03.04 - Intraventricular Tumours

INTRAVENTRICULAR TUMORS
CONTENTS
Preface
Andrew T. Parsa and Mitchel S. Berger
Epidemiology and Pathology of Intraventricular Tumors
James S. Waldron and Tarik Tihan
ix
469
Intraventricular tumors present a diagnostic challenge to the clinician because of a broad

differential diagnosis with significant variability in tumor type between adult and
pediatric populations. This expansive differential diagnosis includes choroid plexus
papillomas and carcinomas, ependymomas, subependymomas, subependymal giant cell
astrocytomas, central neurocytomas, meningiomas, and metastases as well as a number
of cysts, inflammatory lesions, and other rare neoplasms. Posterior fossa ependymomas,
subependymal giant cell astrocytomas, and choroid plexus tumors are more likely to
appear in childhood, whereas subependymomas, central neurocytomas, intraventricular
meningiomas, and metastases are more frequent in adults. This article reviews the
epidemiology, the pathologic characteristics, and the primary diagnostic considerations
of each tumor type.
Intraventricular Neurocytomas
Janet Lee, Susan M. Chang, Michael W. McDermott, and Andrew T. Parsa
483
Central neurocytomas (CNCs) are World Health Organization II benign central nervous
system (CNS) neoplasms first described in 1982 by Hassoun and his colleagues. Hallmark features of CNC include (1) occurrence in the lateral ventricle of young adults,
(2) a well-circumscribed isodense to hyperdense mass with contrast enhancement on
CT and isointense to hyperintense compared with normal brain parenchyma on T1and T2-weighted MRI, (3) resemblance to oligodendroglioma on light microscopy, (4)
neuronal origin seen in electron microscopy and immunohistochemistry, and (5) favorable prognosis with benign biologic behavior. CNCs comprise 0.1% to 0.5% of all CNS
neoplasms based on pathologic review at several neurosurgery centers. A populationbased incidence has not been established, in part because of the paucity of cases. Given
its recent distinction as a unique tumor and its low incidence, most reports of CNC are
from the pathologic literature with little data regarding its management. Furthermore,
many early cases of CNC were misdiagnosed, and treatment was based on the presumed
diagnosis of oligodendroglioma or ependymoma. Accordingly, this article presents a
comprehensive review of the literature and proposes a management paradigm for the
treatment of CNC.
VOLUME 14
NUMBER 4 OCTOBER 2003
Surgical Approaches to Tumors of the Lateral Ventricle

Richard C.E. Anderson, Saadi Ghatan, and Neil A. Feldstein
509
Tumors of the lateral ventricles comprise a relatively rare heterogeneous group of lesions
in children and adults. They arise from the ependyma and subependyma that line the
ventricles, from the choroid plexus arachnoid and epithelium, or from ectopic tissue
rests that have become trapped within the ventricle or its lining. Although the lateral
ventricles are among the most surgically inaccessible areas of the brain, numerous operative approaches to the ventricles have been developed. This article first discusses the
clinical manifestations and differential diagnosis of lateral ventricular tumors. Relevant
regional anatomy and general operative strategies for these lesions are then discussed,
with particular focus on the following approaches: frontal, temporal, and parietal transcortical approaches and anterior and posterior interhemispheric approaches.
Surgical Approaches to Posterior Third Ventricular Tumors

Alan P. Lozier and Jeffrey N. Bruce
527
Advanced microsurgical techniques combined with improved neuroanesthetic and postoperative critical care have made aggressive surgical resection a mainstay in the management of posterior third ventricular and pineal region tumors. Although a variety of
approaches to the posterior third ventricle have been designed, three are in common
use. The infratentorial-supracerebellar approach takes advantage of a natural corridor
between the cerebellum and the tentorium. Supratentorial approaches include the
interhemispheric-transcallosal and occipital-transtentorial approaches. Refinements in
surgical technique have led to a more favorable outlook for patients with these uncommon tumors.
Endoscopic Adjuncts to Intraventricular Surgery

Sandeep Kunwar
547
Recently, endoscopic intraventricular surgery has been performed successfully in several

clinical series. Although the therapeutic results must be compared with conventional
surgery, neuroendoscopy seems to be a safe surgical technique when performed by surgeons with appropriate experience and refined endoscopic tools. Rigid or flexible endoscopes equipped with various-sized working channels should be selected depending on
the nature of the pathologic findings. The well-proven tenets of microsurgery must not
be sacrificed for the sake of more rapid surgical time and noninvasiveness; thus, endoscopic surgery must adhere to the principles of microsurgery. The improved visualization and lower morbidity have established neuroendoscopy in the management of
specific disease processes, such as obstructive hydrocephalus. Its further use in the management of intraventricular cysts and tumors is dependent on long-term follow-up and
the development of even better instrumentation.
Intraventricular Meningiomas
Michael W. McDermott
559
Meningiomas arising in the ventricular system are rare; yet, when they do present
clinically, they are often large, most often within the atrium, and most frequently on
the left. For all these reasons, they are tumors for which it is difficult to achieve the
perfect surgical result: complete removal of a benign tumor without complications
or new neurologic morbidity. With a thorough understanding of the anatomy of structures around the ventricle, selection of the proper surgical approach, and use of modern neurosurgical techniques, however, modern-day surgical results should be superior
to those of the past.
vi
CONTENTS
Intraventricular Gliomas
Aaron S. Dumont, Elana Farace, David Schiff, and Mark E. Shaffrey
571
Significant progress has been realized in the contemporary understanding and treatment
of intraventricular gliomas. However, there remains a substantial need for continued advancement in the clinical management of patients harboring these lesions, particularly
ependymomas. This article addresses the specific types of intraventricular gliomas with
emphasis on each tumors defining characteristics and the specific nuances of management in each variant.
Surgical Resection of Metastatic Intraventricular Tumors

Giacomo G. Vecil and Frederick F. Lang
593
Intraventricular metastases are a unique challenge for neurosurgical oncologists. This

paper describes the clinical features and surgical management strategies of intraventricular metastases based on a review of the literature and an analysis of 35 patients treated
in the Department of Neurosurgery at The University of Texas M.D. Anderson Cancer
Center over the last 10 years. Intraventricular metastases comprise 0.95% of intraparenchymal metastases. Renal cell carcinoma has the highest propensity of all primary tumors to metastasize to the ventricle. The trigone of the lateral ventricular is the most
common location with the ventricle for metastases to occur, presumably due to the high
concentration of choroid plexus in the region. Despite the deep location, surgical resection can be achieved safely in most cases. The survival of surgically treated patients is
comparable to that of patients with intraparenchymal metastases.
Intraventricular Congenital Lesions and Colloid Cysts

Aurelia Peraud, Anna Illner, and James T. Rutka
607
Intraventricular congenital lesions and colloid cysts comprise a rather large spectrum of
different pathologic conditions. In most cases, treatment in not warranted unless there is
progressive ventricular obstruction with hydrocephalus or growth of the lesion itself,
making tissue biopsy and histopathologic diagnosis necessary. Accordingly, a precise
neuroradiologic evaluation is of the utmost importance, because most lesions, if not
symptomatic, only require clinical and radiologic follow-up.
Choroid Plexus Tumors in Children

Nalin Gupta
621
Choroid plexus tumors represent a well-defined subset of brain tumors that occur
mainly in young children. Surgical resection for papilloma is usually curative, although
careful surgical planning is required to minimize the potential risks. Although adjunctive
therapy for carcinoma includes chemotherapy or radiation, the long-term survival for
carcinoma remains poor.
Cumulative Index 2003
CONTENTS
633
vii
FORTHCOMING ISSUES
January 2004
Endoscopy
Rick Abbott, MD, Guest Editor
April 2004
Traumatic Neurovascular Surgery
J. Paul Elliot, MD, Guest Editor
July 2004
Pain Treatment
Gary Heit, MD, Guest Editor
RECENT ISSUES
July 2003
Neuroaugmentation for
Chronic Pain
Jaimie M. Henderson, MD, Guest Editor
April 2003
Surgery for Psychiatric Disorders
Ali R. Rezai, MD, Steven A. Rasmussen, MD,
Benjamin D. Greenberg, MD, PhD
Guest Editors
January 2003
Pituitary Surgery
Martin H. Weiss, MD, and
William T. Couldwell, MD, PhD, Guest Editors
THE CLINICS ARE NOW AVAILABLE ONLINE!

Access your subscription at
http://www.TheClinics.com
Neurosurg Clin N Am 14 (2003) ix
Preface
Intraventricular tumors
Andrew T. Parsa, MD, PhD

Mitchel S. Berger, MD
Guest Editors
Neurosurgical oncology is an evolving discipline that continually benets from the translation
of scientic advances into clinical treatment paradigms. For the past 30 years at the University
of CaliforniaSan Francisco, basic scientists and
clinicians have been working together to rapidly
implement new discoveries for the benet of our
patients. Several examples of these collaborations
can be found here and at other premier neurosurgery departments around the world. A rened
understanding of the molecular pathways that
contribute to tumor development has yielded
new targets for chemotherapy, while our increasing experience with radiosurgery has broadened
treatment options for patients. The advent of surgical adjuncts such as functional mapping techniques, computerized frameless stereotaxy, and
endoscopy has signicantly decreased surgical
morbidity. In addition, surgical procedures are
now in place to facilitate local delivery of chemotherapeutic agents with unprecedented tumor
specicity.
Intraventricular tumors epitomize the challenges faced by neurosurgical oncologists in the
twenty-rst century. These lesions have a complex

biology and require signicant skill to excise
without attendant morbidity. A comprehensive
understanding of ventricular anatomy, surgical
approaches, and nonsurgical treatment options
is requisite for the neurosurgeon. Successful treatment of patients with these lesions requires a
dedicated team of pathologists, oncologists, and
neurosurgeons. In this issue of the Neurosurgery
Clinics of North America, we draw upon the
experience of several colleagues to facilitate a
better understanding of intraventricular tumors
in children and adults.
Andrew T. Parsa, MD, PhD
Mitchel S. Berger, MD
Department of Neurological Surgery
University of CaliforniaSan Francisco
505 Parnassus Avenue M-779
San Francisco, CA 94143, USA
E-mail address: parsaa@neurosurg.ucsf.edu
(A.T. Parsa)
1042-3680/03/$ - see front matter 2003 Elsevier Inc. All rights reserved.
doi:10.1016/S1042-3680(03)00059-7
Neurosurg Clin N Am 14 (2003) 469482
Epidemiology and pathology of intraventricular tumors

James S. Waldron, MDa, Tarik Tihan, MD, PhDb,*
a
Department of Neurological Surgery, University of California at San Francisco, 513 Parnassus, HSW 511,
San Francisco, CA 941430511, USA
b
Neuropathology Unit, Department of Pathology, University of California at San Francisco, 513 Parnassus,
HSW 408, San Francisco, CA 941430511, USA
Intraventricular tumors present a diagnostic

challenge to the clinician because of a broad
dierential diagnosis with signicant variability in
tumor type between adult and pediatric populations. This expansive dierential includes choroid
plexus papillomas (CPCs) and choroid plexus
carcinomas (CPCs), ependymomas, subependymomas, subependymal giant cell astrocytomas
(SEGAs), central neurocytomas, meningiomas,
and metastases as well as a number of cysts,
inammatory lesions, and other rare neoplasms.
Posterior fossa ependymomas, SEGAs, and choroid plexus tumors are more likely to appear in
childhood, whereas subependymomas, central
neurocytomas, intraventricular meningiomas,
and metastases are more frequent in adults. Each
of these tumor types involves the ependymal
lining and subependymal plate of the ventricular
wall, the septum pellucidum, or the highly
vascular choroid plexus. This article reviews the
epidemiology, the pathologic characteristics, and
the primary diagnostic considerations of each
tumor type.
Choroid plexus papilloma and carcinoma
Epidemiology
Choroid plexus tumors are epithelial neoplasms with a prevalence of 0.3 cases per million
[1]. In two large series, choroid plexus tumors
accounted for 0.4% [2] and 0.6% [3] of all
reported intracranial tumors. The tumor pre-
* Corresponding author.
E-mail address: tihan@itsa.ucsf.edu (T. Tihan).
dominantly occurs in childhood, although it can

be seen at any age. The median age of onset is 3.5
years [4], with 20% of patients presenting in the
rst year of life and almost 50% in the rst decade
[5]. The most common locations for choroid
plexus tumors are the lateral and fourth ventricles,
followed by the third ventricle. Cerebellopontine
angle examples are less common and are caused
by extension of tumor through the foramen of
Luschka [6]. In addition, rare suprasellar cases
have been reported [7]. Tumor location is closely
correlated with patient age. The most common
location in children is the lateral ventricle,
whereas the fourth ventricle is the most frequent
site in adults. Choroid plexus tumors are divided
into the CPP (World Health Organization [WHO]
grade I) and the more aggressive CPC (WHO
grade III). CPCs make up a small proportion of
choroid plexus tumors, primarily present in
children less than 3 years of age, and are most
commonly found in the lateral ventricles [8]. CPPs
and CPCs have been shown to spread through the
cerebrospinal uid, and rare metastatic cases have
been documented outside the CNS [4].
Macroscopic and microscopic features
Choroid plexus tumors are often soft to
rubbery and may have a gritty texture because
of calcication. The tumors are frequently shades
of orange-brown. During surgery, an anchoring
pedicle can be seen attached to the normal
choroid plexus or the ventricular wall. Some
papillomas have a cauliower-like appearance.
CPPs and CPCs exhibit features akin to
many papillary neoplasms in other organs. CPPs
have well-developed brovascular cords within
doi:10.1016/S1042-3680(03)00060-3
470
J.S. Waldron, T. Tihan / Neurosurg Clin N Am 14 (2003) 469482
papillary structures and do not exhibit architectural or cytologic atypia (Fig. 1A). Epithelial and
stromal cells contain many characteristics of the
normal choroid plexus, such as calcications and
xanthomatous change [9,10]. In some cases, there
is a striking nuclear monomorphism without
other aggressive features, such as mitoses or
vascular proliferation. Rarely, geographic necrosis without the pseudopalisading that is suggestive
of an infarct can be seen in an otherwise typical
CPP. Osseous or cartilaginous metaplasia and
acinar or tubular dierentiation are reported in
choroid plexus neoplasms [1113]. In addition,
a number of studies report a pigmented variant
that contains neuromelanin and lipofuscin [14].
CPPs with marked oncocytic transformation as
well as glial dierentiation are rare [12]. Transitional zones between the normal and neoplastic
choroid plexus can be found in CPPs and CPCs.
Carcinomas of the choroid plexus are tumors
that exhibit all the histologic hallmarks of
aggressiveness (see Fig. 1B). A typical CPC is
a neoplasm with increased architectural complexity demonstrating partially solid and partially
nonpapillary growth. Most tumors have marked
cytologic atypia, atypical mitotic gures, and
frank necrosis. Some high-grade tumors have
cytologic and architectural features that resemble
anaplastic oligodendrogliomas. Invasion into
neuropil is characteristic of CPCs, although some
CPPs can occasionally exhibit invasion into
surrounding parenchyma. Rare CPCs resemble
undierentiated carcinomas without any distinguishing features [15].
Immunohistochemical features
Cytokeratins and vimentin are expressed by
virtually all CPPs and most CPCs. Glial brillary
acidic protein (GFAP) can be found focally in
about 25% to 55% of CPPs and in 20% of CPCs
[16]. Most of the GFAP-positive cells are simul-
taneously positive for cytokeratin [17]. S-100

protein is present in almost all cases of CPP and,
less frequently, in CPCs. The staining for S-100 is
often stronger and more diuse than GFAP
staining. Synaptophysin has been suggested as
a possible marker for choroid plexus epithelium,
but staining of tumors with this marker produces
variable results. Epithelial membrane antigen
(EMA) is positive in tumor cells only focally, if
at all. A recent study suggested that immunohistochemical staining for prealbumin and carcinoembryonic acid (CEA) is of signicant value for
the dierentiation of CPPs and CPCs [18].
Staining for insulin-like growth factor-II (IGFII) is also a potentially useful marker to distinguish normal choroid plexus and CPP from CPC
[19]. Additionally, indirect indices of proliferation,
such as the Ki-67/MIB-1 antibody, have been used
to distinguish CPP from CPC [20,21]. The mean
Ki-67/MIB-1 labeling index is often less than 2%
in CPPs and greater than 10% in CPCs. Immunohistochemical staining for p53 protein is found
more often in carcinomas than in CPPs [22].
Ultrastructural features
Most CPPs exhibit apical microvilli with
scattered cilia, junctional complexes, interdigitating lateral cell borders, basement membrane, and
fenestrated capillaries. Cilia contain the 9 + 0
microtubule conguration characteristic of neuroepithelial cells. Some tumors have irregularly
shaped structures containing lipid droplets, lamentous material, and structures that resemble
the silver bodies of Biondi seen in normal
choroid plexus [23]. CPCs are often more varied in
their ultrastructural appearance and can show
epithelial features as well as cilia and microvilli,
although such ndings are focal in many cases.
CPCs can also display immature cellular features,
such as polyribosomes, glycogen granules, and
hypertrophied rough endoplasmic reticulum [24].
c
Fig. 1. (a) Choroid plexus papilloma: low magnication showing well-formed papillae composed of uniform small
epithelial type cells. Mitotic gures and necrosis are rare. (b) Choroid plexus carcinoma: a tumor with irregular
architecture, the presence of marked pleomorphic cells with a less prominent papillary pattern, and frequent mitoses and
necrosis. (c) Ependymoma: medium magnication showing uniform cells arranged in a perivascular fashion. (d )
Ependymoma: high magnication of an ependymal pseudorosette, an angiocentric arrangement of cells with brillary
processes perpendicular to the luminal axis. (e) Subependymoma: a paucicellular tumor showing a multinodular compact
architecture without mitotic gures. ( f ) Subependymal giant cell astrocytoma: a tumor composed of gemistocytic
astrocytes, scattered inammatory cells, and dystrophic calcications. (g) Central neurocytoma: a tumor typically
described as oligodendroglioma-like with clear cells (fried-egg cells) and a delicate vasculature (chicken-wire
vasculature). (h) Meningioma: a typical meningioma in the lateral ventricle. The tumor shows multiple whorl formation as
well as calcications known as psammoma bodies.
471
472
Molecular and genetic features

In CPP, recurrent abnormalities, including
partial gains of chromosome 7, have been
reported [25]. A comparative genomic hybridization study of a large number of choroid plexus
tumors showed that +5q, +6q, +7q, +9q,
+15q, +18q, and 21q were signicantly more
common in CPPs, whereas CPCs were characterized by +1, +4q, +10, +14q, +20q, +21q,
5q, 9p, 11, 15q, and 18q [26].
Choroid plexus neoplasms have been associated with the Li-Fraumeni syndrome as well as
the Aicardi syndrome [2729]. Several reports
have identied SV-40 virus genetic material within
tumor cells; however, the contribution of this
virus to the formation of choroid plexus tumors is
unknown [3032].
Pathologic dierential diagnosis
The most frequent challenge during pathologic
evaluation is the distinction of CPP from normal
choroid plexus. The normal choroid plexus has
regular single-layered cells with hobnail luminal
surfaces, whereas CPP displays a more crowded
epithelium with signicant nuclear variability. The
diagnosis of CPP by the pathologist is unreasonable in the absence of clinical and radiologic
ndings, especially without a distinct contrastenhancing intraventricular mass.
The second diagnostic challenge is the exclusion of a rare papillary ependymoma [33].
Papillary ependymomas can form epithelial surfaces but retain a brillary background. Ependymomas with focal papillary change can be
distinguished by their predominantly glial appearance. In cases where the distinction cannot be
made in routine stains, immunohistochemistry
and ultrastructural studies are helpful. Large
partially intraventricular tumors in young patients
with poorly dierentiated morphology should also
raise the possibility of an atypical teratoid/
rhabdoid tumor (AT-RT). Some AT-RTs have
been misdiagnosed as CPC in the past. Distinction
of AT-RT can be made by using a panel of
immunohistochemical markers as well as genetic
studies to conrm the presence of characteristic
abnormalities.
CPCs are extremely rare in adults, and
metastatic carcinoma should be viewed as a more
likely cause for an intraventricular papillary
carcinoma. Metastases from the pulmonary and
genitourinary systems have been shown to mimic
CPC [3436]. Distinction may be dicult because
of overlapping histologic, ultrastructural, and

immunohistochemical features. Typical CPC immunohistochemistry reveals positivity for cytokeratin cocktail and absent or only faint EMA
and CEA immunoreactivity. If positive, synaptophysin can also be used to distinguish CPC. In
addition, BerEp4 staining is considered a reliable
marker for most metastatic carcinomas, and its
presence may exclude a CPC (Marc K. Rosenblum, MD, personal communication, 1999).
Ependymomas
Epidemiology
Ependymomas are neoplasms derived from the
ependymal layer lining the ventricular system and
can occur intracranially and in the spine. Intracranial ependymomas account for 2% to 8% of
all primary CNS neoplasms [37], with more than
half presenting in the rst two decades of life. In
a series of 467 pediatric intracranial neoplasms
reviewed by Farwell et al [38], ependymomas
made up 9% of all intracranial tumors, making it
the third most common pediatric intracranial
tumor. Within the pediatric population, ependymomas favor young patients, with more than 50%
occurring within the rst 3 years of life. No
consistent gender predilection has been identied.
Intracranial ependymomas can be divided by
location into those appearing infratentorially
and those appearing supratentorially. Infratentorial ependymomas make up approximately two
thirds of all cases [39], comprise most pediatric
cases, and most frequently occur in the fourth
ventricle [40]. Supratentorial ependymomas occur
more frequently in older children and adults. In
addition to the lateral ventricles, approximately
50% of supratentorial ependymomas involve the
parenchyma [41].
Ependymomas are often sharply demarcated,
eshy, hemorrhagic, soft, and sometimes rubbery
masses. Rare examples are heavily calcied, giving
the tumor a gritty texture. Intraventricular examples of ependymomas are often lobulated and
display a discrete interface with surrounding
brain. Some tumors may exhibit a delicate overlying ependymal layer that gives them a shiny
texture.
Ependymomas are glial neoplasms composed
of a monomorphous proliferation of neoplastic
cells with typical perivascular pseudorosettes
(see Fig. 1C, D). Some ependymomas are predominantly glial in appearance and may not have
distinct perivascular pseudorosettes, whereas
others may be predominantly epithelial. The latter
may present as a tumor with oval to round nuclei,
discrete cytoplasmic borders, frank papillary
structures, and well-formed brovascular cores.
Other tumors may show true ependymal rosettes distinguished by their well-dened lumina
and cells forming pseudoglandular structures.
Most ependymomas show a substantial number
of nuclear grooves that can be identied in
intraoperative smears and help with the rapid
interpretation of frozen sections [42]. This feature,
however, needs to be interpreted in the context of
other histologic ndings, because many other
tumors, such as meningiomas and other gliomas,
can exhibit nuclear grooves. The tumor nuclei are
uniform, round to oval, and often feature
a distinct nucleolus.
Clear cell change in ependymoma is a rare but
signicant nding [43]. Intraventricular ependymomas may exhibit focal or predominant clear
cell change. When clear cell change is predominant, the hematoxylin-eosin appearance of an
oligodendroglioma is recapitulated. It is likely
that many tumors previously reported as intraventricular oligodendroglioma are examples
of clear cell ependymoma [21]. Clear cell ependymomas are usually higher grade and exhibit
increased mitotic activity and vascular proliferation. The so-called tanycytic ependymoma is
remarkably similar to a pilocytic astrocytoma.
This highly brillary tumor has moderate cell
density, spindled cells, and a fascicular architecture. It has also been described as a piloid tumor
with ependymal nuclei [44]. The tanycytic
ependymoma often lacks nuclear pleomorphism
or aggressive features, such as mitoses or vascular
proliferation. Perivascular pseudorosettes are rudimentary and sometimes absent.
Ependymomas are commonly calcied and
rarely exhibit cartilaginous and osseous metaplasia. Rare ependymomas contain cytoplasmic
eosinophilic granules, clear vacuoles, lipid, or
melanin [45,46].
The current WHO classication denes grade
II ependymomas as tumors with mild cellular
pleomorphism, pseudorosettes, or true ependymal
rosettes. The tumors can have occasional mitotic
gures and necrosis without pseudopalisading.
Occasional foci of hypercellularity and increased
mitoses are allowed. Anaplastic, high-grade,
or grade III ependymomas have moderate to high
473
cellularity, increased mitotic gures, and vascular

proliferation. Necrosis is often present, either in
the form of geographic necrosis or, rarely, in the
pseudopalisading form. Perivascular pseudorosettes or occasional true ependymal rosettes can
be found in most high-grade ependymomas. There
is controversy around whether focal atypia or
anaplasia should elevate a lesion to grade III
anaplastic ependymoma. Some require atypia
and anaplasia to predominate in the tumor tissue,
whereas others report a less favorable prognosis
even for tumors with focal anaplastic features.
Ependymomas are variably positive for
GFAP, which highlights the brillary processes
around vessels. Tumors are diusely positive for
vimentin and stain less avidly with S-100 protein
and neurospecic enolase (NSE). Positive staining
for epithelial markers, such as EMA and cytokeratins, has been reported in most posterior fossa
and spinal cord ependymomas [47]. Rare tumor
cells, true rosettes, and occasional papillary
structures are EMA-positive.
Studies suggest that high Ki-67/MIB-1 and p53
protein positivity might be reliable indicators of
high-grade ependymomas [48]. Even though there
seems to be a positive correlation between highgrade features and the Ki-67/MIB-1 index [49],
none of the immunohistochemical variables signicantly correlate with tumor grade. Conversely,
Ki-67/MIB-1 and p53 were reported to correlate
with patient survival [50]. Currently, there is no
clear evidence for the utility of these markers in
determination of tumor grade or behavior.
The acellular zones around pseudorosettes are
composed of large numbers of closely packed,
lament-rich, cytoplasmic processes. Microlumina
are often present, even though they may not be
observed by light microscopy [51]. These microlumina contain slender curving microvilli and
a variable number of cilia. Bordering cells are
connected by unusually long tight junctions. This
triad (cilia, intracytoplasmic intermediate laments, and cell junctions) makes up the typical
ultrastructural components. The epithelioid cells
found in ependymomas and true rosettes are
characterized by intracellular lumina, cilia, and
microvilli. Clear cell ependymomas reveal densely
packed polyhedral cells with clear cytoplasm and
well-developed intercellular junctions. Abundant
474
hyaloplasmic lipid vacuoles can also contribute to

the clear appearance of the tumor cells [46].
There is a body of evidence suggesting the
presence of a tumor suppressor gene on the long
arm of chromosome 22 that plays a role in the
pathogenesis of ependymomas [52]. In one study,
the most frequent copy number abnormality in
ependymomas was 22q loss, followed by gain of
chromosome 9 and occasional loss of 6q, 3p, 10q,
and 15q [25]. A heterozygous mutation in the
MEN1 gene has also been reported in ependymomas [53].
issue. Another diagnostic consideration is the

central neurocytoma. The central neurocytoma is
a highly cellular neoplasm that may show
perivascular pseudorosettes. The cells appear
more neurocytic, and the brillar areas resemble
neuropil. The tumor strongly reacts with synaptophysin and only weakly (if at all) with GFAP.
Electron microscopy can distinguish the two
entities. Papillary ependymomas may resemble
CPP. The overall immunohistochemical prole
and ultrastructural features can be used to
separate the two entities.
Subependymoma
Epidemiology
Formulation of the dierential diagnosis for

ependymoma is dependent on the location of the
lesion. In the posterior fossa, medulloblastoma
needs to be considered rst in the dierential
diagnosis, although its architecture is more
reminiscent of a small blue round cell tumor than
that of a glioma. Pilocytic astrocytoma of the
cerebellum or brain stem is a second possibility
but can be easily excluded when classic features of
pilocytic astrocytomas, such as Rosenthal bers,
eosinophilic granular bodies, and a fairly paucicellular appearance, are present. Inltrating
astrocytomas or the so-called brain stem gliomas may have an exophytic quality and may
resemble ependymoma. They are easily distinguished by their invasive quality, lack of epithelial
features or pseudorosettes, and marked nuclear
pleomorphism.
Supratentorial intraventricular ependymomas
need to be distinguished from subependymomas.
Such distinction is often subjective and may not
always translate into a signicant change in
clinical outcome. Nevertheless, based on the
overall clinical behavior of ependymomas and
the likelihood of supratentorial examples being
higher grade, one is compelled to make the
distinction. The distinction is usually not dicult,
and the dierential diagnosis is confounded by
limited tissue sample size. A second yet more
important dierential diagnosis is oligodendroglioma, which can easily be confused with clear
cell ependymoma. Clear cell ependymomas are
noninltrating, solid, and distinct from the
surrounding brain. Purely intraventricular neoplasms are not likely to be oligodendrogliomas,
but when a question exists, immunohistochemistry and electron microscopy readily settle the
Subependymomas are slow-growing, benign

intraventricular lesions rst identied as a separate
entity in 1945 by Scheinker [54]. They originate in
the subependymal glial matrix and typically project into the ventricular lumen. Intracranial
subependymomas frequently remain asymptomatic and are documented on autopsy or as an
incidental nding on imaging. A prevalence of
0.4% has been reported in a series of 1000
necropsies of asymptomatic patients reviewed by
Matsumura and colleagues [55]. Subependymomas have been reported over a wide age range, but
generally occur in middle-aged to older adults.
The fourth ventricle, followed by the lateral
ventricles, is the most common site of presentation. Less common locations include the third
ventricle, the septum pellucidum, and the cerebral
aqueduct.
Subependymomas are solid nodular tumors
rmly attached to the ventricular surface. Tumors
are typically soft but can be rubbery and, rarely,
cystic and occasionally have a gritty texture.
Subependymomas are typically paucicellular,
brillar, and markedly nodular neoplasms (see
Fig. 1E). Tumor nuclei cluster within the nodular
regions. Supratentorial tumors, especially those
near the foramen of Monro, are predominantly
microcystic and focally myxoid. The cells are
often spindled with oval nuclei and brillary
processes. The tumor exhibits an extensive brillary background on intraoperative smear preparations. Nuclear pleomorphism is rare, and
mitoses are typically absent. Tumor vessels show
focal hyalinization with occasional hemosiderin
deposition. There is some evidence that the
histologic features in larger and symptomatic

subependymomas may be dierent. Larger symptomatic tumors more frequently demonstrate cyst
formation, microcalcication, and vessel degeneration accompanied by hemorrhage [56]. Subependymomas in the posterior fossa are usually smaller
without signicant microcystic change. The tumors are prominently nodular and show nuclear
clustering and calcications. Mitoses are rare, and
vascular proliferation and necrosis are absent.
Subependymomas are strongly GFAP-positive
in accord with their high content of intermediate
glial laments. Vimentin staining is often strong,
and S-100 protein stains the cytoplasm and the
nuclei. Compared with other ependymal tumors,
subependymomas have the lowest rate of cell
proliferation, as evidenced by a Ki-67/MIB-1
index of less than 1% [57]. In contrast to
ependymomas, staining with epithelial markers,
such as EMA, is usually not observed.
Subependymomas display an abundance of
closely packed cell processes lled with intermediate laments. This meshwork of processes
widely separates small clusters of tumor cells.
Larger cells lacking specialized features and
resembling ependymal precursor cells are often
found. Other cells with transitional forms between
these two types can be identied [58]. Pockets of
microvilli are present, but they dier from
ependymal type rosettes because of the lack of
tight junctions.
There is limited information on the cytogenetics and molecular genetics of subependymomas. A
few case reports have demonstrated a normal
karyotype in subependymomas investigated with
conventional cytogenetic techniques [59].
The main component of the dierential diagnosis for intraventricular subependymoma is
the classic ependymoma. It may not be possible to
distinguish all cases, especially if the amount of
tissue available for pathologic analysis is limited.
Both neoplasms appear remarkably similar, and
foci identical to subependymoma are commonly
475
seen in ventricular ependymomas. In general,

ependymomas can be distinguished by occurrence
primarily in children, hypercellularity, perivascular pseudorosettes, and true ependymal rosettes.
Most supratentorial ependymomas have a solid
cystic appearance and are symptomatic. Subependymomas may also be confused with the tanycytic
variant of ependymomas. Often, tanycytic ependymomas are more cellular and resemble pilocytic
astrocytomas. Ultrastructural examination of
tanycytic ependymoma reveals characteristic
ependymal features, including intracytoplasmic
intermediate laments, prominent intercellular
junctions, numerous slender surface microvilli,
and microvilli-lined lumina.
Subependymal giant cell astrocytoma
Epidemiology
SEGAs are intimately associated with the
tuberous sclerosis complex, an autosomal dominant dysgenetic syndrome that is associated with
the classic triad of seizures, mental retardation,
and papular facial lesions. In the CNS, the
complex is characterized by cortical tubers,
subependymal nodules, and SEGAs. The incidence of tuberous sclerosis is approximately
1:10,000 [60] in the general population. Approximately 6% [61] of these patients develop SEGAs.
Almost all SEGAs arise near the foramen of
Monro and typically present with hydrocephalus
or increased seizure frequency, most commonly
within the rst two decades of life [61].
SEGAs are well-dened typically pedunculated
intraventricular masses that can be soft to rubbery
and often have a broad base on the ventricular
surface. The tumor can be easily removed from its
base. Tumors may be friable, pink as a result of
vascularization, and occasionally gritty from
calcication.
SEGAs are characteristically solid and have
a typical swirling architecture. They exhibit
compact growth with spindled and gemistocytic
cells and are sharply demarcated from the
adjacent normal parenchyma (see Fig. 1F).
Spindled cells are responsible for the swirling
appearance of the tumor on low magnication.
The gemistocyte-like cells have round vesicular
nuclei with distinct nucleoli and an eosinophilic
cytoplasm. In addition, they display thick hairlike
processes and have a tendency to form cohesive
476
clusters and occasional pseudorosettes [62]. The

tumor maintains a compact and uniform appearance despite varying tumor cell types. SEGAs
often contain inammatory cells, including occasional mast cells. Even though rare mitotic gures
are occasionally seen, brisk mitotic activity, necrosis, or vascular proliferation is typically absent.
Some tumors undergo focal infarction, which can
appear ominous and be confused with the necrosis
characteristic of a high-grade astrocytic neoplasm.
Calcication is sometimes present.
The gemistocytic and spindle cells are often
strongly positive for GFAP; however, the absolute number of positive cells in each tumor is
highly variable. SEGAs also show strong positivity for S-100 protein. Neurolament epitopes,
class III b-tubulin, and calbindin 28-kDa are
expressed in some cases [63]. Cytoplasmic staining
for somatostatin, met-enkephalin, 5-hydroxytryptamine, b-endorphin, and neuropeptide Y has also
been noted in more than half of cases of SEGA
[63,64]. The divergent glial and neuronal staining
has been shown to colocalize within the same cell.
SEGAs are negative for HMB-45 antibody, and
the Ki-67/MIB-1 labeling index is usually less
than 2% [65].
SEGAs contain numerous intermediate laments, frequent lysosomes, and occasional rectangular or rhomboid membrane-bound crystalloids
that exhibit lamellar periodicity and structural
transition to lysosomes. Microtubules and stacks
of rough endoplasmic reticulum are common, but
true neuronal dierentiation, such as neurosecretory granules or synaptic formations, is often
absent [63]. Rare tumor cells have features
suggestive of neuronal dierentiation, including
stacks of rough endoplasmic reticulum, occasional
microtubules, and a few poorly dened dense core
granules. Gemistocytic cells are characterized by
abundant intermediate laments within the cell
body and the processes. Lysosomes are common
and, rarely, may contain distinctive membranebound crystalloids.
Cytogenetic analysis of SEGAs within the
tuberous sclerosis complex (TSC) reveals clonal
chromosomal changes, resulting in the partial loss
of chromosome 22q in some tumors [66]. TSCassociated tumors also demonstrate loss of
heterozygosity in chromosomes 9 and 16, which
are known to harbor TSC genes [67]. One of two
suspected genes, TSC2, was found in chromosome
16 by positional cloning. The gene product from
TSC2 has been named tuberin. TSC1 was
discovered earlier in chromosome 9 but has not
yet been characterized. Genetic analysis on TSC
families reveals mutations in chromosome 9q34
(TSC1) and chromosome 16p13 (TSC2) as the
only common genetic anomalies [68].
The Eker rat, a naturally occurring animal
model of TSC, provides a powerful tool for
investigations of TSC. In this model, a conserved
linkage group on rat 10q corresponds to human
16p13.3 (TSC2 gene) [69]. Currently, it is believed
that the products of TSC1 and TSC2 genes
interact with each other in the cell.
SEGAs are fairly distinct intraventricular neoplasms that may be confused with gemistocytic
astrocytoma or high-grade glioma if the typical
pathologic and radiologic features are overlooked.
Small biopsies can also potentially be interpreted
as tanycytic ependymoma or subependymoma,
but this is less likely, because SEGAs are invariably more cellular, less brillary, and far more
gemistocytic.
Central neurocytoma
Epidemiology
The term central neurocytoma was rst used by
Hassoun et al [70] in 1982 to describe dierentiated intraventricular neuronal lesions observed in
2 cases. Central neurocytomas are rare neoplasms,
with 127 reported cases through 1993 [71].
Reported rates in series of pathologically conrmed primary CNS neoplasms range from 0.1%
to 0.5% [7274]. Central neurocytomas are
primarily tumors of young adults, with 45%
occurring in the third decade of life and almost
75% between the ages of 20 and 40 years [71].
Gender distribution is equal. Central neurocytomas arise predominantly from the septum
pellucidum or, less frequently, from the lateral ventricular wall. The anterior lateral ventricle is the
most frequent site (77%), followed by lateral and
third ventricle involvement (21%) [71]. Bilateral
lateral ventricular involvement is uncommon.
Rare cases have been reported in the third and

fourth ventricles.
The tumor forms a soft to gritty, tan, discrete
mass that may be solid or partly cystic.
Central neurocytomas are histologically and
cytologically uniform neoplasms. The cells are
strikingly monomorphous with nely distributed
chromatin and a ne brillary matrix. The tumor
is one of several neuroepithelial neoplasms with
salt and pepper chromatin. Central neurocytoma joins the list of oligodendroglioma-like
tumors because of a striking preponderance of
cells with perinuclear halos that resemble classic
oligodendroglioma (see Fig. 1G). In some cases,
there are perivascular brillary zones reminiscent
of ependymal pseudorosettes. In addition, some
examples resemble nodular medulloblastomas by
exhibiting neurocytic dierentiation with cell
streaming and nodular growth.
Intraoperative frozen sections can sometimes
obscure the histologic and cytologic uniformity
typical of central neurocytomas. The processing
of frozen tissue also adds a degree of nuclear
pleomorphism that can raise the possibility of
a small blue round cell tumor. This is further
confounded in permanent sections of frozen tissue
because of the obscured neuronal/neurocytic
background. Most central neurocytomas are
grade II lesions with minimal nuclear pleomorphism and rare mitotic gures. Tumors with
atypical features and transitional characteristics
between neurocytoma and neuroblastoma have
been reported [75].
Central neurocytomas may show ganglionic
cell dierentiation and have a preponderance of
neuropil with variable numbers of ganglion-like
cells. Such cases have been designated as ganglioneurocytoma or dierentiated neurocytoma. An intraventricular lesion that combines
the features of a neurocytoma with ganglion cells
and a malignant small cell component has been
reported but is extremely rare. It has also been
suggested that some central neurocytomas can
express photoreceptor dierentiation, potentially
relating them to pineocytomas [76]. Rare central
neurocytomas exhibit lipofuscin or neuromelanin
pigment [77].
Central neurocytomas consistently exhibit
immunoreactivity for NSE and synaptophysin,
477
indicating neuronal dierentiation [78]. Synaptophysin antibody stains the brillar zones and, to
a lesser extent, the perinuclear cytoplasm of tumor
cells. Anti-Hu autoantibodies stain neurocyte
nuclei. Tumor cells are also positive for Leu-7 and
S-100 protein, whereas staining for GFAP is predominantly negative and vimentin is conned to the
nonneoplastic mesenchymal elements of blood
vessels [79]. Staining for myelin basic protein,
chromogranin, and neurolament is often negative.
Some studies have shown a small subpopulation of
GFAP-positive neoplastic cells, and glial dierentiation has been suggested in tissue culture. This
mixed phenotype of glial and neuronal marker
positivity in central neurocytoma can be interpreted as a glioneuronal neoplasm, with an overwhelmingly neurocytic component. In rare examples,
a tumor may have an increased Ki-67/MIB-1 index.
Such neoplasms are described as atypical neurocytomas and have a signicantly elevated incidence of local recurrence [80]. Even though no
clear cuto point exists between classic and atypical
neurocytomas, most authors suggest that tumors
with an MIB-1 index of greater than 2% be placed
in the atypical category. Nevertheless, some studies
show no dierence in MIB-1 labeling between
tumors with atypical features and typical central
neurocytomas [81]. Currently MIB-1 labeling is not
used to modify grading of central neurocytomas.
Central neurocytoma is readily recognizable as
neuronal, with microtubules, terminations, clear
vesicles, and dense core granules [79,82,83]. Some
examples may display round cells with abundant
cell processes containing microtubules, cellular
junctions, and lysosome-like structures. Others
contain numerous synaptic vesicles, neuritic processes, and neurosecretory granules. In addition,
rare tumors contain ganglionic cells with welldeveloped processes.
Reported recurrent genetic changes in central
neurocytomas include alterations on chromosomes 2p, 10q, and 18q. The candidate genes in
these loci are currently unknown [84]. Other
studies have suggested gain of chromosome 7 as
a nonrandom genetic alteration in central neurocytomas [85]. Recent studies have demonstrated
that central neurocytomas are genetically distinct
from oligodendrogliomas and that chromosomes
1p and 19q probably do not play an important
478
role in their pathogenesis. In addition, N-myc and

epidermal growth factor receptor amplications
are rare or absent in these tumors [86].
The critical dierentiation for central neurocytoma is from oligodendroglioma. Some cases of
central neurocytoma perfectly recapitulate oligodendroglioma in routine microscopic examination. In such cases, it is important to use a panel
of immunohistochemical stains and to perform an
ultrastructural examination to establish the correct diagnosis. Furthermore, radiologic information should be critically interpreted and the
diagnosis of oligodendroglioma challenged in
purely intraventricular tumors. A second entity
in the dierential diagnosis is the clear cell
ependymoma, which also exhibits a remarkable
resemblance to oligodendroglioma. The presence
of ependymal features as well as immunohistochemical analysis should distinguish a clear cell
ependymoma from central neurocytoma. Small
biopsies from a dysembryoplastic neuroepithelial
tumor may also mimic central neurocytoma, but
exclusive intraventricular location, the absence of
oating neurons, and the immunohistochemical
prole should distinguish between the two. Lastly,
the presence of an intraventricular clear cell
neoplasm in older patients should raise the
possibility of a metastatic lesion, especially a renal
cell carcinoma. Often, the highly anaplastic
histologic features are sucient to distinguish
a renal carcinoma metastasis from a classic central
neurocytoma. Additional immunohistochemical
studies can be used to provide further support.
tricular clear cell meningioma [93] and malignant

meningioma [94] have been reported.
Intraventricular metastases from epithelial
malignancies are extremely rare but can mimic
a choroid plexus tumor clinically and pathologically. Intraventricular metastases originate from
a number of cancers, including renal cell carcinoma [35,95,96], pulmonary adenocarcinoma [34],
gastric carcinoma [97], adrenocortical carcinoma
[98], and bladder carcinoma [36]. In such cases,
immunohistochemical analysis, including a cytokeratin panel, can help to identify the nature of
the neoplasm and dierentiate such tumors from
CPCs [99].
Rare cases of perineurioma from the choroid
plexus of the third ventricle, malignant schwannoma, solitary brous tumors, and hemangiopericytoma have been reported as purely
intraventricular tumors [100102].
A diverse list of cystic tumor-like lesions can
exist within the ventricular system and can be
confused with a neoplasm [103]. Colloid cysts of
the third ventricle [104], ependymal or glioependymal cysts [105], choroid plexus cysts [106],
arachnoid cysts [107], and cavernous angioma
[108] have been reported as intraventricular
masses. Choroid plexus cysts are more common
in fetuses with chromosomal aneuploidies, particularly trisomy 18.
Inammatory or infectious processes can also
present as purely intraventricular masses that
resemble tumors. Such a presentation is much
less common than the usual parenchymal or
leptomeningeal forms. Reports of infectious or
inammatory processes that present as masses
within the ventricular system include cysticercosis
[109], cryptococci [110], and nocardiosis [111]
among others.
Other tumors and tumor-like lesions within the

ventricular system
Other purely intraventricular tumors and tumor-like lesions are rare. One example includes the
intraventricular meningioma [8789]. The intraventricular location is uncommon, with an approximate incidence of 0.5% to 4.5% among all
intracranial meningiomas [89]. Intraventricular
meningiomas are more common in adults because
of the higher overall frequency of meningiomas but
make up a larger percentage of meningiomas in
the pediatric population [9092]. Meningiomas
can arise anywhere in the ventricular system and
exhibit the histologic features common to all
meningiomas (see Fig. 1H). Rare cases of intraven-
Summary
Tumors that primarily or exclusively involve
the ventricular system constitute a rare and
heterogeneous group. Certain histologic tumor
types predominantly occur in children, whereas
others are more common in adults. Tumor
location provides additional clues to correct
diagnosis. When used in conjunction with clinical
and radiologic data, histopathologic features can
distinguish among this wide range of possibilities
to provide the correct diagnosis for optimal
patient management.
Acknowledgement
J.S. Waldron was supported in part by a grant
from the Khatib Research Foundation as a Khatib
Fellow 20022003.
[18]
References
[19]
[1] Janisch W, Staneczek W. Primary tumors of the

choroid plexus. Frequency, localization and age
[in German]. Zentralbl Allg Pathol 1989;135(3):
23540.
[2] Schier D. Brain tumors. Pathology and biological
correlates. Berlin: Springer Verlag; 1993. p. 21721.
[3] Zulch K. Brain tumors. Their biology and pathology. 3rd edition. Berlin: Springer Verlag; 1986.
[4] Wol JE, et al. Choroid plexus tumours. Br J
Cancer 2002;87(10):108691.
[5] Matson DD, Crofton FDL. Papilloma of the
choroid plexus in childhood. J Neurosurg 1960;
17:100227.
[6] Burger PC, Scheithauer BW. Tumors of the central
nervous system. Washington, DC: Armed Forces
Institute of Pathology; 1994.
[7] Kimura M, et al. Primary choroid plexus papilloma located in the suprasellar region: case report.
Neurosurgery 1992;31(3):5636.
[8] Pierga JY, et al. Carcinoma of the choroid plexus:
a pediatric experience. Med Pediatr Oncol 1993;
21(7):4807.
[9] Kepes JJ. Xanthomatous changes in a papilloma
of the choroid plexus. Acta Neuropathol (Berl)
1970;16(4):3679.
[10] Nakashima N, et al. Choroid plexus papilloma.
Light and electron microscopic study. Virchows
Arch A Pathol Anat Histopathol 1983;400(2):
20111.
[11] Cardozo J, et al. Choroid plexus papilloma
containing bone. Acta Neuropathol (Berl) 1985;
68(1):835.
[12] Bonnin JM, Colon LE, Morawetz RB. Focal glial
dierentiation and oncocytic transformation in
choroid plexus papilloma. Acta Neuropathol (Berl)
1987;72(3):27780.
[13] Stefanko SZ, Vuzevski VD. Oncocytic variant of
choroid plexus papilloma. Acta Neuropathol (Berl)
1985;66(2):1602.
[14] Reimund EL, Sitton JE, Harkin JC. Pigmented
choroid plexus papilloma. Arch Pathol Lab Med
1990;114(8):9025.
[15] Paulus W, Janisch W. Clinicopathologic correlations in epithelial choroid plexus neoplasms:
a study of 52 cases. Acta Neuropathol (Berl) 1990;
80(6):63541.
[16] Rickert CH, Paulus W. Tumors of the choroid
plexus. Microsc Res Tech 2001;52(1):10411.
[17] Kouno M, et al. An immunohistochemical study
of cytokeratin and glial brillary acidic protein in
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
479
choroid plexus papilloma. Acta Neuropathol

