features.
Age (<5 years), race (white), place of residence (rural),
bilaterality (rare) all point toward atypical mycobacterial
infection. Purified protein derivative (PPD) skin testing in
742
PART V
HEAD AND NECK
children with atypical mycobacterial lymphadenitis can result

in an intermediate reaction because of cross reactivity, usually
less than 15 mm. Blood interferon-gamma release assay is
emerging as the discriminating test of choice; it was originally
described for pulmonary disease but is now being used
for nodal disease as well.36 Other criteria that point toward
a diagnosis of tuberculous lymphadenitis are (1) a positive
PPD, (2) abnormal chest radiograph, and (3) contact with a
person with infectious tuberculosis. Spyridis and colleagues
have shown that fulfilling two of three criteria results in diagnosis of tuberculous lymphadenitis with a 92% sensitivity.37
Unlike tuberculous adenitis, atypical mycobacterial adenitis generally does not respond to chemotherapy. The treatment
of choice is complete surgical excision with primary wound
closure. In a literature review of the surgical treatment of
atypical mycobacterial cervicofacial adenitis in children, excision, incision and drainage, curettage, and needle aspiration
were compared across 16 studies. The cure rates were 92%,
10%, 86%, and 41%, respectively.38 Incision and drainage
should be avoided, because it often results in a chronically
draining sinus. There have been reports of adequate medical
treatment of atypical mycobacterial lymphadenitis; however,
in a recent multicenter randomized trial comparing surgical excision and antibiotics, surgical excision was superior,
with a 96% cure rate compared with 66% with antibiotic
therapy.39 Multidrug antibiotic therapy, usually including
clarithromycin and rifabutin, may be used as an adjunct for
unresectable or recurrent disease.40 Surgical treatment should
include elliptic excision of the overlying skin when it is
thinned out, debridement of subcutaneous granulation tissue,
and complete excision of the involved node(s) with closure of
the overlying skin; formal lymph node dissection is not
required. Curettage is recommended only if surgical excision
is not possible because of unacceptable cosmetic outcomes or
risk of injuries to adjacent nerves. A nerve stimulator may be
helpful for lesions at the angle of the mandible to avoid injury
to branches of the facial nerve.
Mycobacterial Adenitis
In developed countries, tuberculous adenitis or scrofula is almost exclusively caused by M. tuberculosis. Before control of
bovine tuberculosis, the predominant cause of tuberculous
adenitis was M. bovis. Occasional cases of M. bovis are observed
from underdeveloped regions in which consumption of contaminated raw meat occurs. Patients proven to have human
tuberculous adenitis often report previous exposure to a
known carrier of tuberculosis, but most patients do not have
active disease on a chest radiograph.37 Differentiation between
tuberculosis and atypical mycobacterial adenitis has been
highlighted previously. Tuberculous adenitis is considered
to be a local manifestation of a systemic disease and not an
initial primary focus of tuberculous infection.41
Clinically, children with tuberculous adenitis are usually
older and present with nonsuppurative lymphadenitis, which
may be bilateral.42 A retrospective review of 24 immunocompetent children with tuberculous lymphadenitis showed that
no patient had bilateral disease, and the submandibular (29%)
and the anterior cervical (71%) sites were the only areas of
involvement.37 However, posterior triangle nodal involvement
does occur.
The diagnosis of tuberculous adenitis can be made on the

criteria established by Spyridis and colleagues37 and positive
acid-fast bacteria on stain or culture of nodal tissue. Diagnostic confirmation may be aided by FNA with aspirate culture
and cytologic examination.43 Rapid diagnosis of tuberculous
adenitis by DNA amplification of nodal material using polymerase chain reaction (PCR) has been reported.44 Blood testing and PPD are also used. A negative tuberculin PPD test
essentially excludes the diagnosis of tuberculous adenitis. If
a diagnostic dilemma persists, surgical excisional biopsy is
warranted. Incisional biopsy or incision and drainage should
be avoided to prevent development of chronic, draining sinus
tracts.23,45 Fistula and cheloid formation can be seen in up to
100% of patients who undergo incision and drainage of tuberculous infected lymph nodes.37
Tuberculous adenitis generally responds to medical management that consists of multiple-agent chemotherapy. The
World Health Organization recommends directly observed
short-course therapy, including isoniazid, rifampin, ethambutol, and pyrazinamide for the first 2 months, followed by isoniazid and rifampin for an additional 4 months.46 Nodal
regression usually occurs within 3 months. Although antituberculous chemotherapy remains essential, the role of
complete surgical excision of involved nodes is more controversial.47 Complete excision of involved nodes is prudent
when biopsy is required for diagnosis, when a chronically
draining sinus tract evolves during medical treatment, or
when optimal medical management fails.
CAT-SCRATCH DISEASE
Cat-scratch disease is a common cause of lymphadenitis in
children, with an estimated incidence in the United States
of 9.3 per 100,000 ambulatory pediatric and adult patients
per year.48 The highest age-specific incidence is among children younger than 10 years of age.2 Current microbiologic
and PCR-directed DNA analysis demonstrates that the pleomorphic, gram-negative bacillus Bartonella henselae (formerly
Rochalimaea) is the causative organism of cat-scratch disease.49
Most cases can be directly related to contact with a cat, and the
usual site of inoculation is an extremity. Subsequent adenitis
occurs at regional lymphatic drainage basins (inguinal, axillary, epitrochlear nodes) 5 days to 2 months later.50 Similarly,
cervical lymphadenopathy is observed with scratches in the
head and neck region. Although the primary manifestation
of Bartonella henselae infection is lymphadenopathy, some
series report up to 25% of cases resulting in severe systemic
illnesses.51
Initial infection occurs at the portal of entry in the skin,
such as a scratch or bite. Papule formation may be observed
at the site of inoculation in 3 to 5 days, with development
of subacute lymphadenopathy at regional nodal drainage beginning within 1 to 2 weeks. Early systemic symptoms of fever,
malaise, myalgia, and anorexia are commonly reported.
Although most cases involve the lymph nodes of limbs,
approximately 25% of cases involve the cervical nodes.50
Diagnosis is based on a history of exposure to cats, presence
at a site of inoculation, and regional lymphadenopathy.
Identification of Bartonella henselae from involved lymph
nodes using Warthin-Starry silver impregnation stain has
CHAPTER 57
traditionally been used for diagnosis, but the stain has been
found to be unreliable and lacking specificity. PCR for Bartonella henselae using paraffin sections from lymph nodes or
other tissue is more reliable and specific.52 To confirm diagnosis without obtaining tissue, many centers use serologic testing, which has been available for several years; it has a low
sensitivity but is highly specific.53
Lymphadenitis associated with cat-scratch disease is usually benign, self-limiting, and resolves within 6 to 8 weeks
without specific treatment.54 Antibiotic treatment has thus
been controversial, although azithromycin has been associated with rapid resolution of the adenitis.55 Suppuration is
unusual; however, if it occurs, needle aspiration may provide
symptomatic relief. Excisional biopsy is generally unnecessary
but may be warranted if a draining sinus tract develops or if
the diagnosis is uncertain and the potential for malignancy
cannot be excluded.
MISCELLANEOUS LESIONS
Various other infectious and inflammatory conditions can produce lymphadenopathy in infants and children. Most patients
with these disorders do not require surgical management or,
in particular, excisional biopsy of the lesions. A systematic approach to evaluation of these patients, as outlined previously,
generally leads to the correct diagnosis. Surgical management
of these lesions should be directed to patients who present diagnostic dilemmas and have nodal disease in suspicious areas,
or have persistent adenopathy despite adequate medical therapy.
Infectious lymphadenopathy
Lymphadenopathy caused by infectious agents include toxoplasmosis (caused by Toxoplasma gondii), tularemia (caused by
Francisella tularensis), and mononucleosis (caused by EpsteinBarr virus). Infection with Actinomyces israelii in the head and
neck may lead to cervicofacial actinomycosis that is characterized by a woody indurated cervical mass and development of
chronic, draining fistulas. Direct involvement of the lymph
nodes is uncommon, but the induration can make the clinical
differentiation difficult.56 Infection with HIV can produce
general lymphadenopathy in infants and children.57
743
Castleman disease, also called angiofollicular or giant lymph

node hyperplasia may also occasionally present as a solitary,
enlarged cervical lymph node. The enlarged node appears
hypervascular on US/Doppler or CT scan. Surgical excision
is curative in the localized form.60 The multicentric form of
the disease, often accompanied by visceral involvement, is
considered a type of lymphoproliferative disorder and
requires systemic therapy.
Rosai-Dorfman disease, or sinus histiocytosis with massive
lymphadenopathy, is a rare disorder affecting predominantly
African-American children in the first decade of life. Disease
progresses from unilateral cervical adenopathy to massive bilateral cervical involvement and extension to other nodal
groups or extra nodal sites. The disorder is benign but has a
slow rate of resolution spanning 6 to 9 months. Excisional
biopsy may aid diagnosis.61
MALIGNANT DISORDERS
Although lymphoma is the most common malignant disorder
manifested by cervical adenopathy, neuroblastoma and thyroid carcinoma are other childhood cancers that can present
as enlarged cervical lymph nodes.
Lymphomas are one of the more common malignant conditions in children. They may present as primary neck adenopathy that does not resolve with antibiotics or is enlarging.
Patients with congenital or acquired immunodeficiency states,
including HIV infection, are at greater risk for developing
malignant lymphoproliferative conditions. Excisional biopsy
is often used to help diagnose lymphomas.
In neuroblastoma, adenopathy is usually bilateral. These
patients often have stage 4 disease, and if the primary is not
evident on examination and radiologic evaluation, excisional
biopsy is performed for initial diagnosis of neuroblastoma.
Metastatic thyroid carcinoma may present with unilateral
cervical lymph node enlargement that should not be mistaken for ectopic thyroid gland. If a thorough neck examination does not reveal a thyroid nodule, and a history of
neck irradiation or other high-risk factors is obtained, thyroid ultrasonography should be performed as part of the
evaluation of neck lymphadenopathy.
Inflammatory disorders
Inflammatory disorders include Kawasaki disease, Kikuchi
disease, Castleman disease, and Rosai-Dorfman disease.
Kawasaki disease, or mucocutaneous lymph node syndrome,
is a febrile disorder of childhood that is characterized in part
by the abrupt onset of erythematous changes in the oropharyngeal mucosa, acute vasculitis, and extensive nonsuppurative,
nontender cervical adenopathy.58 Diagnosis is made on clinical
grounds, and the resolution of the nodal disease occurs
relatively quickly in the course of the disease.
Kikuchi disease, or histiocytic necrotizing lymphadenitis,
may present as cervical lymphadenopathy that resolves spontaneously. It typically presents in the older child with bilateral,
painful cervical nodes. There are associated fevers, night
sweats, splenomegaly, leucopenia with atypical lymphocytosis, and elevated erythrocyte sedimentation rate (ESR). This
disease can be clinically confused with malignant disease,
and the patients often appropriately undergo excisional biopsy for definitive diagnosis.59
Summary
------------------------------------------------------------------------------------------------------------------------------------------------
Most adenopathy in children is nonpathologic and spontaneously resolves. Pathologic lymphadenopathy has a large differential diagnosis, with viral lymphadenitis being the most
common. Surgical consultation is often obtained when the
lymph nodes do not spontaneously resolve, if there is concern
for malignancy, or if there is a diagnostic dilemma. Most of the
investigation is usually performed prior to surgical consult,
but the surgeon must be aware of an adequate workup prior
to intervention. The surgeons role is usually limited to excisional biopsy, incision and drainage, and, rarely, aspiration
in children, depending on the pathology suspected. FNA
for diagnosis has a more limited role in children but may be
useful in selected cases.
expertconsult.com.
CHAPTER 58
Childhood Diseases
of the Thyroid and
Parathyroid Glands
Hannah G. Piper and Michael A. Skinner
Diseases of the thyroid and parathyroid glands are relatively