(Berl) 1988;75(3):31720.
Kato T, et al. Clinicopathological study of choroid
plexus tumors: immunohistochemical features and
evaluation of proliferative potential by PCNA and
Ki-67 immunostaining. Noshuyo Byori 1996;13(2):
99105.
Kubo S, et al. Immunocytochemical detection of
insulin-like growth factor II (IGF-II) in choroid
plexus papilloma: a possible marker for dierential
diagnosis. Clin Neuropathol 1999;18(2):749.
Coons S, et al. Choroid plexus carcinoma in
siblings: a study by light and electron microscopy
with Ki-67 immunocytochemistry. J Neuropathol
Exp Neurol 1989;48(4):48393.
Carlotti CG Jr, et al. Evaluation of proliferative
index and cell cycle protein expression in choroid
plexus tumors in children. Acta Neuropathol (Berl)
2002;103(1):110.
Jay V, et al. P53 expression in choroid plexus
neoplasms: an immunohistochemical study. Arch
Pathol Lab Med 1996;120(11):10615.
Navas JJ, Battifora H. Choroid plexus papilloma:
light and electron microscopic study of three cases.
Acta Neuropathol (Berl) 1978;44(3):2359.
Anguilar D, et al. The ne structure of choroid
plexus carcinoma. Histopathology 1983;7(6):
93946.
Grill J, et al. Comparative genomic hybridization
detects specic cytogenetic abnormalities in pediatric ependymomas and choroid plexus papillomas. Cancer Genet Cytogenet 2002;136(2):1215.
Rickert CH, Wiestler OD, Paulus W. Chromosomal imbalances in choroid plexus tumors. Am
J Pathol 2002;160(3):110513.
Yuasa H, Tokito S, Tokunaga M. Primary
carcinoma of the choroid plexus in Li-Fraumeni
syndrome: case report. Neurosurgery 1993;32(1):
1314.
Uchiyama CM, et al. Choroid plexus papilloma
and cysts in the Aicardi syndrome: case reports.
Pediatr Neurosurg 1997;27(2):1004.
Vital A, et al. Astrocytomas and choroid plexus
tumors in two families with identical p53 germline
mutations. J Neuropathol Exp Neurol 1998;
57(11):10619.
Palmiter RD, et al. SV40 enhancer and large-T
antigen are instrumental in development of choroid plexus tumours in transgenic mice. Nature
1985;316(6027):45760.
Van Dyke TA, et al. Relationship between simian
virus 40 large tumor antigen expression and tumor
formation in transgenic mice. J Virol 1987;
61(6):202932.
Chen J, et al. T-antigen mutant activities in vivo:
roles of p53 and pRB binding in tumorigenesis of
the choroid plexus. Oncogene 1992;7(6):116775.
Park SH, Park HR, Chi JG. Papillary ependymoma: its dierential diagnosis from choroid
480
[34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42]
[43]
[44]
[45]
[46]
[47]
[48]
[49]
[50]
[51]
[52]

plexus papilloma. J Korean Med Sci 1996;11(5):
41521.
Tanimoto M, et al. Choroid plexus metastasis of
lung carcinomacase report. Neurol Med Chir
(Tokyo) 1991;31(3):1525.
Mizuno M, et al. Renal cell carcinoma metastasizing to choroid plexus of lateral ventricle; a case
report [in Japanese]. No Shinkei Geka 1992;
20(4):46974.
Qasho R, et al. Choroid plexus metastasis from
carcinoma of the bladder: case report and review
of the literature. J Neurooncol 1999;45(3):23740.
Duncan JAI, Homan HJ. Intracranial ependymomas. In: Kaye AH, editor. Brain tumors.
Edinburgh: Churchill Livingstone; 1995. p. 493
504.
Farwell JR, Dohrmann GJ, Flannery JT. Central
nervous system tumors in children. Cancer 1977;
40(6):312332.
Mork SJ, Loken AC. Ependymoma: a follow-up
study of 101 cases. Cancer 1977;40(2):90715.
Schier D, et al. Histologic prognostic factors in
ependymoma. Childs Nerv Syst 1991;7(4):17782.
Schwartz TH, et al. Supratentorial ependymomas
in adult patients. Neurosurgery 1999;44(4):72131.
Kumar PV. Nuclear grooves in ependymoma.
Cytologic study of 21 cases. Acta Cytol 1997;
41(6):172631.
Kawano N, Yada K, Yagishita S. Clear cell
ependymoma. A histological variant with diagnostic implications. Virchows Arch A Pathol Anat
Histopathol 1989;415(5):46772.
Langford LA, Barre GM. Tanycytic ependymoma.
Ultrastruct Pathol 1997;21(2):13542.
Rosenblum MK, et al. Melanotic ependymoma
and subependymoma. Am J Surg Pathol 1990;
14(8):72936.
Severi B, et al. Ependymoma of the foramen of
Monro: ultrastructural characterization. Ultrastruct Pathol 1989;13(1):3542.
Takeuchi H, et al. Epithelial dierentiation and
proliferative potential in spinal ependymomas.
J Neurooncol 2002;58(1):139.
Rushing EJ, et al. Correlation of bcl-2, p53, and
MIB-1 expression with ependymoma grade and
subtype. Mod Pathol 1998;11(5):46470.
Ritter AM, et al. Ependymomas: MIB-1 proliferation index and survival. J Neurooncol 1998;
40(1):517.
Verstegen MJ, et al. Proliferation- and apoptosisrelated proteins in intracranial ependymomas: an
immunohistochemical analysis. J Neurooncol
2002;56(1):218.
Sara A, Bruner JM, Mackay B. Ultrastructure
of ependymoma. Ultrastruct Pathol 1994;18(1 2):
3342.
Park JP, et al. Constitutional de novo
t(1;22)(p22;q11.2) and ependymoma. Cancer
Genet Cytogenet 1996;86(2):1502.
[53] Urioste M, et al. Complex cytogenetic abnormalities including telomeric associations and MEN1
mutation in a pediatric ependymoma. Cancer
Genet Cytogenet 2002;138(2):10710.
[54] Scheinker I. Subependymoma: a newly recognized
tumor of subependymal derivation. J Neurosurg
1945;2:23240.
[55] Matsumura A, et al. Intracerebral subependymomas. Clinical and neuropathological analyses with
special reference to the possible existence of a less
benign variant. Acta Neurochir (Wien) 1989;96
(1 2):1525.
[56] Scheithauer BW. Symptomatic subependymoma.
Report of 21 cases with review of the literature.
J Neurosurg 1978;49(5):68996.
[57] Prayson RA, Suh JH. Subependymomas: clinicopathologic study of 14 tumors, including comparative MIB-1 immunohistochemical analysis with
other ependymal neoplasms. Arch Pathol Lab
Med 1999;123(4):3069.
[58] Moss TH. Observations on the nature of subependymoma: an electron microscopic study.
Neuropathol Appl Neurobiol 1984;10(1):6375.
[59] Dal Cin P, et al. Cytogenetic investigation in
subependymoma. Cancer Genet Cytogenet 1999;
108(1):84.
[60] Wiederholt WC, Gomez MR, Kurland LT. Incidence and prevalence of tuberous sclerosis in
Rochester, Minnesota, 1950 through 1982. Neurology 1985;35(4):6003.
[61] Shepherd CW, et al. Subependymal giant cell
astrocytoma: a clinical, pathological, and ow
cytometric study. Neurosurgery 1991;28(6):8648.
[62] Altermatt HJ, Scheithauer BW. Cytomorphology
of subependymal giant cell astrocytoma. Acta
Cytol 1992;36(2):1715.
[63] Hirose T, et al. Tuber and subependymal giant cell
astrocytoma associated with tuberous sclerosis: an
immunohistochemical, ultrastructural, and immunoelectron and microscopic study. Acta Neuropathol (Berl) 1995;90(4):38799.
[64] Lopes MB, et al. Immunohistochemical characterization of subependymal giant cell astrocytomas.
Acta Neuropathol (Berl) 1996;91(4):36875.
[65] Gyure KA, Prayson RA. Subependymal giant cell
astrocytoma: a clinicopathologic study with
HMB45 and MIB-1 immunohistochemical analysis. Mod Pathol 1997;10(4):3137.
[66] Debiec-Rychter M, et al. Cytogenetic changes in
two cases of subependymal giant-cell astrocytoma. Cancer Genet Cytogenet 1999;109(1):
2933.
[67] Sampson JR, Harris PC. The molecular genetics
of tuberous sclerosis. Hum Mol Genet 1994;3:
147780.
[68] Janssen B, et al. Rened localization of TSC1 by
combined analysis of 9q34 and 16p13 data in 14
tuberous sclerosis families. Hum Genet 1994;
94(4):43740.

[69] Hino O, et al. The predisposing gene of the Eker
rat inherited cancer syndrome is tightly linked to
the tuberous sclerosis (TSC2) gene. Biochem
Biophys Res Commun 1994;203(2):13028.
[70] Hassoun J, et al. Central neurocytoma. An
electron-microscopic study of two cases. Acta
Neuropathol (Berl) 1982;56(2):1516.
[71] Hassoun J, et al. Central neurocytoma: a synopsis
of clinical and histological features. Brain Pathol
1993;3(3):297306.
[72] Maiuri F, et al. Central neurocytoma: clinicopathological study of 5 cases and review of the literature. Clin Neurol Neurosurg 1995;97(3):21928.
[73] Kim DG, et al. Intraventricular neurocytoma:
clinicopathological analysis of seven cases. J
Neurosurg 1992;76(5):75965.
[74] Yasargil MG, et al. Central neurocytoma: histopathological variants and therapeutic approaches.
J Neurosurg 1992;76(1):327.
[75] Favereaux A, et al. Histopathological variants of
central neurocytoma: report of 10 cases. Ann
Pathol 2000;20(6):55863.
[76] Mena H, et al. Central neurocytomas express
photoreceptor dierentiation. Cancer 2001;91(1):
13643.
[77] Ng TH, et al. Pigmented central neurocytoma: case
report and literature review. Am J Surg Pathol
1999;23(9):113640.
[78] von Deimling A, et al. Patterns of dierentiation in
central neurocytoma. An immunohistochemical
study of eleven biopsies. Acta Neuropathol (Berl)
1990;79(5):4739.
[79] Kubota T, et al. Central neurocytoma: immunohistochemical and ultrastructural study. Acta
[80] Mackenzie IR. Central neurocytoma: histologic
atypia, proliferation potential, and clinical outcome. Cancer 1999;85(7):160610.
[81] Sharma MC, et al. A study of proliferative markers in
central neurocytoma. Pathology 1998;30(4):3559.
[82] Robbins P, et al. Central neurocytoma. A clinicopathological, immunohistochemical and ultrastructural study of 7 cases. Pathol Res Pract
1995;191(2):10011.
[83] Tsuchida T, et al. Neuronal and glial characteristics of central neurocytoma: electron microscopical analysis of two cases. Acta Neuropathol (Berl)
1996;91(6):5737.
[84] Yin XL, et al. Detection of chromosomal imbalances in central neurocytomas by using comparative
genomic hybridization. J Neurosurg 2000;93(1):
7781.
[85] Taruscio D, et al. Nonrandom gain of chromosome 7 in central neurocytoma: a chromosomal
analysis and uorescence in situ hybridization
study. Virchows Arch 1997;430(1):4751.
[86] Tong CY, et al. Central neurocytomas are genetically distinct from oligodendrogliomas and neuroblastomas. Histopathology 2000;37(2):1605.
481
[87] Homan JC, Jr, Bufkin WJ, Richardson HD.

Primary intraventricular meningiomas of the
fourth ventricle. Am J Roentgenol Radium Ther
Nucl Med 1972;115(1):1004.
[88] Criscuolo GR, Symon L. Intraventricular meningioma. A review of 10 cases of the National Hospital,
Queen Square (19741985) with reference to the literature. Acta Neurochir (Wien) 1986;83(34):8391.
[89] Imielinski BL, Kloc W. Meningiomas of the lateral
ventricles of the brain. Zentralbl Neurochir 1997;
58(4):17782.
[90] Vassilouthis J, Ambrose JA. Intraventricular meningioma in a child. Surg Neurol 1978;10(2):1057.
[91] Diaz P, et al. Multiple meningiomas of the fourth
ventricle in infancy: case report. Neurosurgery
1990;26(6):105760.
[92] Sgouros S, Walsh AR, Barber P. Intraventricular
malignant meningioma in a 6-year-old child. Surg
Neurol 1994;42(1):415.
[93] Kakita A, et al. Clear cell variants of intracranial
tumors: meningioma and ependymoma. Noshuyo
Byori 1995;12(2):1116.
[94] Kamiya K, Inagawa T, Nagasako R. Malignant
intraventricular meningioma with spinal metastasis through the cerebrospinal uid. Surg Neurol
1989;32(3):2138.
[95] Killebrew K, et al. Metastatic renal cell carcinoma
mimicking a meningioma. Neurosurgery 1983;
13(4):4304.
[96] Raila FA, Bottoms WT Jr, Fratkin JD. Solitary
choroid plexus metastasis from a renal cell
carcinoma. South Med J 1998;91(12):115962.
[97] Nakabayashi H, et al. Choroid plexus metastasis
from gastric cancercase report. Neurol Med
Chir (Tokyo) 1994;34(3):1836.
[98] Piniella AM, Siatkowski RM. Adrenal cortical
carcinoma metastatic to the brain in a child.
J Neuroophthalmol 2000;20(1):357.
[99] Gyure KA, Morrison AL. Cytokeratin 7 and 20
expression in choroid plexus tumors: utility in
dierentiating these neoplasms from metastatic
carcinomas. Mod Pathol 2000;13(6):63843.
[100] Giannini C, et al. Intraventricular perineurioma:
case report. Neurosurgery 1998;43(6):147882.
[101] Jung JM, et al. Malignant intraventricular schwannoma. Case report. J Neurosurg 1995;82(1):1214.
[102] Abrahams JM, et al. Hemangiopericytoma of the
third ventricle. Case report. J Neurosurg 1999;
90(2):35962.
[103] Hirano A, Hirano M. Benign cystic lesions in the
central nervous system. Light and electron microscopic observations of cyst walls. Childs Nerv Syst
1988;4(6):32533.
[104] Jeree RL, Besser M. Colloid cyst of the third
ventricle: a clinical review of 39 cases. J Clin
Neurosci 2001;8(4):32831.
[105] Tillich M, et al. Symptomatic neuroepithelial (ependymal) cyst of the fourth ventricle: MR appearance.
AJR Am J Roentgenol 1999;172(2):5534.
482
[106] Gratton RJ, Hogge WA, Aston CE. Choroid plexus

cysts and trisomy 18: risk modication based on
maternal age and multiple-marker screening. Am J
Obstet Gynecol 1996;175(6):14937.
[107] Lee KS, Bae HG, Yun IG. Intraventricular arachnoid cyst. J Neurosurg 1989;70(1):
1545.
[108] Suess O, Hammersen S, Brock M. Intraventricular
cavernoma: unusual occurrence in the region of
the foramen of Monro. Br J Neurosurg 2002;16(1):

789.
[109] Madrazo I, et al. Intraventricular cysticercosis.
[110] Vender JR, et al. Intraventricular cryptococcal
cysts. AJNR Am J Neuroradiol 1996;17(1):1103.
[111] Mogilner A, et al. Nocardia abscess of the choroid
plexus: clinical and pathological case report.
Intraventricular neurocytomas
Janet Lee, MS, Susan M. Chang, MD, Michael W. McDermott, MD,
Andrew T. Parsa, MD, PhD*
Department of Neurological Surgery, University of California at San Francisco, 505 Parnassus Avenue,
M-779, San Francisco, CA 94143, USA
Background
Epidemiology and clinical presentation
A review of 385 reported neurocytoma cases
[196] shows that, in general, central neurocytomas (CNCs) are well-dierentiated intraventricular tumors that aect young adult men and
women equally. Most commonly, CNCs occur in
the anterior portion of the lateral ventricle around
the foramen of Monro and can attach to either the
septum pellucidum or the lateral wall of the
ventricle [19,58,71,97]. Whereas 75% of cases
occur in patients between the ages of 20 and
40 years [15,28,98], CNC occurring in patients
18 years of age or younger [3,5,14,19,32,36,42,
44,58,68,69,71,74,76,78,88,94] and in patients 50
years of age or older [3,4,7,19,40,44,45,53,69,
76,78,93,99] have been reported. The term
extraventricular neurocytoma is used to describe
histologically similar tumors not found in intraventricular locations [43]. Extraventricular
locations include the occipital lobe [8,22,54,72],
parietal lobe [3,22,54], frontal lobe [8,22,24,58,96],
temporal lobe [8,9,22,58], thalamus [8,72,100],
hypothalamus [8,22,69], cerebellum [7,17], pons
[77], spinal cord [3,12,18,44,47,79,82,83], cauda
equina [81], retina [52], and pelvis [101] as well as
mature cystic teratoma of the ovary [102]. Cases
J. Lee was supported in part by a grant from the

Khatib Research Foundation as a Khatib Scholar
(20032004).
E-mail address: parsaa@neurosurg.ucsf.edu
(A.T. Parsa).
of neurocytoma have previously been reported as

intraventricular oligodendroglioma, dierentiated
cerebral neuroblastoma [34,95], primary cerebral
neuroblastoma [64,103,104], and intraventricular
neuroblastoma [105].
A review of the available data from clinical
reports has provided some insight into common
signs and symptoms associated with CNC. The
clinical presentation usually involves signs
and symptoms of increased intracranial pressure
(ICP) of a few weeks to several months as a
result of noncommunicating hydrocephalus. As
shown in Table 1, reported symptoms include
headache, visual disturbance, motor disturbance,
altered mental status, sensory disturbance, seizure, dizziness, and nausea or vomiting without
associated headache. Not all reported cases of
CNC include a detailed clinical history, however.
Table 1 also reports signs elicited on physical
examination; however, this analysis is limited by
the lack of detailed information in some reports.
Other authors have reported headache, nausea
and vomiting, and visual disturbance as the most
common symptoms, with papilledema present in
most patients [46,69]. Signs like ataxia [32,42,
46,58,69,94], altered level of consciousness [16,
46,62,69,87,91], hemiparesis [16,32,40,46,63], and
seizures [3,8,22,42,46,58,63,89] were less common. Patients presenting with intraventricular
hemorrhage (IVH) [12,23,37,62,76,84,91] and
sudden death [4] have also been reported. In
addition, many cases have been discovered incidentally in patients undergoing neuroimaging
for unrelated reasons [15,22,32,58,87,91]. Neurologic examination often yields no focal neurologic
ndings other than those caused by increased
ICP [46].
doi:10.1016/S1042-3680(03)00064-0
484
J. Lee et al / Neurosurg Clin N Am 14 (2003) 483508
Table 1
Signs and symptoms
Signs and Symptoms
No.
HA
N/V without HA
Dizziness
Visual disturbance
Altered mental status
Seizure
Motor disturbance
Sensory disturbance
Papilledema
178/202
2/202
4/202
51/202
22/202
9/202
40/202
14/202
46/51
88.1
1.0
2.0
25.2
10.9
4.5
19.8
6.9
90.2
Abbreviations: HA, headache or reported as signs and

symptoms of increased intracranial pressure; N/V,
nausea and/or vomiting without headache. Dizziness
includes dizziness or vertigo; visual disturbance includes
blurring, diplopia, decreased acuity, intermittent vision
loss, photophobia, blindness, or abducens nerve palsy;
altered mental status includes altered level of consciousness, loss of memory, apathy, disorientation to mild
dementia, loss of concentration, mood swings, syncope,
personality change, depression, psychosis, irritability, or
mental change; motor disturbance includes spasticity of
extremities, hemiparesis, hemiplegia, atrophy, clonus,
hypotonia, increased DTR, imbalance, left hemisyndrome, gait disturbance, ataxia, loss of balance, weakness, pyramidal signs unsteady gait, or gait dysfunction;
sensory disturbance includes pain, paresthesia, left
hemisyndrome, hemihypaesthesia; seizure includes all
types of seizures and papilledema includes unilateral and
bilateral types. Data from references [18,10,11,14,15,18,
23,26,27,29,31,33,34,39,41,42,45,5658,6266,68,70,71,
7476,82,85,87,89,91,93,95,96,99,104,108,114, and 118].
Neuroimaging
CT/MRI/angiography
Appropriate diagnostic imaging studies for
patients with CNC may include CT, MRI, or
angiography. Relative to the brain parenchyma,
CNC appears as a well-circumscribed, isodense,
hyperdense, or mixed isodense/hyperdense mass
with slight to moderate contrast enhancement on
CT examination. Noncommunicating (obstructive) hydrocephalus is often present. Calcications
and cyst-like areas may also be seen on CT
images. Compared with the surrounding white
matter, T1-weighted and proton-weighted images
typically appear isodense. T2-weighted images
appear heterogeneous with areas of calcication
and cysts appearing hyperintense and the tumor
appearing isointense to hyperintense. Variable
enhancement with gadolinium is common, reecting the heterogeneous vascularity of CNC (Fig. 1).
For tumor localization and visualization of the
attachment site, MRI is preferred [15,46,97,106].

Angiography has also been used to assess
vascularity; however, the results are nonspecic,
ranging from avascular [3,36,40,71] to highly
vascular [40,46,53,85]. Feeding arteries are reported to include anterior choroidal [1,15,40,
53,71,76], posterior choroidal [1,31,40,53,57,87],
lenticulostriate [40,53], and branches of pericallosal arteries [40,53].
Magnetic resonance spectroscopy/single photon
emission computed tomography/positron
emission tomography
Magnetic resonance spectroscopy (MRS)
[33,35,38,54,94], single photon emission computerized tomography (SPECT) [35], and positron
emission tomography (PET) [35,53,85] have all
been used in preoperative evaluation. The experience using these modalities is limited, however,
and there is no consensus on the characteristics of
CNC. Many investigators suggest that prominent
glycine and choline with low N-acetyl aspartate
(NAA) peaks are characteristic markers of CNC
on MRS [33,38,94]. Others have found an increased choline peak and decreased NAA signal
but failed to consistently nd the 3.55-ppm peak
characteristic of glycine [35,54]. SPECT analysis
shows increased uptake of 201T on delayed images,
indicating a high activity of sodium potassium
triphosphate on cell membranes. PET analysis has
demonstrated decreased O2 extraction fraction,
cerebral metabolic rate of O2, and cerebral
metabolic rate of glucose in CNC. Cerebral blood
ow and blood volume were increased in three of
four cases, correlating with the angiographic
ndings in these patients [53]. The authors suggest
that CNC metabolism is more oxidative than that
of other brain tumors and that a decreased rate of
glucose metabolism may predict a favorable
prognosis [52]. Further study is necessary to
determine whether these combined ndings are
truly characteristic of CNC.
Pathologic examination
After reviewing such factors as presenting
symptoms, patient age, and location of the tumor,
a focused dierential diagnosis for CNC includes
subependymoma, astrocytoma, ependymoma, intraventricular meningioma, intraventricular oligodendroglioma, and subependymal giant cell
astrocytoma [43,46,58,79,97]. Adding information
from imaging studies usually narrows the dierential diagnosis to intraventricular meningioma,
485
Fig. 1. (A) Axial CT image with contrast. (B) Contrast-enhanced coronal T1-weighted MRI. (C) Axial T1-weighted
MRI. (D) Sagittal T1-weighted MRI.
ependymoma, and CNC. A denitive diagnosis

requires tissue analysis with light microscopy,
immunohistochemistry, and, in some cases, ultrastructural examination.
Light microscopy
On light microscopy, CNC appears similar to
oligodendroglioma (Fig. 2). Clinically, they can be
distinguished based on their location. CNCs are
typically intraventricular and centrally located,
whereas oligodendrogliomas arise more peripherally. Light microscopy shows a honeycomb architecture with uniform small round cells with central
nuclei and clear cytoplasm dispersed within a
brillary stroma. The chromatin typically has a
salt and pepper appearance. Microcalcications
or microcysts may be present, and mitoses, endothelial proliferation, and necrosis are rare [15,
43,46,98,106]. The presence of neuroblastic rosettes and nuclei with a ganglionic appearance is
suggestive of neurocytoma; however, immunohis-
tochemistry or electron microscopy is required to

conrm the diagnosis [46].
Immunohistochemistry
The hallmark characteristic of CNC is positivity for synaptophysin, a calcium-binding membrane protein of presynaptic vessels. In addition,
CNC is usually positive for neuronal specic
enolase (NSE) and negative for glial brillary
acidic protein (GFAP) and neurolament protein
(NFP) (Fig. 3A) [15,43,46,51,98,106]. These characteristics dierentiate CNC from oligodendroglioma and ependymoma.
Ultrastructural examination is only required in
the diagnosis of CNC if synaptophysin is lacking
or equivocal or if extraventricular neurocytoma is
suspected. CNC shows neuronal dierentiation,
microtubules, dense core and clear vesicles, and
abortive or typical synapses (see Fig. 3B)
[15,27,28,46,51,98,106]. Occasional mitochondria,
486
Fig. 1 (continued )
moderate free ribosomes, and variable endoplastic

reticulum may also be present [15,43,46,98].
Histogenesis/genetics
CNC is thought to be derived from bipotential
progenitor cells from the subependymal plate that
are capable of neuronal and glial dierentiation
[30,89,93,97,107]. Indeed, on cell culture, CNC
dierentiates into neuronal and glial cells [30,89].
In addition, CNC is capable of ependymal differentiation [75]. CNCs are genetically distinct
from oligodendrogliomas and neuroblastomas, as
evidenced by a lack of association with specic 1p
and 19q loss of heterozygosity and rarity of Nmyc amplication [75].
Atypical neurocytoma
Atypical neurocytomas are a rare variant of
CNC, with cellular pleomorphism, mitotic activity, necrosis, or vascular proliferation (Fig. 4)
[3,19,54,62,71,73,77,92,108111]. Although most
CNCs appear as uniform small round cells on
light microscopy, atypical CNCs may show peri-
Fig. 2. Histological stain of a central neurocytoma

demonstrating sheets of monomorphic pale cells with
small delicate capillaries in the background. The tumor
cells have round uniform nuclei with ne chromatin and
inconspicuous nucleoli. There is no evidence of necrosis
or mitotic activity. (From Anderson RC, Elder JB,
Parsa AT, Issacson SR, Sisti MB. Radiosurgery for the
treatment of recurrent central neurocytomas. Neurosurgery 2001;48(6):12318; with permission.)
487
Fig. 4. Central neurocytoma with atypical histologic

features, including necrosis, vascular endothelial proliferation, and cellular pleomorphism (hematoxylineosin, original magnication 10). (From Mackenzie
IR. Central neurocytoma: histologic atypia, proliferation potential, and clinical outcome. Cancer 1999;85(7):
160610; with permission.)
ation potential correlates with poor outcome,

histologic grade does not seem to have prognostic
value [77,110].
Treatment
Overview
Fig. 3. (A) Central neurocytoma showing immunoreactivity for synaptophysin (synaptophysin immunohistochemistry, original magnication 20). (B) Electron
microscopy of central neurocytoma. The tumor cells
have round nuclei and clear cytoplasm. The cytoplasm
contains microtubules, dense core vesicles, and synapses
(*) (original magnication 10,500). (From Hara M,
Aoyagi M, Yamamoto M, Maehara T, Takada Y, Nojiri
T, et al. Rapid shrinkage of remnant central neurocytoma after gamma knife radiosurgery: a case report.
J Neurooncol 2003;62(3):26973; with permission.)
vascular pseudorosettes, neuropil islands, multinucleate cells, or ganglion cells. Mitotic activity
can be as high as 30 mitoses per high-power eld,
and evidence of necrosis may range from focal to
extensive. Although the correlation between
histologic atypia and proliferation potential in
atypical CNC was poor [110], vascular proliferation showed a signicant correlation with the
MIB-1 labeling index (LI) (P = 0.0006) [77]. It is
unclear how the histology of atypical CNC relates
to biologic behavior. Although elevated prolifer-
Treatment strategies for CNC are based on

retrospective case series (Table 2), case reports, and
analysis of pooled data. There are no randomized
clinical trials and few prospective studies. In many
earlier reported cases, initial management may
have been based on a diagnosis that was revised on
retrospective review [40,42,46,65,67,68,70,71,89].
Most authors agree that, when possible, complete tumor resection for symptomatic CNC is the
treatment of choice [19,40,44,46,60,70,96]. The
addition of adjuvant radiation therapy (RT) in
the immediate postoperative period is controversial. Some authors routinely use RT after subtotal
tumor resection (STR) [1,5,19,39,46,59,89]. Although some have used RT after gross total
resection (GTR) as well [3,7,15,19,27,32,40,45,53,
58,70,96], several authors state that RT after GTR
is not indicated [3,5,27,41,46,56,57,64,70,71,90].
Given the potential for long-term radiation side
eects, some advocate for adjuvant RT only for
recurrent or progressive CNC [24,46,57], because
the subependymal and subventricular zone is
sensitive to radiation. More recent reports of
stereotactic radiosurgery address the concerns of
488
Table 2
Larger case series
MIBY
labeling
n index
Location
Primary
treatment
Recurrence
Average
months to
recurrence
(range)
Radiation
Timing
of RT
Average
FU in
months
(range)
Local
control
Survival
rate
Outcome
Reference
Not
reported
3 LLV, 1
LLV/3rd
4 CTR
4/4
17.25
(925)
GKS, 1620
Gy to tumor margin
4 salvage
20.25 p
0% p initial
GKS, 54.5 surgery,
p CTR
100% p GKS
[2]
100.0% Returned to work with
full fxn, 3/4 neurologically
nl, 1/4 on dilatin for
postoperative seizure
3.6375
9 intraventricular
6 ITR, 3
CTR, 3 RT
2/9
618
55 Gy (2 adjuvant, 1
salvage for
asymptomatic
progression)
2 adjuvant,
1 salvage
45.33
(689)
83.30%
91.7%
Not
reported
3 LLV,
1 BLV
1 CTR with 1/4

GKS for
recurrence, 3
ITR/GKS
53
GKS, 913 Gy (3
adjuvant, 1 salvage)
3 adjuvant,
1 salvage
44
(1299)
100%
100.0% 3 asymptomatic, 1
neurologically intact
18 Not
reported
2 3rd, 3 3rd/
LV, 2 CC, 1
CC/SP, 8
FOM/LV,
2 SS
8/18
4 biopsy/
VPS/RT, 7
CTR, 7 ITR,
1 chemo,
7RT
NS
NS
7 adjuvant
46.4
(684)
44.40%
83.0%
10 1.9
(0.15.6)
Not
reported
8 ITR/RT, 1 1/10
ITR, 1 ITR/
RT then
CTR
At least
12
5 RT, 2650 Gy whole 9 adjuvant,

brain, 2032 Gy local, 1 salvage
1 RT, 50 Gy whole
brain, 4 RT, 5060 Gy
local
90
(23160)
100%
100.0% 10/10 no evidence or

recurrence
2 LLV, 1
midline, 3
LLV/3rd
NS
NS
NS
15
(0.536)
NS
100.0% 2 asymptomatic, returned [23]

to work; 1 mild residual
hemiplegia; 1 behavioral,
cognitive, and dexterity
problems; 1 short-term
memory decit; 1 recovering well after surgery
Not
reported
NS
NS
7 KPS 100, 1 KPS 50, 1

died secondary to
hemorrhage
[3]
[11]
2 dead, tumor recurrence; [19]

7 alive, no tumor
recurrence; 6 alive, tumor
recurrence; dead no
tumor recurrence; 1 dead
4 days after surgery; 1
dead, 2 years after surgery
[20]
5 RLV, 1
LLV, 2
BLV, 2
BLV/3rd, 1
RLV/3rd, 1
LLV/3rd, 3
NS
7 CTR, 8
ITR 7 RT
2/15
821
6 RT, 50.459.8 Gy, 1 7 adjuvant,

2 salvage
RT, 54.6 Gy/21 Gy
spine, 1 radiosurgery
with recurrence 1 with
14 Gy with recurrence,
then refused Tx, reoperation 10 years later
66.1
(18168)
75% surgery
only, 100%
surgery/RT
100.0% 12 KPS 100, 1 KPS 90,

1 KPS 90 before suicide,
1 KPS 60
[39,
40]
Not
reported
2 BLV, 3
LLV, 1
LLV/3rd, 1
RLV
3 CTR, 4
ITR, 2 RT
NA
60 Gy
Adjuvant
52.7
NS
87.5%
2 KPS 100, 3 KPS 90, 1

KPS 80, KPS 0 (died 1
day after surgery)
[41]
Not
reported
7 LV, 1 3rd
2 stereotactic 1/8
Bx/RT, 5
stereotactic
Bx/RT/
chemo
15
50 Gy whole
brain (180 cGY
56 weeks)
Primary
78 (15
108)
88%
86.0%
[42]
1 lost to follow-up; 5
asymptomatic, employed,
KPS >90; 1 shunt
surgery at 9 years after
initial RT, employed,
KPS >90; 1 died as a
result of shunt
dysfunction at 5 years
after initial RT
Not
reported
3 RLV, 2
LLV/3rd
1 ITR, 2
ITRNPS/
RT, 2
ITRNPS/
chemo/RT
NA
1, RT, 52.5 Gy to
tumor, 3 RT, 54.0
Gy to tumor/30
Gy to axis
Adjuvant
23.5 (11
78)
NS
100.0% 5 alive and well
2 Bx/obs, 8 n
CTR, 5 ITR,
5 RT
NA
NS
NS
72.8 (13
255)
71.40%
93.3%
[45]
1 lost to follow-up, 10
alive and well, 3
asymptomatic recurrence,
1 died at 1 month
NA
NA
NA
NS (up to NS
50)
80.0%
[46]
1 recovered with slight
hemiparesis at 5 y, 1 lost
to FU p 3 months, 1 died
after surgery, 2 totally
recovered after surgery
15 1.9 (0.16) 15
ventricular
[44]
Not
reported
2 LLV/3rd, 1 3 CTR, 2
ITR
LLV, 1
BLV/3rd, 1
BLV
Not
reported
2 RLV, 2
LLV/3rd
1 Bx/obs, 2 n
CTR, 1 ITR,
1 RT
NA
50 Gy
Adjuvant
50.25 (4
135)
80%
100.0% 2 no recurrence, 2 no
regrowth
[53]
Not
reported
6 LLV, 1
BLV/3rd, 1
BLV
6 ITR, 2
ITR, 3 RT
NS
NS
3 adjuvant
120.5
(67.2
181.2)
NS
100.0% 3 KPS 100, 2 KPS 90, 2

KPS 70, 1 KPS 50
[58]
NS
489
15 9
proliferating cell
nuclear
antigen
<1%
490
Table 2 (continued)
Radiation
Timing
of RT
2 LV, 1 BLV/ 2 ITR, 2

3rd, 1 LV/3rd ITR, 3 RT
NA
NS
3 adjuvant
2 BLV, 2
RLV, 3 LLV
2 GTR, 4
STR, 1
BxNPS/RT,
2 RT
NA
NS
32 Not
reported
10 LLV, 10
RLV, 11 BLV,
1 hypothalamus
NS
5 GTR,
5 GTR/RT,
14 STR,
8 STR/RT
NS
1 RLV, 1
LLV, 1 SS
3 ITR,
2 RT
Location
4* Not
reported
7
6 <1%
Not
reported
Primary
treatment
1/3
Average
FU in
months
(range)
Local
control
Survival
rate
Outcome
Reference
81.75 (15 100%

227)
100.0% KPS 100, 90, 80, 50; 5/6

no recurrence
[60]
3 adjuvant
10122
100%
85.7%
5 alive, asymptomatic; 1
death; 1 alive with
considerable neurologic
decits
[99]
11 RT, 48.661.2 Gy
to tumor bed, 2 RT,
3036 Gy to
craniospinal axis for
adjuvant RT
13 adjuvant,
3 salvage
56.4
(28184)
79%
81.0%
NS
[70]
NS
1 salvage
at 9 mo
30, 34 ys,
10
66.70%
66.7%
1 died at 10 mo, 1 alive

at 34 y, 1 alive at 2.5 ys,
shunt after surgery
[71]
NA
4060 Gy over 6
weeks
Adjuvant
32 (672)
100%
75.0%
15 alive, 5 died after

[74]
surgery as a result of IVH
[78]
29 alive, 4 NS, 1 dead
with cerebral edema after
surgery, 1 dead with
massive IVH 6 weeks after
surgery, dead with brain
stem infarct 1 month after
recurrence noted at 4
months
20 0.75
(0.13)
14 CTR, 6
n (4 lost
10 RLV, 4
LLV, 4 LLV, 1 ITR, 15 RT to FU)
LLV/3rd, 1
3rd
36 3.4
(0.121)
36 LV
8/36
34 CTR,
1 stereotactic
Bx/RTCTR
for residual,
1 biopsy/RT
27.75
(473)
2 RT, 3460 Gy whole 7 adjuvant,

brain/2030 Gy spine, 1 primary,
6 RT, 6.855 Gy
3 salvage
46.9 (0.5
204)
NS
89.7%
Not
reported
1 LLV, 1 LV/
3rd, 1 3rd/
hypothalamus, 1
3rd/thalamus
1/4
2 biopsy/
GKS, 1
CTR, 1 ITR
2/GKS
48
GKS, 1420 Gy to
tumor margin, 2840
Gy maximum
2 primary,
1 salvage,
1 adjuvant
45.75
100%
p GKS
100.0% No neurologic, visual or, [88]

endocrine problems;
employed; well, no new
neurologic problems, no
new clinical symptoms
Not
reported
1 BLV, 2 LLV, 2 CTR, 3

1 RLV/3rd, 1 ITR, 2 RT
LLV/3rd
18
Dose not recorded
1 salvage,
3 adjuvant
109.2 (12 50% sur150)

gery, 100%
surgery/RT
1/5
100.0% 3 with memory

disturbance, all tumorfree, 3 results good
[89]
Recurrence
Average
months to
recurrence
(range)
MIBY
labeling
n index
10 Not
reported
2 LV, 4 SP, 1
CC, 4 FOM
NS
1/10
24
NS
6 LV, 1 LV/
3rd/4th
6 CTR, 1
ITR, 1 RT
3/7
36, 38, 72
54 Gy, 50 Gy salvage 1 adjuvant,

1 salvage
Not
reported
NS
NS
NS
NS
NS
[98]
61.8 (5
143)
50%
[96]
100.0% 6 returned to work;
1 residual hemiparesis;
3 shunt at 2 mo, 1 mo and
1 y; 1 asymptomatic
recurrence, obs
Cases 4 and 6 included in reference 113.