uncommon in the pediatric population. The incidence of thyroid nodules is estimated to be 20 per 1000 children, and the
incidence of hyperparathyroidism is approximately 4 per
100,000 children.1,2 However, although rare, there is a wide
range of pathology, both benign and malignant, with which
pediatric surgeons must remain familiar. Surgery plays a
central role in the management of many of these disorders,
including thyroid nodules, parathyroid adenomas, and, occasionally, goiter and hyperthyroidism.
Thyroid Embryology
and Physiology
------------------------------------------------------------------------------------------------------------------------------------------------
The thyroid begins to form 24 days after fertilization and is

both the first and the largest of the endocrine glands in the developing embryo. The thyroid commences as an endodermal
outpouching on the floor of the primordial pharynx and becomes the thyroid diverticulum. This diverticulum then descends from the pharynx, passing anterior to the hyoid
bone and maintains a connection to the base of the tongue,

known as the thyroglossal duct. The migration is usually complete by 7 weeks of gestation, at which point the thyroglossal
duct is obliterated. Initially, the thyroid diverticulum is hollow
but then becomes solid cellular parenchyma with both a right
and left lobe. The cells are arranged in spherical units called
follicles, which are filled centrally with proteinaceous colloid.
Parafollicular cells, or C cells, derived from neural crest cells,
are found between the follicles and are the source of calcitonin. In approximately 50% of people, a pyramidal lobe extends superiorly from the isthmus. Infrequently, there will
be an ectopic thyroid gland found along the normal route
of descent. A lingual thyroid is the most common ectopic location, representing 90% of cases.3 When present, a lingual
thyroid is usually the only thyroid tissue and can sometimes
be mistaken for accessory thyroid tissue or a thyroglossal duct
cyst. Ectopic tissue will often produce insufficient amounts of
hormones and can become secondarily enlarged; therefore care
must be taken to avoid inadvertent removal of the only viable
thyroid tissue.4,5 Accessory thyroid tissue can also be found
anywhere along the normal pathway of thyroid descent, as seen
in Figure 58-1, but these remnants are usually insufficient in
size to have any normal function.
Thyroid follicular cells are responsible for the synthesis of
thyroid hormones. This begins when tyrosine molecules
within thyroglobulin are iodinated to form monoiodotyrosine
(MIT) and diiodotyrosine (DIT), neither of which is biologically active. Active hormone synthesis occurs when either
two DIT molecules couple to form thyroxine (T4), accounting
for 90% of excreted hormone, or one DIT and one MIT molecule combine to form triiodothyronine (T3), accounting for
about 9% of excreted hormone. Thyroid-stimulating hormone
(TSH) is the main stimulus for hormone production and release and is produced by the anterior pituitary gland. TSH release, in turn, is controlled by thyrotropin-releasing hormone
(TRH) produced in the hypothalamus. Circulating thyroid
hormone is reversibly bound to carrier proteins, most commonly thyroxine binding globulin (TBG) and has a wide range
of physiologic effects. Children with congenital hypothyroidism are at risk for significant neurologic impairment, delayed
bone development, and decreased metabolism. In contrast, infants and children with hyperthyroidism may have tachycardia with increased cardiac output, excessive sweating,
weight loss, and tremors.
The parafollicular cells produce calcitonin, a 32amino
acid polypeptide. Calcitonin lowers serum calcium and phosphate by inhibiting bone resorption and likely plays a role in
calcium deposition after a postprandial serum rise. Interestingly, there are no known definitive complications in humans
from either excess or deficient calcitonin.
Evaluation of the Thyroid Gland

------------------------------------------------------------------------------------------------------------------------------------------------
Proper evaluation of a child with potential thyroid disease begins with a focused history inquiring about symptoms of hyperthyroidism or hypothyroidism as well as any family history
of thyroid disease or multiple endocrine neoplasia. This is followed by a detailed examination of the neck. The thyroid
should be palpated to evaluate size, consistency, symmetry,
and whether there are any nodules or associated enlarged
745
746
PART V
HEAD AND NECK
Lingual thyroid
Foramen cecum of tongue
Accessory
thyroid tissue
Hyoid bone
Tract of thyroglossal duct

Cervical
thyroid
Pyramidal lobe
of thyroid gland
Normal position
of thyroid gland
FIGURE 58-1 Usual sites of ectopic thyroid tissue. The broken line demonstrates the path of normal thyroid descent. (Reprinted from The Pharyngeal (Branchial) Apparatus: In Moore K, Persaud TVN: The Developing
Human: Clinically Oriented Embryology, ed 6. Philadelphia, WB Saunders,
1998, p 233; with permission from WB Saunders.)
lymph nodes. Certain features on physical exam are suggestive

of specific disease processes. For example, diffuse enlargement can be seen in Graves disease and Hashimoto and endemic goiter. A tender gland can be found with an acute
inflammatory process, multiple nodules are more common
in a metabolic or inflammatory process, and a single nodule
is more likely to be neoplastic. Other worrisome signs suggestive of a neoplasm are rapid growth, hardness, fixation to
surrounding structures, and enlarged cervical nodes.
Useful laboratory tests include measurement of TSH,

which when elevated, suggests hypofunction and when suppressed suggests elevated hormone circulation. To estimate
circulating thyroid hormone, a plasma free T4 level can be
measured. Plasma total T4 and T3 can be measured along with
thyroglobulin to properly calculate unbound, active hormone.
Finally, thyroid imaging may be necessary. Ultrasonography (US) can be used to evaluate cysts and nodules and can
be helpful in following these lesions over time. In addition,
US can be used to help guide FNA for diagnostic purposes.
Features on US suggestive of malignancy include hypoechogenicity, microcalcifications, irregular margins, and hypervascularity.6 Radionuclide scintigraphy is also occasionally useful
for thyroid imaging. In patients with suppressed TSH
iodine-131 (131I) or technetium-99m (99mTc) can be used to
identify hyperfunctioning areas of the thyroid gland. Scintigraphy can also be used to assess for recurrent disease or metastatic disease in the setting of malignancy. An emerging
imaging technique is single-photon emission computed
tomography (SPECT), which gives a more three-dimensional
view of increased uptake. This can also be combined with
traditional CT (SPECT/CT) to add anatomic detail, increasing both the sensitivity and specificity over conventional
scintigraphy,7 as seen in Figure 58-2.
Non-neoplastic Thyroid
Conditions
------------------------------------------------------------------------------------------------------------------------------------------------
GOITER
A goiter is an enlargement of the thyroid gland that can be diffuse or nodular in nature, and patients can be euthyroid, hypothyroid, or hyperthyroid. The most common cause
worldwide is iodine deficiency, resulting in hypertrophy secondary to substrate deficiency. However, in locations with
FIGURE 58-2 SPECT/CT for nodal localization: cervical nodal metastasis in a 12-year-old girl. A, Planar image shows a single focus in the right upper neck
(arrowhead) and two foci of activity in the central neck. B and C, Computed tomography (CT) and fused single-photon emission computed tomography
(SPECT)/CT localizes the right upper neck focus to an enlarged lymph node (arrow). (Reprinted from Wong KK, Zarzhevsky N, Cahill JM, et al: Hybrid
SPECT-CT and PET-CT imaging of differentiated thyroid cancer. Br J Radiol 2009;82:860-876, 2009; with permission from the British Journal of Radiology.)
CHAPTER 58
CHILDHOOD DISEASES OF THE THYROID AND PARATHYROID GLANDS
adequate dietary iodine intake, most patients will have simple

colloid goiter and maintain normal thyroid function. For the
most part, this type of goiter requires no specific therapy, because the rate of regression is no different whether or not the
patient is treated with hormone suppression.8 Surgery is indicated if the goiter grows so large that there is significant airway
compression or if there is a nodule that is concerning for
possible malignancy.
THYROIDITIS
Inflammation of the thyroid gland can have several different
etiologies, including autoimmune, viral, bacterial, or an infiltrative fibrous process. The most common type of thyroiditis
in children is chronic lymphocytic or Hashimoto thyroiditis.
Hashimoto thyroiditis is an autoimmune process in which antibodies against thyroid antigen are produced, resulting in a
lymphocytic infiltrate within the gland. It often occurs in association with other autoimmune disorders, such as type I diabetes, Addison disease, systemic lupus and juvenile arthritis,
as well as in children with Down and Turner syndromes. Initially, patients may be euthyroid or hyperthyroid, but this can
be transient, and eventually hypothyroidism may ensue. Usually, the management is expectant with about 30% of patients,
demonstrating resolution over time. Exogenous thyroid hormone should be provided to patients who are hypothyroid
but does not seem to have any meaningful effect on reducing
the size of the goiter or decreasing progression of disease in
euthyroid patients.9
Other forms of thyroiditis, although rare, can also be seen
in children, as listed in Table 58-1. Subacute thyroiditis is
thought to be viral in nature, and patients often present with
fever and gland tenderness in the setting of a recent upper respiratory infection. Initially, there may be excessive hormone
release, followed by transient hypothyroidism. About 10%
of patients develop permanent hypothyroidism. In most cases,
nonsteroidal anti-inflammatory medication or low-dose corticosteroids is the only treatment required. Acute suppurative
thyroiditis is bacterial in origin, can present with hard nodules,
and be diagnosed on ultrasonography or with fine-needle aspiration (FNA). Treatment is with antibiotics, drainage if necessary, and, very rarely, thyroid lobectomy. Many cases
formerly called acute suppurative thyroiditis are the result of
a third or fourth branchial pouch remnant, also called pyriform sinus fistula. This should be suspected, especially when
the infection presents in the left superior pole (see
Chapter 59). Ideally, treatment is with antibiotics, followed
by excision of the fistula tract to the pyriform sinus.10
747
HYPOTHYROIDISM
The primary cause of congenital hypothyroidism is abnormal
thyroid gland development rather than from a problem with
the hypothalamic-pituitary-thyroid axis. Most commonly, it
is secondary to either thyroid dysgenesis or agenesis and, less
commonly, from defects in thyroid hormone synthesis or from
the transfer of maternal thyroid blocking antibodies. Dietary
iodine deficiency in utero can also lead to hypothyroidism,
and in those cases, a palpable goiter may be appreciated. T4
is essential for myelinization of the central nervous system
during the first 3 years of life. Deficiencies in T4 can lead to
intellectual disability, which is completely preventable if recognized.11 Newborn screening for hypothyroidism is essential
and involves measuring TSH. Some states also require measuring T4, which will allow for rare cases of central hypothyroidism to be detected. However, if these tests are normal but
symptoms persist, it is important to maintain a high index
of suspicion, because up to 50% of cases of central hypothyroidism will have a normal newborn screen.12 In older children with acquired hypothyroidism, presenting signs and
symptoms include a decline in linear growth, fatigue, constipation, and poor school performance. Preteens or teenagers
may complain of dry skin, thin hair, weight gain, and menstrual irregularities. The most common causes of pediatric
hypothyroidism can be seen in Table 58-2.
HYPERTHYROIDISM
Graves disease, or diffuse toxic goiter, is the most common
cause of hyperthyroidism in children, with an incidence of
0.02%,13 and it is approximately 5 times more common in
girls than in boys. Thyroid gland hypertrophy occurs because
antibodies against the TSH receptor bind and mimic the effect
of TSH. Patients usually present with a firm, smooth goiter and
symptoms of hyperthyroidism. Occasionally, there can be fibroblast deposition in the eyes, leading to exophthalmos, or
in the skin, leading to pretibial myxedema, although these
findings are less common in children. Severe hyperthyroidism
is sometimes seen with associated hyperthermia and tachycardia, referred to as thyroid storm, and initial treatment includes
active cooling and propanolol.
Usually, first-line therapy for Graves disease is antithyroid
medication (methimazole or propylthiouracil), and improvement in symptoms can occur within 1 month of treatment.
Treatment is maintained for 12 to 18 months, during which
time thyroid function is monitored routinely.14 Remission is
achieved in 30% of children after the first course of medication, and risk factors for relapse include young age and severe
TABLE 58-1
Types of Thyroiditis
Histopathology
Eponym
Etiology
Goiter
TSH
T4
Thyroid Function
Chronic lymphocytic
Subacute granulomatous
Subacute lymphocytic
Acute suppurative
Invasive fibrous
Hashimoto
De Quervain
Silent
Bacterial
Reidel
Autoimmune
Viral (mumps, Coxsackie virus, EBV)
Autoimmune
Bacterial, fungal, parasitic
Unknown
Yes
Variable
Yes
Variable
No
Variable
Low
Low
Normal
Normal or low
Variable
High
High
Normal
Normal or low
Hyper or hypo
Hyper then hypo
Hyper then hypo
Variable
Hypothyroid
EBV, Epstein-Barr virus; T4, thyroxine; TSH, thyroid-stimulating hormone.

Data from Arici C, Clark OH: Thyroiditis: Cameron J (ed): Current Surgical Therapy, ed 7. Philadelphia, Elsevier, 2001, p 597, with permission from Elsevier.
748
PART V
HEAD AND NECK
TABLE 58-2
Pediatric Causes of Hypothyroidism
No
Goiter
Newborn to
childhood
Adolescence
Goiter
Newborn
Childhood to
adolescence
Gland dysgenesis
Deficiency of the hypothalamic/
pituitary axis
Postsurgical
Inborn error in hormone synthesis
Maternal ingestion (propylthiouracil,
methimazole, iodides)
Severe endemic iodine deficiency
Ingestion of goiter-inducing drugs
Inborn error in hormone synthesis
Hashimoto thyroiditis
Infiltrative disease (lymphoma)
biochemical hyperthyroidism at the time of diagnosis.15,16 Because of the relatively low remission rate, some advocate the
use of radioactive iodine (I131). The radioactive iodine is
ingested and then incorporated into the thyroid. The radiation
to thyroid cells leads to gland ablation over the following 6 to
18 weeks. However, this is not recommended for children less
than 5 years of age, and most patients will be made hypothyroid after treatment requiring hormone replacement therapy.
Surgery is also an option for some children with Graves
disease, usually for those who have very large goiters, are unable to take antithyroid medication, or cannot tolerate radioactive iodine. Surgery is also the treatment of choice if there is
any concern of underlying malignancy in the gland. The preferred operation continues to be debated. In a large metaanalysis looking at more than 7000 adult patients, there were
no cases of recurrent disease after total thyroidectomy versus
an 8% recurrence rate after subtotal thyroidectomy. The incidence of other complications, including injury to the recurrent laryngeal nerve or hypoparathyroidism, did not differ
between the two groups.17
Neoplastic Thyroid Conditions

------------------------------------------------------------------------------------------------------------------------------------------------
MANAGEMENT OF THYROID NODULES

Thyroid nodules are less common in children than in adults,
with an incidence of 1% to 2%.2 However, when found, there
is a 16% to 27% chance that the nodule will be malignant,
which is far greater than the estimated 5% in the adult population.18,19 In general, discrete lesions that are distinct from the
surrounding thyroid tissue and are equal to 1 cm should be investigated. This begins with a history and physical exam, as well
as measuring serum T4 and TSH levels. The utility of imaging is
less clear, especially in children. Some advocate that if a nodule
is functional on scintigraphy, no further workup is required;

however, this must be practiced with caution because, although
rare, functional nodules can still harbor malignancy. Ultrasonography can be useful for measuring the size of a nodule, determining if it is solid or cystic, and locating it within the gland.
Certain features on ultrasonography can raise or lower the suspicion of malignancy; for example, an entirely cystic lesion has a
very low probability of being malignant. However, ultrasonography cannot definitively make the diagnosis of a malignancy.6
The use of FNA for cytologic evaluation is now the most accurate
diagnostic intervention in adult patients20 and is standard in the
workup of a thyroid nodule. FNA is also commonly used in children and adolescents, with the goal of trying to reduce diagnostic thyroid surgery for benign lesions. The question that is raised
is whether this test has a low enough false-negative rate to justify
its use in the pediatric population. Most would agree that for adolescents FNA can be used reliably, because it has an accuracy of
at least 90%, and any potential delay in diagnosis will not likely
result in decreased survival. Because younger children have a
higher incidence of malignancy in any thyroid nodule, there
has been some reservation in relying on FNA.21 However,
FNA results are useful for planning a resection (thyroidectomy
versus hemithyroidectomy) in this population. A recent metaanalysis supporting the use of FNA included 12 studies collectively reviewing 183 patients with malignant nodules and 347
patients with benign nodules, the age range being 1 to 21 years.
The analysis found that FNA has a sensitivity of 94% and
specificity of 81% for detecting malignancy.22 A selection of
individual studies included in the analysis can be seen in
Table 58-3.18,2327 In part, FNA is not very specific because
when the cytology reveals follicular cells, malignancy cannot
be determined because the diagnosis hinges on the presence
of capsular invasion, which can only be seen on histology.28,29
Overall, FNA is a useful tool in the workup of pediatric
patients with thyroid nodules but should not overshadow
other important clinical information, such as a history of
radiation or prior malignancy; the family history; enlarging,
fixed nodules; or associated cervical lymphadenopathy.
If FNA is performed, the results are reported as benign, malignant, or indeterminate. Resection is indicated for malignant
or indeterminate nodules, or for benign nodules that continue
to grow or have other worrisome features on follow-up.
WELL-DIFFERENTIATED THYROID
CARCINOMA
Well-differentiated thyroid cancer (WDTC) includes papillary
and follicular cell tumors, accounting for approximately 1%
of malignancies in prepubertal children and up to 7% in
TABLE 58-3
Sensitivity and Specificity of FNA in Children and Adolescents
Study
Type of Study
Age Range (Years)
Chang and Joo, 2006