Abbreviations: LV, lateral ventricle (otherwise not specied); LLV, left lateral ventricle; BLV, bilateral lateral ventricles; RLV, right lateral ventricle; 3rd, third ventricle;
4th, fourth ventricle; SP, septum pellucidum; CC, corpus callosum; FOM, Foramen or Monroe; FU, follow up; NS, not specied; NA, not applicable; BX, biopsy; Obs,
observation; p, post. Data from Meningiomas 1991:56981.
491
492
radiation side eects; however, long-term followup data are lacking [2,6,11,36,48,65,88]. Chemotherapy has also been used in a limited number of
cases [7,14,18,42,44,71,88]. Outcome measures
used to assess ecacy of therapeutic interventions
include local control, time to progression or
recurrence, survival, and functional performance.
Although most patients with intraventricular
tumors present with symptoms of obstructive
hydrocephalus, the advent of CT and MRI has
increased detection of asymptomatic tumors [15,
22,32,53,60,87,91]. Management of patients with
asymptomatic CNC is largely unexplored in the
published literature. Indications for treatment
usually include signs and symptoms caused by
the tumor. Consequently, a practical approach to
patients harboring an intraventricular tumor with
characteristic features of a CNC may be to delay
intervention until symptoms occur [58]. By extension, judicious observation may be appropriate
for patients with asymptomatic recurrence or
progression.
Given that initial management in many CNC
patients was based on an incorrect diagnosis [2,
3,14,40,42,46,60,65,68,89,90,93], there is no clear
management or treatment paradigm for primary,
recurrent, or progressive CNC in the literature
[36]. In addition, long-term follow-up for patients
is not standardized. The following section discusses treatment management for symptomatic
CNC, including prospective observation, surgery,
RT, radiosurgery, and chemotherapy for primary
and recurrent CNC.
Observation
Because the indications for surgery usually
include signs and symptoms caused by the tumor
[58], one approach to asymptomatic tumor
management is observation with close follow-up.
One of the rst reported cases of CNC was
initially treated with a ventriculoperitoneal shunt
(VPS), followed by observation for an unspecied
period [27]. The patient later underwent biopsy
and GTR after symptomatic progression. Mackenzie [45] has reported two cases of CNC that
were observed after biopsy, and both patients are
alive and well with no progression at 35 and 255
months of follow-up, respectively. The MIB-1 LI
for these biopsies showed MIB-1 LIs of 0.1% and
1.8%, respectively. Giangaspero and colleagues
[22] have reported one case of a parietal neurocytoma discovered incidentally after biopsy. This
patient had stable disease at the 6-month follow-
up examination but showed progression at the 26month follow-up examination. The LI of this
tumor was 1% to 1.5%. Mineura and colleagues
[53] also have reported one case of CNC
discovered incidentally after biopsy. No progression was noted at the 51-month follow-up
examination. Yasargil and colleagues [96] have
reported two cases of asymptomatic recurrent
CNC 3 and 6 years after GTR. The patients were
observed and were alive and well at last follow-up
at 58 and 92 months, respectively.
Takao and colleagues [82] have reported a case
of CNC that was observed after biopsy because
of pregnancy. STR was performed 11 months
later because of symptomatic progression. As the
result of postoperative MRI revealing residual
tumor with spinal cord dissemination, the patient
underwent RT (66 Gy, cone technique, extended
eld at 40 Gy, reduced eld at 20 Gy, with use of
limited eld size for a nal boost of 20 Gy to the
tumor bed and 46 Gy to the spinal cord, 2 Gy/d,
4 fractions per week), resulting in a decrease in
tumor size and stable disease at the 3-month
follow-up examination. Agranovich and colleagues [1] have reported a case of CNC that
presented as IVH on imaging and was followed
with serial CT scans because the patient declined
surgery. After subsequent biopsy and MRI 3
years later because of symptomatic progression,
the patient received RT (50 Gy in 25 fractions
over 5 weeks, isocentric technique [6.5 cm 5
cm, 7 5 eld size] with a 6-MV energy linear
accelerator [LINAC]) and was asymptomatic
with stable tumor size at the 3-year follow-up
examination.
Because of the benign clinical course of CNC,
prospectively observing patients after biopsy until
symptomatic progression of the tumor may be
a reasonable approach. The MIB-1 LI may help
to stratify patients into high-risk and low-risk
groups. Observation may also play a role in tumor
management after asymptomatic progression or
recurrence; however, this approach should be used
with caution, because recurrent tumors demonstrate proliferation that may indicate a more
clinically aggressive tumor (Table 3).
Acute management
Because many patients present with symptoms of increased ICP, acute management may
necessitate the use of temporizing measures
before a more denitive treatment is administered. Treatment of acute noncommunicating
493

Table 3
Observation
Recurrence
or
progresson
Months to
recurrence
or progression
Treatment for
recurrence
or progression
36
Not stated
Biopsy/observation
Biopsy/observation
Biopsy/observation
Biopsy/observation
us a resulted
pregnancy
GRT
n
n
n
y
11
36
Observation
asymptomatic
58 p GRT
GRT
72
Observation
asymptomatic
92 p GTR
Primary
treatment
Observation with CT
patient refused
surgery
VPS/observation
Follow-up
in months
Outcome
Reference
Stereotactic biopsy
and GTR
36 p GTR
Asymptomatic
[1]
Stereotactic
biopsy and RT
STR/RT
Not stated
Alive, no
recurrence
Alive and well
Alive and well
No regrowth
Alive and well
[27]
35 p biopsy
255 p biopsy
51 p biopsy
3 p STR/RT
Returned to work,
asymptomatic
recurrence
Returned to work,
asymptomatic
recurrence
[45]
[53]
[82]
[96]
Abbreviations: GTR, gross total resection; n, no; RT, radiation therapy; STR, subtotal resection; VPS,
ventriculoperitoneal shunt; y, yes; p, post.
hydrocephalus on an emergent basis through

extraventricular drain placement [11,14,15,27,42,
46,62,71], intravenous steroids [32,63,68], and
hyperosmolar therapy [68] has been reported in
cases of CNC.
Surgical treatment
Improvements in neuroanesthesia, surgical
techniques, and postoperative care have helped to
decrease morbidity and mortality after resection
of deep brain tumors. Therefore, aggressive
resection of tumor should be attempted when
possible [70,106]. The goal of neurosurgery is
complete tumor removal with minimal morbidity
[15]. Complete tumor resection for patients is the
treatment of choice [19,40,44,46,58,70,96]; however, this may not be possible given the vascular
nature of the tumor [38,39,57] or adherence to
adjacent structures [40,58]. GTR is achieved in
only one third to one half of cases [28,40,70].
Particular care should be taken to avoid damage
to the fornices [37,56,58,89]. The surgical approach is variable, depending on the tumor
location, size, and surgeon preference. Transcallosal, transcortical, transventricular, and combined approaches have all been used with success
(Table 4). Because of the fact that most CNCs
arise in the septum, the fornices and thalami are
pushed inferiorly by large tumors. The key to
determining a safe plan for resection is identication of the ependymal surface of the oor of the
ventricle anterior and posterior to the tumor. The
choroid plexus and ependymal veins can then be
used as guides for dissection along the inferior
aspect of the tumor.
Surgical resection: gross total resection versus
subtotal resection
Because of the benign nature of the disease,
GTR and STR have resulted in a stable long-term
outcome. After GTR, disease-free survival has
been reported up to 12.5 years [41]. Similarly,
after STR, stable disease has been reported up to
18.5 years [45]. Recurrences have been reported
after STR and GTR, however [18,39,40,72,96].
There is an inherent bias in the STR group,
because those tumors may be more extensive.
In a retrospective review of 32 cases of patients
who received multimodality therapy, Schild and
colleagues [70] compared GTR RT and
STR RT and found a 5-year local control rate
of 70% for patients after STR RT compared
with 100% for patients after GTR RT (log rank
rest of Kaplan-Meier product limit method projection, P = 0.08). Adjuvant conventional external beam RT in initial treatment ranged from 48.6
to 61.2 Gy in 180- to 200-cGy fractions delivered
to the tumor bed in 11 patients. Two patients
received adjuvant craniospinal and whole-brain
494
Table 4
Surgical approaches
Surgical approach
No. patients
References
Transcallosal
Transcortical
Transventricular
Combined
Biopsy
Not specied
31
30
1
17
28
194
[5, 11, 14, 15, 26, 35, 39, 63, 74, 76, 85, 88, 89, 91, 96]
[1, 7, 14, 15, 32, 35, 36, 40, 43, 46, 53, 62, 68, 71, 75, 82, 88, 89]
[57]
[2, 7, 31, 40, 53, 60, 63]
[7, 14, 19, 28, 30, 42, 45, 53, 78, 82, 88, 99, 105, 108]
[3, 6, 10, 11, 19, 20, 23, 27, 30, 33, 41, 44, 45, 48, 58, 60, 65, 70, 78, 85,
87, 9395, 99]
irradiation using 30 to 36 Gy. In addition, 3 patients received RT (50.4 Gy in 28 fractions to the

tumor bed, 30 Gy in 15 fractions to the craniospinal axis with a boost to a total dose of 60 Gy in
30 fractions to the tumor bed, and 15 Gy in 1
fraction with stereotactic gamma knife surgery
[GKS]) as a part of salvage therapy for tumor
progression. The 5-year survival rate was 77% for
patients after STR RT compared with 90%
for patients after GTR RT (log rank rest of
Kaplan-Meier product limit method projection,
P = 0.44). After GTR without postoperative RT,
the 5-year local control and survival rates were
100% and 80%, respectively. This case series
demonstrates a trend for better local control with
GTR versus STR and raises the question of the
need for adjuvant RT after GTR.
Rades and Fehlauer [112] have reported a 3year local control rate of 95% after GTR and
55% after STR in a retrospective analysis of 310
CNC cases. At 5 years, local control rates were
85% after GTR and 46% after STR (log rank rest
of Kaplan-Meier projection, P<0.0001). Median
time to progression was 36 months after GTR and
20 months after STR. The 5-year survival rates
were 99% after GTR and 86% after STR (log rank
rest of Kaplan-Meier projection, P = 0.0028).
These data show that GTR yields signicantly
better local control rates and survival than STR. In
addition, no dierences of surgery-related morbidity were reported between the two groups.
When possible, GTR is the treatment of choice
because overall local control and survival are
high. In addition, functional outcome after GTR
and STR using the Karnofsky performance scale
[35] or other scales seems to be high [40,42,112].
Because many CNC reports are from the pathologic literature, however, postoperative and longterm follow-up data were not always described in
detail. Complications, such as hydrocephalus
requiring VPS placement [2,27,71,96], mild cognitive defects [40,89], transient hemiplegia or hemi-
paresis [11,15,16,32,46,68] meningitis [27], hemorrhage [43,78], and death [41,43,46,70,78], after
surgery have been noted, but most patients have
an uncomplicated postoperative course (Table 5).
Surgery for recurrence/progression
CNC can recur or progress, and some authors
have reported their experience with repeat surgery
[2,10,78,96]. Over the years, the small but demonstrated risk of morbidity associated with surgery
has provided the impetus to pursue alternative
treatments for recurrence. Accordingly, surgery
for recurrence or progression has decreased in
frequency over the years as other therapeutic
options have become available [40,62,71].
Stereotactic biopsy
In cases where initial treatment does not
include surgical resection, a tissue biopsy must
be obtained to establish a denitive diagnosis.
Although histologic atypia has not been shown to
correlate with clinical outcome [39,40,99], proliferation potential has been correlated to prognosis. Evaluation of proliferation potential using
the MIB-1 LI can be used to guide further
therapy. MIB-1 is a monoclonal antibody that is
more durable than the original Ki-67 antibody
and is used as a nuclear proliferation marker.
Nuclei positive for MIB-1 are easier to count than
nuclei positive for proliferating cell nuclear
antigen (PCNA), and the results are comparable
to those of bromodeoxyuridine (BUDR) analysis
(Fig. 5) [51].
In a retrospective case series evaluating proliferation potential using the MIB-1 LI, Soylemezoglu and colleagues [78] reported that tumors
with an MIB-1 LI greater than 2% (39% of cases)
have a signicantly greater chance of relapse
(63%) over an observation period of 150 months
compared with cases with a lower MIB-1 LI (22%
relapse) (v2 test of Kaplan-Meier analysis using
one degree of freedom, P = 0.08). An MIB-1 LI
495
Table 5
Postoperative complications within the rst 3 months
No.
After gross total resection
(n = 127)
None
Death
Transient hemiparesis
Persistent hemiparesis
Shunt within 1 month
Meningitis
Cognitive dysfunction
Comatose <3 weeks
Decreased vision
108
8
1
6
2
1
5
1
2
85.0
6.3
0.8
4.7
1.6
0.8
3.9
0.8
1.6
After subtotal resection

(n = 99)
None
Death
Transient hemiparesis
Persistent hemiparesis
Hydrocephalus
Cognitive defect
Decreased vision
Seizures
Extradural hematoma
85
3
2
4
3
4
1
1
1
84.8
3.0
2.0
4.0
3.0
4.0
1.0
1.0
1.0
After biopsy
(n = 28)
None
28
100
Cognitive dysfunction includes memory disturbance,

confusion, neuropsychologic dysfunction, cognitive defection stupor.
greater than 2% showed a close correlation with

the presence of vascular proliferation (P = 0.006,
test not specied) [78]. The MIB-1 LI is particularly useful, because only cells in G0 show no
immunoreactivity to MIB-1 [78]. Accordingly, the
MIB-1 LI may help to predict the clinical outcome
of CNC [20,45,74,78].
Radiation therapy
The histopathologic features of CNC, such as
neuronal dierentiation, low mitotic activity,
absence of vascular endothelial proliferation,
and lack of tumor necrosis, suggest a relative
resistance to ionizing radiation [96,113]. CNC has
been shown to respond to RT, however.
The eect of RT on CNC has been explained
by Kim and colleagues [40], who suggest that RT
causes hyalinization of arterioles feeding the
tumor rather than by lethal reproductive damage
or induction of apoptosis in tumor cells. CNCs
are hypervascular tumors, as demonstrated by
angiography and enhancement on MRI and CT.
Fig. 5. (A) MIB-1 labeling index (LI) of 0.6% (MIB-1

immunohistochemistry, original magnication20). (B)
MIB-1 LI of 6.0% (MIB-1 immunohistochemistry,
original magnication20). (From Mackenzie IR.
Central neurocytoma: histologic atypia, proliferation
potential, and clinical outcome. Cancer 1999;85(7):
160610; with permission.)
The decrease in tumor size after RT is often not

immediate; rather, it usually occurs 1 year after
RT. This interval is similar to the that of changes
of small to medium-sized vessels reported in
arteriovenous malformations after radiosurgery
[6,9,40,113].
RT is associated with acute, subacute, and
delayed central nervous system (CNS) toxicities.
Acute toxicities include transient worsening of
symptoms usually caused by peritumoral edema,
nausea, vomiting, alopecia, and dermatitis. These
eects usually subside within the rst 4 to 6 weeks
after completing RT. Subacute toxicity of neurologic deterioration in the 6 to 12 weeks after
therapy may be attributed to changes in capillary
permeability or transient demyelination. Delayed
CNS toxicities may include parenchymal or focal
necrosis with associated impairments of recent
memory, abstraction, problem solving, and learning ability [111,114]. There is also a small risk of
iatrogenic tumor development after RT [111].
496
Toxicities caused by RT correlate with the volume

of brain irradiated and dose [113,114].
Conventional radiation therapy
As primary therapy after biopsy. Conventional
RT has been used as a primary treatment after
biopsy in a limited number of cases. Ishiuchi and
Tamura [30] have reported a patient who received
RT using 50 Gy (dosing schedule not specied)
after biopsy and was free of tumor for 23 years,
the longest reported disease-free follow-up period.
This patient subsequently underwent GTR and
had no evidence of disease at last follow-up.
Kulkarni and colleagues [42] have reported a case
series of eight patients who underwent stereotactic
biopsy followed by conventional whole-brain RT
at a total dose of 50 Gy in 180-cGy fractions for 5
to 6 weeks. Six of these patients had an initial
diagnosis of oligodendroglioma and later also
received chemotherapy with lomustine (seven to
nine doses, total dose not specied). The remaining two patients received no other treatment and
were asymptomatic at 36 and 105 months, respectively. Follow-up CT examination showed
decreased or no contrast enhancement as well as
a decrease in tumor size. Figarella-Branger and
colleagues [19] have reported four cases of
ventroperitoneal cerebrospinal uid (CSF) diversion and biopsy followed by RT (dose not
specied). Two of these patients were alive with
residual tumor at 6 and 7 years. The other two
patients died after 11 months and 2 years of
follow-up. Louis and colleagues [44] have reported one case of biopsy followed by 43.2-Gy RT
to the tumor (dosing schedule not specied). This
patient received no other therapy and was alive
and well at 54 months of follow-up. Soylemezoglu
and colleagues [78] reported two patients who
received RT (55- and 34-Gy whole-brain RT plus
20-Gy spinal RT, respectively [dosing schedule
not specied]) after biopsy. One patient underwent surgery for residual tumor soon after,
whereas the other underwent surgery 4 years
later. Brandes and colleagues [7] have reported
one case of RT (limited eld RT with 54.4 Gy in
30 fractions) after stereotactic biopsy of a tumor
initially diagnosed as an oligodendroglioma. A
partial response of a greater than 50% reduction
in tumor size was achieved, and the patient had
stable disease for 5 years. This patient later
received chemotherapy for recurrence. Although
conventional RT as a primary therapy has been
used with success in the initial treatment of CNC,
the experience is limited and the reported followup data are brief. If the tumor is small and
patients are closely observed for tumor progression, focal RT after biopsy may be a possible
treatment option in centers that lack stereotactic
radiosurgery.
Role of radiotherapy after gross total resection.
Rades and Fehlauer [112] have reported that the
3-year local control rate showed no statistical
dierence between GTR (95%) and GTR/RT
(96%) (RT dose not specied). Similarly, local
control at 5 years showed no statistical dierence
between these two groups (85% for GTR and
89% for GTR/RT). Data from Schild and
colleagues [70] support these ndings and demonstrate that GTR resulted in 5-year local control
and survival rates of 100% and 80%, respectively.
Median time to progression was 36 months after
GTR and 39 months after GTR/RT with 5-year
survival rates of 99% for GTR and 95% for
GTR/RT (dose not specied by patient). Although the number of patients in the GTR/RT
group was small (n = 35), these data suggest that
RT should not be used as an adjuvant in patients
after GTR, because there is no proven benet.
Role of radiotherapy after subtotal resection for
treatment of residual tumor. Schild and colleagues
[70] have reported that among patients who
underwent STR, the 5-year local control rate
was 100% for patients who received postoperative
RT and 50% for patients who did not (log rank
rest of Kaplan-Meier product limit method projection, P = 0.02). The 5-year survival rate after
STR also showed a trend for longer survival in
patients who received postoperative RT (88%)
compared with patients who did not (71%) (logrank rest of Kaplan-Meier product limit method
projection, P = 0.3). Brown and colleagues [9]
also found that crude local control rates for STR
versus STR/RT (48.661.2 Gy in 180200 cGy
fractions or 59-Gy median dose) were 62% and
100% (two-tailed test, P = 0.0008), respectively.
In a larger study, Rades and Fehlauer [112] have
reported a 3-year local control rate of 55%
compared with 89% among patients who underwent STR and STR/RT, respectively. At 5
years, local control rates were 46% for STR and
83% for STR/RT (log rank rest of Kaplan-Meier
projection, P<0.001). Median time to progression
was 20 months after STR and 34 months after
STR/RT; however, 5-year survival rates were 86%
for STR and 90% for STR/RT and showed no
signicant dierence between groups. Overall

survival was good in both groups, and postoperative radiation improved local control but
not survival. Rades and Fehlauer [112] argue that
the benet seen in local control is enough to
warrant the use of postoperative RT in all patients
undergoing STR. Schild and colleagues [70] oer
an alternative approach and suggest that observation until progression may spare approximately
50% of patients with STR related potential
radiation toxicity. Kim and colleagues [40] have
reported that at their institution, because of the
benign clinical course of CNC and potential for
delayed radiation toxicity, routine postoperative
radiation is not given even after STR.
We concur with many clinicians who believe
that adjunctive RT should not be routinely
administered after STR of CNC [5,11,24,27,
41,46,57,58,61,64,71,90,93]. The decision to use
adjunctive RT for residual tumor can also be
based on the MIB-1 LI. Because certain aspects of
histologic atypia do not necessarily correlate with
clinical outcome [39,40,99], the MIB-1 LI currently seems to be the best predictor of proliferative potential. Hence, the MIB-1 LI may help
to predict the clinical outcome of CNC [3,77,110].
The time to progression may be prolonged in
tumors with a high MIB-1 LI treated with
adjuvant RT [7].
Prophylactic radiation therapy. If RT is pursued,
most clinicians agree that only the tumor bed and
directly adjacent areas should be included in the
treatment eld [9,70]. Prophylactic radiation of the
spinal cord and total brain has been reported in
earlier cases [20,27,40,42,44,56,70,78]; however,
the rationale for prophylactic RT was either not
stated, based on a diagnosis that was later changed
[40,42], or the result of a diagnosis of atypical
neurocytoma [20,78]. At this time, prophylactic
RT is not indicated for CNC because of the benign
nature of CNC and risk of radiation toxicity [112].
Radiation dosing. The experience with radiation
dosing is varied. Rades and colleagues [115]
studied the optimal radiation dose in a retrospective analysis of patients receiving conventional RT
after STR. Their analysis included 52% of the
reported cases and 11 unpublished cases from
their institutions with a minimum follow-up of 12
months. The cases were divided into group A (40
53.6 Gy [n = 42]) and group B (54.062.2 Gy
[n = 47]), with a 10-year local control rate of
65% for group A and 89% for group B (log rank
497
rest of Kaplan-Meier projection, P = 0.0066).

The median time to recurrence was 26 months in
group A and 96 months in group B, although
a signicant dierence in the 10-year survival rate
was not found. Therefore, the optimal minimal
dose with high chance of local control is the
equivalent dose in 2-Gy fractions of 54 Gy or
greater. This dose seems to result in better local
control and longer time to recurrence. Late
toxicity after radiation is tolerable if the total
radiation does not exceed 60 Gy [112]. The risk of
severe toxicity at 5 years is 5% if up to one third
of the brain receives 60 Gy. Therefore, using
a target volume including only the preoperative
tumor and a 2-cm margin is recommended [115].
The optimal dose for RT for patients with CNC
seems to be 54 to 60 Gy.
Radiation therapy after recurrence or progression. RT has been used with success after tumor
progression or recurrence. Cases of RT after
recurrence are often part of larger case series and
are often combined with other therapy. Yasargil
and colleagues [96] have reported a case of
asymptomatic recurrence 38 months after initial
GTR that was treated with GTR/RT with recurrence (50 Gy, dosing schedule not specied).
Follow-up at 37 months after completing RT
showed no recurrence. Valdueza and colleagues
[89] have reported a case of recurrence noted 18
months after STR treated with GTR/RT (dose
not specied). Follow-up 4 years later showed no
tumor recurrence. Dodds and colleagues [14] have
reported a case of GTR/RT (54 Gy in 30 fractions
over 6 weeks using a shrinking eld technique
with 5-MeV X-rays) following recurrence after
chemotherapy. No radiation toxicity was noted,
and no recurrence was present on CT 5 years after
completing RT. Kim and colleagues [40] have
reported a case of recurrence 21 months after
GTR. The patient began RT and received 14 Gy
(dose schedule not specied) before refusing to
complete the course. This patient was lost to
follow-up until 10 years later, when he underwent
reoperation for tumor progression. Schild and
colleagues [70] have reported two patients who
received RT (50.4 Gy in 28 fractions to the
primary tumor bed, 30 Gy in 15 fractions to the
craniospinal axis, with boost to total dose of 60
Gy in 30 fractions to the tumor bed) following
recurrence after STR. Brandes and colleagues [7]
have reported one case treated with chemotherapy
and RT (39.6 Gy in 22 fractions to the craniospinal axis plus a boost of 14.4 Gy in 8 fractions on
498
T8) following recurrence 3 years after GTR. No

radiation toxicity was noted, and the patient was
in complete remission at the 36-month follow-up
examination.
Not all cases of RT for recurrent or progressive
CNC have shown a positive outcome. Mineura
and colleagues [53] have reported a case of RT
with 60 Gy (dosing schedule not specied)
following progression 4 months after STR. At
the 3-month follow-up examination, the tumor
showed continued progression, but subsequent
follow-up data were not reported. Ashkan and
colleagues [3] have reported a case of asymptomatic progression 29 months after STR treated with
55 Gy (dosing schedule not specied) for a tumor
with an MIB-1 LI of 11.2%, with no further
follow-up data reported. Sgouros and colleagues
[71] describe two cases of RT (dose and schedule
not specied) following recurrence 12 months and
9 months after STR, which resulted in death at 5
years and 10 months later, respectively. The rst
patient responded well initially, but the tumor
continued to progress; the second patient deteriorated while receiving RT, so it was discontinued. Only one of these cases reported the MIB1 LI; therefore, the true characterization of these
tumors is unclear. More classic CNC tumors seem
to show a positive response to RT for recurrent or
progressive CNC. The determination of whether
RT is benecial in treating atypical neurocytomas
requires further study, however.
Stereotactic radiosurgery
GKS or LINAC stereotactic radiosurgery can
deliver a high dose of radiation with a steep drop
o, thereby limiting the unnecessary exposure of
surrounding brain tissue, which results in a decreased risk of side eects. The goals of radiosurgery are long-term local control, maintenance
of neurologic function, and prevention of new
neurologic decits [6,109]. Although not the goal,
tumor volume reduction often occurs. The eects
of GKS and LINAC radiosurgery on tumor
control are thought to be similar [36,110].
CNCs seem to be ideal targets for stereotactic
radiosurgery [6,36,58]. Depending on their size
and location, CNCs are often surrounded by CSF
and are well demarcated from the adjacent brain.
Patients undergoing stereotactic radiosurgery
typically are discharged the same day of the procedure and experience few side eects [2,6,36,65].
The experience of treatment of CNCs with stereotactic radiosurgery is limited, but early reports are
encouraging (Table 6) [2,6,11,36,65].
Gamma knife surgery

As primary therapy. Tyler-Kabara and colleagues [88] have reported two cases of GKS as
primary therapy (two 14-mm isocenters and two
8-mm isocenters, 15-Gy marginal dose and 30-Gy
maximum dose; one 8-mm isocenter and one 4mm isocenter, 20-Gy marginal dose and 40-Gy
maximum dose) after biopsy. Both patients are
well 40 and 42 months, respectively, after the
procedure, with a reduction in tumor size noted on
follow-up imaging. Javedan and colleagues [108]
have reported a case of third ventricular CNC
treated with GKS (18 Gy at the 50% isodose line,
ve isocenters) after endoscopic biopsy and third
ventriculostomy. No adverse side eects were
reported, and the patient was neurologically intact
with minimal shrinkage of the tumor on imaging
25 months after the procedure. Further study
and longer term follow-up data are needed to
evaluate GKS as a primary treatment option after
biopsy.
For residual tumor (adjuvant). Cobery and
colleagues [11] have reported three cases of GKS
(323 isocenters, 30%50% isodose level, 9- to 13Gy peripheral dose) as adjuvant therapy for
residual tumor after STR. All three patients
tolerated the procedure well and demonstrated
a reduction in tumor size (48%, 72%, and 81%)
on follow-up imaging at 99, 23, and 42 months
after the procedure, respectively. Hara and
colleagues [26] have reported a case of GKS (20Gy marginal dose and 40-Gy maximum dose,
50% isodense gradient) after STR with a marked
decrease in tumor size 2 months after GKS, which
remained unchanged after 1 year. The response to
GKS as adjuvant therapy is encouraging. As with
conventional RT, the MIB-1 LI should play a role
in the timing of therapy for residual tumor. The
robust eect on tumor volume coupled with a low
risk of radiation toxicity makes GKS an attractive
alternative to conventional RT, however.
For recurrence. Bertalany and colleagues [6]
have reported three cases of GKS (327 isocenters,
30%60% isodose line, 9.6- to 16-Gy marginal
dose) for asymptomatic tumor recurrence 5 to 6
years after GTR. The MIB-1 LIs were 2.4%, 7%,
and 8.7%, respectively. No patient developed new
neurologic decits after GKS. In all three cases,
follow-up imaging (1, 2, and 5 years, respectively)
showed a reduction in tumor volume.
Anderson and colleagues [2] have reported four
cases treated using GKS (multiple isocenters with
combinations of 8-, 14-, and 18-mm collimators,
1620 Gy to target area) for evidence of recurrence
Table 6
Stereotactic radiosurgery and outcome
Recurrence
in months
Radiation
GTR
GKS, 16 Gy
GTR
25
GKS, 16 Gy
GTR
21
GKS 20, Gy
GTR
14
GKS, 16 Gy
GTR 2
6072
GKS, 12.5 Gy
Timing of
radiation
Salvage at 9
months
Reoperation with
GKS salvage
at 25 months
Reoperation with
GKS salvage
at 21 months
Salvage at 14
months
Salvage at 5 to 6
years
Initial size
in cc cubic
centimeters
Follow-up
MRI
Follow-up
in months
Decrease in
tumor size
Decrease in
tumor size
14 p GKS, 24
p GTR
28 p GKS, 83
p GTR
1.7
Decrease in
tumor size
12 p GKS, 84
p GTR
7.9
Decrease in
tumor size
58% decrease in
tumor size after
GKS
27 p GKS, 27
p GTR
12 p GKS
40% decrease
in tumor size
after GKS
61% decrease
in tumor size
after GKS
48% decrease
in tumor size
72% decrease in
post op tumor
81% decrease
in tumor size
77% decrease
in tumor size
60 p GKS
6.2
12.3
0.6
GTR
GKS, 13 Gy
5.2
GTR 2
GKS, 9.616
Gy
5.9
STR/GKS
GKS, 9 Gy
STR/GKS
GKS, 13 Gy
STR/GKS
GKS, 10 Gy
GTR
53
GKS, 10 Gy
Adjuvant 8
months later
Adjuvant 18
months later
Adjuvant
Salvage at 53
months
6.5
13
29
10.5
Outcome
Reference
4/4 returned to work

with full fxn
3/4 neurologically nl, 1/4
on dilantin for
postoperative seizure
1/4 on dilantin for post
operative seizure
[116]
Died us a result of
cardiac failure caused
by pericarditis 1 year
p GKS, 2/3 returned
to work, 3/3 no new
neurologic problems,
1/3 had persistent
abducens palsy and
visual impairment
[6]
99 p GKS
Neurologically intact
[11]
23 p GKS
Asymptomatic
42 p GKS
Asymptomatic
12 p GKS
Asymptomatic
24 p GKS
Primary
treatment
499
GKS, 20 Gy
Adjuvant
5.7
STR/GKS
GKS, dose
not stated
Adjuvant
Not stated
STR
Biopsy/GKS
Not
GKS, 15 Gy
GKS, 15 Gy
Salvage
Primary
Not stated
4.2
STR 2/
GKS
Biopsy/GKS
GKS, 14 Gy
Adjuvant
7.9
GKS, 18 Gy
Adjuvant
Not stated
GTR/
radiosurgery
72
Radiosurgery,
50 Gy
Adjuvant
Not stated
GTR
Radiosurgery
Salvage at 8
months
1.2
GTR
36
Salvage at 3 years
2.7
STR
12
Radiosurgery
18 Gy
LINAC, 17.5
Gy
Salvage after 6
months of
observation
Not stated
STR/RT
144,
36, 36
30 Gy at
presentation/
50 Gy at
recurrence 2/
LINAC at
recurrence 3
Adjuvant and
salvage
Not stated
2, 4, 6, 8, 10, and 12
months after GKS,
tumor shrank at
2 months
Shrinkage of tumor
at 21 months
12
[26]
21 p GKS
[35]
Stable disease
Signicant decrease
in tumor size at 40
months
Complete regression
at 18 months
Minimal decrease in
tumor size at 25
months
Complete remission
p initial Tx, 6
years later
recurrence
6 months p
radiosurgery
13
50 p GKS
No progression,
shrinkage of tumor,
KPS full
Stable disease
No new neurologic
problems
Signicantly smaller
tumor p
radiosurgery
Decrease in tumor
size at 6 months p
LINAC,
disappeared by 36
months p LINAC
Not stated
53 p GKS
25 p GKS
87 mo p
radiosurgery
No new clinical
symptoms
Normal neurologic
examination
[70]
[89]
[8]
Stable disease no
progression (received
chemo 72 months p
radiosurgery)
KPS 100
[7]
[65]
51 p LINAC
[36]
60 p LINAC,
240 p initial
ITR
Alive with disease
[30]
10 p
radiosurgery,
18 p GTR
34 p
radiosurgery
[40]
Abbreviations: chemo, chemotherapy; fxn, function; GTR, gross total resection; KPS, Karnofsky performance scale; LINAC, linear accelerator; nl, normal; p, post; RT,
radiation therapy; STR, subtotal resection; Tx, therapy.
500
STR/GKS
on MRI 9 to 14 months after GTR and reoperation

for recurrence. At the 12- to 28-month follow-up
examinations, all tumors demonstrated a reduction
in tumor size and the patients have all returned to
their previous employment with full function.
Tyler-Kabara and colleagues [88] also have reported two patients treated with GKS (ve 14-mm
isocenters, 14-Gy marginal dose and 28-Gy maximum dose; three 8-mm and two 4-mm isocenters,
16-Gy marginal dose and 32-Gy maximum dose)
following recurrence after surgery. Both patients
remain well, and follow-up imaging demonstrates
complete regression at 53 months and marked
tumor regression at 38 months, respectively, after
the procedure.
An additional three cases have been reported
in the literature. One patient treated with GKS
(nine isocenters, 18-Gy marginal dose and 36-Gy
maximum dose) following asymptomatic recurrence 3 years after GTR experienced no adverse
side eects and was neurologically intact at the 34month follow-up examination with a decrease in
tumor size noted on MRI [65]. Another patient
treated with GKS (29 isocenters, 50% isodose
level, 10-Gy peripheral dose) for asymptomatic
recurrence 53 months after GTR remained
asymptomatic following the procedure, and
follow-up imaging at 12 months showed a 77%
reduction in tumor volume [11]. The third patient
was treated with GKS (15 Gy in 1 fraction [dose
not otherwise specied]) for tumor progression
after STR, and follow-up imaging at 13 months
showed stable disease [70].
Overall, the reported response of CNC to GKS
is favorable, with 15 of 16 cases demonstrating
a reduction in tumor size and 1 of 16 cases
showing stabilization of disease. Most patients
tolerated GKS without complication and were
discharged the next day. These reported cases
describe GKS as the primary therapy and
treatment of residual tumor or asymptomatic
tumor recurrence identied on neuroimaging.
Because the follow-up period after GKS is limited,
future study is needed to demonstrate the longterm ecacy of this treatment modality.
Linear accelerator. Treatment of CNC using
LINAC radiosurgery is limited to three cases in
the literature. Maruyama and colleagues [48] have
reported a patient with CNC treated with LINAC
radiosurgery (10-MV photons with multiple-arc
noncoplanar method, 50% isodose line, 24 Gy to
central target) after STR. Radiographic follow-up
at 6 months showed no change in tumor size. Kim
501
and colleagues [36] have reported a patient with

CNC who showed residual tumor on MRI 6
months after resection. Because the patient was
asymptomatic at the time and the residual tumor
was small, the patient was observed for 6 months.
After tumor progression on follow-up imaging,
LINAC radiosurgery (three 2-cm isocenters, 70%
isodose line, 17.5-Gy marginal dose and 25.0-Gy
maximum dose) was initiated. Imaging at 6
months after LINAC radiosurgery showed a decrease in tumor size and complete disappearance
by 36 months. At 51 months of follow-up, the
patient was neurologically intact with no signs of
recurrence on MRI. Ishiuchi and Tamura [30] have
reported a patient treated with LINAC radiosurgery (dose not specied) following ventricular
dissemination after repeated surgery and RT. At 5
years after LINAC radiosurgery, this patient was
alive with disease. Although the experience with
LINAC radiosurgery in the treatment of residual
or progressive CNC is limited, these three cases
demonstrate a favorable response.
CyberKnife. The frameless image-guided stereotactic radiosurgery used with the CyberKnife
(Accuray, Sunnyvale, California) is a recently
developed technology that may be used in the
future to treat CNC. This device couples an Xband LINAC to a computer-controlled robotic
arm that aligns the radiation beams with the
target based on frequent input from the radiographic tracking system. Paired orthogonal highresolution digital images are coregistered with
digitally reconstructed radiographs from preoperative CT images to provide spatial positioning
with six degrees of freedom. The CyberKnife
automatically adjusts for patient movements
of up to 1 cm. In phantom testing, the secondgeneration CyberKnife system demonstrates high
accuracy, with a spatial error of 1.1 0.3 mm,
which is comparable to frame-based radiosurgical
systems, such as GKS and LINAC systems
[116,117]. This technology is untested in the
treatment of CNC to date; however, the successful
use of the CyberKnife in treating other benign
tumors [110] suggests a possible role for this
technology in treating CNC.
Chemotherapy
Chemotherapy is an appealing alternative to
radiation for CNC, because these patients tend to
be younger and, as a result, more susceptible to the
long-term side eects of radiation. Incorporation of chemotherapy into treatment strategies
502
involves several challenges: (1) inherent or acquired mechanisms of resistance, (2) eective drug
delivery, and (3) altered drug metabolism caused
by interactions with anticonvulsants and steroids.
Chemotherapeutic regimens often involve combining drugs with dierent mechanisms of action
and nonoverlapping toxicities. Some chemotherapeutic agents may also cause radiosensitization.
In general, histopathologic grade plays a key prog-
nostic role in choosing a chemotherapy regimen

[21]. The goal of adjuvant chemotherapy is to
consolidate tumor reduction after surgery and
decrease the probability of recurrence [21].
Limiting systemic toxicity, such as myelosuppression and end organ and tissue damage, is also
important [7,21].
The experience with CNC and response to
chemotherapy is limited and is often used as a part
Table 7
Chemotherapeutic regimens
Primary
treatment
Chemo
Recurrence in
months
Follow-up
in months
after chemo
Outcome
Adjuvant salvage etoposide

(40 mg/m2/d days 14),
cisplatin (25 mg/m2/d days
14), cyclophosphamide
(1000 mg/m2/d day 4), ve cycles
Same as above, ve cycles
72
15
Stable disease maximum

result 8 months after
chemo, no progression
60
18
Same as above, three cycles
38
36
22
96
Stereotactic Bx/RT/
chemo
Adjuvant four cycles of

carboplatin (500 mg/m2 days
1 and 2 of week 1), etoposide
(100 mg/m2 days 13 of weeks
1 and 3, and ifosfamide
(3 g (m2 days 13 of week 3)
Adjuvant loumustine,
seven dosesa
Stable disease, no
progression
Complete remission for 36
months
Full-time employment, no
progression
60
Stereotactic Bx/RT/
chemo
Stereotactic Bx/RT/
chemo
nine dosesb
seven dosesa
15
15
108
Stereotactic Bx/RT/
chemo
Stereotactic Bx/RT/
chemo
Stereotactic Bx/RT/
chemo
STR/VPS/chemo/RT
STR/VPS/chemo/RT
2 GTR/RT/chemo
eight dosesb
nine dosesa
seven dosesa
Adjuvant cytoxan, cisplatinb
Adjuvant cytoxan, cisplatinb
Adjucant cisplatin plus
lomustineb
Adjuvant lomustine aloneb
Adjuvant lomustine plus
carmustineb
Adjuvant vincristine,
lomustine, prednisoneb
90
114
96
GTR/RT, stereotactic
Stereotactic
biopsy/RT
GTR
Shunt, cells for
cytology, STR/
chemo
2 STR/RT/chemo
Not
stated
14
11
Not
stated
Died because of shunt

dysfunction at 5 years
after initial RT
Lost to follow-up
Shunt surgery at 9 years
after initial RT,
employed, KPS >90
Asymptomatic, employed,
KPS >90
KPS >90
KPS >90
Alive and well
Alive and well
Not stated
Abbreviations: Bx, biopsy; chemo, chemotherapy; GTR, graps total resection; KPS, Kurnofsky performance scale,
RT, radiation therapy; STR, subtotal resection; VPS, ventroperitoneal shunt.
a
Schedule not specied.
b
Dose and schedule not specied.
of multimodality therapy (Table 7). Dodds and

colleagues [14] described a patient who received
primary chemotherapy for a tumor that later was
diagnosed as CNC. Because the tumor was largely
inoperable and some mitotic activity was present
in initial smear preparations, a trial of chemotherapy was started. Four cycles of carboplatin
(500 mg/m2 on days 1 and 2 of week 1), etoposide
(100 mg/m2 on days 13 of weeks 1 and 3), and
ifosfamide (3 g/m2 on days 13 of week 3) were
given before chemotherapy was stopped because
of decreasing renal function. Follow-up CT 1
month after completing chemotherapy showed
regression of the tumor, with a decrease in the
solid component and a corresponding increase in
the cystic component. This patient underwent
successful STR and RT 22 months later because
of progression. At last report, the patient was
back to neurologic baseline with stabilization of
disease. This case demonstrates a potential benet
of chemotherapy. Because chemotherapy was
used in combination with surgery and RT,
however, the individual contribution of chemotherapy cannot be assessed.
Brandes and colleagues [7] have reported three
cases in which patients received chemotherapy
(etoposide, 40 mg/m2 on days 14; cisplatin, 25 mg/
m2 on days 14; and cyclophosphamide, 1000 mg/
m2 on day 4 [repeated every 4 weeks]) for recurrent
or progressive CNC. Two patients showed recurrence on MRI 5 to 6 years after STR and RT.
These patients received ve cycles of chemotherapy
and achieved a 40% to 60% reduction in tumor
size. One patient showed recurrence on MRI 3
years after GTR. This patient received three cycles
of chemotherapy and later received RT, resulting
in complete regression of the tumor. At last followup, all three patients had maintained their clinical
response over 15 to 36 months.
Schild and colleagues [70] also described four
patients who received chemotherapy after STR or
GTR followed by RT. Chemotherapy regimens
(doses not specied) included lomustine alone;
cisplatin plus lomustine; lomustine plus carmustine; and vincristine, lomustine, and prednisone.
Kulkarni and colleagues [42] have reported six
patients who received chemotherapy with lomustine (dose not specied) after stereotactic biopsy
and RT as the result of an initial diagnosis of
oligodendrocytoma. Patient-specic outcome data
were not reported, however, and the contribution
of chemotherapy alone could not be meaningfully
assessed. In these two studies combined, all patients
maintained local control as documented by follow-
503
up CT or MRI, except for one patient who showed

subependymal spread [42,70]. Sgouros and colleagues [71] have reported only a temporary reduction in tumor size associated with carboplatin
(dose not specied) in one patient, however.
Although these reported cases suggest the
potential benet of chemotherapy in the treatment
of CNC, surgery and RT have shown a proven
benet in larger cohorts of patients. If RT and
surgery are not appropriate or possible, chemotherapy in the setting of clinically aggressive
behavior or a high proliferation index may be
considered [14]. Of note, none of the reports
describing chemotherapy for patients with CNC
used the MIB-1 LI to assess proliferation potential before treatment. The limited data concerning
long-term prognosis after chemotherapy may
warrant further study.
Follow-up
There is no established standard for follow-up
of CNC, but most clinicians include components
of postoperative imaging to determine if there is
residual tumor, serial neuroimaging at variable
intervals, and regular clinical examinations. In the
sole reported prospective follow-up study, patients were clinically evaluated three times per
month initially and twice yearly thereafter [3].
Some studies suggest a yearly clinical examination
[11,96] or a clinical examination at 6 months after
surgery [15], whereas others suggest a regular
clinical examination without specifying a time
period [14,88]. Most authors recommend serial
imaging; however, the interval varies from yearly
[6,10,11], to biennially [57], to twice a year initially
[59] and then gradually building up to once every
other year, to periodically with the interval unspecied. Evaluation of postoperative outcome
also varies (Table 8); however, the use of a validated scale, such as the Karnofsky performance
scale [118], should be a goal of all practitioners.
Other factors, such as type of treatment, high
MIB-1 LI, presence of a VPS, and older patient
age, may modify the follow-up intervals. When
patients are left with residual tumor after STR or
biopsy, more frequent initial follow-up may be
warranted. In tumors with a high MIB-1 LI,
closer follow-up should be instituted because of
the higher likelihood of recurrence [78]. Patients
with a VPS may also need closer follow-up
because of the high failure rate over time
[13,38,92,119]. Several deaths as a result of shunt
failure have been reported in patients with CNC
504
Table 8
Outcome measures
Outcome measure
No. patients
References
Survival
181
Shunt failure
Local control (recurrence,
progression, stable disease)
Change in neurologic examination
Functional descriptiona
5
153
[17, 10, 11, 14, 15, 18, 19, 20, 23, 26, 27, 31, 35, 36, 4046, 53, 56
58, 60, 6265, 68, 70, 74, 78, 82, 85, 8789, 9496, 99, 104, 108]
[27, 42, 70]
[24, 6, 7, 10, 11, 14, 15, 18, 19, 20, 26, 27, 30, 31, 35, 36, 40, 42, 45,
53, 57, 58, 60, 62, 6365, 68, 70, 78, 82, 85, 8789, 9496, 99, 104]
[1, 2, 5, 6, 11, 15, 23, 25, 36, 46, 57, 65, 68, 88, 96, 99]
[1, 2, 5, 6, 10, 14, 15, 23, 35, 4042, 44, 45, 56, 58, 60, 63, 68, 96, 99]
26
98
Includes Karnofsky performance scale score, ADL, well, employment status.
with a VPS [42,70]. A reasonable postoperative

management strategy may include immediate
postoperative imaging to conrm GTR or STR,
followed by repeat imaging between 3 and 6
months and then yearly for at least 5 years. With
no tumor recurrence, longer intervals between
follow-up imaging may be considered. The goal of
follow-up should be early identication of recurrent or progressive CNC.
Given that most patients present with signs
and symptoms of noncommunicating hydrocephalus, long-term treatment of hydrocephalus is an
important consideration. Surgery to debulk the
tumor is often sucient for relief of outlet ow
obstruction. Some patients may still require shunt
placement as a part of the postoperative treatment of hydrocephalus, however. A review of
the literature shows that 27 of 303 patients
underwent VPS placement (type not specied) as
part of the management of CNC. For those cases
reporting the timing of CSF diversion, shunt
placement usually occurred either at surgery or
within the 4-month postoperative period (17 of 20
cases). Shunt-related complications included infection (2 of 27 cases) or death (5 of 27 cases).
Third ventriculostomy to treat noncommunicating hydrocephalus has been described in only
1 patient [108].
Management paradigm
Based on the literature, CNC usually follows
a clinically benign course, with most symptoms
caused by increased ICP as a result of noncommunicating hydrocephalus. In general, appropriate management of patients with CNC results
in a favorable clinical outcome; however, more
aggressive variants have also been reported. The
clinical treatment and outcome for reported cases
of CNC for which detailed information was
available.
Initial treatment should be based on symptoms

rather than on incidental ndings on neuroimaging. Emergent treatment and evaluation of noncommunicating hydrocephalus should include
CT imaging and standard therapy, such as intravenous steroids, hyperosmolar uids, and extraventricular drain (EVD) when indicated. Nonemergent cases may also undergo initial CT
imaging. MRI is necessary to characterize the
tumor better and to localize the attachment within
the ventricle.
In all cases, biopsy is necessary to establish
a denitive diagnosis. The MIB-1 LI should also
be determined so as to establish the risk of
progression or recurrence. Large tumors are best
treated with surgery to debulk the tumor and
attempt GTR. Smaller tumors may be treated
with biopsy followed by radiosurgery or conventional RT if the MIB-1 LI is not greater than 2%.
With an MIB-1 LI greater than 2% or with severe
neurologic symptoms, surgery is recommended.
After radiosurgery or RT, patients should be
evaluated with neuroimaging at least every 6
months for the rst 2 years and should also have
regular clinical examinations to monitor for
recurrence and assess the ecacy of treatment.
For patients undergoing surgery for large
tumors or tumors with a high proliferation index,
complete tumor resection should be the goal when
possible. Immediate postoperative imaging should
be performed to evaluate the extent of tumor
resection. Tumors with a high MIB-1 LI should
also receive adjuvant radiosurgery or RT. After
these treatments, patients should be clinically and
radiologically followed for recurrence or progression. Patients with low MIB-1 LI tumors
should also be observed for recurrence or progression. The follow-up interval depends on the
presence of a VPS, patient age, and amount of
residual tumor. If symptomatic recurrence or
progression does occur, patients may be treated
with radiosurgery, RT, or reoperation. If malignant behavior is suspected, reoperation is favored.