Amrikachi et al, 2005
Arda et al, 2001
Khurana et al, 1999
Raab et al, 1995
Gharib and Goellner, 1993
Retrospective
Retrospective
Prospective
Retrospective
Retrospective
Retrospective
2-21
10-21
Children
9-20
1-18
<17
No. of Patients
37
31
44
57
63
41
Minimum Follow-up
23
24
24
24
24
24
months
months
months
months
months
months
Sensitivity
100%
100%
100%
92.9%
88.9%
90%
Specificity
85.7%
64.7%
95%
81.4%
92.6%
96.8%
Data from Stevens C, Lee JK, Sadatsafavi M, Blair GK: Pediatric thyroid fine-needle aspiration cytology: A meta-analysis. J Pediatr Surg 2009;44:2184-2191; with
permission from Elsevier.
CHAPTER 58
adolescents, making it the most common endocrine cancer in

the pediatric population.13 Thyroid cancer occurs at least
4 times as often in females as males and is most common in
white patients.30 The overall incidence does appear to be increasing in children at a rate of approximately 1.1% per year.
This is potentially because of a rise in the use of radiotherapy
for other malignancies.31 Head and neck radiation has been
widely recognized as a significant risk factor for the development of WDTC. Patients exposed to as little as 50 cGy prior to
the age of 4 years have presented 10 to 30 years later with thyroid cancer.32 Malignancies with the strongest correlation
with a secondary thyroid cancer include Hodgkin and nonHodgkin lymphoma, neuroblastoma, and Wilms tumor.33
It is therefore important to provide careful follow-up for
children who have been successfully treated for cancer.
On a molecular level, there is increasing evidence that the
RET (rearranged during transfection) proto-oncogene plays a
role in WDTC. The RET proto-oncogene is a tyrosine kinase
receptor that, when exposed to radiation, can fuse to another
gene to form a hybrid oncogene (RET/PTC), resulting in increased tyrosine kinase activity.34 Even more common is for
WDTC to be caused by an activating mutation in the B isoform
of the Raf kinase (BRAF). These cancers tend to be larger at the
time of presentation and may have a poorer prognosis.35
Papillary cancer is the most common type of thyroid cancer
seen in children but also has the best survival, with estimates
of 98% at 5 years. Follicular cancer is about 6 times less
common, with approximately 96% survival at 5 years.30,36
In general, the treatment of WDTC in children is similar to
that of adults. However, it must be taken into consideration
that children often present with larger tumors, are more likely
to have metastatic spread to cervical nodes, have a higher recurrence rate, and have a longer overall survival.37 The primary therapy is surgical resection of the gland with or
without lymph node dissection, depending on whether there
is clinical evidence of nodal disease.38 There continues to be
some debate regarding how much of the gland should be
resected. The surgical options include total thyroidectomy,
near-total thyroidectomy (leaving less than 1 g of tissue near
the ligament of Berry), subtotal thyroidectomy, and hemithyroidectomy. The argument for more aggressive resection revolves around decreased recurrence rates, the ability to treat
residual disease with radioiodine ablation therapy, and the
ability to assess for recurrence by following thyroglobulin
levels. In addition, some surgeons prefer total or near-total
thyroidectomy in the setting of malignancy, because at least
50% of children will have bilateral or multifocal disease.39
In a retrospective review of 68 children who were less than
19 years of age and undergoing thyroid surgery for malignancy, 75% had a total thyroidectomy, 9% had a lobectomy,
and 6% had a subtotal thyroidectomy. Forty-four percent of
patients who had less than a total thyroidectomy needed further surgery for recurrence compared with only 12% of patients who had a total thyroidectomy, which was a
significant difference.36 Similarly, a larger, recent review of
215 children with papillary thyroid cancer also found that recurrence rates were significantly higher in children having undergone lobectomy compared with total or near-total
thyroidectomy (35% vs. 6%).40 However, mortality from thyroid cancer in children is very low, with 98.8% overall survival
at 10 years, and therefore it is unclear whether the increased
recurrence rate affects mortality.
749
Support for less-than-total thyroidectomy is based upon

minimizing potential morbidity from the operation. Estimates
of complications after total thyroidectomy vary widely,
depending on the series and the time frame of the study, ranging anywhere from 12% to 20%.36,41,42 Two of the most significant complications are injury to the recurrent laryngeal
nerves and permanent hypoparathyroidism. For an experienced surgeon, the incidence of recurrent laryngeal nerve injury is less than 1%, and it is less than 2% for permanent
hypoparathyroidism.43 However, a recent study found that incidental removal of parathyroid glands occurs in up to 21% of
thyroid surgeries and is not clearly dependent on the extent of
resection.44 A retrospective review of 1200 children undergoing thyroid and parathyroid surgery found that hypocalcemia
accounted for 68% of the complications, and voice disturbances accounted for another 6%. Interestingly, they also
found that the complication rates were age dependent: 22%
in children less than 6 years, 15% in children aged 7 to 12
years, and 11% in children aged 13 to 17 years, and these differences were statistically significant.45 It does appear that surgeons with more experience and a higher case volume of
thyroid surgery have fewer complications. In a study by Tuggle
and colleagues,46 surgeons were classified as high volume
(>30 cases per year), pediatric (>90% of cases were children),
or other. A total of 607 patients were included in the study,
and the authors found that high-volume surgeons, on average,
performed 72 thyroid procedures per year, pediatric surgeons
performed 2 thyroid procedures per year, and other surgeons performed 7 thyroid procedures per year. The complication rates were 8.7%, 13.4%, and 13.2%, respectively.
One strategy to reduce the incidence of hypoparathyroidism is to autotransplant one or two of the glands into the sternocleidomastoid, which can be done immediately during the
dissection if the blood supply is thought to be compromised.47
To minimize injury to the recurrent laryngeal nerve, intraoperative nerve stimulation can be used, which can help
identify its course during dissection.48,49
Unfortunately, the recurrence rate for thyroid cancer in
children after surgical resection is up to 32% when followed
for 40 years.40 For this reason, long-term follow-up for these
children is required. Current recommendations for children
who have had a total or near-total thyroidectomy include a
131
I whole-body scan 6 weeks after thyroid resection to assess
for any residual disease, and treatment with the radionuclide
can be administered as needed, at this time, for remnant ablation.50 These children can then be monitored for recurrence
by checking annual plasma thyroglobulin and antithyroglobulin antibody levels, as well as obtaining an annual neck ultrasound scan. Further whole-body radionuclide scanning is
unnecessary for children with low-risk tumors, undetectable
thyroglobulin levels, and negative neck US. Annual scans
can be considered for patients with intermediate- or high-risk
tumors but should be done with a low-activity radionuclide.51
There is now evidence that radiation exposure from radioiodine remnant ablation (RRA) may predispose children to
other malignancies. This potential risk must be carefully considered given the low mortality (<2%) associated with thyroid
cancer. In a recent review of 215 patients less than 21 years
of age with papillary thyroid cancer, there were no diseaseassociated deaths within the first 20 years following surgery.
In addition, none of the patients with distant metastases died
of their disease. It was found that recurrence rates of local and
750
PART V
HEAD AND NECK
TABLE 58-4
ATA Risk Category Guidelines for Prophylactic Thyroidectomy and Screening for Medullary Thyroid Cancer
ATA Risk Level
Age at RET Testing
Age at First US
Age at First Serum Calcitonin
Age of Prophylactic Surgery
D
C
B
A
Within first year

<3-5 years
<3-5 years
<3-5 years
Within first year

>3-5 years
>3-5 years
>3-5 years
6 months unless already postoperative

>3-5 years
>3-5 years
>3-5 years
Within first year

Before age 5 years
Consider before age 5 years; may delay*
May delay after age 5 years*
*Criteria for delay must be met: normal basal and stimulated calcitonin, normal annual neck US, less aggressive MTC family history.
ATA, American Thyroid Association; MTC, medullary thyroid cancer; RET, rearranged during transfection (proto-oncogene); US, ultrasonography.
Adapted from Kloos RT, Eng C, Evans DB, et al: Medullary thyroid cancer: Management guidelines of the American Thyroid Association. Thyroid. 2009;19:565-612;
with permission from Mary Ann Liebert, Inc.
distant disease were lower in patients treated with RRA, but

this did not reach statistical significance. Interestingly, they
did report a statistically higher mortality rate from secondary
malignancy compared with an aged-matched control group.
Of the patients in the study who died of a secondary malignancy, 73% had received RRA. This does not prove the
association, but it does raise some concern.40
MEDULLARY THYROID CANCER

Medullary thyroid cancer (MTC) arises from the parafollicular
cells of neuroectodermal origin and accounts for approximately 5% of thyroid cancers. MTC that arises sporadically
usually involves only one lobe, whereas cases of familial inheritance usually involve both lobes. Inherited MTC includes the
multiple endocrine neoplasia type 2 syndromes (MEN2A and
MEN2B) as well as familial medullary thyroid cancer. In general, hereditary medullary thyroid cancer begins with parafollicular cell hyperplasia and then progresses to invasive
microcarcinoma, followed by macroscopic disease if left
untreated. The RET proto-oncogene also plays an important
role in medullary thyroid cancer, including not only familial
but also 40% of sporadic cases,5255 and RET testing is used
to make the diagnosis of MEN2. When there is a germline
RET mutation, the aberrant protein is expressed in all of the
tissues in which it is produced, leading to MEN2 or familial
medullary thyroid cancer. Sporadic MTC occurs when there
is a somatic mutation with aberrant protein expression only
in the thyroid.56 Although essentially all patients with MEN2A
will eventually develop medullary thyroid cancer, MTC in
MEN2B tends to develop earlier and is in general more aggressive.35 Because of this, when families are known to carry the
RET mutation, genetic testing should be performed before 5
years of age in families with MEN2A and even earlier for
MEN2B. Less clear is the age at which to perform prophylactic
thyroidectomy. On the one hand, the goal is to perform thyroidectomy well before the onset of metastatic disease, after
which point cure can be difficult,56 but on the other hand
it is also important to minimize the risks to the recurrent
laryngeal nerve and parathyroid glands, which are at higher
risk in smaller children. Determining the ideal age for thyroidectomy has been based upon numerous criteria, including
the age of the youngest family member to develop cancer, the
mean age of onset for a particular genotype, and yearly
neck ultrasound findings.5759 It is known that for MEN2A,
microinvasive carcinoma can be seen in children as young
as 5 years of age.60 The most current recommendations have
been established by the American Thyroid Association (ATA)
and are guided by the fact that the risk of developing MTC
correlates with specific RET mutations where different codons