Otherwise, RT and radiosurgery seem to have less
associated risk and morbidity. Clinical and
radiologic follow-up for symptomatic recurrence
should resume after treatment.
Summary
The literature to date on the treatment of CNC
reects an evolution of clinical practice in neurooncology. The advent of sophisticated tools, such
as MRS and molecular pathology, has facilitated
more ecient diagnosis of CNC. Decreased
morbidity associated with surgical intervention
has resulted in better outcomes in patients undergoing resection of CNC. Prospective monitoring of treated patients with MRI coupled with
judicious use of radiosurgery will likely further
decrease treatment-related morbidity.
References
[1] Agranovich AL, Ang LC, Fryer CJ. Central
neurocytoma: report of 2 cases and literature
review. J Neurooncol 1993;16:4753.
[2] Anderson RC, Elder JB, Parsa AT, Issacson SR,
Sisti MB. Radiosurgery for the treatment of
recurrent central neurocytomas. Neurosurgery
2001;48:12318.
[3] Ashkan K, Casey AT, DArrigo C, Harkness WF,
Thomas DG. Benign central neurocytoma. Cancer
2000;89:111120.
[4] Balko MG, Schultz DL. Sudden death due to
a central neurocytoma. Am J Forensic Med Pathol
1999;20:1803.
[5] Barbosa MD, Balsitis M, Jaspan T, Lowe J.
Intraventricular neurocytoma: a clinical and pathological study of three cases and review of the
literature. Neurosurgery 1990;26:104554.
[6] Bertalany A, Roessler K, Dietrich W, Aichholzer
M, Prayer D, Ertl A, et al. Gamma knife radiosurgery of recurrent central neurocytomas: a preliminary report. J Neurol Neurosurg Psychiatry
2001;70:48993.
[7] Brandes AA, Amist P, Gardiman M, Volpin L,
Danieli D, Guglielmi B, et al. Chemotherapy in
patients with recurrent and progressive central
neurocytoma. Cancer 2000;88:16974.
[8] Brat DJ, Scheithauer BW, Eberhart CG, Burger
PC. Extraventricular neurocytomas: pathologic
features and clinical outcome. Am J Surg Pathol
2001;25:125260.
[9] Brown DM, Karlovits S, Lee LH, Kim K, Rothfus
WE, Brown HG. Management of neurocytomas:
case report and review of the literature. Am J Clin
Oncol 2001;24:2728.
505
[10] Christov C, Adle-Biassette H, Le Guerinel C.

Recurrent central neurocytoma with marked increase in MIB-1 labelling index. Br J Neurosurg
1999;13:4969.
[11] Cobery ST, Noren G, Friehs GM, Chougule P,
Zheng Z, Epstein MH, et al. Gamma knife surgery
for treatment of central neurocytomas. Report of
four cases. J Neurosurg 2001;94:32730.
[12] Coca S, Moreno M, Martos JA, Rodriguez J,
Barcena A, Vaquero J. Neurocytoma of spinal
cord. Acta Neuropathol (Berl) 1994;87:53740.
[13] Cozzens JW, Chandler JP. Increased risk of distal
ventriculoperitoneal shunt obstruction associated
with slit valves or distal slits in the peritoneal
catheter. J Neurosurg 1997;87:6826.
[14] Dodds D, Nonis J, Mehta M, Rampling R. Central
neurocytoma: a clinical study of response to
chemotherapy. J Neurooncol 1997;34:27983.
[15] Dodero F, Alliez JR, Metellus P, Hassan H,
Hassoun J, Alliez B. Central neurocytoma: 2 case
reports and review of the literature. Acta Neurochir (Wien) 2000;142:141722.
[16] Elek G, Slowik F, Eross L, Toth S, Szabo Z, Balint
K. Central neurocytoma with malignant course.
Neuronal and glial dierentiation and craniospinal dissemination. Pathol Oncol Res 1999;5:
1559.
[17] Enam SA, Rosenblum ML, Ho KL. Neurocytoma
in the cerebellum. Case report. J Neurosurg
1997;87:1002.
[18] Eng DY, DeMonte F, Ginsberg L, Fuller GN,
Jaeckle K. Craniospinal dissemination of central
neurocytoma. Report of two cases. J Neurosurg
1997;86:54752.
[19] Figarella-Branger D, Pellissier JF, Daumas-Duport C, Delisle MB, Pasquier B, Parent M, et al.
Central neurocytomas. Critical evaluation of
a small-cell neuronal tumor. Am J Surg Pathol
1992;16:97109.
[20] Fujimaki T, Matsuno A, Sasaki T, Toyoda T,
Matsuura R, Ogai M, et al. Proliferative activity of
central neurocytoma: measurement of tumor
volume doubling time, MIB-1 staining index and
bromodeoxyuridine labeling index. J Neurooncol
1997;32:1039.
[21] Galanis E, Buckner JC. Chemotherapy of brain
tumors. Curr Opin Neurol 2000;13:61925.
[22] Giangaspero F, Cenacchi G, Losi L, Cerasoli S,
Bisceglia M, Burger PC. Extraventricular neoplasms with neurocytoma features. A clinicopathological study of 11 cases. Am J Surg Pathol
1997;21:20612.
[23] Goergen SK, Gonzales MF, McLean CA. Interventricular neurocytoma: radiologic features
and review of the literature. Radiology 1992;182:
78792.
[24] Hamilton R. Case of the month. August
1996frontal lobe tumor in 11 year old girl. Brain
Pathol 1997;7:7134.
506
[25] Hara A, Araki Y, Shinoda J, Hirayama H,

Niikawa S, Sakai N, et al. Central neurocytoma:
proliferative assessment by nucleolar organizer
region staining. Surg Neurol 1993;39:3437.
[26] Hara M, Aoyagi M, Yamamoto M, Maehara T,
Takada Y, Nojiri T, et al. Rapid shrinkage of
remnant central neurocytoma after gamma knife
radiosurgery: a case report. J Neurooncol 2003;62:
26973.
[27] Hassoun J, Gambarelli D, Grisoli F, Pellet W,
Salamon G, Pellissier JF, et al. Central neurocytoma. An electron-microscopic study of two
cases. Acta Neuropathol (Berl) 1982;56:1516.
[28] Hassoun J, Soylemezoglu F, Gambarelli D,
Figarella-Branger D, von Ammon K, Kleihues P.
Central neurocytoma: a synopsis of clinical and
histological features. Brain Pathol 1993;3:297306.
[29] Hsu PW, Hsieh TC, Chang CN, Lin TK. Fourth
ventricle central neurocytoma: case report. Neurosurgery 2002;50:13657.
[30] Ishiuchi S, Tamura M. Central neurocytoma: an
immunohistochemical, ultrastructural and cell
culture study. Acta Neuropathol (Berl) 1997;94:
42535.
[31] Jamshidi J, Izumoto S, Yoshimine T, Maruno M.
Central neurocytoma presenting with intratumoral
hemorrhage. Neurosurg Rev 2001;24:4852.
[32] Jay V, Edwards V, Hoving E, Rutka J, Becker L,
Zielenska M, et al. Central neurocytoma: morphological, ow cytometric, polymerase chain reaction, uorescence in situ hybridization, and
karyotypic analyses. Case report. J Neurosurg
1999;90:34854.
[33] Jayasundar R, Shah T, Vaishya S, Singh VP,
Sarkar C. In vivo and in vitro MR spectroscopic
prole of central neurocytomas. J Magn Reson
Imaging 2003;17:25660.
[34] Jerdan M, White C, Solomon D, Scheithauer BW,
Clark A. Dierentiated cerebral neuroblastoma in
adults. J Neuropathol Exp Neurol 1983;42:305.
[35] Kanamori M, Kumabe T, Shimizu H, Yoshimoto
T. (201)T1-SPECT, (1)H-MRS, and MIB-1 labeling index of central neurocytomas: three case
reports. Acta Neurochir (Wien) 2002;144:15763.
[36] Kim CY, Paek SH, Kim DG. Linear accelerator
radiosurgery for central neurocytoma: a case report. J Neurooncol 2003;61:24954.
[37] Kim DG, Chi JG, Park SH, Chang KH, Lee SH,
Jung HW, et al. Intraventricular neurocytoma:
clinicopathological analysis of seven cases. J Neurosurg 1992;76:75965.
[38] Kim DG, Choe WJ, Chang KH, Song IC, Han
MH, Jung HW, et al. In vivo proton magnetic
resonance spectroscopy of central neurocytomas.
Neurosurgery 2000;46:32933.
[39] Kim DG, Kim JS, Chi JG, Park SH, Jung HW,
Choi KS, et al. Central neurocytoma: proliferative
potential and biological behavior. J Neurosurg
1996;84:7427.
[40] Kim DG, Paek SH, Kim IH, Chi JG, Jung HW,
Han DH, et al. Central neurocytoma: the role of
radiation therapy and long term outcome. Cancer
1997;79:19952002.
[41] Kubota T, Hayashi M, Kawano H, Kabuto M,
Sato K, Ishise J, et al. Central neurocytoma:
immunohistochemical and ultrastructural study.
Acta Neuropathol (Berl) 1991;81:41827.
[42] Kulkarni V, Rajshekhar V, Haran RP, Chandi
SM. Long-term outcome in patients with central
neurocytoma following stereotactic biopsy and
radiation therapy. Br J Neurosurg 2002;16:12632.
[43] Lee HY, Chuah KL, Chan LP. An intraventricular
brain tumour in a young woman. Part 1. Central
neurocytoma. Pathology 2002;34:1858.
[44] Louis DN, Swearingen B, Linggood RM, Dickersin GR, Kretschmar C, Bhan AK, et al. Central
nervous system neurocytoma and neuroblastoma
in adultsreport of eight cases. J Neurooncol
1990;9:2318.
[45] Mackenzie IR. Central neurocytoma: histologic
atypia, proliferation potential, and clinical outcome. Cancer 1999;85:160610.
[46] Maiuri F, Spaziante R, De Caro ML, Cappabianca P, Giamundo A, Iaconetta G. Central
neurocytoma: clinico-pathological study of 5 cases
and review of the literature. Clin Neurol Neurosurg 1995;97:21928.
[47] Martin AJ, Sharr MM, Teddy PJ, Gardner BP,
Robinson SF. Neurocytoma of the thoracic spinal
cord. Acta Neurochir (Wien) 2002;144:8238.
[48] Maruyama I, Sadato N, Waki A, Tsuchida T,
Yoshida M, Fujibayashi Y, et al. Hyperacute
changes in glucose metabolism of brain tumors after
stereotactic radiosurgery: a PET study. J Nucl
Med 1999;40:108590.
[49] McConachie NS, Worthington BS, Cornford EJ,
Balsitis M, Kerslake RW, Jaspan T. Review
article: computed tomography and magnetic resonance in the diagnosis of intraventricular cerebral
masses. Br J Radiol 1994;67:22343.
[50] McCutchen TQ, Smith MT, Jenrette JM, Van
Tassel P, Patel SJ, Thomas CR Jr. Interparenchymal hemorrhagic neurocytoma: an atypical presentation of a rare CNS tumor. Med Pediatr Oncol
1999;32:4406.
[51] McKeever PE. Insights about brain tumors gained
through immunohistochemistry and in situ
hybridization of nuclear and phenotypic markers.
J Histochem Cytochem 1998;46:58594.
[52] Metcalf C, Mele EM, McAllister I. Neurocytoma
of the retina. Br J Ophthalmol 1993;77:3824.
[53] Mineura K, Sasajima T, Itoh Y, Sasajima H,
Kowada M, Tomura N, et al. Blood ow and
metabolism of central neurocytoma: a positron
emission tomography study. Cancer 1995;76:
122432.
[54] Moller-Hartmann W, Krings T, Brunn A, Korinth
M, Thron A. Proton magnetic resonance spectros-
[55]
[56]
[57]
[58]
[59]
[60]
[61]
[62]
[63]
[64]
[65]
[66]
[67]
[68]
[69]
[70]
[71]
copy of neurocytoma outside the ventricular

regioncase report and review of the literature.
Neuroradiology 2002;44:2304.
Mrak RE. Malignant neurocytic tumor. Hum
Pathol 1994;25:74752.
Nakagawa K, Aoki Y, Sakata K, Sasaki Y,
Matsutani M, Akanuma A. Radiation therapy of
well-dierentiated neuroblastoma and central neurocytoma. Cancer 1993;72:13505.
Namiki J, Nakatsukasa M, Murase I, Yamazaki
K. Central neurocytoma presenting with intratumoral hemorrhage 15 years after initial treatment by partial removal and irradiation. Neurol
Med Chir (Tokyo) 1998;38:27882.
Nishio S, Morioka T, Suzuki S, Fukui M.
Tumours around the foramen of Monro: clinical
and neuroimaging features and their dierential
diagnosis. J Clin Neurosci 2002;9:13741.
Nishio S, Takeshita I, Kaneko Y, Fukui M. Cerebral neurocytoma. A new subset of benign neuronal
tumors of the cerebrum. Cancer 1992;70:52937.
Nishio S, Tashima T, Takeshita I, Fukui M.
Intraventricular neurocytoma: clinicopathological
features of six cases. J Neurosurg 1988;68:66570.
Ojeda VJ, Jacobsen PF, Papadimitriou JM.
Primary cerebral neuroblastoma. Case report with
light microscopy, tissue culture and electron
microscopy study. Pathology 1980;12:26974.
Okamura A, Goto S, Sato K, Ushio Y. Central
neurocytoma with hemorrhagic onset. Surg Neurol
1995;43:2525.
Patil AA, McComb RD, Gelber B, McConnell J,
Sasse S. Intraventricular neurocytoma: a report of
two cases. Neurosurgery 1990;26:1404.
Pearl GS, Takei Y. Cerebellar neuroblastoma:
nosology as it relates to medulloblastoma. Cancer
1981;47:7729.
Pollock BE, Staord SL. Stereotactic radiosurgery
for recurrent central neurocytoma: case report.
Porter-Grenn LM, Silbergleit R, Stern HJ, Patel SC,
Mehta B, Sanders WP. Intraventricular primary
neuronal neoplasms: CT, MR, and angiographic
ndings. J Comput Assist Tomogr 1991;15:3658.
Rabinowicz AL, Abrey LE, Hinton DR, Couldwell WT. Cerebral neurocytoma: an unusual cause
of refractory epilepsy. Case report and review of
the literature. Epilepsia 1995;36:123740.
Ruppert J. Central neurocytoma: a case study.
J Neurosci Nurs 2002;34:2014.
Schild SE. Benign central neurocytoma: a double
misnomer? Cancer 2002;94:284.
Schild SE, Scheithauer BW, Haddock MG, Schi
D, Burger PC, Wong WW, et al. Central neurocytomas. Cancer 1997;79:7905.
Sgouros S, Carey M, Aluwihare N, Barber P,
Jackowski A. Central neurocytoma: a correlative
clinicopathologic and radiologic analysis. Surg
Neurol 1998;49:197204.
507
[72] Sgouros S, Jackowski A, Carey MP. Central

neurocytoma without intraventricular extension.
Surg Neurol 1994;42:3359.
[73] Sgouros S, Walsh AR, Barber P. Central neurocytoma of thalamic origin. Br J Neurosurg
1994;8:3736.
[74] Sharma MC, Rathore A, Karak AK, Sarkar C. A
study of proliferative markers in central neurocytoma. Pathology 1998;30:3559.
[75] Shimura T, Mori O, Kitamura T, Kobayashi S,
Sanno N, Teramoto A, et al. Central neurocytoma expressing characteristics of ependymal dierentiation: electron microscopic ndings
of two cases. Med Electron Microsc 2003;36:
98105.
[76] Smoker WR, Townsend JJ, Reichman MV.
Neurocytoma accompanied by intraventricular
hemorrhage: case report and literature review.
AJNR Am J Neuroradiol 1991;12:76570.
[77] Soontornniyomkij V, Schelper RI. Pontine neurocytoma. J Clin Pathol 1996;49:7645.
[78] Soylemezoglu F, Scheithauer BW, Esteve J,
Kleihues P. Atypical central neurocytoma. J
Neuropathol Exp Neurol 1997;56:5516.
[79] Stapleton SR, David KM, Harkness WF, Harding
BN. Central neurocytoma of the cervical spinal
cord. J Neurol Neurosurg Psychiatry 1997;63:
119.
[80] Steinbok P, Boyd M, Flodmark CO, Cochrane
DD. Radiographic imaging requirements following ventriculoperitoneal shunt procedures. Pediatr
Neurosurg 1995;22:1416.
[81] Stephan CL, Kepes JJ, Arnold P, Green KD,
Chamberlin F. Neurocytoma of the cauda equina.
Case report. J Neurosurg 1999;90:24751.
[82] Takao H, Nakagawa K, Ohtomo K. Central
neurocytoma with craniospinal dissemination. J
Neurooncol 2003;61:2559.
[83] Tatter SB, Borges LF, Louis DN. Central neurocytomas of the cervical spinal cord. Report of two
cases. J Neurosurg 1994;81:28893.
[84] Taylor CL, Cohen ML, Cohen AR. Neurocytoma
presenting with intraparenchymal cerebral hemorrhage. Pediatr Neurosurg 1998;29:925.
[85] Tomura N, Hirano H, Watanabe O, Watarai J,
Itoh Y, Mineura K, et al. Central neurocytoma
with clinically malignant behavior. AJNR Am J
Neuroradiol 1997;18:11758.
[86] Tong CY, Ng HK, Pang JC, Hu J, Hui AB, Poon
WS. Central neurocytomas are genetically distinct
from oligodendrogliomas and neuroblastomas.
Histopathology 2000;37:1605.
[87] Townsend JJ, Seaman JP. Central neurocytomaa rare benign intraventricular tumor. Acta
Neuropathol (Berl) 1986;71:16770.
[88] Tyler-Kabara E, Kondziolka D, Flickinger JC,
Lunsford LD. Stereotactic radiosurgery for residual neurocytoma. Report of four cases. J
Neurosurg 2001;95:87982.
508
[89] Valdueza JM, Westphal M, Vortmeyer A, Muller

D, Padberg B, Herrmann HD. Central neurocytoma: clinical, immunohistologic, and biologic
ndings of a human neuroglial progenitor tumor.
[90] Vaquero J, Coca S, Oya S, Del Pozo JM, Martinez
R, Arias A. Clinicopathological experience with
intraventricular neurocytomas. J Neurosurg Sci
1992;36:318.
[91] Vates GE, Arthur KA, Ojemann SG, Williams F,
Lawton MT. A neurocytoma and an associated
lenticulostriate artery aneurysm presenting with
intraventricular hemorrhage: case report. Neurosurgery 2001;49:7215.
[92] Vinchon M, Fichten A, Delestret I, Dhellemmes P.
Shunt revision for asymptomatic failure: surgical
and clinical results. Neurosurgery 2003;52:
34753.
[93] von Deimling A, Janzer R, Kleihues P, Wiestler
OD. Patterns of dierentiation in central neurocytoma. An immunohistochemical study of eleven
biopsies. Acta Neuropathol (Berl) 1990;79:4739.
[94] Warmuth-Metz M, Klein R, Sorensen N, Solymosi
L. Central neurocytoma of the fourth ventricle.
Case report. J Neurosurg 1999;91:5069.
[95] Wilson AJ, Leaer DH, Kohout ND. Dierentiated cerebral neuroblastoma: a tumor in need of
discovery. Hum Pathol 1985;16:6479.
[96] Yasargil MG, von Ammon K, von Deimling A,
Valavanis A, Wichmann W, Wiestler OD. Central
neurocytoma: histopathological variants and therapeutic approaches. J Neurosurg 1992;76:327.
[97] Koeller KK, Sandberg GD. From the archives of
the AFIP. Cerebral intraventricular neoplasms:
radiologic-pathologic correlation. Radiographics
2002;22:1473505.
[98] Waldron JS, Tihan T. Epidemiology and pathology of intraventricular tumors. Neurosurg Clin N
Am 2003;14(4):46982.
[99] Robbins P, Segal A, Narula S, Stokes B, Lee M,
Thomas W, et al. Central neurocytoma. A
clinicopathological, immunohistochemical and ultrastructural study of 7 cases. Pathol Res Pract
1995;191:10011.
[100] Cheung YK. Central neurocytoma occurring in the
thalamus: CT and MRI ndings. Australas Radiol
1996;40:1824.
[101] Friedrichs N, Vorreuther R, Fischer HP, Wiestler
OD, Buettner R. Neurocytoma arising in the
pelvis. Virchows Arch 2003;443:2179.
[102] Hirschowitz L, Ansari A, Cahill DJ, Bamford DS,
Love S. Central neurocytoma arising within
a mature cystic teratoma of the ovary. Int J
Gynecol Pathol 1997;16:1769.
[103] Ojeda VJ, Spagnolo DV, Vaughan RJ. Palisades in
primary cerebral neuroblastoma simulating socalled polar spongioblastoma. A light and electron
microscopical study of an adult case. Am J Surg
Pathol 1987;11:31622.
[104] Poon T, Mangiardi J, Matoso I, Weitzner I. Third

ventricular primary cerebral neuroblastoma. Surg
Neurol 1988;30:23741.
[105] Pearl GS, Takei Y, Bakay R, Davis P. Intraventricular primary cerebral neuroblastoma in adults:
report of three cases. Neurosurgery 1985;16:8479.
[106] DeAngelis L, Gutin P, Leibel S, Posner J. Intracranial tumorsdiagnosis and treatment. London: Martin Dunitz Ltd; 2002.
[107] Tsuchida T, Matsumoto M, Shirayama Y, Imahori
T, Kasai H, Kawamoto K. Neuronal and glial
characteristics of central neurocytoma: electron
microscopical analysis of two cases. Acta Neuropathol (Berl) 1996;91:5737.
[108] Javedan SP, Manwaring K, Smith KA. Treatment
of posterior third ventricular central neurocytoma
with endoscopic biopsy, endoscopic third ventriculostomy and stereotactic radiosurgery. Minim
Invasive Neurosurg 2003;46:1658.
[109] Kondziolka D, Lunsford LD, McLaughlin MR, Flickinger JC. Long-term outcomes after radiosurgery for
acoustic neuromas. N Engl J Med 1998;339:142633.
[110] Luxton G, Petrovich Z, Jozsef G, Nedzi LA,
Apuzzo ML. Stereotactic radiosurgery: principles
and comparison of treatment methods. Neurosurgery 1992;32:24159.
[111] Ron E, Modan B, Boice JD Jr., Alfandary E,
Stovall M, Chetrit A, et al. Tumors of the brain
and nervous system after radiotherapy in childhood. N Engl J Med 1988;319:10339.
[112] Rades D, Fehlauer F. Treatment options for
central neurocytoma. Neurology 2002;59:126870.
[113] OConnor MM, Mayberg MR. Eects of radiation
on cerebral vasculature: a review. Neurosurgery
2000;46:13849.
[114] Leibel SA, Sheline GE. Radiation therapy for
neoplasms of the brain. J Neurosurg 1987;66:122.
[115] Rades D, Schild SE, Ikezaki K, Fehlauer F.
Dening the optimal dose of radiation after
incomplete resection of central neurocytomas. Int
J Radiat Oncol Biol Phys 2003;55:3737.
[116] Adler JR Jr, Murphy MJ, Chang SD, Hancock SL.
Image-guided robotic radiosurgery. Neurosurgery
1999;44:1299306.
[117] Chang SD, Main W, Martin DP, Gibbs IC,
Heilbrun MP. An analysis of the accuracy of the
CyberKnife: a robotic frameless stereotactic radiosurgical system. Neurosurgery 2002;52:1407.
[118] Karnofsy D, Burchenal J. Evaluation of chemotherapeutic agents. In: Macleod C, editor. Evaluation of chemotherapy agents. New York:
Columbia University Press; 1949. p. 191205.
[119] Hanlo PW, Cinalli G, Vandertop WP, Faber JA,
Bogeskov L, Borgesen SE, et al. Treatment of
hydrocephalus determined by the European Orbis
Sigma Valve II survey: a multicenter prospective 5year shunt survival study in children and adults in
whom a ow-regulating shunt was used. J Neurosurg 2003;99:527.
Surgical approaches to tumors of the lateral ventricle

Richard C.E. Anderson, MD*, Saadi Ghatan, MD,
Neil A. Feldstein, MD
The Neurological Institute, Department of Neurosurgery, Columbia University College of Physicians and Surgeons,
710 West 168th Street, Unit 167, New York, NY 10032, USA
Tumors of the lateral ventricle comprise

between 0.8% and 1.6% of all brain tumors and
occur more commonly in children [13]. They are
considered primary intraventricular tumors
when they develop from the ventricular lining
(ie, the ependyma and subependymal glia), from
the epithelium of the choroid plexus and its
arachnoid supporting tissue, or from misplaced
tissue. They are secondary or paraventricular
tumors when they develop from the brain
parenchyma and more than two thirds of their
surface bulges into the lateral ventricle [3]. The
most frequent lateral ventricular tumors are
ependymoma, astrocytoma, choroid plexus papilloma, and meningioma. Less common tumors in
this location are subependymal giant cell astrocytoma, oligodendroglioma, subependymoma,
pilocytic astrocytoma, neurocytoma, choroid
plexus carcinoma, teratoma, choroid plexus cyst,
xanthogranuloma, hemangioblastoma, cavernous
malformation, epidermoid, metastatic carcinoma,
and primary melanoma [2]. Although malignant
tumors are encountered occasionally, most tumors of the lateral ventricle are histologically
benign or have slow growth potential [1].
The clinical manifestations of lateral ventricular tumors are most often a consequence of the
hydrocephalus they produce, either by obstruction
of the normal pathways of cerebrospinal uid
(CSF) ow or by its overproduction rather than
from compression of eloquent or autonomic
regions of the brain [1]. Because they often grow
E-mail address: rca24@columbia.edu
(R.C.E. Anderson).
to a considerable size before they reach clinical

attention, the best treatment option for these
tumors is invariably surgery [3].
Any one or a combination of surgical approaches can be used for lateral ventricular
tumors. Because a variety of eloquent structures
surround the lateral ventricles, surgical approaches to tumors in this region must address
their relatively deep location in the brain,
appropriate pathways through the surrounding
neural structures, the large size of the tumors,
their vascular supply and drainage, and the presence or absence of hydrocephalus. Historically,
the transcortical approaches have been favored
over the interhemispheric routes by many surgeons [13]. More recently, however, the interhemispheric transcallosal approaches have
become more popular in that no cortical brain
tissue has to be violated to provide direct access
to the ventricular system [46]. In general, for
tumors in the anterior portion of the lateral
ventricles, either an anterior interhemispheric
transcallosal approach or a frontal transcortical
approach is used. Approaches to more posterior
intraventricular tumors include the transcortical
temporal or superior parietal lobule approach
and the posterior interhemispheric transcallosal
approach.
Common clinical presentations
Overall, intraventricular tumors are more
common in children than in adults, accounting
for approximately 3% of adult brain tumors and
16% of childhood and adolescent brain tumors
[7]. Half of the intraventricular tumors in adults
and one quarter of the intraventricular tumors in
doi:10.1016/S1042-3680(03)00054-8
510
R.C.E. Anderson et al / Neurosurg Clin N Am 14 (2003) 509525
children are found in the lateral ventricles [3].

Although tumors in the lateral ventricle present
in either gender at any age, there is a slight
male predilection and a denite trend toward the
younger population, with the average patient age
at presentation typically between 20 and 30 years
[13].
Because of their location and their often
benign character, tumors of the lateral ventricle
tend to grow quite large before they become
clinically apparent. As these tumors grow slowly
within the ventricular space, they either obstruct
normal CSF ow, leading to hydrocephalus and
resulting signs and symptoms of increased intracranial pressure (ICP), or they compress the
adjacent structures, causing a focal neurologic
decit. The most frequently seen clinical symptoms and signs are papilledema (42.9%), headache (35.7%), motor disturbance (25%), sensory
disturbance (25%), nausea and vomiting (22.3%),
visual eld decit (17.8%), loss of vision (17.8%),
and mental status changes (17.8%) [2,3]. Mass
eect, either directly from the tumor or indirectly
from hydrocephalus, often impairs the fornices
and leads to short-term memory loss. Infrequently, intraventricular tumors may bleed spontaneously, leading to acute clinical deterioration.
Because of the variable location of these tumors
within the ventricular system, however, no stereotypic neurologic or behavioral signs or symptoms
can be expected.
Tumors of the lateral ventricles: dierential
diagnosis
In one of the largest reported series of patients
with lateral ventricular tumors (112 patients), the
most common pathologic diagnoses were ependymoma (25%), astrocytoma (21.4%), oligodendroglioma (7.1%), choroid plexus papilloma (6.3%),
meningioma (5.3%), and subependymal giant
cell astrocytoma (5.3%) [2]. These tumors were
located throughout the lateral ventricles with the
following distribution: frontal horn and foramen
of Monro (42.8%), body and septum pellucidum
(22.3%), atrium or trigone (19.7%), temporal
horn (8.9%), and occipital horn (6.3%). This
series is slightly atypical, however, in that several
other authors have reported the atrium as the
most common location for tumors of the lateral
ventricle [1,3,8,9].
The diagnosis of lateral ventricular lesions
varies signicantly with age and location within
the ventricle (Fig. 1). In young children, lateral
ventricular neoplasms are most often choroid

plexus papillomas or carcinomas, and they
typically arise in the trigonal region [10]. If arising
from the body of the ventricle in this age group,
a primitive neuroectodermal tumor (PNET) must
also be considered. Tumors arising in the frontal
horn or body of the ventricle in older children are
typically low-grade gliomas, including subependymal giant cell astrocytomas, pilocytic astrocytomas, and ependymomas [3,10]. In adults, the most
common pathologic diagnoses and locations are
central neurocytomas and malignant astrocytomas, both of which localize to the frontal horn or
body of the lateral ventricle, and meningiomas,
which typically occur in the atrium [3]. The
temporal and occipital horns are the least
common sites for tumor occurrence within the
lateral ventricles [8]. More commonly, the temporal and occipital horns become trapped and
encysted by growing neoplasms in the atrium or
adjacent brain parenchyma.
Relevant regional anatomy
The lateral ventricles are paired C-shaped
structures, each subdivided into the frontal horn,
body, atrium (trigone), occipital horn, and temporal horn. Anteriorly, the lateral ventricles
communicate with the third ventricle by way of
the foramina of Monro; posteriorly, they curve
around the thalami and diverge from the midline.
In the absence of hydrocephalus, each lateral
ventricle has a volume of approximately 8 mL [8].
Each lateral ventricle extends anteriorly from
the foramen of Monro into the frontal lobe as the
frontal horn (cornu anterius) and extends posteriorly from the foramen over the thalamus as the
body of the ventricle (cella media). Once behind
the thalamus, the ventricles curve in a lateral and
inferior direction and then anteriorly into the
temporal lobe as the temporal horn (cornu
inferius). The occipital horn (cornu posterius)
extends posteriorly from the junction of the body
and temporal horn. The triangular expansion of
the ventricle between the occipital and temporal
horns is known as the trigone or atrium. The
atrium thus opens anteriorly above the thalamus
into the body of the lateral ventricle, anteriorly
below the thalamus into the temporal horn, and
posteriorly into the occipital horn [11].
In normal-sized ventricles, the average length
of the frontal horn is around 3.2 cm, with
approximately 3 to 4 cm of frontal cortex lying
anterior to it. The body of the lateral ventricle
511
Fig. 1. Anatomic locations and tumor pathology by age group. PNET, primitive neuroectodermal tumor; CPP, choroid
plexus papilloma; GBM, glioblastoma multiforme; F, foramen; SGCA, subependymal giant cell astrocytoma. (From
Jelinek J, Smirniotopoulos JG, Parisi JE, Kanzer M. Lateral ventricular neoplasms of the brain: dierential diagnosis
based on clinical, CT, and MR ndings [Fig. 1]. AJNR Am J Neuroradiol 1990;11(3):56774; with permission.)
from the foramen of Monro to the trigone spans

nearly 4.0 cm, and the anterior-posterior diameter
of the trigone adds another 2.1 cm to the overall
length of the ventricle. The length of the temporal
horn from the trigone averages 4.0 cm, leaving
about 2.5 cm of temporal cortex anterior to the
ventricle. The occipital horn is the most variable,
ranging in size from 0 to 3.4 cm [8].
The septum pellucidum, thalamus, corpus
callosum, caudate nucleus, and fornix form the
walls of each of the lateral ventricles (Fig. 2). The
frontal horns and bodies of the lateral ventricle
are separated by the septum pellucidum, a thin,
translucent, triangular membrane consisting of
two glial laminae with a potential space (cavum)
in between [12]. The inferior leaets of the septum
hold the bodies of the fornices, another set of
paired C-shaped structures. The fornices originate
in the hippocampus on the oor of the temporal
horn, pass posteriorly around the thalamus along
the anterior wall of the atrium, curve superomedially to travel in the medial wall of the body of
the lateral ventricle, and nally separate into two
columns that form the anterior and superior

margins of the foramen of Monro in their course
toward the mamillary bodies.
The head of the caudate nucleus forms the
lateral wall of the frontal horn of the lateral
ventricle, and the rostrum of the corpus callosum
forms the oor medial to the caudate. The medial
wall is formed by the septum pellucidum, and the
roof is formed by the body and genu of the corpus
callosum. At the foramen of Monro, the head of
the caudate nucleus is situated laterally, whereas
the columns of the fornix curve ventrally and
inferiorly to outline the medial and anterior
borders of the foramen [11].
The caudate nucleus courses posteriorly to
form the lateral wall of the body of the ventricle,
and the superior surface of the thalamus forms the
oor. The roof of the body of the lateral ventricle
is formed by the corpus callosum, and the medial
wall is formed by the septum pellucidum above
and the body of the fornix below.
The atrium and occipital horn together form
a triangular cavity with the apex posteriorly in the
512
Fig. 2. Structures in the walls of the lateral ventricles. The central diagram shows the level of the cross-sections through
the frontal horn (A), body (B), atrium (C), and temporal horn (D). (A) Frontal horn. The genu of the corpus callosum is
in the roof, the caudate nucleus is in the lateral wall, the rostrum of the corpus callosum is in the oor, and the septum
pellucidum is in the medial wall. (B) Body of the lateral ventricle. The body of the corpus callosum is in the roof, the
caudate nucleus is in the lateral wall, the thalamus is in the oor, and the septum pellucidum and fornix are in the medial
wall. The choroidal ssure, the site of the attachment of the choroid plexus in the lateral ventricle, is situated between the
fornix and the thalamus. (C) Atrium. The lateral wall and roof are formed by the tapetum of the corpus callosum, and
the oor is formed by the collateral trigone, which overlies the collateral sulcus. The inferior part of the medial wall is
formed by the calcar avis, the prominence that overlies the deep end of the calcarine sulcus, and the superior part of the
medial wall is formed by the bulb of the corpus callosum, which overlies the forceps major. (D) Temporal horn.
The medial part of the oor of the temporal horn is formed by the prominence overlying the hippocampal formation,
and the lateral part of the oor is formed by the prominence called the collateral eminence, which overlies the deep end of
the collateral sulcus. The roof is formed by the caudate nucleus and the tapetum of the corpus callosum, the lateral wall
is formed by the tapetum of the corpus callosum, and the medial wall of the temporal horn is little more than the cleft
between the mbria of the fornix and the inferolateral aspect of the thalamus. Call., callosum; Coll., collateral; Corp.,
corpus; Hippo., hippocampus; Nucl., nucleus; Pell., pellucidum; Sept., septum; Sulc., sulcus; Trig., trigone. (From
Rhoton AL, Jr. The lateral and third ventricles [Fig. 5.4]. Neurosurgery 2002;51(4 Suppl 1):S20771; with permission.)
occipital lobe and the base anteriorly on the

posterior thalamus [13]. The lateral wall and roof
of the atrium are formed by the body, splenium,
and tapetum of the corpus callosum, and the oor
is formed by the collateral trigone, which overlies
the collateral sulcus. The medial wall is formed by
a combination of the calcar avis inferiorly and the
forceps major superiorly.
The superior and lateral walls of the temporal
horn are formed primarily from the tail of the
caudate nucleus and the tapetum of the corpus
callosum. The bulge of the hippocampus lies in
the oor of the temporal horn, and the only
structure in the medial wall is the choroidal

ssure.
Once the lateral ventricle has been entered,
certain landmarks are usually visible and should be
identied if not obstructed by tumor (Figs. 3 and
4). The characteristic white smooth ependymal
surface is lined with veins that serve as guidelines
for the surgeons orientation. The thalamostriate
vein, coursing in a groove separating the caudate
nucleus from the thalamus, joins the septal vein to
form the internal cerebral vein just proximal to the
foramen of Monro beneath the choroid plexus.
The choroid plexus lying within the choroidal
513
Fig. 3. Frontal transcortical approach into the right lateral ventricle. This opening exposes the caudate nucleus, fornix,
foramen of Monro, thalamus, thalamostriate vein, and choroid plexus. The inset on the lower right shows the site of the
cortical incision. (From Rhoton AL, Jr. The lateral and third ventricles [Fig. 5.25]. Neurosurgery 2002;51(4 Suppl 1):
S20771; with permission.)
ssure is arguably the single most valuable intraventricular landmark for the neurosurgeon.
The choroid plexus runs parallel and lateral to
the fornix and is attached by the taenia at the
choroidal ssure, which lies between the fornix
and the thalamus. The choroid plexus is thus
found in the medial aspect of the body, atrium,
Fig. 4. Intraoperative photograph exposing the right

lateral ventricle with the caudate nucleus, fornix, foramen of Monro, thalamus, thalamostriate vein, septal
vein, and choroid plexus.
and temporal horns of the lateral ventricles. It

courses anterior and inferior through the foramina of Monro to lie in the roof of the third
ventricle. The glomus is a prominent tuft of
choroid plexus found in the atrium. From the
atrium, the choroid plexus extends forward and
inferior into the temporal horn, where it terminates immediately posterior to the amygdala. The
arterial supply to the choroid plexus derives from
the anterior and posterior choroidal arteries,
which, in turn, arise from the internal carotid
and posterior cerebral arteries, respectively. The
choroidal arteries enter into the choroid plexus
through the choroidal ssure.
The venous drainage of the ventricular system
consists of the symmetric, midline, bilateral,
internal cerebral veins formed by the thalamostriate and septal veins on each side. The internal
cerebral veins help to dene the oor of the lateral
ventricles and roof of the third ventricle as they
course from anterior to posterior to join and form
the vein of Galen.
Most lateral ventricular tumors receive their
blood supply from the anterior and posterior
choroidal arteries. Tumors situated in the body of
the lateral ventricles derive most of their blood
supply from the posterior lateral choroidal arteries.
514
Tumors of the temporal horn are supplied

primarily by the anterior choroidal artery, and
tumors of the atrium usually derive their supply
from both branches. In addition, supply from the
lenticulostriate arteries or from penetrating
branches can occur in tumors originating from
the oor or wall of the lateral ventricle. Most
ventricular tumors have venous drainage toward
the deep cerebral veins via dilated subependymal
branches [2].
Operative intervention
Preoperative considerations
MRI is the preferred modality for preoperative
evaluation because it provides the best threedimensional images of the tumor and the surrounding neurovascular structures. It can also give
clues regarding the vascularity of the tumor. CT
provides reasonable detail when MRI is contraindicated, and areas of calcication or hemorrhage
are more easily seen with CT. In select cases of
large intraventricular tumors, angiography may be
helpful to provide more detailed information
about the vascular supply of the lesion [3] and to
provide the opportunity for preoperative embolization. The absence of either large dural feeding
vessels or dedicated tumor vessels often makes
superselective embolization impossible, however,
and risks injury to the anterior and posterior
choroidal arteries supplying eloquent areas of the
brain.
A number of adjuncts for preoperative and
intraoperative planning that have facilitated removal of intraventricular lesions previously
thought to be resectable only with considerable
morbidity have become available during the last
several years [14,15]. Frameless stereotactic guidance systems allow precise localization of the
tumor, which permits the surgeon to choose an
approach to the lesion that minimizes manipulation of functionally critical cortex, targets vascular
pedicles, and potentially increases the safety of
aggressive tumor removal. Critics of these systems
argue that the accuracy and resolution of tumor
position and size are compromised if the brain is
retracted, diuretics are used, or the brain shifts
with tumor removal. The ventricular system,
however, is deep, based on the midline, and less
likely to shift than other areas of the cortex [14,15].
Moreover, technical developments linking stereotactic systems with real-time intraoperative ultrasound may improve accuracy and curtail these
limitations. Regardless, these stereotactic systems

are helpful not only in determining the margins of
otherwise poorly dened tumors but in aiding the
planning of surgical approaches, designing limited
craniotomies, and avoiding eloquent cortex or
vascular structures.
It is important to consider the operative
position of the patient at the time of the localizing
study, because standard ducial placement can
lead to diculty in perioperative registration and
poor intraoperative accuracy. If the prone or
lateral position is to be used, for example, standard
ducial markers across the front of the head often
cannot be seen by the navigation camera, which
leads to increased registration time, ducial
marker omission, and decreased accuracy. Furthermore, ducials placed on the back of the head
are more mobile and are often distorted when the
patient lies supine in the headholder at MRI, also
leading to decreased accuracy. Careful preoperative placement of ducials across the vertex or
appropriate side of the head, however, usually
leads to quicker registration and increased intraoperative accuracy. It is often necessary to shave
small areas of hair for ducial placement, and we
always mark the center and edges of the ducials
with a marker in case they move before coregistration. Of note, the appropriate side of the head may
not always be the operative side of the head; if an
interhemispheric approach is used and the patient
is positioned lateral with the tumor side down and
the head angled 45 to the oor so that the falx
holds up the cortex, optimal ducial placement
may be on the side of the head contralateral to the
tumor.
As previously mentioned, ultrasound is another
tool that is becoming more useful during surgery.
Ultrasound provides real-time feedback on the
location of the lesion, which avoids problems with
intraoperative brain movements that limit the
accuracy of stereotactic techniques after the tumor
resection has been initiated. Other intraoperative
techniques that are increasingly popular with
temporal approaches include cortical stimulation
and functional language mapping.
Intraoperative considerations
Common features of the surgical preparation
include insertion of a urinary drainage catheter, an
arterial line, and in cases where signicant blood
loss is anticipated, a central line and sizable peripheral intravenous lines to facilitate expeditious
replacement of blood and clotting factors if
needed. Mild hyperventilation (to a pCO2 of 30

35) is used if increased ICP is a concern. Prophylactic antibiotics are administered 30 minutes
before skin incision and repeated every 4 to 6
hours during the procedure. Corticosteroids and
anticonvulsants are also given during surgery and
continued after surgery.
The positioning used for resection of lateral
ventricular tumors is commonly supine or lateral
decubitus, depending on the operative trajectory.
Occasionally, the prone position is used for
posteriorly located lesions. A three-pinhead xation device is generally used for adults and children
older than 2 years, whereas a horseshoe headrest is
commonly used in younger patients. Care must be
taken to avoid excessive tightening of the headholder, because this can lead to iatrogenic skull
fractures and CSF leaks in younger children.
Attention must also be given to padding all
potential pressure points and avoiding extremes
of neck rotation or exion, which may place
traction on the brachial plexus or cause compression of the jugular veins.
If frameless stereotactic guidance is to be used,
the device is registered to the patients craniofacial
surface anatomy and any ducial markers that
may have been placed during preoperative MRI.
Stereotactic guidance allows the skin incision and
craniotomy to be tailored to the specic approach.
To avoid shaving large areas of the head, we
routinely shave 1- to 2-cm wide strips of hair
along the planned incision line. The incision line is
marked, and routine skin preparation is conducted, including a 5-minute prescrub with an
antibacterial soap and formal povidone iodine
(Betadine) preparation. The skin incision is determined by the location of the lesion and is
discussed in the section on specic operative
approaches. After formal draping, 0.5% local
anesthetic with 1:100,000 epinephrine is inltrated
along the incision line. If operating on a small
child, the local anesthetic can be diluted accordingly (maximum dose is 1 mL/kg of 0.5%
anesthetic). Although not necessary in the adult
population, we perform the incision in layers in
children, using a number 15 blade to incise the
skin and a needle tip cautery to incise the dermis,
galea, and periosteum. Although care must be
taken to avoid thermal injury to the skin, this
approach facilitates scalp exposure with minimal
blood loss, and skin clips are rarely required. The
scalp ap is then reected subperiosteally. To
achieve low-prole scalp retraction, we commonly
set up the Greenberg retractor system and use skin
515
hooks with rubber bands axed to the bars to

retract the skin edges. The craniotomy site is
determined by the location of the tumor and is
discussed with specic approaches elsewhere in this
article. Burr holes are made, and the underlying
dura is carefully stripped using a ball-ended or
other blunt dissecting instrument. The burr holes
are then connected using a small craniotome bit
(eg, Midas Rex B5 drill bit, Medtronic, Fortworth,
TX), and the bone ap is removed. Bone edges are
waxed, and epidural bleeding is controlled with
surgical and 4-0 tenting sutures if needed.
If the dura is tense, mannitol (up to 1 g/kg) can
be used to help reduce ICP and prevent brain
herniation on dural opening. In young children, we
prefer to use boluses of thiopental (starting at 5 mg/
kg up to a maximum of 10 mg/kg) rather than
mannitol to reduce perioperative uid and electrolyte shifts and to minimize the reduction in brain
volume after resection of large tumors. Dural
opening depends on the location of the craniotomy
and is discussed with each specic operative
approach elsewhere in this article. When possible,
we prefer to use linear dural openings to reduce the
chances of outward brain herniation in the setting
of increased ICP and to facilitate dural opening and
closing. The dural edges are reected using 4-0 silk
sutures attached to hanging mosquito clamps, and
the brain surface is protected using Surgicel
(Johnson and Johnson, New Brunswick, NJ), Telfa
(Kendall, Lynn, MA), or cottonoids.
For intraventricular tumors, the most direct
trajectory to the lesion is usually appropriate. At
this stage, stereotactic or ultrasonic localization
can be helpful to guide the cortical incision. Once
the entry site has been selected, a 3- to 4-cm pial
surface is coagulated and incised. Although some
surgeons prefer a transsulcal approach, we generally enter the cortex through a gyrus to avoid
injury to sulcal vessels. The corticectomy is deepened through the gray and white matter until
the ventricle is reached. It should be possible to
identify the ependyma without rupturing it. It will
appear darker and more blue than the adjacent
white matter. Careful dissection of the white
matter overlying it should be continued with the
sucker until the ependyma is exposed over the full
length of the cortical incision. Before incising the
ependyma and entering the ventricle, all bleeding
points on the walls of the cortical incision are
controlled with cautery and the walls are protected with Surgicel and cottonoid strips. In many
cases, the cortical mantle is stretched quite thin
from chronic ventriculomegaly and the ventricle
516
is entered quickly. The length of the cortical

exposure is extended as needed, and self-retaining
retractors are placed to provide adequate exposure of the tumor.
Familiarity with multiple-arm self-retaining
retractor systems is essential. We routinely use the
Greenberg retractor system because of its independently functioning components. We have found
that at least three separate retractors are needed to
expose deep intraventricular lesions optimally. The
retractor blades are placed just within the ventricle
using a curve at the tip of the blade to lift the brain
from its collapsed position. The retractors should
be adjusted so that the brain is simply being
supported and not forcibly retracted.
After ventricular entry, cottonoids are placed
behind the tumor over the dependent portion of the
ventricle to avoid letting blood or tumor fragments
travel into the remaining ventricular system. For
highly vascular intraventricular tumors, it is important to avoid attempts at resecting the tumor
until the arterial supply has been coagulated and
divided. If the large size of the tumor precludes
early control of the vascular pedicle, we routinely
use bipolar cautery along the capsule to devascularize progressively and shrink the tumor until it
can be mobilized away from its feeding vessels.
Tumors can then be debulked using cautery and
microscissors, removing the tumor in lobules or
with an ultrasonic aspirator. The tumor should be
gently delivered by piecemeal dissection, allowing
complete removal through a small cortical incision. En bloc tumor removal increases the surgical morbidity [2]; piecemeal removal reduces the
tumor in size so that it can be mobilized to expose
the vascular supply and removed. Frozen specimens are sent to the pathologist for preliminary
identication of tumor type.
After the tumor has been resected and preliminary hemostasis has been achieved, the
ventricle should be gently lled with saline and
emptied several times to seek any bleeding vessels
from the walls of the ventricular system. Furthermore, a Valsalva maneuver can be performed to
conrm that the eld is dry. Minor bleeding from
the walls can often be controlled with serial
packing of cotton balls or cottonoids. The remainder of the ventricular system is then inspected to be certain that large blood clots or
tumor fragments have not accumulated in remote
sites. The cavity can be lined with Surgicel if
needed. Because Surgicel not only complicates
postoperative imaging studies but can also potentially become dislodged and obstruct CSF ow, it
is avoided if possible. The resection cavity is lled

with saline to prevent collapse of the brain after
the retractors have been removed.
In most cases, we perform a septal fenestration
and leave a ventricular drain in place to drain
postoperative blood and debris and allow for
emergent CSF diversion in case of hydrocephalus.
After exiting thorough a lateral burr hole, the
drain is tunneled subcutaneously away from the
skin incision and secured to the skin with multiple
2-0 silk sutures. The subdural space should be
inspected and irrigated thoroughly to be sure no
bridging veins have been torn by collapse of the
cortex. The dura is then closed with 4-0 silk
sutures or approximated and covered with pericranium or another dural substitute. The bone
ap is secured using either titanium plates and
screws or 2-0 absorbable sutures if the bone is too
thin to accommodate 3-mm screws. The scalp is
then closed in layers with absorbable galeal
sutures and either staples, nylons, or a running
plain gut suture to approximate the skin edges. A
sterile dressing and head wrap are applied, usually
with a chin strap in children to prevent immediate
removal of the head wrap by an agitated child.
The ventricular catheter is connected to a drainage
bag with a volurimeter.
Postoperative considerations
Patients are monitored in the intensive care unit
the night of surgery. Antibiotics are continued for
24 hours after surgery, and corticosteroids are
gradually tapered over 10 to 14 days. Antiepileptic
medications are continued for at least 1 week after
surgery and are then discontinued according to the
individual surgeons preference. Externalized ventricular drainage is weaned over 3 to 5 days by
elevating and clamping the drain. After clamping
the drain and verifying patient tolerance, head CT
is usually done to document the absence of severe
hydrocephalus. Although most patients tolerate
removal of the ventricular drain, either an endoscopic third ventriculostomy or a shunting procedure may eventually be required to control
hydrocephalus. Further postoperative imaging
generally consists of MRI within 48 hours to conrm the extent of tumor resection.
Selective surgical approaches

The choice of approach to the lateral ventricle
depends on several factors, including (1) localization of the tumor within the ventricle, (2) presence
or absence of hydrocephalus, (3) whether the

hemisphere involved is dominant, (4) size of the
tumor, (5) origin of the blood vessels supplying
the tumor, and (6) histopathologic features.
Ideally, the approach chosen should ensure
sucient exposure to permit piecemeal removal
of the tumor, allow rapid identication of the
supplying vessels so they can be divided promptly,
avoid excessive brain retraction, and limit damage
to functional cortex. It is impossible to reach
a lateral ventricular tumor without opening some
neural structures, but some important structures
must be preserved during surgery. At the least,
these include the Rolandic area, the language
area, the fornix at least on one side, the internal
cerebral veins, and the vein of Galen [2]. The
anterior approaches are the anterior transcallosal,
anterior transcortical, and anterior frontal. The
posterior approaches are the posterior transcallosal, posterior transcortical, and occipital.
517
The inferior approaches are the temporal and

posterior frontotemporal (Fig. 5) [11,13]. Because
they are used more commonly, this report focuses
on the following approaches: anterior, temporal,
and parietal transcortical approaches and anterior
and posterior interhemispheric approaches. As
with many areas of tumor surgery, the choice of
approach has much to do with a particular
surgeons experience with a given approach.
Transcortical approaches
Frontal transcortical (middle frontal gyrus)
approach
The frontal transcortical corridor can be used
to access large tumors of the frontal horn and the
anterior portion of the body of one lateral ventricle (Fig. 6). This approach is also eective for
tumors originating in the anterior third ventricle
and extending up into one lateral ventricle
Fig. 5. Surgical approaches to the lateral ventricles. The site of the skin incision (solid line) and the bone ap (broken
line) are shown for each approach. The anterior part of the lateral ventricle may be reached by the anterior transcallosal,
anterior transcortical, and frontal approaches. The posterior routes to the lateral ventricle are the posterior transcallosal,
posterior transcortical, and occipital approaches. The inferior part of the lateral ventricle can be reached using the
frontotemporal and temporal approaches. Ant., anterior; Post., posterior. (From Rhoton AL, Jr. The lateral and third
ventricles [Fig. 5.21]. Neurosurgery 2002;51(4 Suppl 1):S20771; with permission.)
518
Fig. 6. T1-weighted axial (A) and sagittal (B) MRI without contrast demonstrates a large mass in the left lateral ventricle
that was resected via a frontal transcortical approach.
through the foramen of Monro. This approach is

greatly facilitated if there is hydrocephalus or the
involved ventricle has been partially obstructed
and is enlarged. The primary advantage of this
approach is that with large tumors, transcortical
exposure provides better access to the tumor than
a transcallosal route [3]. Compared with a transcallosal approach, disadvantages of the transcortical approach are that (1) it is more dicult
to expose the lateral ventricle on the opposite side
and (2) the incidence of postoperative epilepsy
after any transcortical approach is higher than
after a transcallosal approach [2].
Because the lateral ventricles cannot be entered
from a transcortical approach without an incision
through the overlying cortex, functional cortex
may be injured during the approach. Although
a middle frontal gyrus entry is least likely to cause
noticeable neurologic injury, transient postoperative attention decits can result from an incision in
the middle frontal gyrus of either hemisphere.
A more posteriorly located incision in the supplemental motor or premotor area in either hemisphere can cause transient hemiparesis, and if an
incision is made in the dominant hemisphere,
transient speech diculties can occur. A frontal
transcortical approach is contraindicated for tumors of the midventricular body because it would
involve making a lesion in motor or sensory cortex.
The patient is positioned supine with gentle
neck exion. Either a bicoronal or unilateral
curvilinear skin incision is most often employed,
and a free bone ap is fashioned. The bone ap is
designed with its center slightly anterior to the
coronal suture and overlying the middle frontal
gyrus, allowing a familiar trajectory to the foramen
of Monro. The dura is then opened in a linear,
cruciate, or U-shaped fashion with the base medial.
A 3- to 4-cm cortical incision is made along

the long axis of the middle frontal gyrus anterior
to the coronal suture and well anterior to the
premotor gyrus. If available, stereotactic guidance
or intraoperative ultrasound can be used to direct
the angle of transit to the site of the tumor.
Deepening of the corticectomy and placement of
retractors are as previously described. Once the
medial, lateral, and inferior intraventricular anatomic landmarks have been identied, resection of
the tumor can proceed with condence. The
extent of inferior and inferolateral resection of
these tumors should be tempered by the knowledge that the subependymal white matter immediately lateral to the foramen of Monro overlies
the genu of the internal capsule [13].
Temporal transcortical (middle temporal gyrus)

approach
Although the temporal horn and trigonal
regions can be approached from the posterior
frontotemporal, temporal, or subtemporal route,
the approach through the posterior portion of the
middle temporal gyrus is preferred by many
surgeons, especially for choroidal tumors. Occasionally, an approach through the inferior temporal gyrus is possible if the lesion is small and
located primarily in the temporal horn of the
lateral ventricle. A temporal transcortical approach is advantageous, because the anterior
choroidal vessels can be identied early, coagulated, and divided before internal decompression
of the tumor. It oers poor visualization, however, of the posterior choroidal artery. This approach also has the advantage of providing a short
trajectory to the lesion. As with all transcortical
approaches, the temporal transcortical approach
is greatly facilitated in the presence of hydrocephalus or a trapped dilated temporal horn.

The most common complications associated
with a transcortical entry through the middle
temporal gyrus are visual eld loss from injury
to the optic radiations and speech dysfunction,
especially if approaching through the dominant
hemisphere [9,16]. Despite extreme variability,
Haglund and colleagues [17] demonstrated no
language localizations below the middle temporal
gyrus. Nevertheless, some surgeons use cortical
mapping when operating in the dominant temporal lobe, and others argue that these complications
can be reduced if an incision through the inferior
temporal gyrus is used instead [9]. In addition,
choroidal artery territory infarcts can occur if the
artery is taken too proximally.
The patients head is set in the lateral position.
An inverted U-shaped incision above the ear or
a reverse question mark incision is made, and
a temporal craniotomy is performed. The craniotomy should be low enough to expose the oor
of the middle fossa, with its posterior extent above
the asterion to minimize inadvertent injury to the
transverse sinus. The dura is opened in a Ushaped fashion with its base inferiorly or linearly
along the middle temporal gyrus. A cortical
incision is made in the middle temporal gyrus
anterior to the optic radiations. The middle
temporal gyrus is then traversed to expose the
ventricle and tumor. Some authors advocate
a transsulcal incision between the middle and
inferior temporal gyri because it reduced hemianoptic complications, with preservation of the
superior quadrant bers, as well as the risk of
inducing speech disturbances [9].
519
Parietal transcortical (superior parietal lobule)

approach
The parietal transcortical approach may be
used for removal of tumors in the trigonal region
or for tumors of the posterior body of the lateral
ventricle (Fig. 7). The shortest distance between
a trigonal lesion and the surgeon is often transcortical in the parietal cortex. This approach is
particularly well-suited for large masses in these
regions with associated ventricular dilatation.
Compared with the middle temporal gyrus approach, the parietal transcortical approach may be
associated with a reduced incidence of language
impairment, although this approach does not
provide for access to supplying anterior choroidal
arteries before tumor removal. The primary
limitation of this approach is delayed visualization
of the aerent vessels, which can only be controlled once the tumor has been debulked, because
they often lie underneath the lesion [9].
Possible complications of the superior parietal
lobule approach are permanent visual eld
impairment; although the cortical incision is
performed away from the optic radiations running
lateral and inferior to the trigonal region, large
tumors may destroy the ventricular ependyma and
inltrate the white matter so that the optic
radiations are damaged during dissection [9].
Furthermore, apraxia and acalculia or a complete
Gerstmann syndrome can occur when the lesion
arises in the dominant hemisphere [18]. Only
rarely, however, is this approach complicated by
neurologic decits [2].
The patient may be positioned prone or lateral
decubitus. A linear or U-shaped incision is made,
and a bone ap centered over the superior parietal
Fig. 7. (A) T1-weighted sagittal MRI with contrast demonstrates an enhancing mass in the posterior lateral ventricle
that was resected via a parietal transcortical approach. (B) Intraoperative photomicrograph demonstrating the surgeons
view of the exposed lateral ventricle and tumor. Note the use of three self-retaining retractors for optimal exposure.
520
lobule is removed above the lambdoid suture. A

dural opening is made, followed by a cortical
incision through the superior parietal lobule into
the ventricle. The corticectomy is made medial
enough to avoid the optic radiations and anterior
enough to allow access to the oor of the body of
the lateral ventricle but well behind the sensorimotor cortex. Self-retaining retractors are then
placed to maintain operative exposure.
Interhemispheric approaches
Anterior transcallosal approach
The anterior transcallosal approach may be
used to approach tumors of the midbody or
anterior horn of the lateral ventricle (Fig. 8) [5].
This approach is preferred if the tumor is small,
located near the midline, and does not require
excessive hemispheric retraction. Many authors
advocate this approach for tumors located in the
midbody of the ventricle, because a transcortical
approach through the Rolandic area is contraindicated [9]. With small tumors, this approach
may be chosen from the opposite site if the tumor
lies in the dominant hemisphere [19,20]. Compared with a transcortical route, it oers the
advantages of (1) easy access to both lateral
ventricles (which may be dicult to achieve using
the frontal transcortical route) and (2) the absence
of a cortical incision, reducing the incidence of
postoperative seizures. Furthermore, the transcallosal approach is easier to perform than the
transcortical approach if the ventricles are of
normal size or only minimally enlarged [2,20,21].
If entry into the third ventricle is required,
although a transcortical approach allows only
a transchoroidal entry, a transcallosal approach
allows either an interforniceal trajectory or a
transchoroidal entry.
The complications of interhemispheric transcallosal approaches to the ventricles are well
reported and include hemiparesis, aphasia, mutism, confabulation, memory decits, astereognosis, and alexia without agraphia [5,6,20,22]. Entry
into the lateral ventricles through an interhemispheric approach cannot be achieved without
a callosal resection. There is substantial evidence
from the study of patients with sections of their
cerebral commissures that the tracts of the corpus
callosum have an essential role in the transfer of
sensory information from one cerebral hemisphere to the other [16]. Extensive neuropsychiatric testing has shown that after sectioning of the
corpus callosum, patients do exhibit signicant
Fig. 8. T1-weighted axial (A) and coronal (B) MRI with

contrast demonstrates an enhancing mass in the left
lateral ventricle that was resected via an anterior
transcallosal approach. (C ) Intraoperative photograph
demonstrating the size and position of the bone ap for
this approach. (D) Intraoperative photograph showing
the use of three self-retaining retractors (two laterally
and one medially) to establish the interhemispheric
corridor.
decits in tests designed to demonstrate interhemispheric transfer of tactile information [23].

Even when callosal sections have involved the
entire body of the corpus callosum, with preservation of the genu and rostrum, the immediate
and long-term eects of surgery have not been
disconcerting, however [23]. There is no change in
the patients ordinary behavior, and the patients
do not appear to suer any inconvenience from
this subtle incapacity [23]. Furthermore, when
dividing only the anterior third of the body of the
corpus callosum, no signicant clinical decits
could be demonstrated [5]. The interhemispheric
dissection may also result in injury to the bridging
veins. The issue of venous drainage is an
important one, because excessive venous compromise can lead to venous hypertension and infarction of the basal ganglia and internal capsule
[6]. Although we and others have sacriced
bridging veins anterior to the coronal suture
without adverse eects [20], we try to preserve
all bridging veins whenever possible.
The patient is positioned supine with the head

straight and the neck gently exed. Either a Cshaped or bicoronal incision is made. Burr holes
are made over the midline at the anterior and
posterior limits of the craniotomy. An additional
burr hole is made over the coronal suture laterally
to facilitate dissection of the underlying dura. We
commonly make a bone ap that is 3 cm wide and
approximately 6 cm in anterior-posterior length,
situated two thirds anterior and one third
posterior to the coronal suture. This provides
a familiar trajectory to the lateral ventricle and
foramen of Monro and minimizes the risk of
venous infarction from interruption of draining
veins from the cortex. A rectangular-, trapezoid-,
or half-ellipticshaped bone ap is then removed. In the pediatric population, we prefer to
make the medial cut of the bone ap at the edge
of the sagittal sinus to minimize potential occlusion or injury during medial retraction. In the
adult population, however, the thicker bone often
makes it necessary to cross the midline and expose
the entire sinus for an adequate operative corridor. The dura is opened in a U-shaped fashion
with its base toward the sinus, taking care not to
injure bridging veins in the subdural space.
Although we attempt to preserve all medially
draining veins, at least one anterior vein occasionally must be sacriced to allow adequate
access to the interhemispheric ssure. The falxcortical interface is identied, and the interhemispheric plane is developed. Telfa or cottonoid
strips are laid down medially and laterally,
followed by placement and progressive advancement of self-retaining retractors. We nd it
optimal to use one retractor blade medially and
two laterally to establish an uncluttered operative
corridor and prevent excessive retraction on the
hemisphere laterally. With the medial retractor
blade, a gentle hook is placed at the tip to ensure
that the sagittal sinus is not occluded. After the
retractors are in place, the operating microscope is
used to help identify the glistening white corpus
callosum and its associated pericallosal arteries. It
is easy to mistake the cingulate gyrus for the
corpus callosum, but the former has the typical
brown-gray appearance of the cortical-pial surface, whereas the callosum is strikingly white.
Once the callosum is identied, cotton balls or
rolled cottonoids can be placed at the deep
anterior and posterior poles to help maintain
hemispheric retraction. The callosum is then
traversed over a length of approximately 2.0 to
2.5 cm to allow entry into the ventricular system.
521
To limit the potential for functional injury, the

callosal section is performed in the anterior third of
the body, sparing the genu [5]. Depending on the
location of the callosal incision, the surgeon may
enter into the left or right lateral ventricle or into
a cavum septum pellucidum. If the anatomy is not
excessively distorted from the tumor, orientation
can be achieved by identifying the major anatomic
landmarks within the ventricle: the choroid plexus,
the thalamostriate vein, and the septal vein [5].
Entry into a cavum septum pellucidum may be
confusing until the surgeon realizes that no
intraventricular structures are present. Regardless
of the side of initial ventricular entry, it is generally
possible to achieve adequate biventricular exposure by extending the callosotomy or fenestrating
the septum pellucidum. Both maneuvers are often
needed for lateral ventricular tumors that extend
bilaterally via the foramen of Monro. If needed,
additional exposure of tumor in the third ventricle
can be obtained using either a transchoroidal or
interforniceal exposure.
Posterior transcallosal approach

A posterior interhemispheric transcallosal approach may be used to approach tumors of the
trigone or posterior body of the lateral ventricle
(Figs. 9 and 10). The primary motor and sensory
gyri are situated on the cortical surface overlying
the body of the lateral ventricle, precluding
a transcortical approach. Medially positioned
tumors or tumors with blood supply primarily
from the posterior choroidal arteries are preferably approached from a posterior interhemispheric
approach [3]. Larger tumors of the trigone are not
suitable for transcallosal removal, because the
tumor itself prevents the hemispheric retraction
that is required to eect tumor removal. Some
authors insist that in patients with a preexisting
hemianopsia, a transcallosal approach must be
avoided, because a lesion of the splenium of the
corpus callosum can lead to alexia and a syndrome
of verbal-visual disconnection (inability to name
objects situated in the left half of the visual eld)
[18,24]. Additional complications of an interhemispheric transcallosal approach have been described previously.
The patient is placed in the lateral decubitus
position with the tumor side down and the head
elevated 45 . With this method, gravity causes the
hemisphere to fall away from the falx once the
dura is opened, thus minimizing brain retraction.
Alternatively, the prone position can be used,
522
Fig. 9. Illustration of the posterior transcallosal approach. (A) The sagittal view depicts the corridor of approach to the
posterior corpus callosum. (B) View into the lateral ventricle, showing route of entry. (From Abosch A, McDermott
MW, Wilson CB. Lateral ventricular tumors [Fig. 64.4]. In: Kaye AH, Black P, editors. Operative neurosurgery.
London: Churchill Livingstone; 2000. p. 799812; with permission.)
preserving the orientation to the midline. The

incision and bone ap are performed as previously
described. The posterior portion of the corpus
callosum is resected, but the splenium is preserved
to minimize the potential for functional injury [2].
Tumor resection and closure are then performed
as previously described.
Outcome and complications

The risks of intraventricular surgery are
potentially catastrophic. The relative rarity of
these lesions has precluded large series examining
postoperative complication rates and the prognosis for recovery. Postoperative complication rates
Fig. 10. (A) T1-weighted sagittal MRI with contrast demonstrates a large cavernous malformation throughout the body
and atrium of the lateral ventricle that was resected via a posterior transcallosal approach with the patient in the prone
position. (B) Intraoperative photomicrograph showing the surgeons view into the lateral ventricles from this approach.
Both ventricles were widely opened, and the septum was resected with the mass.
vary, but they typically approach 20% [2].

Surgical mortality has decreased tremendously
with advancements in microsurgery, falling from
35% in older series [25] to 6% to 10% in more
recent series [13]. The most common complications are severe brain edema, intraventricular
hemorrhage, subdural hematoma, epidural hematoma, and additional neurologic decits [2]. A
review by Piepmeier and colleagues [26] reported
rates of occurrence for specic postoperative
decits. The most commonly reported neurologic
complications include visual eld decits (20%
64%), hemiparesis (8%30%), speech decit
(8%36%), subdural hematoma (11%), seizures
(29%70%), persistent hydrocephalus (12%
33%), and death (12%75%). As with all
craniotomies, the particular postoperative neurologic decit risked depends on the surgical
approach selected (Fig. 11). As previously mentioned, the risk of postoperative seizures has been
523
noted to be higher with transcortical routes than

with transcallosal approaches [18], especially in
children [23].
Tumors of the lateral ventricles may lead to
functional forniceal lesions either directly or as
a result of increased ICP. Furthermore, the
fornices are often manipulated to some degree
with intraventricular surgery. As a result, some
patients experience severe memory diculties
[6,27]. Fortunately, this problem is usually transient [16,23]. When the eects on memory were
compared between transcortical and transcallosal
removal of intraventricular tumors, it was concluded that transcallosal approaches did not seem
to be responsible for any increased memory
decits [16].
Because of the deep-seated nature of these
tumors, the proximity of eloquent cortex, and the
histologic nature of some lesions, gross total
removal of lateral ventricular tumors is not always
Fig. 11. Illustration summarizing approaches to tumors of the lateral ventricles and some of the potential attendant
neurologic decits. (From Abosch A, McDermott MW, Wilson CB. Lateral ventricular tumors [Fig. 64.4]. In: Kaye AH,
Black P, editors. Operative neurosurgery. London: Churchill Livingstone; 2000. p. 799812; with permission.)
524
possible without unacceptable morbidity. In a large

series, total removal was accomplished in 38.4% of
lesions, subtotal removal in 56.2%, and biopsy in
2.7% [2]. Fortunately, most intraventricular lesions
are benign or of low malignancy, and a complete or
even subtotal resection leads to a cure or long
survival in most cases [1].
[8]
[9]
Summary
[10]
Lateral ventricular tumors are rare lesions of

the central nervous system, and because most
tumors are benign or low grade, permanent cure
can be achieved with complete removal. After
adequate preoperative imaging discloses a lateral
ventricular mass, the neurosurgeon has several
options to choose from when determining the
ideal surgical approach to the tumor. The surgical
approach cannot be standardized, because the
specic location, size, and vascularization of these
deep-seated tumors are fundamental elements
inuencing the choice of surgical approach.
Although access to the lateral ventricles may
require additional preoperative considerations
and planning, the combination of proper knowledge of the cortical and intraventricular anatomy
with the familiarity and selection of an appropriate surgical approach will optimize the surgical
outcome.
[11]
[12]
[13]
[14]
[15]
[16]
References
[1] Delni R, Acqui M, Oppido PA, Capone R,
Santoro A, Ferrante L. Tumors of the lateral
ventricles. Neurosurg Rev 1991;14:12733.
[2] Gokalp HZ, Yuceer N, Arasil E, Deda H, Attar A,
Erdodan A, et al. Tumours of the lateral ventricle.
A retrospective review of 112 cases operated upon
19701997. Neurosurg Rev 1998;21:12637.
[3] Pendl G, Ozturk E, Haselsberger K. Surgery of
tumours of the lateral ventricle. Acta Neurochir
(Wien) 1992;116:12836.
[4] Vogel S, Meyer R, Lehmann R, Woiciechowsky C.
Transcallosal removal of lesions aecting the third
ventricle: an anatomic and clinical study. J Neurosurg 1995;83:9235.
[5] Shucart WA, Stein BM. Transcallosal approach to
the anterior ventricular system. Neurosurgery
1978;3:33943.
[6] Apuzzo ML, Chikovani OK, Gott PS, Teng EL, Zee
CS, Giannotta SL, et al. Transcallosal, interforniceal
approaches for lesions aecting the third ventricle:
surgical considerations and consequences. Neurosurgery 1982;10:54754.
[7] Koos WT, Miller MH. Tumors of the ventricular
system. In: Koos WT, Miller MH, editors. In-
[17]
[18]
[19]
[20]
[21]
[22]
tracranial tumors of infants and children. Stuttgart:

Verlag; 1970. p. 23353.
Abosch A, McDermott MW, Wilson CB. Lateral
ventricular tumors. In: Kaye A, Black P. Operative
neurosurgery, vol. 1. 1st edition. London: Churchill
Livingstone; 2000. p. 799812.
Santoro A, Salvati M, Frati A, Polli FM, Delni R,
Cantore G. Surgical approaches to tumours of
the lateral ventricles in the dominant hemisphere.
J Neurosurg Sci 2002;46:605.
Jelinek J, Smirniotopoulos JG, Parisi JE, Kanzer
M. Lateral ventricular neoplasms of the brain:
dierential diagnosis based on clinical, CT, and MR
ndings [erratum appears in AJNR Am J Neuroradiol 1990;11(4):734]. AJNR Am J Neuroradiol
1990;11:56774.
Timurkaynak E, Rhoton AL Jr, Barry M. Microsurgical anatomy and operative approaches to the
lateral ventricles. Neurosurgery 1986;19:685723.
Sarwar M. The septum pellucidum: normal and
abnormal. AJNR Am J Neuroradiol 1989;10:
9891005.
Rhoton AL Jr. The lateral and third ventricles.
Neurosurgery 2002;51(Suppl):S20771.
Giorgi C, Riva D. Stereotactically guided transfrontal removal of intraventricular midline tumors in
children. Neurosurgical and neuropsycho logical
considerations. J Neurosurg 1994;81:37480.
Morita A, Kelly PJ. Resection of intraventricular
tumors via a computer-assisted volumetric stereotactic approach. Neurosurgery 1993;32:9206.
Geen G, Walsh A, Simpson D, Jeeves M.
Comparison of the eects of transcortical and
transcallosal removal of intraventricular tumours.
Brain 1980;103:77388.
Haglund MM, Berger MS, Shamseldin M, Lettich
E, Ojemann GA. Cortical localization of temporal
lobe language sites in patients with gliomas.
Fornari M, Savoiardo M, Morello G, Solero CL.
Meningiomas of the lateral ventricles. Neuroradiological and surgical considerations in 18 cases.
J Neurosurg 1981;54:6474.
Spencer DD, Collins W, Sass KJ. Surgical management of lateral intraventricular tumors. In: Schmidek HH, Sweet WH, editors. Operative neurosurgical
techniques, vol. 1. Orlando, NY: Grune and Stratton;
1988:58396.
Benes V. Advantages and disadvantages of the
transcallosal approach to the III ventricle. Childs
Nerv Syst 1990;6:4379.
Long DM, Chou SN. Transcallosal removal of
cranio-pharyngiomas within the third ventricle.
Villani R, Papagno C, Tomei G, Grimoldi N,
Spagnoli D, Bello L. Transcallosal approach to
tumors of the third ventricle. Surgical results and
neuropsychological evaluation. J Neurosurg Sci
1997;41:4150.

[23] Jeeves MA, Simpson DA, Geen G. Functional
consequences of the transcallosal removal of intraventricular tumours. J Neurol Neurosurg Psychiatry 1979;42:13442.
[24] Levin HS, Rose JE. Alexia without agraphia in
a musician after transcallosal removal of a left
intraventricular meningioma. Neurosurgery 1979;4:
16874.
[25] Lapras C, Deruty R, Bret P. Tumors of the lateral
ventricles. Adv Tech Stand Neurosurg 1984;11:10367.
525
[26] Piepmeier JM, Spencer DD, Saas KJ, George TM.

Lateral ventricular masses. In: Apuzzo ML, editor.
Brain surgery: complication avoidance and management. New York: Churchill Livingston; 1993.
p. 58199.
[27] Winkler PA, Weis S, Buttner A, Raabe A, Amiridze
N, Reulen HJ. The transcallosal interforniceal
approach to the third ventricle: anatomic and
microsurgical aspects. Neurosurgery 1997;40:
97381.
Surgical approaches to posterior third ventricular

tumors
Alan P. Lozier, MD, Jerey N. Bruce, MD*
The Neurological Institute of New York, New York Presbyterian Hospital, College of
Physicians and Surgeons, Columbia University, 710 West 171 Street, Box 174, New York, NY 10009, USA
Advanced microsurgical techniques combined

with improved neuroanesthetic and postoperative
critical care have made aggressive surgical resection a mainstay in the management of posterior
third ventricular and pineal region tumors.
Although a variety of approaches to the posterior
third ventricle have been designed [17], three are
in common use. The infratentorial-supracerebellar
approach takes advantage of a natural corridor
between the cerebellum and the tentorium. Supratentorial approaches include the interhemispherictranscallosal and occipital-transtentorial approaches. Choosing the optimal surgical approach
depends on the anatomic extent of the tumor as
well as on the preference and experience of the
surgeon. Renements in surgical technique have
led to a more favorable outlook for patients with
these uncommon tumors.
Spectrum of pathology
The list of possible posterior third ventricular
and pineal region lesions is extensive because of
the histologic diversity of the area. Lesions may
arise from within the posterior third ventricle
itself, the pineal gland, the velum interpositum, the
surrounding parenchyma (thalamus, midbrain, or
splenium), the tentorium, the quadrigeminal cistern, or the posterior cerebral vasculature. The
most common posterior third ventricular mass is
tumor emanating from the pineal gland. Tumors
of the pineal gland proper can be grouped into
three principal categories: germ cell tumors, pineal
E-mail address: jnb2@columbia.edu (J.N. Bruce).
cell tumors, and glial cell tumors [8,9]. A continuum of histopathologic subtypes exists within each
of these categories, ranging from benign to highly
malignant. Additionally, tumors of germ cell
origin may exhibit a mixture of cell types. Rarely,
miscellaneous lesions, such as metastatic tumors
[10,11], lymphoma [10], chemodectoma [12], and
primary melanocytic tumors [13], present in the
pineal gland. Benign pineal cysts are being
discovered with increasing frequency as people
undergo MRI scans for routine neurologic complaints [14].
In addition to tumors of the pineal gland,
posterior third ventricular tumors include thalamic astrocytomas and ependymomas, choroid
plexus tumors [1517], epidermoids [18,19], craniopharyngiomas [20], and meningiomas arising
from the velum interpositum [2123]. Large
falcotentorial meningiomas that extend anteroventrally may also impinge on the third ventricle
[24]. Arachnoid cysts [25], arteriovenous malformations [26], vein of Galen malformations [27,28],
and cavernous malformations [29,30] may also be
found in the region.
Aims of the operation
The goals of surgery for posterior third
ventricular tumors vary depending on the clinical
circumstances of an individual situation. The rst
objective is to establish a histologic diagnosis
[3134]. Surgery may be avoided altogether in the
presence of either serum or cerebrospinal uid
(CSF) a-fetoprotein or b-human chorionic gonadotropin. These markers are pathognomonic
for malignant germ cell elements; aected patients are treated nonoperatively with radio- and
doi:10.1016/S1042-3680(03)00061-5
528
J.N. Bruce, A.P. Lozier / Neurosurg Clin N Am 14 (2003) 527545
chemotherapy [35,36]. In the absence of positive

markers, a tissue sample is generally required.
Although a limited amount of tissue can provide
a histologic diagnosis, extensive sampling from
various portions of the tumor is desirable to
reduce the risk of sampling error. Most patients
benet from a more extensive resection of the
tumor rather than a simple biopsy. Resection can
provide immediate relief of mass eect exerted by
large lesions and improve response to adjuvant
therapy for malignant lesions. Gross total resection is a feasible goal in all patients with benign
or encapsulated tumors.
Stereotactic biopsies of pineal region tumors
are more complicated and have a reduced margin
for error compared with tumors at other locations.
The potential for misdiagnosis exists when limited
sampling of a heterogeneous tumor occurs [37].
The risk of hemorrhage is increased because of
the often vascular nature of the lesion and the
proximity of the deep venous system [38,39]. The
mortality rate of stereotactic biopsy of a pineal
region mass was recently reported to be 1.3%,
representing a two- to fourfold increase in risk of
death compared with unselected stereotactic biopsies [40,41]. Stereotactic biopsy is generally
reserved for patients with multicentric or disseminated disease or for those individuals with medical
contraindications to open surgery. In the event that
a stereotactic biopsy is warranted, a low frontal
trajectory that crosses the posterior limb of the
internal capsule and avoids the lateral ventricle is
preferred. This approach comes inferior and lateral
to the internal cerebral veins and reduces the risk of
hemorrhage. Alternatively, a superior parietal
lobule approach may be used. This approach
crosses two ependymal surfaces in the lateral
ventricle and is only recommended for tumors
with signicant lateral or superior extension.
Radiographic features
MRI with gadolinium enhancement is the
diagnostic test of choice for posterior third
ventricular tumors (Fig. 1) [4246]. MRI reveals
the degree of hydrocephalus and allows for the
evaluation of tumor size, vascularity, homogeneity, and proximity to surrounding structures. The
extent of tumor invasiveness can be estimated
from the margination and irregularity of the
tumor border; however, the true degree of tumor
encapsulation only becomes apparent at surgery.
The relation of the tumor to the deep venous
system may yield a diagnostic clue and inuence
the choice of surgical approach. Pineal tumors

generally displace the deep venous system superiorly along the dorsal periphery of the tumor.
Notable exceptions are velum interpositum meningiomas arising from the superior leaf of the
tela choroidea, anteroventrally directed falcotentorial meningiomas, and dermoids or other lesions
arising near the corpus callosum. These tumors
displace the deep venous system inferiorly. Despite the high resolution of MRI, tumor histology
cannot reliably be predicted based on radiographic features alone. CT can complement MRI by
allowing for an assessment of intralesional calcication [47]. Angiography is not necessary unless
a vascular anomaly is suspected.
Pineal cell tumors, malignant germ cell tumors,
and ependymomas of the pineal region are all
prone to CSF dissemination. As such, preoperative spinal MRI screening is suggested, because
blood products or operative debris may mimic
spinal metastases in the early postoperative
period.
Management of hydrocephalus
Pineal region tumors often present with
obstructive hydrocephalus resulting either from
tectal and aqueductal compression or from direct
aqueductal plugging by the lesion. Consideration
for CSF diversion is the initial step in clinical
management. Patients with only mild asymptomatic hydrocephalus may be managed without
a preoperative CSF diversion procedure if it is
expected that complete resection of the tumor will
restore aqueductal patency. In these instances,
a ventricular drain is advisable at the time of
surgery. In patients with symptomatic hydrocephalus, an endoscopic third ventriculostomy is
desirable because it avoids shunt-related complications, such as infection, overshunting, and
peritoneal seeding by malignant cells. Further, if
aqueductal patency is restored and hydrocephalus
remits, the patient is not rendered shunt dependent.
Anatomy
A thorough understanding of the anatomy of
the third ventricle and posterior incisural space is
required to plan a surgical approach to lesions
situated in and around the posterior third ventricle
appropriately (Fig. 2). The third ventricle is divided
into anterior and posterior portions in a coronal
plane extending through the massa intermedia and
529
Fig. 1. Midsagittal gadolinium-enhanced MRI. (A) Germinoma with dorsal displacement of internal cerebral veins. (B)
Pineocytoma with dorsal displacement of internal cerebral veins. The two lesions are best approached via the
infratentorial-supracerebellar or occipital-transtentorial corridor. (C) Epidermoid tumor with ventral displacement of
the internal cerebral veins. This lesion is best approached via the posterior-interhemispheric corridor. (D) Tectal glioma
with extension into the cerebellomesencephalic ssure. The occipital-transtentorial approach provides the best trajectory
to the inferior pole of the tumor.
the mammillary bodies. The posterior third ventricle is bounded by a roof, a oor, two lateral
walls, and a posterior wall. The oor is formed by
mesencephalic structures extending between the
mammillary bodies and the cerebral aqueduct.
Viewed from within the ventricle, this surface is
smooth and concave. The posterior oor overlies
the posterior perforate substance anteriorly and
the medial part of the cerebral peduncles and the
tegmentum of the midbrain posteriorly [48,49].
The lateral walls of the posterior third ventricle
are formed inferiorly by the posterior hypothalamus and superiorly by the thalamus. The
thalamic and hypothalamic surfaces are demarcated by the hypothalamic sulcus, an indistinct
groove that extends from the foramen of Monro
to the aqueduct of Sylvius. The massa intermedia
often projects into the upper half of the posterior
third ventricle (present in approximately 75% of
specimens). The stria medullaris thalami extend
from the foramen of Monro to the habenulae
along the superomedial surface of the thalamus
and mark the superior limit of the lateral wall of

the third ventricle. The teniae thalami, raised
ridges along the surface of the stria medullaris,
serve as the site of attachment for the inferior leaf
of the tela choroidea. The habenulae are small
prominences on the dorsomedial surface of
the thalamus just anterior to the pineal gland.
They receive the stria medullaris thalami and
are connected by the habenular commissure in
the rostral half of the stalk of the pineal gland
[50,51].
The posterior wall of the third ventricle
consists of the suprapineal recess, the habenular
commissure, the pineal body and recess, the
posterior commissure, and the aqueduct of
Sylvius. The suprapineal recess extends posteriorly between the pineal gland and the inferior
layer of tela choroidea. The pineal gland projects
posteriorly into the quadrigeminal cistern from its
stalk, which has cranial and caudal laminae. The
habenular commissure crosses in the cranial
lamina, whereas the posterior commissure crosses
530
Fig. 2. Sagittal (A) and dorsal (B) views of the anatomy of the third ventricle and pineal region. (From Apuzzo MLJ.
Surgery of the third ventricle. 1st edition. Baltimore: Williams & Wilkins; 1987. p. 612; with permission.)
in the caudal lamina. The triangularly shaped

cerebral aqueduct has its base on the posterior
commissure; the other two sides are formed by the
gray matter of the midbrain. When viewed
posteriorly, only the pineal body is visible in the

quadrigeminal cistern. The gland and the rest of
the posterior wall are concealed by the splenium
of the corpus callosum above, the thalami
laterally, and the quadrigeminal plate and cerebellar vermis inferiorly [51,52].
The roof of the posterior third ventricle has
four layers: the fornices superiorly, two thin
membranous layers of tela choroidea, and a potential space between the layers of the tela
choroidea called the velum interpositum. The
body of the fornix is suspended from the body
of the corpus callosum by the septum pellucidum.
Posteriorly, the septum wanes and the body
divides into crura that are directly attached to
the undersurface of the corpus callosum. The
superior layer of the tela choroidea is attached to
the fornix. The inferior layer of the tela choroidea
is attached to the teniae thalami and the pineal
body. The velum interpositum harbors the internal cerebral veins and the medial posterior
choroidal arteries. Although usually a potential
space, it may occasionally communicate with the
quadrigeminal cistern. The choroid plexus of the
third ventricle is suspended from the inferior leaf
of the tela choroidea. The lateral margins of the
roof of the third ventricle are composed of the
choroidal ssure superiorly and the stria terminalis thalami inferiorly [51,53,54].
Two of the three principal approaches to the
posterior third ventricle pass through the posterior incisural space. This anatomically complex
area is synonymous with the terms pineal region
and quadrigeminal cistern. The posterior incisural
space has an anterior wall, a roof, a oor, two
lateral walls, and a posterior apex at the level of
the tentorium. The superior portion of the
anterior wall is formed by the habenular trigone,
the habenular commissure, and the pineal body.
The pineal body overlies the quadrigeminal plate
of the midbrain, which is formed by the paired
superior and inferior colliculi. Inferiorly, the
anterior wall is formed by the lingula of the
vermis in the midline and the superior cerebellar
peduncles laterally. The roof of the posterior
incisural space is formed by the splenium of the
corpus callosum, the crus of the fornix, and the
hippocampal commissure. The oor of this space
is delimited by the anterior-superior surface of the
cerebellum. This space extends inferiorly into the
cerebellomesencephalic ssure. Each lateral wall is
formed by the pulvinar anteriorly, the crura of the
fornices, and the medial surface of the cerebral
hemisphere posteriorly. At the tentorial apex, the
quadrigeminal cistern is separated from the
superior cerebellar cistern by a thick sheet of
arachnoid that contains the precentral cerebellar
vein [2,52,55,56].
531
The posterior incisural space houses the conuence of the deep venous system. The paired
internal cerebral veins exit the velum interpositum
along the superolateral surface of the pineal body.
The union of the internal cerebral veins, forming
the great cerebral vein of Galen, may be located
above or posterior to the pineal body and inferior
or posterior to the splenium. The basal veins of
Rosenthal emanate from the ambient cisterns and
may be received either by the internal cerebral
veins or by the great vein directly. The precentral
cerebellar vein emanates from the cerebellomesencephalic ssure and drains directly into the vein of
Galen. It often receives the superior vermian vein.
After the great vein is formed, it ascends in
a superoposterior direction to join the straight
sinus at the falcotentorial junction [52,54,55].
Selection of a surgical approach
The typical pineal region mass is located in the
midline below the tentorial apex and displaces the
deep venous system superiorly. These anatomic
features give the infratentorial-supracerebellar
approach several natural advantages (Fig. 3)
[57]. The approach provides a midline trajectory
that is ventral to the velum interpositum and deep
venous system. The corridor between the cerebellum and the tentorium does not violate any
normal tissue, and the ability to use the sitting
position minimizes venous bleeding and facilitates
dissection of the tumor from the deep venous
system. In cases where the torcula is low lying or
the lesion has signicant lateral or supratentorial
extension, the occipital-transtentorial approach in
either the sitting or three-quarter prone position is
preferred. This approach provides the widest
exposure of both the supra- and infratentorial
compartments. This approach is also useful if
the tumor has signicant inferior extension into
the cerebellomesencephalic cistern; tumor in this
cistern would be out of view if approached via
the infratentorial corridor. For lesions that are
truly in the posterior third ventricle and do not
extend posterior to the splenium of the corpus
callosum, the interhemispheric-transcallosal approach in the lateral position is useful. This
approach follows the shortest route to the lesion
and has the advantage of working anterior to the
conuence of the deep venous system.
Patient positioning
Numerous patient positions have been described for approaching the pineal region and
532
Fig. 3. Principal approaches to the posterior third ventricle and pineal region. Midsagittal (A) and trajectory (B) views
along the operative corridors (arrows) provided by the occipital-transtentorial (1), infratentorial-supracerebellar (2), and
interhemispheric-transcallosal (3) approaches. (Adapted from Apuzzo MLJ. Surgery of the third ventricle. 1st edition.
Baltimore: Williams & Wilkins; 1987. p. 600; with permission. Adapted from Kaye A, Black P. Operative neurosurgery.
London: Churchill Livingstone; 2000. p. 832; with permission.)
posterior third ventricle. Each position oers

advantages and disadvantages in terms of surgeon
comfort, possible complications, and compatibility with the proposed operative corridor.
The sitting position
The sitting position is preferred for the supracerebellar-infratentorial approach and is also
suitable for the occipital-transtentorial approach
(Fig. 4). The position is initiated with the patient
placed supine on a reversed motorized operating
table. The head is grasped bitemporally with
a three-point pin xation device. The surgeon
supports the head, and the table is manipulated to
obtain a C-shaped conguration between the
patients head and knees. To do so, the table is
brought into the Trendelenburg position with the
back elevated and the foot slightly reclined, which
achieves hip exion. The patients legs are slightly
elevated to assist with venous return. The neck is
then exed to bring the tentorium parallel to the
oor while maintaining two ngers breadth
between the patients chin and sternum. The
headholder is then rigidly xed to a U-shaped
bar extension mounted to the midsection of the
operating table.
The operating microscope is vital to the success
of the procedure. The microscope is balanced with
the objective parallel to the oor. A variable focal
length objective is recommended, but a 275-mm
xed focal length lens provides adequate space
533
between the objective and the eld to permit

insertion of the long instruments that are required. The eye pieces should be rotated slightly
upward to avoid cervical hyperextension for the
surgeon.
The principal advantage of the position is that
gravity aids with cerebellar retraction and dissection of the tumor o the deep venous system.
Additionally, blood and CSF ow out of the
operative eld rather than pool in a eld that is in
a dependent position. Surgeons who are unfamiliar
with the position may nd the elevated operative
eld uncomfortable to work in. A freestanding
table or chair-mounted armrest is essential to
reduce arm fatigue, because the surgeons forearms
are in a nearly vertical position. Most surgeons rest
their elbows or proximal forearms on the armrest
and their wrists on the edge of a suitable selfretaining retractor frame.
Potential complications of the position include
air embolism [58], pneumocephalus [59,60], supratentorial hemorrhage [61], midcervical exion
myelopathy [62,63], and shunt malfunction [64].
Precordial Doppler monitoring and constant
monitoring of end-tidal PCO2 can detect small
amounts of entrained air before it becomes
problematic [65]. The risk of air embolism is
greatest during the craniotomy and dural opening.
All cut bone surfaces must be meticulously
waxed, and venous bleeding must be controlled
before proceeding with deeper dissection. Some
Fig. 4. The sitting position. (From Kaye A, Black P. Operative neurosurgery. London: Churchill Livingstone; 2000.
p. 770; with permission.)
534
anesthesiologists prefer a central venous catheter

for attempting to aspirate air from the
right catrium in the event that a signicant air
embolism is detected, but the utility of this
maneuver is quite limited [66].
The lateral position
Several variations of the lateral position exist
(Figs. 5 and 6). In the lateral decubitus position,
the nondominant right hemisphere is placed in
a dependent position [67]. For the interhemispheric-transcallosal approach, the falx is positioned
parallel to the oor and the head is then brought
into 30 of left lateral exion. This maneuver
allows gravity to retract the nondominant hemisphere while the falx supports the dominant
hemisphere. For the occipital-transtentorial approach, the head is rotated with the nose 30
toward the oor. A more desirable variation of
this position is the three-quarter prone position
[68]. To obtain this position, the patient is rotated
from the supine position to rest on a supporting
roll for the left hemithorax. The right arm is
supported in a sling-like fashion beneath the
patient. A right axillary roll is used to avoid
traction on the brachial plexus. A three-point pin
xation device supports the head in a slightly
extended position with the nose pointing 45 to
the left of perpendicular to the oor. The patient
is securely strapped to the table so that it can be
rotated to gain additional exposure when necessary. The patients legs are elevated slightly to
improve venous return. Although this position is
cumbersome to set up, many surgeons nd it more
comfortable than the sitting position, because the
surgeons hands work in the horizontal plane and
do not need to be extended.
The prone position