are known to have different clinical behavior.61 There are four
risk categoriesATA-A having the lowest risk and ATA-B
through ATA-D almost always resulting in MTC.62 Specific
recommendations based on these risk categories are seen in
Table 58-4. In general, a central neck node dissection is not
necessary during prophylactic thyroidectomy for children,
unless there is evidence of nodal disease.63 For patients with
a suspicion of sporadic medullary thyroid cancer, serum calcitonin levels are useful for screening.62 Treatment after the
diagnosis of established medullary cancer includes total thyroidectomy with central node dissection of all nodal tissue
from the hyoid bone to the sternal notch and to the carotid
sheaths laterally. Preoperative neck ultrasonography is recommended for all children with medullary thyroid cancer to assess for nodal disease.64 Children can then be followed with
serum calcitonin and carcinoembryonic antigen (CEA) levels,
which reliably correlate with disease recurrence. Survival in
patients with established medullary thyroid cancer has been
most recently reported as 96% at 5 years and 86% at 15
years.30 Novel therapy in adults with metastatic medullary
thyroid cancer includes the use of an oral RET inhibitor that
blocks kinase signaling. There is evidence that the inhibitor
may halt disease progression in the adult population.65 To
date, similar studies have not been done in children.
Parathyroid Embryology
and Physiology
------------------------------------------------------------------------------------------------------------------------------------------------
The parathyroid glands arise from the third and fourth pharyngeal pouches, which are paired endoderm-lined structures
between the branchial arches. By the sixth week of gestation,
the dorsal aspects of the third and fourth pouches differentiate
into the inferior and superior parathyroids, respectively. The
ventral aspects of the third pouches form the thymus and
the ventral aspects of the fourth pouches develop into the ultimobranchial body, which eventually fuses with the thyroid to
supply the parafollicular cells that produce calcitonin. The
thymus and parathyroid glands then lose their connection
to the pharynx and descend into the neck. Typically the parathyroid glands migrate to the posterior aspect of the thyroid
gland, where they obtain their blood supply from the thyroid
capsule. However, there is significant variability in the
eventual location of the parathyroid glands, as seen in
Figure 58-3. They are variable in both location and number
and can be found anywhere in the vicinity of the thyroid or
thymus. The superior glands tend to be more consistent in
their location than the inferior glands.
CHAPTER 58
1.2%
14.8%
1.2%
82.7%
Normal position
n = 67 glands
Thymus
n = 12 glands
Carotid sheath
n = 1 gland
Paraesophageal
n = 1 gland
FIGURE 58-3 Location of parathyroid glands found and removed during

primary parathyroidectomy. (Reprinted from Schlosser K, Schmitt CP,
Bartholomaeus JE, et al: Parathyroidectomy for renal hyperparathyroidism
in children and adolescents. World J Surg 2008;32:801-806; with permission
from Springer.)
The parathyroid glands regulate calcium and phosphorous

by secreting parathyroid hormone (PTH). PTH is an 84amino
acid protein with a very short half-life that is primarily metabolized by the liver and kidney. PTH secretion is normally stimulated by a drop in circulating calcium and is then inhibited by
a negative feedback loop. An increase in serum phosphorus
also indirectly stimulates PTH secretion by lowering serum
calcium. PTH also acts directly on bones by increasing osteoclast activity, on the kidneys by increasing renal calcium absorption, and on the gastrointestinal (GI) tract by increasing
vitamin D activation.
Disorders of the Parathyroid

Glands
------------------------------------------------------------------------------------------------------------------------------------------------
DiGEORGE SYNDROME
DiGeorge syndrome is a congenital disorder characterized by
athymia or thymic hypoplasia and absence of the parathyroid
glands due to failure of the third and fourth pharyngeal
pouches to differentiate. Infants are hypocalcemic and also
have increased susceptibility to infection. Additional associated characteristics can include shortened philtrum, low set
ears, nasal clefts, thyroid hypoplasia, and cardiac anomalies,
especially truncus arteriosum. Treatment is largely symptomatic and includes supplementation with calcium and vitamin
D. Emerging therapies include replacement with synthetic
PTH66 and thymus transplantation for children with severe
immunodeficiency.67
HYPERPARATHYROIDISM
There are several forms of hyperparathyroidism, all of which
lead to excessive PTH secretion. In the case of primary hyperparathyroidism, this results from one or more abnormal
751
parathyroid glands. In secondary hyperparathyroidism, elevated PTH is in response to hypocalcemia, and in tertiary hyperparathyroidism, PTH remains elevated despite correction
of the hypocalcemia. Both secondary and tertiary types are
seen in the setting of renal disease. Neonatal severe hyperparathyroidism (NSHPT) is a rare disorder that presents with
severe hypercalcemia resulting from a homozygous loss-offunction mutation with four-gland hyperplasia. Treatment
has traditionally been with a three and one halfgland parathyroidectomy or total parathyroidectomy with autotransplantation. More recently, there has also been some success
with medical management with bisphosphonates.68
In general, hyperparathyroidism presents clinically with
symptoms related to elevated calcium or is suspected when
hypercalcemia is found incidentally. However, the differential
diagnosis of hypercalcemia in children is quite broad, and this
must be kept in mind during the initial evaluation, as outlined
in Table 58-5. For example, familial hypocalciuric hypercalcemia is an autosomal dominant condition resulting from a mutation in the calcium sensing receptor. Children have elevated
serum calcium but can be distinguished from having hyperparathyroidism by detecting decreased 24-hour urinary calcium levels. PTH levels are often within the normal range.
Usually these children are completely asymptomatic, and no
specific treatment is required. This condition should be
excluded prior to considering parathyroid resection for
hypercalcemia.
The incidence of primary hyperparathyroidism in children
is estimated to be 2 to 5 per 100,000.69 In contrast to adults,
children may present with more serious effects of hypercalcemia, including renal failure, cardiac arrhythmias, and osteopenia.2 When suspected, the diagnosis can be confirmed by
measuring serum calcium and PTH. Most commonly, there
will be a solitary parathyroid adenoma, and there are several
localization studies that can be used to identify the abnormal gland. The imaging options include ultrasonography,
TABLE 58-5
Causes of Hypercalcemia in Children
Endocrine
Primary hyperparathyroidism
Secondary hyperparathyroidism
Tertiary hyperparathyroidism
Thyrotoxicosis
Familial hypocalciuric hypercalcemia
Neonatal severe hyperparathyroidism
Ectopic parathyroid hormone production
Granulomatous disease
Sarcoidosis
Tuberculosis
Fungal infection
Pharmacologic
Vitamin D
Vitamin A
Thiazide diuretics
Theophylline
Milk alkali
Lithium
Immobilization
Subcutaneous fat necrosis
752
PART V
HEAD AND NECK
Parathyroid hormone (pmol/L)
INTRAOPERATIVE PTH
800
700
600
500
400
Parathyroid
hormone
300
200
100
0
0
10
Time (minutes)
FIGURE 58-4 A, Parathyroid adenoma in the setting of hypercalcemia. The small arrow demonstrates the parathyroid adenoma, and the large arrow
demonstrates the retracted thyroid gland. B, Intraoperative PTH monitoring for the patient seen in A demonstrating the fall in PTH over 10 minutes after
removal of the parathyroid gland.
99m
Tc-sestamibi planar scintigraphy, SPECT, SPECT/CT, and/

or magnetic resonance imaging (MRI). Because of the rarity
of parathyroid adenomas in children, most of the studies used
to compare techniques have been done in adults. In one study
by Munk and colleagues,70 ultrasonography and 99mTc-sestamibi scans were in agreement 70% of the time and, when in
agreement, identified the correct gland in 97% of cases. When
these tests were not in agreement, MRI was used, and if consistent with either an ultrasound or 99mTc-sestamibi scan, the correct gland was identified 100% of the time. There were six
patients (11%) in whom there was no definitive agreement
among the three tests. More advanced scintigraphy is now being used (SPECT), which allows for a three-dimensional reconstruction, and this can also be combined with traditional CT
(SPECT/CT), allowing for more anatomic detail.71 However,
whether the added detail improves accurate localization is still
debated.72 If localization is successful, a directed resection can
then be used instead of the traditional four-gland exploration.
However, in the hands of an experienced surgeon, this remains
a very reliable approach. To confirm resection of the adenoma,
intraoperative PTH measurement can be used. The half-life of
PTH is approximately 3 to 4 minutes, and at least a 50% decrease from baseline should be observed to confirm removal
of the abnormal parathyroid.73 An example of a typical decline
can be seen in Figure 58-4. It is important to follow the PTH
level far enough out to ensure the decline is not temporary,
which can be seen with multiple-gland disease.74 If the PTH
remains elevated, a complete cervical exploration should be
pursued. If a normal parathyroid is resected or devascularized
during thyroid surgery, the gland should be autotransplanted
into the forearm or sternocleidomastoid.75 Hyperparathyroidism is also seen in the setting of multiple endocrine
neoplasia and is usually the result of four-gland hyperplasia.
The surgical options include resection of only visibly

enlarged glands, three and one halfgland parathyroidectomy,
or total parathyroidectomy with autografting.76,77
Secondary and tertiary hyperparathyroidism are most commonly seen with end-stage renal disease and affect all four
glands. Usually, this can be managed with dietary modifications, dialysis, phosphate binders, and vitamin D. However,
vitamin D analogs and calcimimetics have not been approved
for long-term use in children because of concern for interference with longitudinal growth and the possible impact on timing of puberty.78 There is some debate as to whether the
optimal surgical management is total parathyroidectomy with
autotransplantation or subtotal parathyroidectomy, but the
preferred management in the United States tends to be the
former.75,79
Parathyroid Carcinoma
------------------------------------------------------------------------------------------------------------------------------------------------
Parathyroid carcinoma is exceedingly rare in children, with

only seven case reports in the literature to date. In all cases,
the patients presented with a neck mass and severe hypercalcemia with extremely elevated PTH (3 to 10 times normal).
The first step in management is to control the hypercalcemia,
followed by surgical excision if appropriate. The recommended resection includes en-bloc hemithyroidectomy, parathyroidectomy, and lymph node dissection. If completely
resected, 90% long-term survival can be achieved. However,
with incomplete resection, the recurrence rate is up to 50%.80
expertconsult.com.
lesions to minimize the risks of recurrence, infection, or malignancy. Knowledge of the relevant local anatomy and
adjacent structures is crucial to safe surgical dissection.
Embryology
------------------------------------------------------------------------------------------------------------------------------------------------
CHAPTER 59
Neck Cysts and