The prone position is simple, safe, and suitable
for supratentorial approaches (Fig. 7). The steep
angle of the tentorium makes the position
impractical for the infratentorial approach. The
prone position is particularly useful in the
pediatric population. The position is generally
comfortable for the surgeon, although the operative eld is considerably elevated, making it
dicult for the surgeon to be seated. Using
a microscope in a bridge conguration allows
two surgeons to work together with simultaneous
binocular vision. The Concorde position [69,70],
a variation of the prone position in which the
head is rotated 15 away from the side of the
craniotomy, is particularly useful for the occipital
transtentorial approach. Disadvantages of prone
approaches are that venous drainage is not
facilitated because of the dependent operative
eld and the brain tends to collapse into the tumor
bed during the dissection.
Operative approaches
Infratentorial-supracerebellar approach
The infratentorial-supracerebellar approach
was rst described by Krause [71] in 1926. The
technique was abandoned for many years because
of unacceptable morbidity but was successfully
reintroduced by Stein [57] in 1971 after the advent
of the microsurgical era.
The infratentorial-supracerebellar approach is
usually performed with the patient in the sitting
position (Fig. 8) [34,72]. If a ventriculostomy is
indicated, a catheter may be passed into the
trigone of the lateral ventricle through a burr hole
Fig. 5. (A) The lateral position with 30 of lateral exion. (B) Coronal section demonstrates retraction of the dependent
hemisphere by gravity and support of the opposite hemisphere by the falx. (Adapted from Apuzzo MLJ. Surgery of the
third ventricle. 1st edition. Baltimore: Williams & Wilkins; 1987. p. 6245; with permission.)
535
Fig. 6. The three-quarter prone position. (Adapted from Apuzzo MLJ. Surgery of the third ventricle. 1st edition.
Baltimore: Williams & Wilkins; 1987. p. 627; with permission.)
placed at the intersection of the midpupillary line

and the lambdoid suture. The exposure begins
with a midline incision extending from approximately 6 cm above the external occipital protuberance to the level of the C4 spinous process.
The suboccipital musculature is dissected through
the avascular nuchal ligament and retracted
laterally, usually with a single curved self-retaining retractor placed from above. The suboccipital
plate is exposed to a level just above the foramen
magnum; the suboccipital muscles inserting on the
spinous processes of C1 and C2 are left undisturbed. The craniotomy is centered just below
the torcula. The bony opening must be adequate
to allow sucient illumination from the operating
microscope and to provide access for the long
surgical instruments that are employed. A craniotomy is preferred over a craniectomy because it
reduces the incidence of aseptic meningitis, uid
collections, and posterior fossa syndrome [73].
Slots are drilled over the superior sagittal sinus
approximately 2 cm above the external occipital
protuberance, over the transverse sinus 3 to 4 cm
o the midline, and over the occipital sinus 1 to 2
cm above the foramen magnum. A craniotome is
used to connect the slots, and the rectangular
bone ap is elevated. There should be sucient
bone removal above the transverse sinus to ensure
that the view along the tentorium is not obscured.
Bone edges should be waxed meticulously, and all
venous bleeding should be controlled to protect
against air emboli. The dura is then palpated to

assess the intracranial pressure in the posterior
fossa. If the dura is tense, CSF may be released
from the ventricular drain or the foramen
magnum and mannitol may be administered.
A semilunar durotomy is created by beginning at the lateral aspects of the exposure and
working toward the falx cerebelli and occipital
sinus. If the occipital sinus is patent, it may be
divided between titanium clips. The falx cerebelli
is then divided, and the occipital sinus is secured
with silk ligatures. The dural ap is reected
superiorly and placed in slight tension with
tenting sutures suspended from rubber bands
attached to the retractor frame. Excess retraction
may occlude the transverse sinuses and should
be avoided. Cauterizing and dividing arachnoid
adhesions and bridging veins between the dorsal
surface of the cerebellum and the tentorium
open the infratentorial corridor to the pineal
region. Extensive collateral circulation in the
posterior fossa minimizes the risk of complications from venous sacrice [74]. Care must be
taken not to divide bridging veins too close to
the tentorial surface, because subsequent dural
bleeding can be dicult to control with cautery.
When these attachments are divided, the cerebellum drops away from the tentorium, providing
an excellent corridor with minimal retraction.
The dorsal surface of the cerebellum is protected
with padding (eg, Telfa, The Kendall Compant,
536
Fig. 7. The prone position. (From Kaye A, Black P. Operative neurosurgery. London: Churchill Livingstone; 2000.
p. 613; with permission.)
LTD., Manseld, MA). A small malleable

retractor is used to provide additional cerebellar
retraction in a posterior and inferior direction.
Additional adhesions and bridging veins near the
anterior vermis may be encountered as the
cerebellum is retracted; they, too, are cauterized
and retracted.
Deeper dissection is performed under the
microscope. A freestanding armrest is useful at
this point to support the surgeons arms in an
outstretched position. The opalescent arachnoid
of the quadrigeminal cistern is recognized and
opened using an arachnoid knife, long bayonet
scissors, and irrigating bipolar cautery. The
precentral cerebellar vein is encountered in the
midline extending from the anterior vermis up to
the vein of Galen; it can be cauterized and divided
with minimal risk. In the case of a pineal region
mass, the posterior aspect of the tumor can now
be visualized. Small branches of the choroidal and
superior cerebellar arteries often cover pineal
region tumors. Although these vessels may be
taken with impunity, it is essential to avoid
damaging the vessels of the deep venous system.
The trajectory up to this point has been in line
with the vein of Galen. Because further pursuit of
this trajectory leads to damage of the vein of
Galen and the conuence of the deep venous
system, the trajectory must be redirected several
degrees inferiorly such that it is in line with the
center of the lesion. With the quadrigeminal
arachnoid opened, cerebellar retraction can be
maximized. Typically, it is only necessary to
retract the cerebellum until the inferior portion
of the tumor is visualized. This goal is facilitated
by widely opening the arachnoid laterally where
the cerebellum is tethered.
The tumor is initially internally debulked
through its posterior surface. Specimens are taken
from within the capsule and sent for frozen

section examination. The accuracy of frozen tissue
diagnosis may be unreliable, and this fact should
be taken into consideration during intraoperative
decision making regarding the extent of resection.
Internal debulking then continues with a variety
of instruments, such as suction, cautery, cupped
forceps, and ultrasonic aspiration. Most tumors
are of a soft consistency that is amenable to
removal with pressure-adjustable macrosuction.
After an adequate internal decompression has
been achieved, the capsule can be separated from
the surrounding thalamus. Most vessels along the
capsule wall are of choroidal origin and need not
be preserved. The lateral or central dissection
eventually leads to the third ventricle, which
becomes apparent as CSF oods the operative
eld. The tumor is then dissected o the midbrain.
This maneuver is often the most dicult portion
of the tumor dissection and can be facilitated by
retracting the tumor superiorly and bluntly
dissecting the mass o the collicular plate under
direct vision. The nal tumor attachment is
superiorly along the velum interpositum and deep
venous system. Although these attachments are
cauterized and sharply divided, a rent in the
deep venous system must be avoided. Bleeding
from the deep venous system is profuse, usually
only controllable by direct tamponade, and often
heralds disastrous consequences for the patient.
An experienced surgeon can appreciate the
degree to which a tumor is invasive or encapsulated. Invasive tumors cannot be completely
resected without signicant morbidity. Some
malignant lesions exhibit pseudoencapsulation
that can be exploited for radical removal.
Although radical removal reduces the risk of
postoperative hemorrhage and improves the
ecacy of adjuvant therapy with regard to tumor
537
Fig. 8. Infratentorial-supracerebellar approach. (A) Numeric sequence of incisions of regional arachnoid; the precentral
cerebellar vein is sacriced to permit further relaxation of the anterior vermis. (B) Exposure of a pineal tumor with the
basal veins and vein of Galen overlying the mass. (C) View into the posterior third ventricle after tumor removal.
(Adapted from Apuzzo MLJ. Surgery of the third ventricle. 1st edition. Baltimore: Williams & Wilkins; 1987. p. 582, 586;
with permission.)
recurrence, its impact on long-term survival

remains debatable.
After tumor removal, the surgeon has an
excellent view into the posterior third ventricle.
A exible mirror may be used to inspect the
inferior portion of the tumor bed and to examine
the orice of the cerebral aqueduct. Meticulous
hemostasis must be obtained, because there is
little tissue turgor to tamponade bleeding in the
tumor bed [64]. Direct low-intensity cautery is the
method of choice. Extensive use of hemostatic
agents should be avoided because they may oat
into the ventricle and cause obstructive hydrocephalus, shunt malfunction, or chemical meningitis. If necessary, long strips of Surgicel (Ethicon,
Somerville, NJ) draped over the surface of the
cerebellum and onto the tumor bed can provide
hemostasis at minimal risk.
After hemostasis is obtained, an attempt is
made to close the dura in a watertight fashion.
The craniotomy ap is secured with titanium
miniplates. As a general rule, the more complete
the dural and bone ap closure, the lower is the
risk of postoperative headaches and wound
538
complications. To avoid excessive brain shift, the

patient should be extubated and maintained with
a signicant degree of head elevation and exion.
Interhemispheric-transcallosal approach
First described by Dandy in 1921, the interhemispheric-transcallosal approach established
a corridor between the falx and the hemisphere
along the parieto-occipital junction [75]. Dandys
early contributions to this approach recognized
the importance of the deep venous system and the
cortical bridging veins between the hemisphere
and the superior sagittal sinus [76]. This approach
can be used in a variety of patient positions, but
the lateral or three-quarter prone position is
generally preferred (Fig. 9).
The positioning of the bone ap depends on
where the tumor is positioned within the third
ventricle [77]. A wide craniotomy, approximately

8 cm in length, provides exibility in identifying
a corridor that minimizes the sacrice of bridging
veins. Although a variety of skin incisions may be
used, a U-shaped ap extending across the midline and reected laterally provides adequate
exposure. The craniotomy is centered over the
vertex with the inferior margin being 2 cm above
the apex of the lambdoid suture. Slots are drilled
over the superior sagittal sinus anteriorly and
posteriorly, and a craniotome is used to turn
a generous biparietal craniotomy that is usually
eccentric to the nondominant hemisphere. It is
important that 1 to 2 cm of bone be removed on
the dominant side to facilitate operative illumination. Bleeding from the sinus may be brisk but is
nevertheless under low pressure and easily controlled with hemostatic agents.
Fig. 9. Options for tumor exposure using the posterior interhemispheric corridor. (A) Transcallosal approach. (B)
Retrocallosal and transsplenial variations employing a slightly more posterior trajectory. (Adapted from Apuzzo MLJ.
Surgery of the third ventricle. 1st edition. Baltimore: Williams & Wilkins; 1987. p. 617; with permission.)
A U-shaped durotomy is reected medially

with its base on the sinus. The bridging veins are
inspected, and an approach is chosen that minimizes the number of veins scaried. The depth of
the operative corridor allows for several possible
angles of approach through a small supercial opening; the optimal corridor for a given patient
features as a trajectory that is as close to perpendicular to the corpus callosum as possible. It is
unlikely that sucient exposure can be achieved
without sacricing one bridging vein, but sacricing more than one increases the risk of cortical
venous infarction and should be avoided if
possible. The exposed hemisphere is protected with
Bicol (Codman, Piscataway, NJ) or Telfa, and two
malleable retractors are used to mobilize the parietal lobe laterally in a gentle arc. A third retractor
may be used along the falx, which may be divided
inferiorly to gain additional exposure. There are
generally no adhesions between the medial hemispheric surface and the falx within the interhemispheric ssure, although the cingulate gyri may
adhere to each other below the falx. The remainder
of the dissection is performed under the microscope.
The cingulate gyri are separated, and the
corpus callosum is identied by its striking white
appearance. The paired pericallosal arteries are
noted and, depending on their location, may be
retracted to one side or to each side with separate
retractors. The corpus callosum may be attenuated over the tumor, and a bulge in its dorsal
surface may reveal the location of the tumor.
MRI-based frameless stereotactic guidance may
be helpful for positioning the callosotomy. A
2-cm opening in an anterior-posterior direction
is sucient to remove most tumors. The lateral
extent of the opening is less critical and need only
be limited by what is sucient to resect the tumor
without injuring the pericallosal arteries. In most
cases, the corpus callosum is relatively avascular
and can be opened with cautery and gentle
suction. A 2-cm opening in the posterior body
of the corpus callosum is not likely to lead to
cognitive impairment or a disconnection syndrome. Section through the splenium, although
acceptable in most patients, produces a left hemialexia in some. Splenial section combined with
a left occipital injury or any lesion producing
a right hemianopsia leads to alexia without
agraphia, a severely disabling disconnection syndrome for patients with any degree of literacy [78].
The retractors are now shifted to retract the
medial and lateral margins of the callosotomy. The
dorsal surface of the tumor is visualized through
539
the tela choroidea. This layer is opened, and a small

biopsy is taken for frozen tissue diagnosis. It is
important to identify and preserve the internal
cerebral veins early in the dissection. Lesions in this
region tend to displace the internal cerebral veins
o the midline favorably, although they are
occasionally found overlying the tumor in the
midline separated by only a few millimeters of
connective tissue. It is usually possible to separate
the two veins and displace them laterally; however,
if it is absolutely necessary, one may be divided.
Once the deep venous system has been appreciated,
internal debulking may proceed as described for
the supracerebellar-infratentorial approach. In patients with large masses or signicant hydrocephalus, surprisingly little dissection may be needed to
enter the third ventricle; with small lesions and little
hydrocephalus, the third ventricle is encountered
deep within the operative corridor. As the tumor is
removed, the entire content of the third ventricle is
available for inspection; the foramen of Monro and
the aqueduct of Sylvius provide familiar landmarks. Care must be taken not to damage the
anterior roof of the third ventricle, where the bodies
of the fornices are located. The degree of diculty
in attempting a complete resection is determined in
part by the tumors point of origin. Exophytic
tumors of the thalamus must be pursued with
caution. Tumors stemming from the pineal gland
itself or from the velum interpositum may be
resected more aggressively.
Occipital-transtentorial approach
The occipital-transtentorial approach was
originally described by Jamieson [79] in 1971.
Although the procedure may be performed in the
sitting position, the three-quarter prone position
is generally preferred, because gravity helps to
retract the occipital lobe. In almost all cases, the
exposure is between the right occipital lobe and
falx cerebri. A right-sided approach protects the
dominant visual cortex from the potential for
retraction injury. The principal anatomic advantage of this approach is that no bridging veins
cross from the occipital lobe into the superior
sagittal sinus [80]. This fact limits the risk of
cortical venous infarction that accompanies the
interhemispheric approach as long as the inferior
cerebral vein is preserved. Although this vein
drains much of the occipital lobe, it usually lies in
a safe harbor lateral to the operative exposure.
Division of the tentorium provides excellent
exposure of the collicular plate, thus making the
approach well suited for tumors with substantial
540
Fig. 10. Occipital-transtentorial approach. The location of the deep venous system when the tumor arises form the
pineal gland (A), the falcotentorial junction (B), and the quadrigeminal plate (C). (Adapted from Apuzzo MLJ. Surgery
of the third ventricle. 1st edition. Baltimore: Williams & Wilkins; 1987. p. 603; with permission.)
inferior extension. A potential disadvantage of the

approach is that it uses an oblique trajectory for
lesions that are essentially midline, creating the
potential for the inexperienced surgeon to become
disoriented. As with other approaches to the
pineal region, mannitol and ventricular drainage
are useful for gaining brain relaxation.
The approach is initiated with a U-shaped right

occipital scalp ap that is reected inferiorly. The
medial limb of the incision is placed just to the left
of the midline, beginning at the level of the torcula.
A slot is drilled across the superior sagittal sinus
just above the torcula, and a second slot is drilled
6 to 10 cm higher in the midline. A craniotome is
used to turn a generous craniotomy extending 1 to

2 cm across the midline. The dura may be opened
as a U-shaped ap based on the sagittal sinus or as
a pair of triangular leaves based on the sagittal
sinus and the inferior margin of the craniotomy.
Dissection within the occipital corridor is
performed under the microscope. With gravity
retracting the occipital lobe, the straight sinus is
identied and the tentorium is divided adjacent to
it. The initial incision is made near the junction of
the straight sinus and the torcula and is carried
forward toward the incisura. The tentorium may
contain venous channels that can be controlled
with bipolar cautery or hemoclips. A retractor can
be placed on the falx to gain additional exposure;
the falx and inferior sagittal sinus may be divided
to gain further exposure. Retention sutures may be
placed on the cut tentorial edge to retract it
laterally. At this point, the arachnoid overlying the
tumor, the galenic system, and the quadrigeminal
cistern are visualized.
The arachnoid over the pineal region is dense,
and great care must be taken not to injure the
deep venous system during the dissection. The
dissection is performed sharply with an arachnoid
knife and microscissors. The vein of Galen is
usually encountered rst, followed by the right
basal vein of Rosenthal, the internal cerebral
veins, and the precentral cerebellar vein. A rightsided arachnoid dissection is often sucient to
resect the lesion; left-sided dissection and division
of the precentral cerebellar vein may be added for
a more generous exposure. Once the arachnoid
dissection is complete, lesions of the superior
vermis, collicular plate, posterior third ventricle,
pineal gland, and splenium are readily accessible.
Most of the aforementioned lesions displace
the deep venous system posteriorly and superiorly
toward the surgeon. This arrangement facilitates
identication of the deep veins and necessitates
that tumor dissection proceed between them. In
contrast, lesions that arise from the posterior leaf
of the tela choroidea, the free edge of the
tentorium, or the falcotentorial junction displace
the galenic system anteriorly or inferiorly and out
of the surgeons view. In this case, the tumor must
be internally debulked, knowing that these critical
veins are located just beyond the deep capsule.
After an adequate decompression is achieved, the
tumor capsule may be cautiously dissected and the
underlying veins identied. Generally speaking,
capsular dissection should begin laterally and
inferiorly and proceed medially and superiorly.
The tumor type, specic location, and degree of
541
invasiveness determine the degree of resection that

can be achieved (Fig. 10).
Intraoperative tumor management
For the typical pineal region tumor, the lesion
should be biopsied and a frozen section diagnosis
obtained. When a germinoma is identied, the
value of aggressive resection is controversial,
given the radiosensitivity of this tumor, and this
must be considered when deciding how much
tumor to remove. For invasive tumors, such as
glioma of the brain stem or thalamus, debulking
should proceed cautiously, because a gross total
resection may be associated with a high likelihood
of incurring a neurologic decit. Re-establishing
the CSF circulation is a reasonable goal.
With benign tumors (eg, teratoma, well-dierentiated pineocytoma, pilocytic astrocytoma, dermoid), the surgeon must strive for a gross total
resection in either piecemeal or en bloc fashion. If
a total resection is achieved, the third ventricle
should be inspected to ensure that there is no
obstruction of the CSF circulation.
If the lesion is a meningioma of the velum
interpositum or the tentorium, a complete resection is desirable. These tumors are usually
benign in their behavior; accordingly, the surgeon
must rely on good judgment when eorts to achieve
a total resection may threaten the patients
neurologic function.
If the lesion extends inferiorly under the
superior cerebellar vermis, this structure may be
divided to facilitate dissection and removal of the
mass. If the tumor extends superiorly into the
splenium, it may be possible to mobilize the capsule
without sacricing the splenium. If unavoidable,
the splenium may be divided, but restraint should
be exercised, because a partial disconnection syndrome may follow.
Dissection may proceed toward a goal of total
resection as long as the surgeon can identify
a plane between the tumor and surrounding
normal structures. The adequacy of a subtotal
resection is a matter of judgment and experience.
Possible goals of a subtotal resection beyond
diagnosis include cytoreduction in preparation for
adjuvant therapy, relief of mass eect, and reestablishment of the CSF circulation.
Postoperative care
The most signicant immediate problems include bleeding within the tumor bed, hydrocephalus, shunt malfunction, pneumocephalus, and
542
subdural hematoma [64]. All these are potentially

reversible problems; thus, vigilant clinical examination and a low threshold for obtaining a postoperative CT scan are warranted. Tumor bed
hemorrhage is usually only seen in patients with
invasive tumors that are subtotally resected. De
novo hydrocephalus may arise as a result of
operative debris within the third ventricle; likewise, such debris may lead to a shunt malfunction
or failure of a third ventriculostomy. Pneumocephalus is a frequent consequence of the sitting
position that may lead to postoperative confusion,
but its course is generally self-limiting [59]. Subdural hematomas may occur from shifting of the
brain within the cranial vault and tearing of
bridging veins. Most often associated with the
sitting position, acute postoperative subdural
hematomas generally need to be evacuated. In
contrast, hygromatous collections can usually be
managed conservatively.
Patients should remain on high-dose steroids
until their clinical condition is stable. Generally
a 2-week Dexamethasone (Decadron) taper is
advisable. A slow steroid taper also mitigates
against the development of aseptic ventriculitis.
Patients who have undergone supratentorial
approaches benet from perioperative seizure
prophylaxis, although long-term use of antiepileptic agents is not warranted.
Complications
Most patients display some degree of impairment of extraocular movements, particularly upgaze and diculty with convergence [81]. Any
preoperative oculomotor decit should be expected
to be exacerbated by the surgery. Fortunately,
these problems are generally transient and tend to
resolve over weeks to months. Permanent decits
are rare, although a mild limitation of upgaze may
persist, usually with limited clinical consequences.
Complications resulting from brain retraction
may occur with all three approaches. After the
supracerebellar-infratentorial approach, a brief
period of ataxia may occur. Rarely, cerebellar
infarction has been observed [74]. In the parietal
region, brain retraction during the transcallosal
approach may lead to contralateral sensory decits
or, rarely, to hemiparesis. Sacrice of bridging
veins may lead to cortical infarction, particularly if
more than one vein is taken. Occipital lobe
retraction during the transtentorial approach
may result in cortical visual eld decits [81].
Disconnection syndromes may occur if the splenium is divided [78].

Summary
A variety of surgical approaches to the
posterior third ventricle and pineal region exist.
The choice of approach is inuenced by the exact
location of the lesion, its expected pathologic
ndings, and the comfort level of the operating
surgeon with the approach that is being considered. For most pineal region masses that are
situated in the midline below the deep venous
system, we favor the supracerebellar-infratentorial
approach in the sitting position. For pineal region
lesions that displace the deep venous system
inferiorly or have signicant lateral extension,
we prefer the occipital-transtentorial approach in
the three-quarter prone or sitting position. For
lesions that are truly in the posterior third
ventricle without extension posterior to the
splenium, we prefer the interhemispheric-transcallosal approach in the lateral position.
References
[1] Horrax G. Extirpation of a huge pinealoma from
a patient with pubertas praecox. Arch Neurol
Psychiatry 1937;37:38597.
[2] Kawashima M, Rhoton AL Jr, Matsushima T.
Comparison of posterior approaches to the posterior incisural space: microsurgical anatomy and
proposal of a new method, the occipital bi-transtentorial/falcine approach. Neurosurgery 2002;51:
120820.
[3] Ogata N, Yonekawa Y. Paramedian supracerebellar
approach to the upper brain stem and peduncular
lesions. Neurosurgery 1997;40:1014.
[4] Poppen JL. The right occipital approach to a pinealoma. J Neurosurg 1966;25:70610.
[5] Sekhar LN, Goel A. Combined supratentorial and
infratentorial approach to large pineal-region meningioma. Surg Neurol 1992;37:197201.
[6] Van Wagenen WP. A surgical approach for the
removal of certain pineal tumors. Surg Gynecol
Obstet 1931;53:21620.
[7] Yonekawa Y, Imhof HG, Taub E, et al. Supracerebellar transtentorial approach to posterior temporomedial structures. J Neurosurg 2001;94:33945.
[8] Bruce JN. Management of pineal region tumors.
Neurosurg Q 1993;3:10319.
[9] Hirato J, Nakazato Y. Pathology of pineal region
tumors. J Neurooncol 2001;54:23949.
[10] Freedman SJ, Pantanowitz L, Joseph JT, et al.
Unusual locations for lymphomas. Case 2. Pineal
lymphoma. J Clin Oncol 2001;19:29603.

[11] Ramina R, Coelho NM, Mariushi WM, et al. Pineal
metastasis as rst clinical manifestation of colorectal adenocarcinoma. Case report. Arq Neuropsiquiatr 1999;57:925.
[12] Smith WT, Hughes B, Ermocilla R. Chemodectoma
of the pineal region, with observations on the pineal
body and chemoreceptor tissue. J Pathol Bacteriol
1966;92:6976.
[13] Suzuki T, Yasumoto Y, Kumami K, et al. Primary
pineal melanocytic tumor. Case report. J Neurosurg
2001;94:5237.
[14] Fetell MR, Bruce JN, Burke AM, et al. Nonneoplastic pineal cysts. Neurology 1991;41:103440.
[15] Do HM, Marx WF, Khanam H, et al. Choroid
plexus papilloma of the third ventricle: angiography, preoperative embolization, and histology.
[16] Nakano I, Kondo A, Iwasaki K. Choroid plexus
papilloma in the posterior third ventricle: case
report. Neurosurgery 1997;40:127982.
[17] Tsumoto T, Nakai E, Uematsu Y, et al. Choroid
plexus papilloma in the posterior third ventricle in
infancy: a case report [in Japanese]. No Shinkei
Geka 1999;27:6738.
[18] Konovalov AN, Spallone A, Pitzkhelauri DI. Pineal epidermoid cysts: diagnosis and management.
[19] MacKay CI, Baeesa SS, Ventureyra EC. Epidermoid cysts of the pineal region. Childs Nerv Syst
1999;15:1708.
[20] Usanov EI, Hatomkin DM, Nikulina TA, et al.
Craniopharyngioma of the pineal region. Childs
Nerv Syst 1999;15:47.
[21] Konovalov AN, Spallone A, Pitzkhelauri DI.
Meningioma of the pineal region: a surgical series
of 10 cases. J Neurosurg 1996;85:58690.
[22] Mallucci CL, Obukhov S. Successful removal of
large pineal region meningiomas: two case reports.
[23] Rozario R, Adelman L, Prager RJ, et al. Meningiomas of the pineal region and third ventricle.
[24] Matsuda Y, Inagawa T. Surgical removal of pineal
region meningiomathree case reports. Neurol
Med Chir (Tokyo) 1995;35:5947.
[25] Engel U, Gottschalk S, Niehaus L, et al. Cystic
lesions of the pineal regionMRI and pathology.
[26] Kojima Y, Kuwana N. Progressive diuse arteriovenous malformationcase report. Neurol Med
Chir (Tokyo) 1993;33:2425.
[27] Konovalov AN, Pitskhelauri DI, Arutiounov NV.
Surgical treatment of the thrombosed vein of Galen
aneurysm. Acta Neurochir (Wien) 2002;144:90915.
[28] McComb JG, Apuzzo ML. Operative management
of malformations of the vein of Galen. In: Apuzzo
ML, editor. Surgery of the third ventricle. 2nd
edition. Baltimore: Williams & Wilkins; 1998.
p. 77986.
543
[29] Kobayashi S, Kamagata M, Nakamura M, et al.

Pineal apoplexy due to massive hemorrhage associated with cavernous angioma: case report. Surg
Neurol 2001;55:36571.
[30] Vhora S, Kobayashi S, Okudera H. Pineal cavernous angioma presenting with Parkinsonism. J Clin
Neurosci 2001;8:2636.
[31] Chapman PH, Linggood RM. The management of
pineal area tumors: a recent reappraisal. Cancer
1980;46:12537.
[32] Linggood RM, Chapman PH. Pineal tumors.
J Neurooncol 1992;12:8591.
[33] Neuwelt EA. An update on the surgical treatment
of malignant pineal region tumors. Clin Neurosurg
1985;32:397428.
[34] Stein BM, Bruce JN. Surgical management of pineal
region tumors (honored guest lecture). Clin Neurosurg 1992;39:50932.
[35] Itoyama Y, Kochi M, Kuratsu J, et al. Treatment of
intracranial nongerminomatous malignant germ cell
tumors producing alpha-fetoprotein. Neurosurgery
1995;36(3):45964.
[36] Vervenne WL, Bakker PJ, Stalpers LJ, et al.
Malignant intracranial germ cell tumor treated with
chemotherapy and radiotherapy without histopathological conrmation [in Dutch]. Ned Tijdschr
Geneeskd 2000;144:52731.
[37] Chandrasoma PT, Smith MM, Apuzzo ML.
Stereotactic biopsy in the diagnosis of brain
masses: comparison of results of biopsy and resected surgical specimen. Neurosurgery 1989;24:
1605.
[38] Field M, Witham TF, Flickinger JC, et al. Comprehensive assessment of hemorrhage risks and outcomes after stereotactic brain biopsy. J Neurosurg
2001;94:54551.
[39] Peragut JC, Dupard T, Graziani N, et al. Prevention of risk in stereotaxic biopsy of various
tumors of the pineal region. Apropos of 3 cases [in
French]. Neurochirurgie 1987;33:237.
[40] Krieger MD, Chandrasoma PT, Zee CS, et al. Role
of stereotactic biopsy in the diagnosis and management of brain tumors. Semin Surg Oncol 1998;
14:1325.
[41] Regis J, Bouillot P, Rouby-Volot F, et al. Pineal
region tumors and the role of stereotactic biopsy: review of the mortality, morbidity, and diagnostic rates in 370 cases. Neurosurgery 1996;39:
90712.
[42] Korogi Y, Takahashi M, Ushio Y. MRI of pineal
region tumors. J Neurooncol 2001;54:25161.
[43] Liang L, Korogi Y, Sugahara T, et al. MRI of
intracranial germ-cell tumours. Neuroradiology
2002;44:3828.
[44] Muller-Forell W, Schroth G, Egan PJ. MR imaging
in tumors of the pineal region. Neuroradiology
1988;30:22431.
[45] Satoh H, Uozumi T, Kiya K, et al. MRI of pineal
region tumours: relationship between tumours and
544
[46]
[47]
[48]
[49]
[50]
[51]
[52]
[53]
[54]
[55]
[56]
[57]
[58]
[59]
[60]
[61]
[62]
[63]

adjacent structures. Neuroradiology 1995;37:624
30.
Tien RD, Barkovich AJ, Edwards MS. MR imaging
of pineal tumors. AJR Am J Roentgenol 1990;155:
14351.
Ganti SR, Hilal SK, Stein BM, et al. CT of pineal
region tumors. AJR Am J Roentgenol 1986;146:
4518.
Rhoton AL Jr. Microsurgical anatomy of the third
ventricular region. In: Apuzzo ML, editor. Surgery
of the third ventricle. 2nd edition. Baltimore:
Williams & Wilkins; 1998. p. 89158.
Rhoton AL Jr. The lateral and third ventricles.
Neurosurgery 2002;51(Suppl):S20771.
Wen HT, Rhoton AL Jr, de Oliveira E. Transchoroidal approach to the third ventricle: an anatomic study of the choroidal ssure and its clinical
application. Neurosurgery 1998;42:120517.
Yamamoto I, Rhoton AL Jr, Peace DA. Microsurgery of the third ventricle:. part I. Microsurgical
anatomy. Neurosurgery 1981;8:33456.
Ono M, Ono M, Rhoton AL Jr, et al. Microsurgical
anatomy of the region of the tentorial incisura.
J Neurosurg 1984;60:36599.
Fujii K, Lenkey C, Rhoton AL Jr. Microsurgical
anatomy of the choroidal arteries: lateral and third
ventricles. J Neurosurg 1980;52:16588.
Ono M, Rhoton AL Jr, Peace D, et al. Microsurgical anatomy of the deep venous system of the
brain. Neurosurgery 1984;15:62157.
Rhoton AL Jr. Tentorial incisura. Neurosurgery
2000;47(Suppl):S13153.
Rhoton AL Jr, Yamamoto I, Peace DA. Microsurgery of the third ventricle:. part 2. Operative
approaches. Neurosurgery 1981;8:35773.
Stein BM. The infratentorial supracerebellar approach to pineal lesions. J Neurosurg 1971;35:
197202.
Porter JM, Pidgeon C, Cunningham AJ. The sitting
position in neurosurgery: a critical appraisal. Br J
Anaesth 1999;82:11728.
Di Lorenzo N, Caruso R, Floris R, et al. Pneumocephalus and tension pneumocephalus after posterior fossa surgery in the sitting position: a
prospective study. Acta Neurochir (Wien) 1986;
83:1125.
Kishan A, Naidu MR, Muralidhar K. Tension
pneumocephalus following posterior fossa surgery
in sitting position. A report of 2 cases. Clin Neurol
Neurosurg 1990;92:2458.
Seiler RW, Zurbrugg HR. Supratentorial intracerebral hemorrhage after posterior fossa operation.
Haisa T, Kondo T. Midcervical exion myelopathy
after posterior fossa surgery in the sitting position:
case report. Neurosurgery 1996;38:81921.
Nitta H, Yamashita J, Nomura M, et al. Cervical
spinal cord infarction after surgery for a pineal
[64]
[65]
[66]
[67]
[68]
[69]
[70]
[71]
[72]
[73]
[74]
[75]
[76]
[77]
[78]
region choriocarcinoma in the sitting position: case

report. Neurosurgery 1997;40:10825.
Bruce JN, Stein BM. Complications of surgery for
pineal region tumors. In: Post KD, Friedman ED,
McCormick PC, editors. Postoperative complications in intracranial neurosurgery. New York:
Thieme Medical Publishers; 1993. p. 7486.
Morcos JJ, Heros RC. Distal basilar artery
aneurysm: surgical techniques. In: Batjer HH,
Caplan LR, Friberg L, et al, editors. Cerebrovascular disease. Philadelphia: Lippincott-Raven; 1997.
p. 105577.
Motomatsu K, Adachi H, Uno T, et al. Evaluation
of catheter placement for treatment of venous air
embolism in the sitting position. Fukuoka Igaku
Zasshi 1979;70:6671.
Apuzzo ML, McComb JG. The lateral decubitus
position for the surgical approach to pineal location tumors. Concepts Pediatr Neurosurg 1988;8:
18699.
Ausman JI, Malik GM, Dujovny M, et al. Threequarter prone approach to the pineal-tentorial
region. Surg Neurol 1988;29:298306.
Kobayashi S, Sugita K, Tanaka Y, et al. Infratentorial approach to the pineal region in the prone
position: Concorde position. Technical note. J
Neurosurg 1983;58:1413.
Kyoshima K. Arm-down Concorde position: a technical note. Surg Neurol 2002;57:4435.
Krause F. Operative Frielegung der Vierhugel,
nebst Beobachtungen uber Hirndruck und Dekompression. Zentralbl Chir 1926;53:28129.
Bruce JN, Stein BM. Surgical management of pineal
region tumors. Acta Neurochir (Wien) 1995;134:
13035.
Gnanalingham KK, Lafuente J, Thompson D, et al.
Surgical procedures for posterior fossa tumors in
children: does craniotomy lead to fewer complications than craniectomy? J Neurosurg 2002;97:
821826.
Ueyama T, Al Mefty O, Tamaki N. Bridging veins
on the tentorial surface of the cerebellum: a microsurgical anatomic study and operative considerations. Neurosurgery 1998;43:113745.
Dandy WE. An operation for the removal of pineal
tumors. Surg Gynecol Obstet 1921;33:1139.
Dandy WE. Operative experience in cases of pineal
tumor. Arch Surg 1936;33:1946.
McComb JG, Levy M, Apuzzo ML. The posterior
intrahemispheric, retrocallosal and transcallosal approaches to the third ventricle region. In: Apuzzo
ML, editor. Surgery of the third ventricle. 2nd
edition. Baltimore: Williams & Wilkins; 1998. p.
74377.
Bogen JE. Physiological consequences of complete
or partial commissural section. In: Apuzzo ML,
editor. Surgery of the third ventricle. 2nd edition.
Baltimore: Williams & Wilkins; 1998. p. 16786.

[79] Jamieson KG. Excision of pineal tumors. J Neurosurg 1971;35:5503.
[80] Lazar ML, Clark K. Direct surgical management of
masses in the region of the vein of Galen. Surg
Neurol 1974;2:1721.
545
[81] Nazzaro JM, Shults WT, Neuwelt EA. Neuroophthalmological function of patients with pineal region tumors approached transtentorially in
the semisitting position. J Neurosurg 1992;76:
74651.
Endoscopic adjuncts to intraventricular surgery

Sandeep Kunwar, MD
Department of Neurosurgery, University of California at San Francisco School of Medicine, 505 Parnassus Avenue,
Box 0112, San Francisco, CA 94143, USA
The cerebral ventricles, located deep within the

brain substance and lled with clear cerebrospinal uid (CSF), are ideally suited for endoscopic surgery. Endoscopic neurosurgery was rst
introduced in the early 1900s as a means of
diagnosing and treating hydrocephalus. It was
abandoned for many years because of poor
instruments and poor results. Recently, there has
been a resurgence of interest in neuroendoscopy
for the management of ventricular pathology
as the result of a variety of technical advances,
including improved optics, illumination, and
miniaturized instruments. Minimally invasive
neurosurgery is a concept that holds great appeal,
and with the availability of improved instrumentation and surgical experience, this concept is
currently practical in the clinical setting. The
indications and limitations for endoscopic neurosurgery are still being dened, however. Endoscopes have been used successfully within the
cerebral ventricles to treat ependymal and arachnoid cysts, simplify complex loculated hydrocephalus, and treat obstructive hydrocephalus as well
as to biopsy and even resect a variety of intraventricular cysts and tumors. This article reviews the
history of endoscopic surgery and focuses on its
current uses in the management of intraventricular pathology.
History
The rst recorded endoscopic neurosurgical
procedure was in 1910 by a urologist in Chicago,
who introduced a small cystoscope in the ventricle
E-mail address: kunwars@neurosurg.ucsf.edu
and cauterized the choroid plexus bilaterally in

two infants [1]. Walter Dandy, who is considered the father of neuroendoscopy, began using
neuroendoscopy in 1918 for the management of
hydrocephalus. The initial instrument was a speculum to visualize the lateral ventricles. By 1922,
Dandy [2] reported using a Kelley cystoscope to
inspect the lateral ventricle in two patients and to
coagulate the choroid plexus in one. He coined the
term ventriculoscope, an instrument that provided
clear visualization of the ventricle, but limitations
in the primitive instruments available at that time
resulted in little advancement in the following
decades. Dandy was also the rst to perform third
ventriculostomy, but he used an open procedure
to fenestrate the oor of the third ventricle. In
1923, Mixter [3] reported the rst successful
endoscopic third ventriculostomy (ETV). After the
development of the rst valved shunt for the
control of hydrocephalus in 1949 by Nulsen and
Spitz [4], there was little development in the
science of neuroendoscopy.
In 1970, led by the eorts of the physicist
Harold Hopkins at the University of Reading in
England, the rst major advance was made in
neuroendoscopy. Hopkins developed the solid rod
lens system, which was the prototype of the rigid
endoscopes in use today. This signicantly improved the illumination, eld of view, and image
resolution, leading to the development of smaller
endoscopes. Hopkins was also credited with the
development of the exible endoscope. Fukushima et al [5] introduced the ventriculoberscope
in 1973, which had a 4-mm outer diameter and
a exible tip that could be navigated around
corners within the ventricles. The miniaturization of optical technology has made it possible to
use endoscopic techniques within the ventricles
doi:10.1016/S1042-3680(03)00063-9
548
S. Kunwar / Neurosurg Clin N Am 14 (2003) 547557
and subarachnoid space through exposures small

enough to cause minimal disruption of the brain.
The indications for neuroendoscopy in the treatment of various disease processes are currently
being dened. Neuroendoscopy can be used as
an adjunct to conventional open microsurgical
procedures (endoscope-assisted procedure) or has
some dened and developing roles as a standalone tool (endoscopic neurosurgery).
Anatomy
Understanding the endoscopic anatomy is
critical, because the view diers from that of
conventional open surgical approaches. In an
endoscopic exposure, only a small portion of the
anatomy can be visualized at any one time. Because
of the limited eld of view, the surgeon must keep in
mind that important neural and vascular structures
may be located adjacent to the endoscope but
outside the exposed eld. The absence of stereoscopic vision results in loss of depth perception.
The surgeon must learn to use dierent cues, such
as shadows and image brightness, as well as acquire
an understanding of the predicted anatomy so as to
estimate depth and navigate the ventricular system.
In many cases, particularly for fenestration of cysts
and where prior surgery has taken place, the use of
neuronavigational systems with endoscopy becomes helpful.
The most common approach to the lateral or
third ventricle is through a standard coronal burr
hole in the midpupillary line, typically on the right
side. Often, to improve the angle of approach,
if visualization of the posterior lateral horn or
middle to posterior portion of the third ventricle is
necessary, a burr hole placed 1 to 3 cm anterior to
the coronal suture and approximately 3 cm from
midline is used. Before introducing the endoscope
into the lateral ventricle, the orientation of the
camera is adjusted such that movements of the
endoscope are concordant to the direction of
movement on the monitor. Entrance into the
ventricle can be done freehand using anatomic
landmarks (following a perpendicular trajectory
from the brain surface) or with the use of
stereotaxy (frame or frameless). In cases where
the ventricles are small, stereotaxy is important in
minimizing the complications of passing a 4- to
6-mm introducer sheath. In most cases, I prefer
to use frameless CT-based stereotaxy to conrm
the trajectory to the target site.
On entering the ventricle, the rst landmark to
identify is the foramen of Monro. The foramen can
often be found by following the choroid plexus

anteriorly to where it turns at the foramen to sit on
the roof of the third ventricle. If the choroid plexus
is not seen, the endoscope may be pointed far
anterior into the frontal horn. The foramina are
paired structures, each serving to connect the
lateral ventricle to the third ventricle. Each
foramen has important neural and vascular relations. The head of the caudate nucleus is situated
laterally, and the septum pellucidum is located
medially. The choroid plexus of the lateral
ventricle projects forward to the foramen of
Monro and passes through the foramen before
turning posteriorly to lie under the roof of the third
ventricle. The septal vein, located anteromedially,
merges with the thalamostriate vein, located
posterolaterally, at the foramen of Monro. The
joined veins then pass through the foramen and
bend posteriorly to form the internal cerebral vein,
which runs in the tela choroidea of the third
ventricle. Within the body of the lateral ventricle, if
the surgeon becomes disoriented, following the
veins that become larger toward the foramen or the
choroid plexus will lead the surgeon to the foramen
of Monro, which will help to reorient the anatomy.
Often with long-standing hydrocephalus, the
septum pellucidum can be widely fenestrated, and
the left foramen may be seen before the right
foramen despite a right-sided entry. Understanding the venous anatomy in relation to the choroid
plexus prevents mistaking the left foramen of
Monro for the right. The location of the thalamostriate vein in reference to the choroid plexus
predicts the anatomic side (Fig. 1).
The anterior columns of the fornices curve
ventrally and downward and make up the superomedial and anterior borders of the foramen of
Monro. The fornices are paired C-shaped white
bundles of axons that represent the major eerent
output of the hippocampus, terminating in the
mammillary bodies. Damage to the fornix during
manipulation of the endoscope passing through
the foramen can result in memory loss; however,
this is uncommon in unilateral approaches.
On passing the endoscope through the foramen
of Monro, the oor of the third ventricle is
visualized. There are several recesses and prominences along the oor that serve as landmarks.
The mammillary bodies, which appear as two
whitish prominences, serve as key landmarks
(Fig. 2). Just anterior to the mammillary bodies
is the oor of the third ventricle, the tuber
cinereum. In cases of hydrocephalus, the oor
can be attenuated and translucent, which helps to
549
Posterior to the mammillary bodies lies the

superior portion of the tegmentum leading into
the aqueduct of Sylvius. Further posterior along
the oor of the third ventricle is the posterior
commissure. These structures may be dicult to
visualize with a 0 lens inserted through a coronal
approach. Use of a 30 angled lens improves
visualization. If the target is along the posterior
aspect of the third ventricle, a more anterior
approach should be used (23 cm anterior to the
coronal suture). Alternatively, a exible endoscope can be used and steered to achieve better
visualization of structures that are hard to reach,
such as those in the posterior third ventricle, as in
aqueductoplasty or tumor biopsy.
Endoscopic equipment
Fig. 1. (A) Endoscopic view of the right foramen of

Monro. The septal vein is to the right of the choroid
plexus. The thalamostriate vein runs lateral or adjacent
to the choroid plexus. C, choroid plexus; F, fornix; FM,
foramen of Monro; S, septal vein; T, thalamostriate
vein. (B) View of the left foramen of Monro lled with
a colloid cyst. The septal vein is to the left of the choroid
plexus. C, choroid plexus; F, foramen of Monro; S,
septal vein; T, colloid cyst.
outline the anterior margins of the mammillary

body. Often, the tip of the basilar artery can be
visualized through the translucent membrane.
Staying midline, the anterior portion of the tuber
cinereum is a region of increased vascularity into
a recess representing the infundibular recess,
which descends into the pituitary stalk. Further
anteriorly, a horizontal band of white bers
represents the posterior aspect of the optic chiasm.
Above this is the optic recess and then the lamina
terminalis. It may be dicult to visualize the
anterior margins of the third ventricle with a 0
endoscope; however, the anatomy is clearly visible
with a 30 scope.
The two most common types of endoscopes

used for ventricular access are a rigid rod lens
scope and a rigid or exible beroptic endoscope.
The highest resolution is through a rigid rod lens
endoscope, which provides brilliant images on the
monitor. Lens diameters range from 2 to 4 mm
(Karl Storz GmbH & Company, Tuttlingen,
Germany; Codman, Randolph, MA; Aesculap,
Tuttlingen, Germany). Rod lens scopes are introduced with an endoscope sheath or introducer
containing working channel(s) as well as irrigation
and drainage channels. Illumination is provided
by a high-intensity xenon light source transmitted
to the endoscope through a beroptic cable. A
microchip camera (Stryker, San Jose, CA; Karl
Storz GmbH & Company) is tted over the
eyepiece of the lens. The camera digitizes the
images that are electronically transmitted by cable
to an image processor for projection onto a television monitor. Because the camera is mounted to
the rod lens system, the device can become
quite heavy to hold and maneuver; thus, a xed
camera holder is useful. Hydraulic-based endoscope holders are commercially available (Unitrac, Aesculap; Mitaka, Mitaka Kohki Company,
Tokyo, Japan), and some include micromanipulators for ne adjustments in the angle or position
of the scope. The superior image quality makes
the rod lens system the workhorse for endoscopic
surgery.
Fiberoptic endoscopes can either be rigid
(Medtronics, Goleta, CA) or exible. Rigid beroptic endoscopes are also designed with working
channels. They are introduced into the ventricle
through a sheath, which allows for easy entry and
removal of the scope during surgery. This also
550
Fig. 2. (A) Endoscopic view of the oor of the third ventricle through the right foramen of Monro. The mamillary bodies
are prominences that serve as landmarks for the tuber cinereum, which lies anteriorly. The oor may be translucent as in
(C) or opaque as in (A). The ventriculostomy can be seen anterior to the left mammillary body, with the basilar artery
lying just underneath and visible through the fenestration. B, basilar artery; M, mamillary bodies. (B) Visualizing more
anteriorly, the ventriculostomy is now in the inferior view, with the infundibular recess seen anteriorly associated with
hypervascularity of the oor and the posterior aspect of the chiasm further anterior. C, chiasm; I, infundibular recess. (C)
In another case, the oor of the third ventricle is translucent, allowing immediate visualization of basilar artery and
dorsum sella. B, basilar artery; D, dorsum sella. (D) After fenestration of the oor with bipolar cautery, a Fogarty
balloon is passed through the opening to dilate the opening. F, Fogarty balloon.
provides adequate drainage of uid during continuous irrigation. The main advantage to the
rigid beroptic endoscope is the light weight and
small size of the endoscope, because the images
are carried by beroptic cable to a camera
attached distant to the operative site. In essence,
the surgeon only holds the beroptic cable
traversing through a holder with a working and
irrigation channel and a beroptic light source.
The camera digitizes the images, which are then
processed for projection onto a television monitor
o of the surgical eld. The image quality is
dependent on the density of beroptic cables and
currently does not reach the resolution of rod lens

systems.
Irrigation is also important. I prefer to use
a pump irrigation system with a foot control.
Adjusting the ow rate of irrigation allows debris
removal or, more importantly visualization of
sites of bleeding. Often, minor bleeding can be
controlled with irrigation alone, and the operation
can continue. Whenever an irrigation pump is
used, it is important to conrm free egress of uid
out of the ventricles either through a drainage
channel in the endoscope sheath or around the
endoscope if a peel-away sheath is used to
cannulate the ventricles. Several endoscopic tools

that allow for a greater degree of tissue manipulation are now available. Monopolar and bipolar
cautery and a beroptic laser introduced into
the working channel are available for controlling bleeding or for tissue coagulation. A host of
biopsy and grasping forceps, angled dissectors,
and microscissors are also available for tissue
manipulation. In most cases, the instruments are
brought in through the working channel and
thus are used coaxial to the line of sight. This is
sucient for most neuroendoscopic procedures;
however, in certain cases, such as a complicated
colloid cyst, I have found biportal access useful. A
second sheath (2 mm) is introduced either anterior
or posterior to the endoscope. The endoscopic
instruments can be inserted through the second
portal under direct visual guidance. This allows
for two instruments to work in concert and
permits tissue manipulation without the need to
move the eld of view.
Endoscopic third ventriculostomy
Since Walter Dandy rst recognized and
described the pathologic importance of aqueduct
blockage in the etiology of noncommunicating
hydrocephalus in 1922, third ventriculostomy has
been performed by means of an open approach,
stereotactic approach, and endoscopic direct vision
[3,68]. The development of a biocompatible
shunt at a time when endoscopic instrumentation
was limited led to ventriculoperitoneal shunting as
the primary treatment of noncommunicating
hydrocephalus. The recognition of the long-term
complications of shunts and signicant improvements in optical technology led to a revival of
interest in third ventriculostomy, particularly
fueled by the Hopkins rod lens system in 1975
[9]. Currently, ETV has increasingly become
accepted as rst-line management for patients
with obstructive hydrocephalus.
Patient selection is important in determining
the benets of a third ventriculostomy. Success
of therapy (dened by clinical improvement and
avoidance of a ventriculoperitoneal shunt) has
ranged from 52% to 92%. The most favorable
outcomes occur in patients with aqueductal
stenosis and benign space-occupying lesions of
the midbrain or posterior fossa and in patients
older than 2 years of age [10,11]. Recently, several
reports have documented surprising results in
patients with associated abnormalities, such as
myelomeningocele, and those who had shunt
551
infections, with success in more than 70% of

patients [1013]. There has been some evidence of
ecacy in a subgroup of patients with normal
pressure hydrocephalus [14]. ETV would not be
eective in patients when CSF absorption is compromised. At times, it may be dicult to determine whether a patient with obstructive
hydrocephalus has abnormalities in CSF absorption as well.
The initial workup of patients with progressive
hydrocephalus includes structural studies like CT
and MRI. The third ventricle must be of adequate
size for the endoscope being used so as to avoid
damage to the walls of the third ventricle. The
oor of the third ventricle should be examined;
a downward bulge of the oor of the third
ventricle is essential to permit safe performance
of this procedure [1517]. Only then is the
ventricular oor translucent, permitting safe
perforation. Occasionally, a at third ventricular
oor can be perforated, but this may carry a higher
risk of basilar artery or hypothalamic injury. The
nal question to address is the patency of the
subarachnoid spaces. Homan [17] and Sayers
and Kosnik [18] have previously advocated
a preliminary shunt to open up the subarachnoid
spaces and improve the success rate of ETV. I do
not advocate this, because it seems more logical
for the pulsatile CSF ow to be more eective.
There are not reliable and routinely applicable
clinical tests that can predict whether ETV will be
successful. At this time, I as well as other authors
think that an initial attempt at management for
hydrocephalus from third or fourth ventricle
outow obstruction should be by means of ETV
to avoid becoming shunt dependent [10,19,20].
Nevertheless, it is important that adequate information regarding the lower success rates and
higher surgical risks compared with shunting be
addressed in the preoperative informed consent
form.
The operation is performed through a burr
hole at Kochers point. Unless the ventricles are
markedly enlarged, I use CT-guided frameless
stereotaxy to conrm the trajectory and minimize
injury that can be caused by insertion of the
endoscope sheath. The rigid beroptic endoscope
is used for freehand manipulation and visualization. Once the ventricle is entered, the foramen
of Monro is identied. The choroid plexus is a
useful landmark and can be followed toward the
foramen of Monro. The relation between the
thalamostriate vein and choroid plexus conrms
the left or right side (see anatomy section).
552
The endoscope is then passed through the

foramen of Monro after conrmation of the
ventricular side, and the oor of the third ventricle
is visualized. The techniques available for perforating the oor vary from blunt perforation using
the endoscope itself, a blunt probe such as
a Fogarty catheter, or cautery of the oor and
atraumatic balloon dilation of the perforation
with a Fogarty catheter [13,21]. Special balloons
and instruments have also been developed to
minimize damage to the hypothalamus and
basilar artery. I do not use direct perforation of
the oor with the endoscope, because visualization is lost during this maneuver. The translucent
membrane anterior to the mammillary bodies is
initially perforated by fulguration using bipolar
cautery. It is important to make the initial
perforation as anterior as possible without approaching the infundibular recess so as to avoid
the basilar artery. In cases where the prepontine
cistern is compressed or the oor is not translucent, the posterior clinoid can be palpated. The
space between the posterior clinoid can be used
for penetration, just posterior to the dorsum sella
so as to avoid the basilar artery. The Fogarty
balloon (2 French) is then inserted into the
perforation and expanded such that the diameter
of the balloon atraumatically dilates the perforation. I avoid inating the balloon in the prepontine cistern and withdrawing the catheter through
the perforation, because this causes greater
trauma to the oor with potential injury to the
hypothalamus. Once the perforation is expanded
to the diameter of the balloon, the endoscope can
be passed through the hole into the prepontine
cistern (see Fig. 2). Occasionally, a second membrane beneath the oor may exist (thickened
arachnoid or the membrane of Liliequist), which
needs to be opened by blunt perforation using
the angled probe or Fogarty catheter. Movement
of the ragged margins of the perforation with
CSF pulse wave is often seen, conrming good
ow through the ventriculostomy. If bleeding that
obscures the eld of view is encountered at any
stage and is not cleared completely with irrigation,
the procedure should be terminated and an external ventricular drain placed. A further attempt
at ventriculostomy can be made 1 or 2 days later if
appropriate.
The goal in the management of hydrocephalus
is not the achievement of small ventricles but the
restoration of normal intracranial pressures and
amelioration of the presenting symptoms. In some
patients, it may take several days before the
subarachnoid pathways for CSF reabsorption

become near normal. Even after successful treatment, the perforation may seal closed. A cine
MRI examination can sometimes be useful to
document ow through the ventriculostomy.
Occasionally, monitoring of intracranial pressure
is the most denitive way to assess the success of
a ventriculostomy. Results of ETV are dependent
on the cause of hydrocephalus. In cases of
aqueductal stenosis or benign tectal gliomas, the
success of ETV ranges from 83% to 100%
[10,2224]. In cases of tumors involving the third
ventricle, success decreases to 64% to 83%
[10,23]. Patients presenting with obstructive
hydrocephalus with enlargement of the third
ventricle after subarachnoid hemorrhage or intraventricular hemorrhage (previously shunted)
may also respond to ETV, with a 61% to 63%
success rate in 54 patients [10,25]. Similarly,
patients with obstructive hydrocephalus with
a history of infection can be treated with ETV
with a successful outcome in 64% of patients [25].
Complications of ETV include arterial or
venous bleeding, perforation of the basilar artery,
diabetes insipidus, and rare cases of cerebral
infarction and death. There have been no deaths
as a result of the procedure or arterial injuries at
this institution; however, there has been one patient with hemiparesis associated with insertion of
the endoscope sheath. There is a learning curve
associated with this procedure as with any other,
and complication rates vary between 6% and 11%
[10,23,26]. Potential complications of ETV are
more serious compared with shunting, but fatal
risks are rare. Complications of ETV compared
with ventriculoperitoneal shunt placement need
to be discussed with the patient. Nevertheless,
because of the eectiveness of ETV, the improved
CSF dynamics, and the elimination of foreign
hardware, I think that it should be oered to
patients with occlusive hydrocephalus.
Endoscopic colloid cyst resection
Colloid cysts are benign cystic tumors that
usually arise from the roof of the third ventricle. In
most patients, the lesion presents at the foramen of
Monro, causing obstructive hydrocephalus. They
can also be observed in the posterior aspect of the
third ventricle, or they may involve the septum
pellucidum or fornices [2729]. Patients commonly
present with complaints of headache, nausea,
emesis, memory loss, gait disturbance, or visual
obscuration. Rare cases of sudden death or acute
obstructive hydrocephalus have been reported [30].

More recently, asymptomatic cysts are being
identied through the increased use of MRI.
Treatment of colloid cysts includes conservative observation in lesions less than 1 cm without
hydrocephalus with serial MRI scans, ventricular
shunting to manage the hydrocephalus, and
removal of the lesion. Incidental colloid cysts less
than 1 cm without hydrocephalus can be managed
with serial observation, with treatment reserved
for those lesions that grow, cause ventriculomegaly, or become symptomatic. Colloid cysts have
traditionally been treated by microsurgical resection through either a transcallosal or transcortical approach. Minimally invasive therapies
include stereotactic and endoscopic aspiration of
the cyst, although these are associated with a high
rate of recurrence [3135]. The use of endoscopic
surgical procedures for resection of colloid cysts
has been developing over the past 10 years. With
the further development of endoscopic instruments, more radical resections are achievable
through a minimally invasive approach. This has
the advantage of carrying little morbidity with
a shorter operative time and hospital stay. Longterm results are not available, however, although
the short-term results seem promising [36,37].
The procedure involves positioning the patient
supine, with the neck exed 15 above the
horizontal plane. I routinely use neuronavigation
to project the angle of approach accurately.
Unlike third ventriculostomies, the burr hole
should be placed more anteriorly to better
visualize the posterior aspect of the foramen of
Monro. For single portal access, a burr hole is
place 1 cm anterior to the coronal suture and 3 cm
lateral to midline. A second portal can be placed
1 cm anterior to the rst portal. I have found that
the use of two portals aids in manipulation of the
tumor tissue to allow more radical resection and
improve safety. Often, tissue can be retracted with
one instrument and cauterized or incised with the
other, without the limitation of working with
both instruments in the same trajectory. Operating with instruments in a coaxial fashion limits the
degrees of freedom with which the tissue can be
manipulated and visualized. The additional portal
using a 3-mm sheath is inserted under stereotactic
guidance.
The goal of surgery is complete tumor resection,
but with safety in mind, a remnant along the roof of
the third ventricle or the internal cerebral vein is
often left behind and cauterized. The initial
visualization is through the 0 scope. The relation
553
of the colloid cyst to the foramen of Monro,

choroid plexus, fornix, septum pellucidum, and
deep venous system is determined. Often, choroid
plexus overlies the cyst wall and needs to be
cauterized for better visualization. With large
lesions, the cyst wall is cauterized and punctured
so as to remove as much of the cyst contents as
possible. A silastic catheter can be inserted into the
cyst for aspiration of cyst contents. Because of the
viscous nature of the colloid material, piecemeal
removal of the cyst contents can be performed
using cup or alligator forceps. Once the cyst is
decompressed, the wall can be grasped and
separated from the fornix. The cyst wall can be
cauterized to help shrink the overall contents. The
lesion can then be delivered into the ipsilateral
ventricle for complete extirpation. Care must be
taken not to place excessive traction on the cyst
wall so as to prevent injury to venous structures
from the tumor attachment to the tela choroidea
adjacent to the venous angle. If the cyst does not
separate easily, the free portions are excised with
microscissors and the remaining fragment is
cauterized, taking care not to coagulate the
thalamostriate or internal cerebral veins. A nal
inspection of the tumor resection is performed
using the 30 angled scope. The septum pellucidum
is also fenestrated, and the scope is passed to the
contralateral side to visualize the contralateral
foramen of Monro for any tumor remnants and for
patency of the contralateral foramen. Depending
on the degree of blood or tissue debris in the
ventricles, an external ventricle drain can be left in
place for overnight drainage and evacuation of
ventricular contents. Patients are monitored in the
intensive care unit overnight.
The lower morbidity and shorter operative time
and hospital stay have been documented by several investigators [3639]. Bauer and Hellwig [40]
described the endoscopic treatment of 70 cystic
lesions, including colloid cysts, with a morbidity
rate of 1.4% and no mortality. The rst comparative study of microsurgery and endoscopy was
conducted by Lewis et al [41], who found shorter
operative, hospitalization, and rehabilitation times
and achieved a lower complication rate with the
use of endoscopic treatment. Those results were
conrmed by Kehler et al [42] and King et al [37];
in these reports, the average operating time for
endoscopic procedures was 94 minutes and the
average total hospital stay in the endoscopic group
was 2.3 days. It is important, however, not
to compromise treatment ecacy for less invasiveness. A review of the recent literature with
554
Table 1
Outcome from endoscopic colloid cyst resection in the modern era
Author
Year
No. cases
Extent of resection
Follow-up
Recurrence (time)
Abdou and Cohen
1998
13
7 years
None
Decq et al
1998
15
15.2
months
1 (18 months)
King et al
1999
12
2002
12
24.5
months
56
months
Gaab et al
Hellwig et al
2003
20
10 near-complete
resection
3 subtotal resection
12 aspiration of cyst,
coagulation of cyst wall
3 partial removal
10 gross total resection
1 no resection/observation
3 near-complete resection
(coagulation of cyst wall)
follow-up (Table 1) shows promising results;

however, long-term follow-up is necessary. Although there is a high incidence of recurrence with
simple aspiration, with endoscopic surgery, most
of if not all the cyst contents can be removed safely
and the cyst wall resected. Often, an attachment of
the cyst wall to the roof of the third ventricle is
coagulated and left behind to avoid injury to the
deep venous systems (Fig. 3). In cases of nearcomplete resections, the signicance of a small
remnant cyst wall when related to recurrence
remains unclear, and only with long-term followup will this be understood.
There are several situations in which endoscopic
colloid cyst resection may be dicult or impossible.
The presence of small ventricles poses a technical
diculty and may carry higher morbidity with
diculties in cannulating the ventricles as well as in
manipulating the cyst wall [37]. Second, recurrent
colloid cysts become technically challenging because of the loss of anatomic landmarks, cicatrix
formation within the ventricles, and greater adherence of the recurrent cyst wall to the adjacent
structures. Finally, a small lesion located in the roof
of the third ventricle and posterior to the foramen
of Monro would be hard to remove with the rigid
endoscope and would be associated with an
increased risk of injury to the fornix.
Our early experience and the results of other
centers suggest that rigid endoscopy is an excellent
alternative to the established microsurgical approach. In the hands of well-trained surgeons,
with improved endoscopic instruments, this technique can be considered the rst method of
choice. Long-term follow-up is necessary to
62
months
0 symptomatic
1 asymptomatic
(93 months)
1 requiring
reoperation
(72 months)
determine if endoscopic surgery is a suitable

replacement for microsurgical strategies.
Management of pineal region tumors
Surgery has assumed an important role in the
management of pineal region tumors. Radiographic features are not diagnostic, and therapy
and outcome are dependent on tumor type.
Management of hydrocephalus requires urgent
attention, because patients can develop acute
obstruction and a herniation syndrome. Nearly
all patients with pineal region tumors present with
symptomatic hydrocephalus. The standard of care
has been placement of a ventriculoperitoneal
shunt, at which time CSF can be collected. More
recently, endoscopic management of obstructive
hydrocephalus with a third ventriculostomy, CSF
sampling, and biopsy of the tumor has gained
popularity. Disadvantages of a permanent ventriculoperitoneal shunt for the initial management of hydrocephalus in these patients include
shunt malfunction, infection, and, rarely, peritoneal seeding of tumors. In many cases, if the
tumor can be removed, the patient may not need
the shunt at all. ETV allows for successful
treatment of the obstructive hydrocephalus and
collection of CSF for markers and cytology and
permits multiple biopsies under direct endoscopic
vision as a one-step procedure [43]. I have used
this approach in the initial management of all
patients with pineal region tumors presenting with
hydrocephalus. In patients with a pineal region
mass with hydrocephalus, a third ventriculostomy
is performed as described previously. CSF is also
555
Fig. 3. (A, B) Preoperative sagittal precontrast T1-weighted and coronal uid attenuated inversion recovery (FLAIR)
images demonstrating the colloid cyst associated with hydrocephalus. (C, D) Postoperative imaging after endoscopic cyst
resection. No visible remnant of the cyst is seen, and the ventricles have decreased in size. The patients imaging studies
show no cyst recurrence at 24 months of follow-up.
sampled at this time for tumor markers. If the

anatomy is appropriate, a biopsy can be performed under direct endoscopic visualization. To
biopsy a lesion in this location, a more anterior
burr hole is necessary to reach the posterior third
ventricle. This is a particular case where the
exible endoscope could be benecial. If the
biopsy is adequate and demonstrates a germinoma
or primary malignant glioma without signicant
mass eect, the patient is managed with radiation
and chemotherapy. For non-germinoma germ cell
tumors (NGGCTs), pineal cell tumors, and large
glial tumors, the patient can then undergo a more
elective craniotomy appropriate for the location
and size of the tumor. A concern with endoscopic

biopsy is sampling error. Germinomas can contain nests of malignant germ cell tumors that
would signicantly alter therapy and outcome.
Likewise, a glial tumor may contain mixed cell
types or focal areas of a higher grade neoplasm,
although this is less of a problem, because radiographic appearance can help in choosing the most
appropriate region to biopsy.
Summary
Recently, endoscopic intraventricular surgery
has been performed successfully in several
556
clinical series. Although the therapeutic results

must be compared with conventional surgery,
neuroendoscopy seems to be a safe surgical
technique when performed by surgeons with
appropriate experience and rened endoscopic
tools. Rigid or exible endoscopes equipped with
various-sized working channels should be selected depending on the nature of the pathologic
ndings. The well-proven tenets of microsurgery
must not be sacriced for the sake of more rapid
surgical time and noninvasiveness; thus, endoscopic surgery must adhere to the principles of
microsurgery. The improved visualization and
lower morbidity have established neuroendoscopy in the management of specic disease
processes, such as obstructive hydrocephalus. Its
further use in the management of intraventricular
cysts and tumors is dependent on long-term
follow-up and the development of even better
instrumentation.
References
[1] Davis L. Neurological surgery. Philadelphia: Lea &
Febiger; 1936.
[2] Dandy WE. Cerebral ventriculoscopy. Bull Johns
Hopkins Hosp 1922;33:189.
[3] Mixter WJ. Ventriculoscopy and puncture of the
oor of the third ventricle. Boston Med Surg J
1923;188:2778.
[4] Nulsen FE, Spitz EB. Treatment of hydrocephalus
by direct shunt from the ventricle to the jugular
vein. Surg Forum 1951;2:399402.
[5] Fukushima T, Ishijima B, Hirakawa K, Nakamura
N, Sano K. Ventriculoberscope: a new technique
for endoscopic diagnosis and operation. Technical
note. J Neurosurg 1973;38:2516.
[6] Dandy WE. An operative procedure for hydrocephalus. Bull Johns Hopkins Hosp 1922;33:18990.
[7] Scar JE. The treatment of hydrocephalus. A
historical and critical review. J Neurol Neurosurg
Psychiatry 1963;26:126.
[8] McKnickle HF. Surgical treatment of hydrocephalus. Br J Surg 1947;34:3027.
[9] Grith HB. Technique of fontanelle and persutural
ventriculoscopy and endoscopic ventricular surgery
in infants. Childs Brain 1975;1:35963.
[10] Hopf NJ, Grunert P, Fries G, Resch KD, Perneczky
A. Endoscopic third ventriculostomy: outcome
analysis of 100 consecutive procedures. Neurosurgery 1999;44:795806.
[11] Jones RF, Kwok BC, Stening WA, Vonau M. Third
ventriculostomy for hydrocephalus associated with
spinal dysraphism: indications and contraindications. Eur J Pediatr Surg 1996;6(Suppl 1):56.
[12] Teo C, Jones R. Management of hydrocephalus by
endoscopic third ventriculostomy in patients with
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
myelomeningocele. Pediatr Neurosurg 1996;25:

5763.
Sainte-Rose C. Third ventriculostomy. In: Manwaring KH, Crone KR, editors. Neuroendoscopy.
New York: Mary Ann Liebert; 1992. p. 4762.
Meier U, Zeilinger FS, Schonherr B. Endoscopic
ventriculostomy versus shunt operation in normal
pressure hydrocephalus: diagnostics and indication.
Acta Neurochir Suppl (Wien) 2000;76:5636.
Jones RF, Brazier DH, Kwok BC, Stening WA,
Vonau M. Neuroendoscopic third ventriculostomy.
In: Cohen AR, Haines SJ, editors. Minimally
invasive techniques in neurosurgery. Philadelphia:
Homan HJ, Harwood-Nash D, Gilday DL.
Percutaneous third ventriculostomy in the management of noncommunicating hydrocephalus. Neurosurgery 1980;7:31321.
Homan HJ. The advantages of percutaneous third
ventriculostomy over other forms of treatment of
infantile obstructive hydrocephalus. In: Morley TP,
editor. Controversies in neurosurgery. Philadelphia:
WB Sanders; 1976. p. 671703.
Sayers MP, Kosnik EJ. Percutaneous third ventriculostomy: experience and technique. Childs Brain
1976;2:2430.
Grant JA, McLone DG. Third ventriculostomy:
a review. Surg Neurol 1997;47:2102.
Jones RF, Kwok BC, Stening WA, Vonau M.
Neuroendoscopic third ventriculostomy. A practical alternative to extracranial shunts in non-communicating hydrocephalus. Acta Neurochir Suppl
(Wien) 1994;61:7983.
Vries JK. An endoscopic technique for third
ventriculostomy. Surg Neurol 1978;9:1658.
Wellons JC III, Tubbs RS, Banks JT, et al. Longterm control of hydrocephalus via endoscopic third
ventriculostomy in children with tectal plate gliomas. Neurosurgery 2002;51:638.
Buxton N, Ho KJ, Macarthur D, Vloeberghs M,
Punt J, Robertson I. Neuroendoscopic third ventriculostomy for hydrocephalus in adults: report of
a single units experience with 63 cases. Surg Neurol
2001;55:748.
Grunert P, Charalampaki P, Hopf N, Filippi R.
The role of third ventriculostomy in the management of obstructive hydrocephalus. Minim Invasive
Neurosurg 2003;46:1621.
Siomin V, Cinalli G, Grotenhuis A, et al. Endoscopic third ventriculostomy in patients with
cerebrospinal uid infection and/or hemorrhage.
J Neurosurg 2002;97:51924.
Beems T, Grotenhuis JA. Is the success rate of
endoscopic third ventriculostomy age-dependent?
An analysis of the results of endoscopic third ventriculostomy in young children. Childs Nerv Syst
2002;18:6058.
Antunes JL, Louis KM, Ganti SR. Colloid cysts of
the third ventricle. Neurosurgery 1980;7:4505.

[28] Camacho A, Abernathey CD, Kelly PJ, Laws ER
Jr. Colloid cysts: experience with the management
of 84 cases since the introduction of computed
tomography. Neurosurgery 1989;24:693700.
[29] Ciric I, Zivin I. Neuroepithelial (colloid) cysts of the
septum pellucidum. J Neurosurg 1975;43:6973.
[30] Chan RC, Thompson GB. Third ventricular colloid
cysts presenting with acute neurological deterioration. Surg Neurol 1983;19:35862.
[31] Apuzzo ML, Chandrasoma PT, Zelman V, Giannotta SL, Weiss MH. Computed tomographic
guidance stereotaxis in the management of lesions
of the third ventricular region. Neurosurgery 1984;
15:5028.
[32] Bosch DA, Rahn T, Backlund EO. Treatment of
colloid cysts of the third ventricle by stereotactic
aspiration. Surg Neurol 1978;9:158.
[33] Gutierrez-Lara F, Patino R, Hakim S. Treatment of
tumors of the third ventricle: a new and simple
technique. Surg Neurol 1975;3:3235.
[34] Hall WA, Lunsford LD. Changing concepts in the
treatment of colloid cysts. An 11-year experience in
the CT era. J Neurosurg 1987;66:18691.
[35] Mathiesen T, Grane P, Lindquist C, von Holst H.
High recurrence rate following aspiration of colloid
cysts in the third ventricle. J Neurosurg 1993;78:
74852.
[36] Hellwig D, Bauer BL, Schulte M, Gatscher S,
Riegel T, Bertalany H. Neuroendoscopic treat-
[37]
[38]
[39]
[40]
[41]
[42]
[43]
557
ment for colloid cysts of the third ventricle: the

experience of a decade. Neurosurgery 2003;52:
52533.
King WA, Ullman JS, Frazee JG, Post KD,
Bergsneider M. Endoscopic resection of colloid
cysts: surgical considerations using the rigid endoscope. Neurosurgery 1999;44:110311.
Horvath Z, Veto F, Balas I, Doczi T. Complete
removal of colloid cyst via CT-guided stereotactic
biportal neuroendoscopy. Acta Neurochir (Wien)
2000;142:53946.
Schroeder HW, Gaab MR. Endoscopic resection of
colloid cysts. Neurosurgery 2002;51:14415.
Bauer BL, Hellwig D. Minimally invasive endoscopic neurosurgerya survey. Acta Neurochir
Suppl (Wien) 1994;61:112.
Lewis AI, Crone KR, Taha J, van Loveren HR,
Yeh HS, Tew JM Jr. Surgical resection of third
ventricle colloid cysts. Preliminary results comparing transcallosal microsurgery with endoscopy.
Kehler U, Brunori A, Gliemroth J, et al. Twenty
colloid cystscomparison of endoscopic and microsurgical management. Minim Invasive Neurosurg 2001;44:1217.
Pople IK, Athanasiou TC, Sandeman DR, Coakham HB. The role of endoscopic biopsy and third
ventriculostomy in the management of pineal region
tumours. Br J Neurosurg 2001;15:30511.
Intraventricular meningiomas
Michael W. McDermott, MD, FRCSC
Departments of Neurological Surgery and Radiation Oncology University of California at San Francisco,
505 Parnassus Avenue, M-774, San Francisco, CA 94143, USA
Meningiomas arising in the ventricular system

are rare; yet, when they do present clinically, they
are often large, most often within the atrium, and
most frequently on the left. For all these reasons,
they are tumors for which it is dicult to achieve
the perfect surgical result: complete removal of
a benign tumor without complications or new
neurologic morbidity. With a thorough understanding of the anatomy of structures around the
ventricle, selection of the proper surgical approach, and use of modern neurosurgical techniques, however, modern-day surgical results
should be superior to those of the past. This
article reviews some of the important clinical and
technical considerations for the surgery of intraventricular meningiomas.
in the posterior third ventricle, and 5% in the

fourth ventricle (Table 1). At our own institution,
over a 15-year period, there were only 13 intraventricular meningiomas (all in female patients):
11 with benign histology and 2 anaplastic. When
a meningioma does occur in the lateral ventricle, it
is more often on the left than on the right and
more than 90% are located in the atrium. Within
the third ventricle, meningiomas are more often
posterior than anterior [5].
The incidence of ventricular meningiomas is
higher in pediatric patients. Germano and colleagues [6] reviewed 15 pediatric meningioma
series in the literature totaling 278 meningiomas
and found that 9.4% were intraventricular.
Thirty-nine percent occurred in patients younger
than 10 years of age, and 61% occurred in the
1- to 20-year-old age group.
Incidence
Meningiomas are the second most common
primary brain tumor in adults [1]. In a surgical
series reported by Cushing and Eisenhardt
(N = 295) [2], intraventricular meningiomas accounted for only 1.3% of the total. Guidetti and
Delni [3] found that over a 38-year period, only
22 (1.5%) of 1451 meningiomas were intraventricular: 20 in the lateral ventricles, 2 in the fourth
ventricle, and none in the third ventricle. In an
earlier extensive review of the published literature
up to 1986, Criscuolo and Symon [4] identied 400
intraventricular meningiomas and categorized
their location. Eighty percent of intraventricular
meningiomas occur in the lateral ventricles, 15%
E-mail address: mcdermottm@neurosurg.ucsf.edu
Origin and anatomy

Meningiomas arise within the ventricle from
the choroid plexus or from the tela choroidea
within the ventricular system. As pointed out by
Cushing and Eisenhardt [2], meningiomas within
the ventricle tend to assume the shape of the
ventricle in which they lay. The atrium is the most
common location for lateral ventricular meningiomas. The atrium is formed by the junction of
the temporal horn anteriorly and inferiorly, the
occipital horn posteriorly, and the posterior body
of the lateral ventricle anteriorly and superiorly.
The oor of the atrium is formed by the
hippocampus, the medial wall by the splenium
of the corpus callosum, and the roof and lateral
wall by the splenium and tapetum of the corpus
callosum [7]. Fibers of the geniculocalcarine tract
run lateral and inferior to the atrium and account
doi:10.1016/S1042-3680(03)00055-X
560
M.W. McDermott / Neurosurg Clin N Am 14 (2003) 559569
Table 1
Common sites for intraventricular meningioma
Ventricle
Site
Frequency
Blood supply
Lateral
80%
Third
Atrium >80%
(L > R)
Posterior
15%
Fourth
Midline
5%
AChA
PLChA
PMChA
PLChA
PICA
Abbreviations: AChA, anterior choroidal artery; L,

left; PICA, posterior inferior cerebellar artery; PLChA,
posterior lateral choroidal artery; PMChA, posterior
medial choroidal artery; R, right.
for visual symptoms seen with larger tumors

(Fig. 1).
Meningiomas of the fourth ventricle arise from
the choroid or the inferior tela choroidea [3].
Tumors of the posterior third ventricle are
thought to arise from the tela of the velum
interpositum, the space between the two layers
of the tela choroidea in the roof of the third
ventricle that carry the posterior medial choroidal
arteries and internal cerebral veins [5]. Tumors in
this location are to be distinguished from meningiomas arising from the falcotentorial junction
immediately behind the third ventricle.
Pathology
In the series of Guidetti and Delni [3], 81% of
the meningiomas were broblastic. Meningotheliomatous and psammomatous variants reported
by others as case reports are less common. All three
types are classied as grade I tumors in the World
Health Organization scheme, with a low risk of
recurrence and nonaggressive clinical behavior [8].
The pathologic entities encountered in the lateral
ventricle also include choroid plexus papilloma,
ependymoma, and metastases. The author has
operated on three large meningiomas of the
atrium, which proved pathologically to be metastases from renal cell (1 case) and thyroid (2 cases)
carcinoma. Posterior third ventricular tumors that
may mimic meningiomas radiographically include
pineocytoma and teratoma, whereas choroid plexus papilloma, ependymoma, and hemangioblastoma should be considered in the fourth ventricle.
Clinical presentation
Meningiomas of the lateral ventricles present
primarily with signs of increased intracranial
pressure. Headache, nausea/vomiting, and disturbance of vision are seen in 40% to 80% of patients
Fig. 1. Posterior coronal section of hemispheres. Note position of optic radiations lateral and inferior to atrium. Large
arrow on left indicates trajectory with posterior parietal-occipital approach. SPL, superior parietal lobule; IPL, inferior
parietal lobule; STG, superior temporal gyrus; MTG, middle temporal gyrus; ITG, inferior temporal gyrus; S, splenium;
OR, optic radiations.
(Table 2) [3]. Visual symptoms most often relate

to impaired vision from papilledema but can also
include visual eld decits from large atrial
tumors. Motor, sensory, and speech disturbances
are also seen, and seizures were the presenting
symptom in 27% of patients for one series [3].
Symptoms of impaired memory may relate to
dilatation of a trapped temporal horn or to direct
compression on the hippocampal formation in the
oor of the atrium with larger tumors.
Tumors of the third and fourth ventricles
usually present with symptoms of hydrocephalus
as a result of obstruction of cerebrospinal uid
(CSF) ow. Parinauds syndrome with impaired
upgaze and pupillary light reexes may be seen
with posterior third ventricular masses [5].
Radiology
Frequently, patients still have CT scans as their
rst imaging study. Meningiomas in the lateral
ventricle are slightly hyperdense, may have small
areas of calcication, and usually show homogeneous enhancement [9]. With larger tumors,
obstruction of the temporal horn, which contains
the choroid plexus, results in dilatation of this
portion of the ventricle and low density in the
brain surrounding the atrium, which is partially
related to transependymal ow of CSF and
partially related to tumor-associated vasogenic
edema (Fig. 2). MRI reveals superior anatomic
Table 2
Common symptoms and signs for atrial meningiomas
Frequency
Symptoms
Headaches
Nausea/vomiting
Seizures
Speech disturbance
Motor
Mental disturbance
Visual disturbance
Sensory
Signs
Papilledema
Visual eld decit
Motor disturbance
Dysphasia
Sensory disturbance
80%
40%
35%
30%
25%
20%
20%
15%
60%
50%
50%
40%
20%
From Guidetti B, Delni R. Lateral and fourth

ventricle meningiomas: In: Al-Mefty O, editor. Meningiomas. New York: Raven Press; 1991. p. 56981; with
permission.
561
detail compared with CT scans, with meningiomas being iso- or hypointense on T1-weighted
images and T2-weighted images [10]. It is the
authors experience that broblastic meningiomas,
the most common pathologic subtype of meningioma within the ventricle, are usually prominently
hypointense on T2-weighted images. On T1weighted postcontrast images, there is uniform
contrast enhancement. Thin-cut axial spoiled
gradient recall (SPGR) images are routinely used
for image-guided surgical systems, and these
images can also provide the opportunity for
creating two-dimensional (2D) and three-dimensional (3D) venograms to assist the surgeon with
the preoperative selection of surgical approaches
[11]. The dierential diagnoses for other tumors in
these locations also need to be considered, especially as they relate to patient age (Table 3) [9].
MRI also oers the ability to perform spectroscopy
and blood volume time intensity maps, which,
alone or together, may increase the certainty of the
radiologic diagnosis. A high alanine-to-creatinine
ratio has been reported as a relatively specic MR
spectroscopy nding for meningiomas [10].
Cerebral angiography is rarely used these days,
because intraventricular meningiomas can rarely
be embolized. Angiography can conrm the predominant blood supply (see Table 1) and the
position of prominent parasagittal draining veins
[12]. Catheter cannulation of choroidal arteries is
dicult, however, and the target for occlusion is
distal in the vascular territory. Catheter-related
arterial spasm of the anterior choroidal artery
after attempted embolization of a vascular malformation, with subsequent infarction, has prevented us from considering preoperative embolization.
Thus, in selecting an operative approach, the
surgeon may need to take into consideration the
potential advantage of occluding the arterial blood
supply to large tumors early in the operation.
Options for treatment