Sinuses
Craig Lillehei
Cysts and sinuses of the neck represent a wide variety of

anomalies, both congenital and acquired. The focus of this
chapter is on those of congenital origin. It is a fascinating opportunity to apply our understanding of embryologic development to the spectrum of malformations seen and to guide
appropriate management. The most common lesions arise
from thyroglossal duct or branchial anomalies (particularly
from the second cleft). Nonetheless, one must be cognizant
of the range of usual variants as well as the broad differential
diagnosis. Lymphatic and vascular malformations will be
addressed elsewhere (see Chapter 125). Although present at
birth, congenital lesions may not become evident for weeks,
months, or even years. However, there are recognizable patterns. Accurate diagnosis guides appropriate intervention.
Proper identification is aided by careful history and physical
examination. Age at presentation, evolution, anatomic location, and associated drainage are often important diagnostic
clues. Radiographic studies, such as ultrasonography or
cross-sectional imaging (computed tomography [CT], magnetic resonance imaging [MRI]), may be helpful in selected
cases. Although the exact identity is not always evident preoperatively, an awareness of differential diagnostic possibilities will guide the prepared mind and improve prospects
for optimal outcomes. The emphasis is total excision of the
During the fourth week of gestation, neural crest cells migrate

into the future head and neck region. A series of six paired
branchial or pharyngeal arches begin to develop (Fig. 59-1).
In humans the fifth arch, if present, is only very short lived.
Their mesoderm is covered externally by ectoderm and lined
internally by endoderm. Each arch contains a distinct artery,
nerve, cartilage rod, and muscle. These arches are separated
by depressions that are referred to as clefts on their external
ectodermal surface, and pouches on their internal endodermal
surface. These swellings may give rise to normal cervical
structures, leave pathologic remnants, or involute entirely.
In fish and amphibians, the closing membranes that separate
the branchial pouches and clefts regress with the resultant
connections forming gills.1
Each branchial arch and its components can be traced
to the formation of future anatomic structures as outlined
in Table 59-1. The first branchial arch forms the mandible
and a portion of the maxilla. It is also involved in structures of the inner ear. The first cleft and pouch connect
to form the eustachian tube/middle ear, tympanic membrane, and external auditory canal. The other branchial
cleft components usually regress. However, it is important
to note that during development the second, third, and
fourth branchial clefts share a common external opening,
the cervical sinus of His (Fig. 59-2, A). For this reason
the location of the external opening of a persistent sinus
or fistula from these clefts cannot be used to distinguish
between them.
The remaining pouches give rise to normal glandular structures (Fig. 59-2, B). The palatine tonsil and supratonsillar
fossa originate from the second pouch. The tonsils are the only
structures to remain at their pouch of origin. The third pouch
gives rise to the thymus and inferior parathyroid glands, while
the fourth pouch is responsible for the superior parathyroid
glands. It is believed that the calcitonin-producing cells of
the thyroid gland arise from the ultimobranchial body,
probably a remnant of the ventral portion of the caudal pharyngeal complex formed by the fourth and vestigial fifth
pouch (see Fig. 59-2).
The thyroid gland arises from an endodermal thickening in
the floor of the primitive pharynx called the tuberculum impar
(see Fig. 59-1). A bilobate diverticulum develops between the
anterior and posterior muscle complex of the tongue. As the
embryo elongates this anlage descends anterior to or through
the eventual location of the hyoid bone and fuses with elements of the fourth and fifth branchial pouches to form the
thyroid gland. This descending median thyroid anlage gives
rise to the thyroglossal duct, which usually obliterates by
the fifth week of gestation.2 The proximal remnant of this
pathway is the foramen cecum at the base of the tongue,
whereas the distal remnant is represented by the pyramidal
lobe of the thyroid gland (Fig. 59-3). Cystic remnants or accessory thyroid tissue can remain anywhere along this tract
(Fig. 59-4).
753
754
PART V
HEAD AND NECK
ARCH
Tuberculum impar
Branchial arch
Mandibular n.
Facial n.
Glossopharyngeal n.
Superior laryngeal
branch of vagus
Branchial cleft
FIGURE 59-1 Early development
of branchial apparatus. (From
Donegan JO: Congenital neck masses. In Cummings CW, Fredrickson
JM, Harker LA, et al [eds]:
OtolaryngologyHead and Neck
Surgery, ed 2. St Louis, Mosby-Year
Book, 1993.)
Branchial pouch
Foramen cecum
Branchial nerve
Arch 5 disappears
Branchial artery
6
Branchial cartilage
Recurrent laryngeal
branch of vagus
TABLE 59-1
Derivatives of Branchial Arches, Clefts, and Pouches
I
Arch
External maxillary artery
Nerve V
Cleft
Pouch
II
Arch
Stapedial artery
Dorsal
Ventral
Midline Floor of Pharynx
Incus body
Malleus head
Pinna
External auditory canal
Eustachian tube
Middle ear cavity
Mastoid air cells
Meckel cartilage
Malleus
Body of tongue
Stapes
Styloid process
Hyoid (lesser horn and
part of body)
Root of tongue
Foramen cecum
Nerves VII and VIII

Pouch
Thyroid glands
median anlage
Palatine tonsil
Supratonsillar fossa
III
Arch
Internal carotid artery
Nerve IX
Pouch
IV
Arch
Arch of aorta (L)
Part of subclavian artery (R)
Nerve X
Pouch
V
Arch
Pouch
VI
Arch
Pulmonary artery
Ductus arteriosus (L)
Nerve X (recurrent laryngeal)
Hyoid (greater horn and

part of body)
Part of epiglottis
Thymus
Inferior parathyroid
Piriform fossa
Thyroid cartilage
Cuneiform cartilage
Part of epiglottis
Superior parathyroid (lateral
anlage of thyroid gland)
Thymus (inconstant)
Ultimobranchial body (lateral

anlage of thyroid gland)
Cricoid
Arytenoid
Corniculate cartilage
From Skandalakis JE, Gray SW, Todd NW: The pharynx and its derivatives. In Skandalakis JE, Gray SW (eds): Embryology for Surgeons, ed 2. Baltimore, Williams &
Wilkins, 1994.
CHAPTER 59
NECK CYSTS AND SINUSES
755
Maxillary
process
Mandibular
process
FIGURE 59-2 A, Schematic representation of the development of the

pharyngeal (or branchial) clefts and
pouches. Note that the second arch
grows over the third and fourth
arches, thereby burying the second,
third, and fourth pharyngeal clefts.
B, Remnants of the second, third,
and fourth pharyngeal clefts form
the cervical sinus, which is normally
obliterated. Note the structures
formed by the various pharyngeal
pouches. (From Sadler TW: Head
and neck. In Langman J, Sadler TW
(eds): Langmans Medical Embryology, ed 7. Baltimore, Williams &
Wilkins, 1995.)
Pharyngeal
pouches
Pharyngeal
clefts
1
I
External
auditory
meatus
1
II
Primitive
tympanic
cavity
Auditory
tube
II
Palatine
tonsil
Parathyroid
gland (inferior)
III
III
3
3
Thymus
IV
4
5
IV
Cervical
sinus
Ultimobranchial
body
Epicardial
ridge
Parathyroid gland
(superior)
Body of tongue
Auditory tube
Foramen cecum
Primitive
tympanic
cavity
Thyroglossal cyst
Ventral side of
pharynx
External
auditory
meatus
Epiglottis
Foramen
cecum
Hyoid bone
Palatine tonsil
Thyroglossal cysts
Thyroid cartilage
Superior parathyroid gland

(from 4th pouch)
Inferior parathyroid gland
(from 3rd pouch)
Thyroid
gland
Thyroid gland
Ultimobranchial body
Thymus
Cricoid cartilage
Foregut
FIGURE 59-3 Schematic representation of migration of the thymus,

parathyroid glands, and ultimobranchial body. The thyroid gland originates in the midline at the level of the foramen cecum and descends to
the level of the first tracheal ring. (From Sadler TW: Head and neck. In
Langmans Medical Embryology, ed 7. Baltimore, Williams & Wilkins, 1995.)
Although the exact incidence varies between different pediatric series, thyroglossal duct remnants are typically the
more common etiology of congenital neck cysts or sinuses,
followed closely by branchial cleft remnants.35 In general,
thyroglossal duct lesions lie close to the midline, whereas
branchial remnants present more laterally in the neck,
although atypical locations have been described.6
FIGURE 59-4 Various locations of thyroglossal duct cysts. (From Sadler TW:
Head and neck. In Sadler TW (ed): Langmans Medical Embryology, ed 11.
Baltimore, Lippincott Williams & Wilkins, 2010.)
Thyroglossal Duct Cysts

------------------------------------------------------------------------------------------------------------------------------------------------
Thyroglossal duct remnants are clearly the most common

midline congenital cervical anomalies. As described, they occur along the path of thyroid descent from the foramen cecum
at the base of the tongue to the lower neck. The remnants usually lie in close proximity to the hyoid bone. Given this relationship one can appreciate why the cysts often move
cephalad with swallowing or tongue protrusion. Although
classically described as midline, up to 40% may lie just lateral
to the midline. Most lesions present as cystic masses, but up to
25% have a draining sinus.7 Because the thyroglossal duct
756
PART V
HEAD AND NECK
does not communicate with ectoderm during development,

the sinus is either the result of spontaneous rupture, infection,
or a prior drainage procedure. Approximately 60% of thyroglossal duct cysts are adjacent to the hyoid bone, 24% lie
above the hyoid, and 13% lie below.8 The remaining 8% of
cysts are intralingual and may pose a risk for acute airway
obstruction, particularly in the neonate.9 Most thyroglossal
duct cysts present during the first 5 years of life.2
In view of the potential communication with the oral cavity,
it is not surprising that the cysts may fluctuate in size or that
approximately one third of patients present with an active infection of the cyst or history of prior infection.2 Some patients
may actually report noticing a foul taste. The most common
pathogens are Haemophilus influenzae, Staphylococcus aureus,
and Staphylococcus epidermidis.7,10
The thyroglossal duct remnants are lined by ductal epithelium and may contain solid thyroid tissue. In fact, in roughly
1% to 2% of patients with presumed thyroglossal duct cysts,
the actual lesion is a median ectopic thyroid.8 It may represent
their only functional thyroid tissue in which case excision
would be problematic. Further evaluation may be obtained
with a screening thyroid-stimulating hormone (TSH) level
and neck ultrasonography. If there is evidence of hypothyroidism or the mass appears solid without a visible normal thyroid
gland, one might wish to obtain a thyroid scan to determine
whether there is any additional thyroid tissue. Such patients
are often hypothyroid with an elevated TSH, which is responsible for the hypertrophy of the ectopic tissue. In the setting
of hypothyroidism, hormonal supplementation would be appropriate and might promote shrinkage of the hypertrophic
thyroid tissue, thereby obviating the need for surgery.
FIGURE 59-5 Sistrunk procedure: intraoperative photograph of resection of thyroglossal duct cyst in continuity with central hyoid bone (arrow).
SURGICAL MANAGEMENT
The primary indication for excision of thyroglossal duct remnants is to avoid problems with recurrent infection. However,
malignancy within thyroglossal duct remnants is also well described.11 Such tumors usually present as papillary carcinoma
in adults, but pediatric cases are reported, and multiple cell
types have been encountered.1214
Appropriate surgical management of uncomplicated thyroglossal duct disease involves complete resection of the cyst
and its tract in continuity with the central hyoid bone, as described by Sistrunk.15 One should be aware that in young children the hyoid bone may override the thyroid notch,
potentially placing the larynx at risk. The patient is positioned
supine with the head elevated and neck extended. A transverse
cervical incision is used to carefully mobilize the cyst along
with its tract. The underlying hyoid bone is divided about
1 cm from the midline on either side after dividing the attachments of the mylohyoid and hyoglossus muscles from its
superior border. En-bloc resection is completed with suture
ligation of the proximal tract, prior to removal of the specimen
(Fig. 59-5). Elegant studies of resected surgical specimens
by Horisawa and colleagues, as depicted in Figure 59-6, demonstrate the importance of this strategy to achieve complete excision, thereby reducing the likelihood of recurrence.16
In the setting of acute infection, initial efforts are aimed to
control the infection. If antibiotics alone are insufficient, aspiration or incision and drainage of the cyst/abscess may be required. Once the infection is well-controlled, the described
Sistrunk procedure can be performed using an elliptic skin
b
c
Hyoid bone
Hyoid bone
Cyst
Cyst
FIGURE 59-6 Diagram of the common running pattern of the thyroglossal duct based on anatomic reconstruction. a, Horizontal distance from
midline to the most distant thyroglossal duct; b, length of the single duct
above the hyoid bone; c, point where the diameter of the duct is measured. (From Horisawa M, Niinomi N, Ito T: What is the optimal depth
for core-out toward the foramen cecum in a thyroglossal duct cyst operation? J Pediatr Surg 1992;27:710-713.)
incision around the cutaneous opening to permit excision

of this tract in continuity with the remainder of the specimen.
Recurrences after thyroglossal duct excisions may occur in
up to 10% of cases.17,18 The most likely cause is incomplete
excision of the tract or intraoperative rupture. An association
with preoperative infection has been suggested,19,20 but not
confirmed in more recent analyses. Postoperative infection is
clearly associated with recurrence, but it is uncertain whether
this development represents cause or effect.18 Excision of a recurrent thyroglossal duct remnant has a 20% to 35% risk of
failure.10,21 Wider resection is recommended, including the
pyramidal lobe if present, central strap muscles, additional
hyoid bone, and residual tissue up to the foramen cecum.17,22
CHAPTER 59
757
Branchial Anomalies
------------------------------------------------------------------------------------------------------------------------------------------------
Most branchial cleft anomalies arise from the second cleft/

pouch, with a much smaller proportion from the first. Remnants of the third or fourth pouches are rare. It is the internal
opening of branchial sinuses that best defines their embryologic origin. The anomalies may present as fistulae, cysts, sinus
tracts, or cartilaginous remnants and are thought to arise from
incomplete obliteration during embryogenesis. To clarify,
cysts have mucosal or epithelial lining, but no external openings. Sinuses may communicate either externally with the skin
or internally with the pharynx, whereas fistulae connect to
both. When an external tract is present, branchial anomalies
are usually diagnosed within the first decade of life. However,
when there is no external opening the diagnosis may be
delayed into adulthood. Up to 10% of these lesions are bilateral as depicted in Figure 59-7.23,24 The presence of preauricular pits in patients with branchial anomalies should raise
the suspicion for the branchio-oto-renal (BOR) and branchiooculo-facial (BOF) syndromes. Both are autosomal dominant
conditions with associated hearing loss, ear malformations,
and renal anomalies in the BOR syndrome, while BOF includes
eye anomalies, such as microphthalmia and obstructed lacrimal
ducts, and facial anomalies consisting of cleft or pseudocleft lip/
palate.24,25
It is worth mentioning that short sinus tracts, pedunculated
skin appendages or subcutaneous cartilaginous remnants are
often encountered in the anterior neck and upper chest. These
structures are probably branchial remnants, but cannot usually
be ascribed to a specific arch. Lesions presenting below the
clavicles are more likely epidermoid or dermoid cysts rather
than branchial remnants. Most often elective excision is used.
SECOND BRANCHIAL ANOMALIES

Second branchial cleft anomalies typically lie somewhere between the lower anterior border of the sternocleidomastoid
(SCM) muscle and tonsillar fossa of the pharynx. They may
be in close proximity to the glossopharyngeal and hypoglossal
nerves as well as carotid vessels as the tract travels through the
carotid bifurcation and over the nerves to enter the lateral
pharyngeal wall as depicted in Figure 59-8.
FIGURE 59-7 Child with bilateral second branchial cleft sinuses (arrows).
9
12
10
FIGURE 59-8 Second branchial cleft cyst and sinus tract. (From
Donegan JO: Congenital neck masses. In Cummings CW, Fredrickson JM,
Harker LA, et al [eds]: OtolaryngologyHead and Neck Surgery, ed 2. St
Louis, Mosby-Year Book, 1993.)
The branchial anomalies are lined by epithelium. Overall

cystic lesions are more common than fistulae, but usually present later (e.g., second decade).7 Cysts most often present as
nontender soft tissue masses beneath the SCM muscle. However, they may present with acute infection. Change in size
during upper respiratory infections is noted in up to 25%.26
The anomalies have been classified into four types.5 Type 1
are superficial, but located deep to the platysma and cervical
fascia, along the anterior border of the SCM muscle. Type 2
anomalies are the most common. They course deep to the
SCM muscle and either anterior or posterior to the carotid
artery. Type 3 lesions pass between the carotid bifurcation
and lie adjacent to the pharynx. Type 4 lesions are medial
to the carotid sheath and in close approximation to the
pharynx, usually at the level of the tonsillar fossa.
The most common presentation in infants and young children is a second branchial cleft sinus with drainage from a
small cutaneous pit along the anterior border of the lower sternocleidomastoid muscle. On occasion, a subcutaneous tract
is palpable more cephalad. Less common symptoms include
stridor, dysphagia, odynophagia, or cranial nerve palsies.
Branchiogenic carcinoma has been diagnosed in adults.27
Given the risks of infection, further enlargement or malignancy, elective excision is recommended once the diagnosis
has been made. There is typically no urgency; so, one can defer excision beyond 3 to 6 months of age or to allow treatment
of an acute infection. Systemic antibiotics and aspiration are
generally preferable to incision and drainage, which might
produce more distortion of the surgical planes; however,
when diagnosis is unclear, the latter allows biopsy of the cyst
wall, which can help to distinguish between an infected
branchial cleft cyst and simple bacterial lymphadenitis.
758
PART V
HEAD AND NECK
Type 1
Type 2
FIGURE 59-9 Intraoperative photo showing excision of second branchial

cleft sinus/fistula using second parallel (step-ladder) cervical incision.
The goal is complete excision of the tract without injury to

surrounding nerves or vascular structures. A transverse cervical incision in a skin crease directly over the cyst will aid to
optimize the future cosmetic result. In the case of a sinus or
fistula, precise identification may be facilitated by gently
inserting a probe, catheter, or monofilament suture into the
tract. A lacrimal probe dipped in methylene blue has also been
used to stain the tract and make it easier to identify should
it break during dissection. Excision is best accomplished by
dissection directly on the surface of the lesion. The tract
may be very thin-walled; so, one must be careful to avoid
avulsion with possible loss of the proximal lumen. If the tract
is long, exposure may be improved by a second (so-called
stepladder) incision along a skin crease more cephalad
(Fig. 59-9). Second branchial anomalies presenting as
pharyngeal cysts can be excised by an intraoral approach.28,29
FIRST BRANCHIAL ANOMALIES

First branchial cleft anomalies are rare, but more common in
females. Accurate diagnosis is difficult and may be quite
delayed.30 Remnants may persist anywhere between the external auditory canal and submandibular area. They should be
distinguished from preauricular pits and sinuses, which arise
from failure of the auricular hillocks to fuse. The first cleft
anomalies often lie in close association to the parotid gland
and facial nerve. In 1972, Work classified first branchial anomalies into types 1 and 2 (Fig. 59-10).31 Type 1 lesions are rarer
and considered duplications of the membranous external auditory canal. They are of ectodermal origin and generally course
lateral to the facial nerve. Type 2 lesions contain both ectodermal and mesodermal elements, which may include cartilage.
These anomalies pass medial to the facial nerve but may present
FIGURE 59-10 Type I and type II first branchial cleft abnormalities. (From
Donegan JO: Congenital neck masses. In Cummings CW, Fredrickson JM,
Harker LA, et al [eds]: OtolaryngologyHead and Neck Surgery, ed 2. St
Louis, Mosby-Year Book, 1993.)
in preauricular, infraauricular, or postauricular locations. Sinuses may present with external drainage below the angle of
the mandible or otorrhea, which may become infected. Cysts
present as soft tissue masses in this region which may also become secondarily infected. A communication with the external
auditory canal may be present. A careful otologic examination is
important to define the pathology.
Complete surgical excision is once again recommended,
but great care must be taken given the proximity of the facial
nerve. In infants and children, the nerve is probably even
more susceptible given that it is smaller and more superficial
without well-developed landmarks.32 Many authors recommend initial exposure of the main trunk of the facial nerve
and its peripheral branches with superficial parotidectomy
to reduce the risk of facial nerve injury.33,34 Prior infection
may distort accurate tissue dissection planes. It is necessary
to excise the involved skin and cartilage of the external auditory canal. Furthermore, if the tract extends medially to the
tympanic membrane a second operation may be required to
remove this segment.35,36
THIRD AND FOURTH BRANCHIAL

ANOMALIES
Third and fourth branchial anomalies are very rare and almost
always occur on the left side of the neck. Most present as sinuses or infected cysts rather than congenital fistulae and
drain into the piriform sinus. Although sometimes combined
CHAPTER 59
12
759
12
12
10
10
FIGURE 59-11 Third branchial cleft cyst and sinus tract. Note that these occur much more frequently on the left side (approximately 90%); also, during
surgery the tract is often seen to go straight up from the left upper thyroid lobe
area toward the thyroid cartilage, without passing behind the carotid artery
as the embryologic development would suggest. (From Donegan JO: Congenital neck masses. In Cummings CW, Fredrickson JM, Harker LA, et al
[eds]: OtolaryngologyHead and Neck Surgery, ed 2. St Louis, Mosby-Year
Book, 1993.)
generically as piriform sinus tracts, distinction is possible. As

noted in Figure 59-1 the superior laryngeal nerve represents
the nerve to the fourth branchial arch. Third pouch anomalies
enter the piriform sinus above the superior laryngeal nerve,
whereas fourth pouch anomalies enter below this nerve. A
third branchial cleft fistula theoretically would extend from
the anterior border of the SCM, traversing deep to the internal
carotid artery and glossopharyngeal nerve, piercing the thyroid membrane above the internal branch of the superior
laryngeal nerve and entering the pharynx at the piriform
sinus as depicted in Figure 59-11. A fourth branchial fistula
would course around the subclavian artery on the right or aortic arch on the left to ascend back up over the hypoglossal
nerve and enter the piriform apex or cervical esophagus
(Fig. 59-12). A complete fourth branchial fistula has yet to
be identified in humans,5 and most third branchial fistulae described appear to have been secondary to infection or repeated
surgery.37
Presentation of piriform sinus tracts may be quite subtle
and their diagnosis very challenging. Noncommunicating or
noninfected communicating cysts may present as cold thyroid
nodules, which may be partly or totally intrathyroid.37 A history of repeated upper respiratory tract infections and sore
throats, hoarseness or pain, and tenderness of the thyroid
gland should raise suspicion. Infection may result in suppurative thyroiditis38; any thyroid abscess in a child should raise
the suspicion of a branchial remnant, particularly if closely related to the left upper pole of the thyroid gland. Acute respiratory compromise in neonates has been described.39 Needle
aspiration may be required to temporarily relieve respiratory
symptoms. A contrast esophagogram after resolution of the
acute infection may demonstrate the tract from the piriform
10
FIGURE 59-12 Anatomic relationships of theoretical course of fourth

branchial fistula. Such a complete fistula has never been described in
humans. (From Liston SL: Fourth branchial fistula. Otolaryngol Head Neck
Surg 1981;89:520-522.
FIGURE 59-13 Contrast esophagogram demonstrating contrast within

piriform sinus tract (arrow).
sinus as depicted in Figure 59-13. Other imaging has been

successful if air is visualized within the cyst or tract originating
from the piriform fossa opening.40 Combinations of ultrasonography, CT, MRI, and thyroid scan may help in establishing
a diagnosis.37
760
PART V
HEAD AND NECK
Once again, complete excision is necessary to avoid continued difficulties. Often several previous operations have
been performed before the correct pathology is recognized.41,42 Direct laryngoscopy or rigid pharyngoscopy, using
a Hopkins rod-lens telescope, is recommended for accurate
diagnosis as well as endoscopic cannulation of the opening into
the piriform sinus, if possible, to facilitate accurate dissection.37,43 A standard collar incision is used with identification
of the recurrent laryngeal nerve. Partial or total ipsilateral thyroid lobectomy with excision of the tract to the piriform sinus is
usually required. Partial resection of the thyroid cartilage may
also be necessary to remove the entire tract.44 Cauterization of
the internal opening has been described.45,46
Dermoid Cysts
------------------------------------------------------------------------------------------------------------------------------------------------
Cervical dermoid cysts are thought to arise from elements

trapped during fusion of the anterior branchial arches. They
are composed of ectodermal and mesodermal elements,
but in contrast to teratomas, do not contain any endodermal derivatives.35 These lesions are typically midline and
well-circumscribed. They are lined by squamous epithelium
and usually contain sebaceous debris, which can become secondarily infected. Although they appear echogenic rather than
cystic on ultrasound examination, imaging is useful to differentiate submental dermoids from benign reactive lymph
nodes. The overlying skin is often adherent, and a small cutaneous pit may be visible. If the cyst is adherent to the underlying fascia or lies within the strap muscles, it may move with
swallowing or tongue protrusion, making the distinction from
a thyroglossal duct cyst impossible. Complete excision is appropriate. A yellowish appearance at surgery and the sebaceous
cyst content allow distinction from a thyroglossal duct cyst,
which more often contains a clear viscous fluid. If the cyst
lies adjacent to the hyoid bone and a diagnostic doubt exists,
a formal Sistrunk procedure with in continuity excision of
the central hyoid bone is recommended to ensure complete removal of the pathology.
Congenital Midline
Cervical Clefts
------------------------------------------------------------------------------------------------------------------------------------------------
Congenital midline cervical clefts are very rare anomalies

thought to arise from failure of anterior fusion of the first
two branchial arches.47,48 They typically present as a longitudinal area of thinned or atrophic skin along the anterior midline of the neck. Characteristically, there are skin tags at the
upper end and small sinus tracts at the inferior aspect. Secretions may be noted from accessory salivary glands draining
into the cleft.49,50 Early complete excision is recommended,
both for cosmesis as well as to avoid limitations to neck extension and mandibular growth. Wound closure is accomplished
using a series of Z-plasties, to avoid a contracting linear scar
(Fig. 59-14).51,52
B
FIGURE 59-14 Photographs of infant with midline cervical cleft (A) and
Z-plasty reconstruction after excision (B).
Cervical Thymic Cysts

------------------------------------------------------------------------------------------------------------------------------------------------
Thymic cysts are usually seen within the chest and mediastinum. However, given that the thymus arises from the third,
and sometimes fourth, branchial pouches, one can appreciate
the possibility for a cervical location. Cervical thymic cysts
typically present in the anterior triangle, more commonly
on the left than the right. They occur more frequently in males,
with peak onset at age 5 to 7 years.53 They may be difficult to
distinguish preoperatively from more common cystic lesions,
such as branchial cleft cysts or lymphatic malformations. They
can be unilocular or multilocular. Extension into the mediastinum is common and accounts for the often-described physical
finding of enlargement with a Valsalva maneuver. The precise
diagnosis is usually made postoperatively when elements of
CHAPTER 59
thymus are identified within the cyst wall. The fluid within
these cysts is typically brownish in color. The lesions are almost always benign. Surgical excision is generally quite
straightforward, although one needs to be cognizant of potentially adherent vessels (e.g., carotid artery, jugular vein) or
nerves (e.g., phrenic, recurrent laryngeal). Although the aim
is to completely remove the cyst, one should be careful in very
young children, to avoid removing the entire thymus, which
might have untoward immunologic consequences.
Although the focus of this chapter has been congenital
lesions, the differential diagnosis for neck cysts and sinuses
761
must be much broader. A wide variety of acquired conditions,

including infections and tumors, should also be considered.
The distinction between infection of a congenital cervical remnant and a primary cervical infection with the development of
an abscess or draining sinus may not always be straightforward. Nonetheless, an awareness of the congenital possibilities and their likely anatomic locations will assist the astute
clinician.
expertconsult.com.

PART 2 Coran Pediatric Surgery, 7th Ed PDF

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

PART 2 Coran Pediatric Surgery, 7th Ed PDF

Uploaded by

Copyright:

Available Formats

CHAPTER 28

A number of milestones in the evolution of cancer therapy

of cancers. Since the late 1980s, neuroblastoma has been the

Epidemiology and Survival

Cancer in children is uncommon; it represents only about

MAJOR TUMORS OF CHILDHOOD

Molecular Biology of Cancer

in children, and brain tumors are the most common solid

During normal cellular development and renewal, cells evolve

NORMAL CELL PHYSIOLOGY

PRINCIPLES OF PEDIATRIC ONCOLOGY, GENETICS OF CANCER, AND RADIATION THERAPY

signals for proliferation or quiescence must be integrated by

or the reactivation of telomerase in somatic cells appears to

MAJOR TUMORS OF CHILDHOOD

Growth factors and receptors

Tumor Suppressor Gene

Familial polyposis coli

Intestinal polyposis, colorectal

*WAGR: Wilms tumor, aniridia, genitourinary abnormalities, and mental retardation.

These aberrations affect growth factors and their receptors,

rates compared with those who have near-diploid or neartetraploid tumors.14

PRINCIPLES OF PEDIATRIC ONCOLOGY, GENETICS OF CANCER, AND RADIATION THERAPY

Desmoplastic small round

FIGURE 28-1 Spectrum of gross chromosomal aberrations using

(at 2q35) or PAX7 (at 1p36) with FKHR, a gene encoding a

involved in the pathogenesis of a small percentage of pediatric

MAJOR TUMORS OF CHILDHOOD

and contributes to malignant transformation. The cumulative

PRINCIPLES OF PEDIATRIC ONCOLOGY, GENETICS OF CANCER, AND RADIATION THERAPY

new blood vessels have developed) can grow to only a limited

Angiogenesis is the biological process of new blood vessel

MAJOR TUMORS OF CHILDHOOD

Low resolution analysis of metaphase

Does not require a priori knowledge

Detection of translocations, amplifications,

PCR and RTPCR

Extremely sensitive detection of DNA

Can be applied to chromosomal

Requires fresh, sterile tumor tissue for

only be effective anticancer agents but, because of their

Childhood Cancer and Heredity

Advances in molecular genetic techniques have also improved

PRINCIPLES OF PEDIATRIC ONCOLOGY, GENETICS OF CANCER, AND RADIATION THERAPY

WAGR syndrome57 and also occur in approximately 35% of

Along with an increased understanding of the molecular basis

MAJOR TUMORS OF CHILDHOOD

exquisitely so, the successful use of chemotherapy is often

the body by a specific organ (see Table 28-5). The processes of

Common Toxic Effects

A, H, M, (esp. thrombocytopenia), N/V

A, N/V, R (significant), ototoxicity,

A, N/V, SIADH, M, R, cardiac, cystitis

A, CNS, N/V, M, R, cardiac, cystitis

H, N/V, M, hepatic vein thrombosis

M, N/V, mucositis, diarrhea

M, N/V, diarrhea, CNS

CNS, N/V, M, cardiac, diarrhea,

PRINCIPLES OF PEDIATRIC ONCOLOGY, GENETICS OF CANCER, AND RADIATION THERAPY

Common Toxic Effects

CNS, H, M, R, mucositis, skin

A, H, M, N/V, mucositis, vesicant

A, M, N/V, cardiac, diarrhea, vesicant,