Observation
Intraventricular meningiomas discovered incidentally, those that are not causing obstruction
of CSF ow or hydrocephalus, and those not
associated with vasogenic edema should be
observed [13]. The author usually follows patients
with two successive 3-month scans; if the tumor is
stable, the interval is increased to 6 months for 2
years and then to once a year. It is important to
compare follow-up scans with the original scan
562
563
and not the one immediately preceding it, because

smaller changes in size may not be appreciated.
Anterior-posterior, lateral, and vertical dimensions should be recorded. Documented tumor
growth on serial scans in a medically t individual
or development of symptoms related to secondary
tumor eects, such as hydrocephalus or vasogenic
edema, may warrant surgical intervention. Certainly, for a nondominant right atrial meningioma
with symptomatic trapping of the temporal horn,
surgical removal of the tumor is preferred over
shunting and radiotherapy. In contrast, a slowly
growing meningioma in the left atrium of a 70year-old patient that is not causing hydrocephalus
or vasogenic edema may be considered for radiosurgical treatment.
Surgery
Surgery is the gold standard of treatment for
growing symptomatic meningiomas in patients
who are candidates for general anesthesia. Because most of these tumors are benign and
complete excision of lateral ventricle tumors can
be accomplished, surgical cures can be achieved.
Potential side eects of the various surgical
approaches need to be considered and reviewed
with the patient [12].
Table 3
Common radiographic dierential diagnoses for intraventricular meningioma
Site
Child
Adult
Atrium
CPP
Ependymoma
Metastasis
Ependymoma
Posterior
third
ventricle
Germinoma
Pineocytoma
blastoma
Astrocytoma
Teratoma
Pineocytoma
Astrocytoma
Metastasis
Fourth
ventricle
CPP
Ependymoma
Astrocytoma
Medulloblastoma
Ependymoma
CPP
Hemangioblastoma
Medulloblastoma
decreased in size, obstructive hydrocephalus with

a trapped temporal horn developed and surgical
removal was performed 20 months later. We have
seen exaggerated radiation toxicity in the subependymal periventricular region as well as with
fractionated radiation therapy; thus, even noninvasive treatments should be recommended only
to those who are not candidates for, or refuse,
microsurgical removal.
Radiotherapy
Radiosurgery
Radiosurgery is an eective form of treatment
for selected meningiomas, including intraventricular meningiomas. Reports on the tumor control
rates achieved with radiosurgery of 85% to 98%
cannot be overlooked during the informed consent process [14,15]. A size limit of 3 cm as used
in other brain locations is not acceptable for intraventricular tumors of the atrium, third ventricle, or fourth ventricle at our institution, and we
usually limit tumors to 1.5 to 2.0 cm depending on
associated vasogenic edema. Of the 13 patients
treated at our institution, 4 have had radiosurgery
for their intraventricular meningioma. In 2
patients, the treatment was used for residual or
recurrent disease, and in 2 patients, it was used as
the primary form of therapy. Of these latter 2
patients, 1 developed symptomatic subependymal
radiation toxicity despite a low marginal dose of
12 Gy. In spite of the fact that the tumor
Fractionated 3D conformal radiotherapy offers eective tumor control for residual or recurrent meningiomas in a variety of intracranial
locations [16]. Although there is no reason to
believe that the results should be any dierent for
intraventricular meningiomas, none of the four
patients at our institution who required additional
therapy other than initial surgery (ie, reoperation,
radiosurgery) was treated with this method.
Although the conformality of present-day techniques limits the volume of normal tissue irradiated,
intraventricular meningiomas are such discrete
targets that surgery or radiosurgery has been
recommended instead.
Surgical considerations
Evaluation of the patient for surgery involves
consideration of patient factors, such as age,
medical conditions, and neurologic status, and
b
Fig. 2. (A) Larger intraventricular meningioma presenting with seizure and language disturbance. Trapped temporal
horn (B) and evidence of parietal vasogenic edema (C). (D, E) Postoperative scans showing complete resection using
contralateral interhemispheric approach.
564
tumor characteristics, such as relation to symptoms and signs, growth rate, and resectability [7].
The routine evaluations for patients and specic
discussion of the approaches for intraventricular
meningiomas are reviewed below.
Preoperative studies
MRI without and with contrast is the basic
imaging study needed. For all supratentorial
approaches to the lateral ventricles, we obtain
volumetric, thin-cut, T1-weighted images for use
with image-guided surgical systems. MR venography can also be obtained to look at collapsed
vertex views or 3D reformats. Cerebral angiography is not routinely done, because the blood
supply to tumors in various locations is known
and embolization is not possible without signicant risk.
For patients having a transcallosal procedure
or transcortical superior parietal lobule approach
in the dominant hemisphere, the author always
obtains preoperative neuropsychologic testing as
a baseline. This can be repeated 3 months after
surgery and helps with quantitating patient status
for return to work and other disability issues. It
can also be used to document improvement in
function for patients who undergo successful
resection of large tumors or those associated with
hydrocephalus. If a superior parietal occipital
approach is selected, most anesthesiologists want
the patient to have a bubble echocardiogram done
to rule out a patent foramen ovale and a potential
right-to-left shunt. Humphrey visual elds should
be tested in all patients with tumors of the atrium.
General intraoperative measures
Patients undergoing transcortical approaches
receive anticonvulsants for 1 week around the
operation, beginning the morning of surgery.
Intravenous fosphenytoin or phenytoin is used
to load those patients who are not already on
medications. Standard measures to reduce intracranial pressure and improve blood rheology
are used. Mean arterial blood pressure should be
kept in the normal range, and hyperventilation
should be avoided throughout the case. Depending on the tumor location and hemisphere, every
attempt is made to identify and interrupt the
blood supply to these tumors early on.
Specic surgical approaches
Common surgical approaches by tumor site
are outlined in Table 4.
Middle temporal gyrus approach
Patient position: semilateral
Head position: extended on neck, tilted 20
downward, rotated 90 to opposite side,
parallel to oor
The middle temporal gyrus approach (Fig. 3) is
best suited for meningiomas of the atrium of the
lateral ventricle of the nondominant hemisphere.
A variety of skin incisions can be used, including
U-shaped incisions, reverse question mark incisions, or curvilinear incisions with posterior
hockey stick extension coming o at right angles
posteriorly [9,17]. A transsulcal approach in the
posterior third of the temporal lobe minimizes
tissue disruption, and image-guided systems can
be used to dene the precise trajectory to the
ventricle. The advantage of this approach is the
ability to pick up the anterior choroidal artery
within the temporal horn and eliminate the
predominant blood supply before tumor resection
starts. The choroid plexus can be followed back to
the tumor; internal debulking is then used,
followed by capsular dissection of the tumor from
the walls and oor of the ventricle. The surgeon
should be mindful of visual bers that rst pass
over the roof of the temporal horn and then swing
back lateral and inferior to the atrium in the
periventricular white matter; thus, horizontal
Table 4
Common surgical approaches by tumor site
Site
Approach
Patient position
Atrium
1. Middle temporal gyrus (nondominant)

2. Superior parietal lobule (dominant)
3. Contralateral interhemispheric, transcallosal (dominant)
4. Ipsilateral interhemispheric transcallosal
1. Infratentorial supracerebellar
2. Occipital transtentorial
Midline suboccipital
1. Left lateral
2. Supine or lateral
3. Lateral, tumor side up
4. Supine
1. Prone or semisitting
2. Prone, approach side down
Prone or concord
Posterior third ventricle

Fourth ventricle
565
Fig. 3. Coronal section through atria, anterior cut to Fig. 1. White arrow marks middle temporal transsulcal or middle
temporal gyrus approach. Ipsilateral interhemispheric transcallosal approach (shaded arrow) and contralateral
interhemispheric transcallosal approach (solid arrow) are also marked. SF-P, Sylvian ssure, posterior; OTG,
occipital-temporal gyrus; PHG, parahippocampal gyrus; CC, corpus callosum; A, atrium; STG, superior temporal
gyrus; MTG, middle temporal gyrus; ITG, inferior temporal gyrus.
subcortical dissection planes should be used to

minimize the risk of tract disruption. After
removal of the tumor, an external ventricular
catheter should be left to drain CSF until returns
are clear.
Superior parietal occipital approach
Patient position: supine with back up 15 or
semilateral
Head position: neck exed on chest, head
slightly exed on neck, no rotation; or
laterally exed on neck tilted 10 up, rotated
90 to opposite side
The superior parietal occipital approach can be
selected for either the left or right side but is most
often employed for dominant hemisphere tumors
[9,17]. The author prefers to have the patient
supine, with the back elevated so that the patient
is in a slouching position, with the neck exed on
the chest and the head exed on the neck. Most
anesthesiologists want to exclude a patent foramen ovale and right-to-left shunt before surgery
as noted previously. After pin xation and
registration, the image-guided system is used to

help identify the trajectory to the ventricle and the
midline. A horseshoe-shaped incision is fashioned,
and the brain is exposed to the parietal-occipital
junction. A posterior parietal sulcus is used for the
dissection, with maintenance of the plane using
self-retaining retractors. The corridor to the
tumor with this approach is long, and if the
tumor is rm and not that vascular, this is one of
the few instances where the laser is of particular
utility. One disadvantage of this approach is that
some of the blood supply (anterior choroidal
artery) to the tumor is not identied until late in
the procedure. Internal debulking followed by
marginal dissection is again employed. Care
should be taken not to extend the white matter
dissection too far lateral to the atrium for fear of
damaging visual pathway bers.
Interhemispheric transcallosal
Ipsilateral approach
Patient position: semisitting or prone
566
Head position: neck exed on chest with

head exed on neck or neck neutral with
head exed on neck
Contralateral approach
Patient position: full lateral, tumor side up
Head position: neck neutral, head slightly
exed on neck, laterally exed upward
toward tumor
The original description of the ipsilateral
transcallosal approach was provided by Kempe
and Blaylock in 1976 [18]. The patient can be positioned semisitting or prone for a parietal occipital
craniotomy. The bone ap should cross the
midline, and the interhemispheric ssure should
be opened widely. Jun and Nutik [19] described
a slightly more vertical angle than along the
tentorium so as to preserve part of the splenium.
The posterior callosum is split for 2 cm, leaving the
posterior part of the splenium intact. The tumor is
debulked and removed piecemeal.
We have used the contralateral transcallosal
approach with success for several cases in the
dominant hemisphere (see Fig. 2; Fig. 4) [20]. The
patient is positioned so that the retracted right
hemisphere is toward the oor; thus, retraction is
assisted by gravity. A posterior interhemispheric
approach is taken through a large parietal
craniotomy crossing midline, preserving parasagittal draining veins. Once the interhemispheric
ssure is split along the length of the exposure, the
image-guided system is used to dene the most
appropriate callosotomy, which is not more than
2 cm in length. The inferior two thirds of the falx
are incised to allow for gentle retraction of the
medial aspect of the left hemisphere and, later, the
callosum. The tumor can be removed piecemeal
using the long microsurgery set for the supracerebellar approach. At the completion of tumor
removal, an external ventricular drain is left in
place for several days.
Infratentorial supracerebellar
Patient position: semisitting or prone
Head position: neck exed on chest with head

exed on neck or neck neutral with head
exed on neck
The approaches to meningiomas of the posterior third ventricle are similar to those for pineal
region tumors [5]. Because meningiomas in this
location are so rare, this approach is not discussed
in great detail. The choice of the approach
generally depends on the patients body habitus
and the angle of the tentorium on sagittal MRI.
For obese patients, those with a short neck, or
those with a steeply angled up tentorium, the
semisitting position is preferred after preoperative
cardiac screening. In our experience, massive air
embolism is rare. The author uses a torcular
craniotomy, exposing the transverse sinuses bilaterally and the torcular with a bipartite bone
ap. The posterior fossa craniotomy below the
transverse sinus is done rst, followed by the second
ap exposing the sinuses from above. The rest of the
dissection is as described elsewhere [22].
Occipital transtentorial
Patient position: prone
Head position: neck neutral with head exed
on neck or head exed and rotated downward
toward oor
This is another approach for pineal region
tumors [5]. For the prone position, the craniotomy
is on the side of the sinus toward the oor so that the
retracted occipital lobe is assisted by gravity. The
dura is opened in a U-shaped fashion based
laterally so that the brain is protected from a sharp
dural edge.
Midline Suboccipital
Patient position: prone
Head position: neck extended on chest with
head exed on neck
Again, meningiomas of the fourth ventricle are
rare, but a standard surgical approach is used [3].
A midline incision extending into the occipital
region is used so that pericranium can be
c
Fig. 4. (A) Axial T1-weighted image showing trapped temporal horn in patient treated with radiosurgery for atrial
meningioma 20 months earlier. In spite of reduction in tumor size, the patient developed a subependymal reaction
causing obstructive hydrocephalus. (B) Relatively small contrast-enhancing tumor at atrium. Immediate postoperative
scans with contrast showing reduction in size of temporal horn (C) and complete resection of tumor in axial (D) and
coronal (E) planes. (F) Contralateral interhemispheric transcallosal approach was used.
567
568
harvested for dural repair at the end. Once the

occipital bone is removed and the dura and
cisterna magna are opened, the cerebellomedullary ssure is dissected to gain access to the
ventricle, thus avoiding splitting the inferior
vermis. The ssure is developed by dissecting the
lateral aspect of the tonsil and incising the tela
choroidea along the posterior lateral margin of
the oor of the fourth ventricle to the level of the
lateral recess [21]. The choroidal branch of the
posterior inferior choroidal artery is the main
supply and should be seen entering the tumor in
the roof of the ventricle.
Postoperative care
Fluid is not routinely restricted in these
patients so as to reduce the chance of cortical
venous thrombosis. Low-molecular-weight heparin is started 24 hours after surgery and is
continued for 5 days to reduce the chance of deep
venous thrombosis in the extremities. Excessive
drainage of CSF should be avoided so as to
reduce the chance of subdural hematoma formation. Conversely, adequate clearance of blood and
debris from the ventricle reduces the chance of
hydrocephalus. Patients can be followed with CT
scans to assess ventricle size and debris. Postoperative MRI should be delayed several days
until it is clear that the patient will be safe in the
MRI environment.
Summary
Intraventricular meningiomas are most often
seen in the lateral ventricles of adults, more often
on the left than on the right. A variety of surgical
approaches can be selected depending on patient
and tumor factors as well as on surgeon
preference. Some of these approaches are complex
and demand a thorough understanding of the
ventricular and periventricular anatomy. When
successful, surgery can be curative in most cases
of total removal.
References
[1] Surawicz TS, McCarthy BJ, Kupelian V, et al.
Descriptive epidemiology of primary brain and
CNS tumors: results from the Central Brain Tumor
Registry of the United States, 19901994. Neurooncology 1999;1:1425.
[2] Cushing H, Eisenhardt L. Meningiomas. Their
classication, regional behavior, life history and surgical end results. Springeld: Charles C. Thomas;
1938.
[3] Guidetti B, Delni R. Lateral and fourth ventricle

meningiomas. In: Al-Mefty O, editor. Meningiomas. New York: Raven Press; 1991. p. 56981.
[4] Criscuolo GR, Symon L. Intraventricular meningioma. A review of 10 cases of the National
Hospital, Queen Square (19741985) with reference
to the literature. Acta Neurochir (Wien) 1986;83:
8391.
[5] Tung H, Apuzzo ML. Meningiomas of the third
ventricle and pineal region. In: Al-Mefty O, editor.
Meningiomas. New York: Raven Press; 1991.
p. 583591.
[6] Germano IM, Edwards MS, Davis RL, et al.
Intracranial meningiomas of the rst two decades
of life. J Neurosurg 1994;80:44753.
[7] Abosch A, McDermott MW, Wilson CB. Lateral
ventricular tumors: a practical guide to neurosurgical technique. In: Kaye AH, Black P, editors.
Operative neurosurgery. London: Harcourt Brace;
2000. p. 799812.
[8] Louis DN, Scheithauer B, Budka H, et al. Meningiomas. In: Kleihues P, Cavenee WK, editors.
Pathology and genetics. Tumors of the nervous
system. Lyon: IARC Press; 2000. p. 17684.
[9] Tew JM, Larson JJ. Intraventricular meningioma.
In: Kaye AH, Black P, editors. Operative neurosurgery. Lyon: Harcourt Brace; 2000. p. 57585.
[10] Majos C, Alonso J, Aguilera C, et al. Utility of
proton MR spectroscopy in the diagnosis of
radiologically atypical intracranial meningiomas.
[11] McDermott MW, Gutin PH. Image-guided surgery
for skull base neoplasms using the ISG viewing
wand. Anatomic and technical considerations.
Neurosurg Clin North Am 1996;7:28595.
[12] McDermott MW, Wilson CB. Meningiomas. In:
Youmans JR, editor. Neurological surgery, vol. 4.
4th edition. Philadelphia: WB Saunders; 1996.
p. 2782825.
[13] McDermott MW. Current treatment of meningiomas. Curr Opin Neurol 1996;9:40913.
[14] Lee JY, Niranjan A, McInerney J, et al. Stereotactic
radiosurgery providing long-term tumor control of
cavernous sinus meningiomas. J Neurosurg 2002;
97:6572.
[15] Pollock BE, Staord SL, Utter A, et al. Stereotactic
radiosurgery provides equivalent tumor control to
Simpson grade 1 resection for patients with smallto medium-size meningiomas. Int J Radiat Oncol
Biol Phys 2003;55:10005.
[16] Goldsmith BJ, Wara WM, Wilson CB, et al.
Postoperative irradiation for subtotally resected
meningiomas. A retrospective analysis of 140
patients treated from 1967 to 1990. J Neurosurg
1994;80:195201.
[17] Guidetti B, Delni R, Gagliardi FM, et al.
Meningiomas of the lateral ventricles. Clinical,
neuroradiologic, and surgical considerations in 19
cases. Surg Neurol 1985;24:36470.

[18] Kempe LG, Blaylock R. Lateral-trigonal intraventricular tumors. A new operative approach. Acta Neurochir (Wien) 1976;35:
23342.
[19] Jun CL, Nutik SL. Surgical approaches to intraventricular meningiomas of the trigone. Neurosurgery 1985;16:41620.
[20] Lawton MT, Golnos JG, Spetzler RF. The
contralateral transcallosal approach: experi-
569
ence with 32 patients. Neurosurgery 1996;39:

72934.
[21] Matsushima T, Inoue T, Inamura T, et al. Transcerebellomedullary ssure approach with special
reference to methods of dissecting the ssure.
J Neurosurg 2001;94:25764.
[22] Lozier AP, Bruce JN. Meningiomas of the velum
interpositum: surgical considerations. Neurosurg
Focus 2003;15(1):19.
Intraventricular gliomas
Aaron S. Dumont, MDa,b, Elana Farace, PhDc, David Schi, MDd,
Mark E. Sharey, MDe,*
a
Department of Neurological Surgery, University of Virginia Health Sciences Center,

Box 212, Charlottesville, VA 22908, USA
b
Department of Neuroscience, University of Virginia, Charlottesville, VA 22908, USA
c
Departments of Neurological Surgery, Psychology, and Health Evaluation Sciences, NeuroOncology Center,
University of Virginia, Box 800432, Charlottesville, VA 22908, USA
d
Departments of Neurology and Neurological Surgery, NeuroOncology Center, University of Virginia,
e
Department of Neurological Surgery, NeuroOncology Center, University of Virginia Health Sciences Center,
Gliomas remain the most common symptomatic primary brain tumor in adults. Those arising
within or relating to the ventricular surface
represent a relatively small but important proportion of all gliomas. Intraventricular gliomas
comprise a unique spectrum of histologic subtypes, the most common of which include
ependymomas, subependymomas (SEs), and subependymal giant cell astrocytomas (SEGAs).
Other less common variants, including chordoid
glioma [13], glioblastoma multiforme [4,5], and
mixed glial-neuronal tumors [6] among others,
have been reported. Each type of intraventricular
glioma has its own unique epidemiologic, clinical,
radiologic, and pathologic characteristics. Furthermore, each type commands its own constellation of management considerations, and each is
associated with dierent prognostic indicators and
outcomes.
Considerable advances have been made in the
contemporary understanding and management of
these tumors, particularly over the last several
decades. The advent and widespread implementation of advanced microsurgical technique coupled
with advancements and renements in adjuvant
therapies and their indications for use have helped
to improve the care of patients harboring these
E-mail address: mes8c@virginia.edu (M.E. Sharey).
lesions. Studies into the basic biology of these

tumors using modern molecular biologic techniques are increasingly more common [710].
Despite this, however, the need for continued
progress cannot be overemphasized, particularly
in the management of ependymomas.
The following sections address the individual
types of intraventricular gliomas, focusing on the
unique characteristics and management considerations pertinent to each.
Ependymomas
Intracranial ependymomas refer to tumors of
neuroepithelial tissue arising from ependymal cells
lining the cerebral ventricles or from rests of
ependymal cells situated in the cortical white
matter. They occur most commonly in children
and young adults. Contemporary perspectives on
ependymomas have evolved considerably since the
original reported description of tumors of ependymal cell origin by Virchow [11] in 1863.
General comments and epidemiology
Ependymomas comprise between 3% and 10%
of intracranial tumors in most series [1217]. The
estimated incidence of ependymomas is approximately 0.2 to 0.8 per 100,000 persons per year
[14,18]. The median age at diagnosis is between 3
and 8 years [19,20], with 70% to 80% occurring in
doi:10.1016/S1042-3680(03)00062-7
572
A.S. Dumont et al / Neurosurg Clin N Am 14 (2003) 571591
children less than 8 years of age and 40%

occurring in those less than 4 years of age
[19,2123]. There has been a recent trend reported
by the Childhood Brain Tumor Consortium study
in the relative proportion of older children (>11
years of age) with ependymomas (especially
supratentorial) [21]. Although some series have
shown a slight male predominance [15,19], there
does not appear to be a gender dierence across
most series [20,22,24]. Ependymomas predominantly arise from a ventricular surface (being of
ependymal origin), although they may rarely
develop without any direct association to a ventricular surface [25]. Approximately two thirds are
infratentorial in location (most commonly arising
from the oor of the fourth ventricle in children)
with the remaining one third originating within
the supratentorial compartment [19,20].
Not unexpectedly, the most common presenting signs and symptoms stem from raised intracranial pressure from mass eect or
obstruction of cerebrospinal uid (CSF) ow
and hydrocephalus. Supratentorial ependymomas
most commonly present with signs and symptoms
of raised intracranial pressure but may also cause
seizures from cortical irritation or rarely present
with apoplexy and intracerebral hemorrhage [26].
Patients with infratentorial tumors often present
with a longer clinical history, although this varies.
Common symptoms include headache, emesis,
lethargy, irritability, and poor balance, whereas
clinical signs observed include increasing head
circumference (across percentiles), bulging fontanelle, papilledema, meningismus, ataxia, nystagmus, and cranial nerve palsies [19,22]. The
expeditious diagnosis is more dicult in younger
children, particularly in those with nonspecic
symptoms, such as lethargy and irritability with
occasional emesis. To avoid signicant delay in
diagnosis (and, ultimately, delay in necessary
treatment), the clinician must maintain a high
degree of suspicion, especially in this age group.
Imaging
Clinical suspicion of a possible ependymoma
is conrmed with CT or MRI. On CT scanning
without contrast, ependymomas often appear
hyperdense, with some heterogeneity in approximately 85% of cases, secondary to the presence of
solid and cystic components as well as calcication
(although infratentorial lesions less commonly
have a cystic component [17,27]) [28,29]. Specically, calcication on CT scanning is encountered

in 50% to 80% of cases [27,28,30]. Ependymomas
demonstrate signicant contrast enhancement that
may assume a heterogeneous or homogeneous
pattern [27,28,30]. Additionally, peritumoral edema is present surrounding the typically welldelineated tumor margins [27,29].
MRI of ependymomas most clearly denes the
pathologic ndings in each case. Tumors typically
demonstrate hypo- to isointense signal on T1weighted images relative to parenchyma, with
hyperintense signal on T2-weighted sequences
[31], although signal characteristics may be relatively nonspecic [32]. Calcication is less well
appreciated on MRI but seems to be present in
approximately 50% of cases [31]. As with CT,
tumors appear fairly well demarcated on MRI,
with considerable contrast enhancement after
gadolinium injection (Fig. 1). Again, supratentorial lesions are more often of variable consistency
(solid and cystic) than infratentorial lesions [31].
Evidence of intratumoral hemorrhage may be seen
in nearly 60% of supratentorial lesions and
approximately 30% of infratentorial lesions [32].
All infratentorial lesions are, at least in part,
intraventricular in location on MRI, whereas 50%
of supratentorial lesions appear to be intraventricular, with the remainder appearing to be
entirely intraparenchymal [27,31]. With respect to
the infratentorial lesions, at least half of the tumors
extend from the fourth ventricle out the foramina
of Luschka or Magendie into the cerebellopontine
(CP) angle or cisterna magna, respectively [27,31].
Likely because of their intimate association
with CSF pathways, ependymomas have a propensity for CSF spread, with between 5% and
22% of children at diagnosis appearing to have
documented leptomeningeal spread [19,20,23].
Consequently, it has become well accepted to
obtain preoperative MRI of the entire spine in
combination with sampling of CSF for cytology
for staging, particularly for fourth ventricular
lesions (see Fig. 1). There has been an apparent
increase over the last several decades in the
proportion of cases with documented metastases
at diagnosis, likely secondary to advancements in
imaging and emphasis on thorough staging.
Pathology
The World Health Organization categorizes
ependymal tumors into the following four distinct
categories [33]:
573
Fig. 1. Grade III ependymoma. (A) Head MRI reveals ventricular dilatation and an intensely enhancing mass in the
third ventricle on T1-weighted images after intravenous gadolinium administration. (B) The patient had evidence of
spinal drop metastases at the time of diagnosis (arrow).
Ependymoma (subtypes cellular, papillary,

clear cell, and tanycytic)
Anaplastic ependymoma
Myxopapillary ependymoma
Subependymoma
SEs are discussed in a separate section, and
myxopapillary ependymomas, found almost exclusively in the conus medullarislum terminale
cauda equina region of the spine, are not
discussed further.
Ependymomas are characterized by unique
histologic, immunohistochemical, and ultrastructural features. They typically seem to be well
circumscribed and moderately cellular with monomorphic nuclear morphology [33]. On light
microscopy, important histologic features include
perivascular pseudorosettes and ependymal rosettes with rare or absent mitotic gures. The
perivascular pseudorosettes occur in most ependymomas, with tumor cells being radially arranged around a blood vessel. True ependymal
rosettes are diagnostic, although rare, and consist
of columnar cells arranged concentrically around
a central lumen. Fibrillary elements are commonly
observed, in addition to regressive changes, including evidence of myxoid degeneration, intratumoral hemorrhage, calcications, intratumoral
hemorrhage and foci of cartilage, and bone
formation on occasion [33].
Cellular ependymomas have increased cellularity with an increased mitotic rate. Pseudoro-
settes and ependymal rosettes may be absent [33].

Papillary ependymomas are rare, being characterized by well-formed papillae in which blood
vessels are enveloped by smooth layers of tumor
cells [33]. With cells exhibiting clear perinuclear
halos similar to oligodendroglial tumor cells, clear
cell ependymomas seem to occur disproportionately in supratentorial ependymomas arising in
young patients [33]. The presence of rosettes,
immunoreactivity for glial brillary acidic protein
(GFAP), and electron microscopy studies are
useful in dierentiating this subtype from other
tumors, such as oligodendroglioma, clear cell
carcinoma, hemangioblastoma, and central neurocytoma [33]. Tanycytic ependymomas consist of
arrangement of tumor cells into fascicles of
varying width and cell density [33]. Ependymal
rosettes are often absent, and pseudorosettes are
poorly formed. There is a predisposition for
involvement of the spinal cord in this ependymal
subtype [33].
Electron microscopic examination of ependymomas demonstrates the presence of frequent
glandlike lumina with microvilli and cilia, basal
bodies, intracytoplasmic intermediate laments,
and long distinct junctional complexes [33].
Microrosettes may also be seen.
Anaplastic ependymomas typically exhibit
frequent mitoses, marked cellular polymorphism,
a high nuclear-to-cytoplasmic ratio, necrosis, and
microvascular proliferation [33]. The distinction
574
between low-grade and anaplastic ependymomas

is dicult, however, with dierent tumors displaying a spectrum of pathologic changes. Additionally, a focus or foci of anaplastic tissue may be
seen in an otherwise low-grade lesion, the signicance of which has not been established.
Uniform and accurate diagnosis of an ependymoma is obviously important to management,
especially after surgery. It must be realized,
however, that there is considerable discrepancy
in diagnosis between skilled pathologists. For
instance, a recent prospective randomized trial
conducted by the Childrens Cancer Group
demonstrated a discordant pathologic diagnosis
between an individuals treating institution and
the Central Review Board in 69% of cases [20].
This lack of uniformity has hampered past
studies, and data must be scrutinized in the
present body of literature. Future eorts to
improve the precision and accuracy of diagnosis
should help to rectify this issue.
Treatment and outcome
The mainstay of treatment in the management
of patients harboring ependymomas remains
surgical resection, with the goal of total removal
whenever possible. The use of postoperative
therapy is a current area of active investigation
and is dependent on multiple factors, including
the extent of resection, preoperative staging, histologic type of tumor, age of the patient, and
patient/familys wishes among other factors.
Surgery
The surgical approach is tailored to the
individual patient and depends on all clinical,
radiologic, and pathologic data. It is also inuenced by intraoperative data, such as the tumors
consistency and relation to critical neural structures (ie, adherence to cranial nerves or the oor
of the fourth ventricle). As previously emphasized,
preoperative staging is important in guiding
subsequent therapy. With widespread metastases,
the surgeons enthusiasm for removing the last
minute part of the tumor densely adherent to the
oor of the fourth ventricle should be dampened,
given the already apparent need for potentially
aggressive postoperative therapy.
When patients present with acute symptomatic
hydrocephalus, immediate intervention is imperative. The rst line of treatment is intravenous
high-dose dexamethasone (10-mg initial dose,
followed by 6 mg every 6 hours). Often, improve-
ment is rapidly seen and may obviate the need for

urgent CSF diversion procedures. If high-dose
steroids are not eective, an external ventricular
drainage catheter is placed as a temporizing
measure until the tumor can be removed after
the necessary preoperative evaluation has been
performed. Permanent ventriculoperitoneal shunting is avoided if at all possible.
The authors surgical considerations are briey
detailed. Intraoperative frameless stereotactic
image guidance is frequently used, particularly
with deep supratentorial tumors. Real-time intraoperative ultrasonography can be used to localize
the tumor in the depths below the cortical surface.
Only infrequently would the use of intraoperative
mapping and electrocorticography be contemplated (eg, with tumors resulting in epilepsy). In
all instances, perioperative steroids and antibiotics
are administered. For supratentorial ependymomas, anticonvulsants are administered before
surgery and continued for only 3 months if there
has never been a seizure.
There are dierent surgical considerations for
supratentorial and infratentorial ependymomas.
For supratentorial ependymomas, the patient is
placed in a three-point xation apparatus and
positioned to optimize venous outow (head
above the level of the heart with avoidance of
neck kinking). The skin is inltrated with local
anesthetic (0.2% ropivacaine with 1:100,000 epinephrine) before incision. Mannitol (0.250.5 g/
kg) and mild hyperventilation (end-tidal PCO2 of
2535 mm Hg) may be implemented in cases
associated with raised intracranial pressure. An
approach is chosen that allows optimal exposure
while minimizing potential complications. The
two major approaches, depending on the location
of the tumor, include an interhemispheric transcallosal approach (anterior or posterior) and
a transcortical transventricular approach. In
general, the interhemispheric transcallosal approach is useful with midline lesions, particularly
those of the third ventricle, whereas the transcortical transventricular approach is useful for
more lateral lesions situated within the lateral
ventricles. Frameless stereotactic image guidance
is useful in precisely planning the proposed
craniotomy. Frameless stereotaxy and intraoperative ultrasound are useful in localizing the tumor
or planning the corticotomy where appropriate.
The microscope is used for the tumor resection.
Under high magnication, the tumor is coagulated circumferentially and subsequently debulked
internally. The capsule can then be methodically
dissected and pulled away from its attachments.

Meticulous hemostasis is critical, and avoidance
of intraventricular blood collection is imperative
for the prevention of postoperative chemical
meningitis or hydrocephalus. A temporary ventricular drain may be placed during surgery at the
discretion of the surgeon. If placed, the drain can
usually be removed in the early postoperative
period.
For infratentorial ependymomas with signicant hydrocephalus refractory to steroids, an
external ventricular drain alone or third ventriculostomy with a prophylactic external ventricular
drain can be placed before surgery. If a drain has
not been placed before surgery, a potential site is
prepared and draped into the surgical eld at the
time of the tumor resection but is opened only if
needed. The patient is placed in a prone position
with the neck exed and the head xated in
a three-point apparatus. A midline incision is used
most often, which may be linear or hockey-stick
in nature. A paramedian incision is used in the less
common instance in which a tumor is located
predominantly in the CP angle. We have more
recently preferred posterior fossa craniotomy
rather than craniectomy, especially in pediatric
patients. The bony opening should permit access
to the entire tumor (including exposure to the
transverse-sigmoid sinus junction when exploration of tumor extrusion into the CP angle is
necessary). Depending on the inferior extension of
the tumor, a C1 laminectomy may have to be
performed to expose the tumor completely. The
dura is usually opened in a Y-shaped fashion. At
times, the cerebellum may be under considerable
pressure. Rapid opening of the cisterna magna is
usually sucient to alleviate this, and only in
exceptional circumstances would a ventriculostomy need to be placed if not present before
surgery. The tumor is often seen extruding
through the obex. The posterior-inferior cerebellar arteries are identied bilaterally. Under the
operating microscope, the tonsils are separated
and the vermis is often split in the midline. One
of the most important points of emphasis is to
discern and continue to be keenly aware of the
exact location of the oor of the fourth ventricle.
Once identied, a moist Telfa (Kendall Company,
Manseld, MA) patty may be placed over the
oor to protect it and serve as a landmark. The
tumor is carefully dissected from the oor of
the fourth ventricle; however, excessively aggressive attempts to remove a densely adherent tumor
are tempered by the goal of preserving neurologic
575
function. The lateral extent of the tumor may also

be a signicant challenge because of the potential
intimate association with blood vessels and lower
cranial nerves. Careful sharp dissection under
high magnication is useful, but attempts to
remove minute pieces of tumor at the expense of
neurologic function are to be avoided. The basic
technique of internal debulking, followed by
dissection and pulling in the sides of the tumor,
is useful here as well. After resection, the dura is
usually closed in a watertight fashion with a dural
graft.
After surgery, it is the authors policy for
patients to undergo MRI to determine the extent
of resection within the rst 72 hours. Postoperative imaging is believed to be more uniformly
accurate than the objective intraoperative impression of the extent of resection.
The results of surgical resection reported in the
literature have varied considerably. Earlier series
have reported rates of gross total resection of 22%
to 30% [22,34,35], whereas contemporary series
have reported rates of complete resection of 43%
to 71% [36,37]. When interpreting the data
concerning extent of resection, it is important to
consider the criteria for determining the extent of
resection (surgeons intraoperative impression
versus postoperative CT or MRI). Some series
have only included the extent of resection based
on the surgeons intraoperative impression [36]. In
terms of operative mortality, modern series have
reported mortality rates of 0% to 2% for supratentorial tumor resection [38,39] and 0% to 13%
for infratentorial tumor resection [22,34,35,
3941]. The rates of morbidity for operative
intervention are less frequently carefully recorded
and reported. Visual eld decits associated with
disruption of the optic tract have been published
to occur in 20% to 30% of patients after resection
of supratentorial ependymomas [42,43]. Morbidity associated with infratentorial ependymoma
resection largely stems from injury of the lower
cranial nerves or their nuclei and the brain stem
and may be around 10% to 14% [27]. Future
appraisals of true operative morbidity must
include neuropsychologic assessment, which is
typically lacking in most surgical series.
Radiation therapy
Postoperative therapy after surgical resection
of ependymomas has been an area of much
interest but continues to be a matter of controversy. A signicant amount of data has been
accrued demonstrating that postoperative
576
radiation has a signicant impact on survival

[22,34,35,3941,4446], with increases in 5-year
survival from less than 20% up to 40% to 50% in
selected patients [44,45]. Radiation has become
the primary postoperative adjuvant therapy in
patients harboring ependymomas. Despite this,
the timing, method, and extent of radiation
therapy linger as areas of contention and prospective randomized study of postoperative radiation is yet to be conducted.
Pioneering work on the role of radiation
therapy in patients with ependymomas established
that radiation doses less than 4500 cGy to the
primary tumor site did not seem to be ecacious
[4749]. Subsequent work has generally established a dose of 5000 to 5500 cGy over 5 to 6
weeks for the treatment of subtotally resected
nonanaplastic ependymomas [5052]. For patients
with aggressive ependymomas, some authorities
have recommended escalated doses (55006000
cGy) over 6 to 7 weeks. Recent interest has also
been given to the use of stereotactic radiosurgery
for recurrent ependymoma and has been suggested for use in the initial postoperative treatment of ependymomas [53]. The early results of
stereotactic radiosurgery seem to be somewhat
promising, and future study is warranted.
The target for radiation therapy administration
has also received considerable attention. In general, supratentorial low-grade ependymomas can be
treated with focused radiation targeting the
surgical site, because the incidence of leptomeningeal spread is quite low [50,51,54]. When compared with whole-brain irradiation, local eld
radiation therapy (with or without a boost)
demonstrated a signicant 10-year progressionfree survival rate [42]. The literature concerning
postoperative radiation therapy for infratentorial
ependymomas is more heterogeneous, and the
practice across institutions is quite variable,
especially in those cases with benign histology
and no evidence of leptomeningeal metastases
[42,52,5460]. The authors practice is to administer local radiation therapy in cases of subtotally
resected benign ependymoma without prophylactic craniospinal irradiation. Craniospinal radiation
therapy is reserved for those patients with evidence
of leptomeningeal dissemination by MRI or CSF
cytology and for selected patients with malignant
ependymomas. It is important to note that 80% of
cases with leptomeningeal spread occur in patients
with high-grade ependymomas [50,51,61].
One of the emerging paradigms with postoperative radiation therapy is to avoid its
administration to children less than 3 years of

age because of its potential deleterious long-term
neurologic and neuropsychologic sequelae. We
have also chosen not to administer postoperative
radiation therapy in selected patients with lowgrade ependymomas in whom complete resection
was achieved and conrmed by postoperative
MRI, particularly with young patients and supratentorial tumors. These patients are followed
closely with serial imaging for any sign of
recurrence. Other reports of this practice have
demonstrated long-term survival in patients who
underwent total resection without postoperative
radiation therapy [43,62].
Chemotherapy
Another facet of postoperative care that
remains controversial is the administration of
adjuvant chemotherapy. The overall ecacy of
chemotherapy has been disappointing, although
most reports on the treatment of patients with
ependymomas are based on anecdote and small
numbers [19,20,39,6376]. The most extensively
studied and potentially active agent in patients
with ependymomas has been cisplatin [7679].
These data, taken collectively, demonstrate an
overall response rate of 33% with cisplatin, of
which 18% were complete responses. Other less
well-studied agents that have demonstrated at
least some ecacy include carmustine, lomustine,
etoposide, cyclophosphamide, dibromodulcitol,
and carboplatin [64,8083]. Other agents, including 1-(2-chloroethyl)-3(2,6 dioxo-1-piperidyl)-1nitrosource (PCNU), thiotepa, ifosfamide, and
idarubicin, have been investigated, with a paucity
of demonstrated ecacy [8492]. Anecdotal unpublished data also suggest occasional partial
response or stable disease with temozolomide and
procarbazine.
Overall, the response rate to single chemotherapy agents in patients with ependymomas seems
to be 11%, with less than 5% complete responses
[76]. Combination drug regimens have also been
investigated, with limited ecacy, including
mechlorethamine, vincristine, procarbazine, and
prednisone (MOPP) and eight agents in 1-day
regimens [20,67,6975,78,93,94]. The use of
high-dose chemotherapy with bone marrow/stem
cell rescue has been disappointing thus far in
patients with ependymomas [95,96]. Nevertheless,
an area of emerging promise and recent interest
has been in the use of adjuvant chemotherapy in
place of radiation therapy in young children to
circumvent the deleterious eects of radiation
exposure to the immature and rapidly developing

nervous system [37,65,70,71,75,97]. In particular,
the Pediatric Oncology Group presented interesting data demonstrating success with the administration of chemotherapy, especially in patients
aged 24 to 36 months, in terms of safely delaying
adjuvant radiation therapy and improving survival [75].
In summary, based on the available data,
chemotherapy has not proven to be particularly
ecacious in the treatment of ependymomas.
Chemotherapy should probably be used largely
in clinical trials and protocols. Future appropriately powered prospective investigations are
warranted to delineate the role of chemotherapy
in patients harboring ependymomas. Of special
interest remains the delineation of the role of
chemotherapy in young patients, in whom radiation therapy is highly undesirable.
Prognostic factors and outcome
The ability to predict prognosis for patients
with ependymomas is important from both patients and clinicians perspectives. The establishment of predictive factors has been an arduous
task, in part, because of signicant diculty in
comparing outcomes between dierent studies and
even within studies at times. Ependymoma is
a relatively rare tumor, and signicant experience
with treatment has accrued slowly over time.
Hence, an individual institutions experience is
relatively limited and has developed over dierent
eras, thereby introducing heterogeneity within and
between series. Furthermore, as previously alluded
to, there has been considerable diculty in
arriving at consensus pathologic diagnoses between dierent observers, thereby introducing
additional variability. The nature of most series
has also been retrospective case series with its own
inherent biases, also limiting the quality of available data.
Extent of resection. The preponderance of available data demonstrates that the completeness of
resection (especially gross total resection) is
correlated with improved prognosis [20,22,24,36,
37,39,46,75,98100]. From these data, 5-year
survival has ranged from 60% to 93% after gross
total resection compared with 21% to 46% after
subtotal resection. As mentioned, the means by
which assessment of the extent of resection is
made is clearly important, and the present gold
standard is establishment by postoperative MRI.
In fact, there have been data demonstrating that
577
radiologic imaging rather than the surgeons

intraoperative impression is statistically predictive
[101]. Furthermore, even a small amount of
residual tumor (<1.5 cm2) may predict improved
survival [20].
Age. In children, age seems to be an important
prognostic factor in most studies, with children
older than 3 or 4 years of age seeming to have
longer survival, with 5-year survival rates ranging
from 55% to 83% compared with 12% to 48%
for their younger counterparts [19,22,24,39,
59,100]. Even in subgroups of patients less than
3 years old, those older than 2 years of age seem to
fare better, with 5-year survival rates of 63%
versus 26% between groups [75]. It should be
noted that not all reports have conrmed age as
a prognostic factor [20,61,98,102].
Histologic grade. Histologic grade as a prognostic
factor has been controversial across dierent
series. This may be due, at least in part, to diculty in establishing consensus diagnoses between
pathologists. A diagnosis of higher grade/anaplastic ependymoma portending a poorer prognosis
makes intuitive sense but has not been universally
borne out. Considerable data exist to suggest that
anaplastic ependymomas or those with higher
grade features are associated with a poorer
prognosis [22,24,36,38,43,61,99,102107]. There
exists a signicant body of literature in which
histologic grade was not established as an independent prognostic factor, however [19,20,37,
39,75,108]. Of note, two of these studies were
prospective in nature, including a prospective
cohort study from the Pediatric Oncology Group
[75] and a randomized trial from the Childrens
Cancer Group [20]. These studies are still limited
in sample size, however, and disagreement in the
pathologic diagnosis occurred in a disappointing
69% of cases in an independent review process in
the Childrens Cancer Group study [20], which
makes it dicult to draw rm conclusions.
Tumor location. Numerous studies suggest that
an infratentorial location is associated with better
prognosis compared with supratentorial lesions,
with 5-year survival rates for infratentorial tumors
ranging from 35% to 59% compared with 22% to
46% for supratentorial tumors [19,24,98,109,110].
Furthermore, certain features of infratentorial
tumors may help to predict prognosis. A significant lateral extension into the CP angle seems to
be a poor prognostic factor, which may be
578
ascribed to additional diculty in complete removal because of intimate association with lower
cranial nerves and critical vascular structures [34].
Miscellaneous factors. Although the aforementioned prognostic factors have been best characterized, several other factors have surfaced as
putative prognostic indicators. Gender, race, and
duration of symptoms have been suggested as
possible prognostic factors [19,99,100], although
they are questionable in their predictive ability.
Additional evidence is needed before any of these
latter factors could be considered a rm prognostic indicator.
Despite considerable progress in the management of patients with ependymomas, overall
prognosis remains relatively poor. Most patients
eventually develop recurrence, and the most
common site for recurrence seems to be the primary
tumor site [24,61,71,109]. Local recurrence remains
the primary cause of progressive neurologic decits
[24]. Tumor recurrence manifesting as metastases
without local recurrence is quite rare, occurring in
7% to 8% of cases [20,104]. The most important
factors in preventing recurrence are thus the same
factors as those for establishing local tumor
control. Unfortunately, treatment of tumor recurrence is quite limited. Repeat surgery is contemplated for local recurrence. Radiation therapy
has usually been administered; hence, further
radiation therapy is not usually an option. Chemotherapy may be administered, but its ecacy
has been disappointingly poor.
The overall 5-year survival rate for children
with ependymomas is approximately 39% to 93%
[19,20,22,34,36,37,99,101,104,111]. The 10-year
survival rates range from around 45% to 75%
[34,36,37,100,101]. The absolute survival rates do
not reveal the burden of neurologic morbidity
with which these patients may live. In children,
survivors have relatively low IQs and impaired
academic and psychosocial functioning that limit
their ability to interact with their environment.
Much of this morbidity may be iatrogenic, and
future eorts directed toward minimizing insult by
treatment on the developing nervous system are
clearly necessary.
Subependymomas
SEs are relatively uncommon well-dierentiated tumors associated with the ventricular
system; they are generally characterized by slow

growth and an indolent clinical course. SEs are
thought to originate from the subependymal glial
matrix, consisting of a mixture of astrocytic,
ependymal, and transitional cell clusters surrounded by neuroglial bers [112]. Scheinker
[113] is generally given credit for the description
of SE as a distinct entity in 1945, although tumors
had been described previously that contained both
astrocytic and ependymal features. The exact
etiology of these tumors remains in doubt. SEs
have been reported to occur in families and twins,
leading to speculation of a specic genetic
mechanism, although, to date, a mechanism has
not been clearly identied [114116]. Some
suggest that SEs might represent hamartomas, in
part, because of the fact that they have been
associated with heterotopic neuroglial tissue in the
leptomeninges [117]. A reactive origin has also
been postulated as a result of reports of concurrent presentation of SE with hydrocephalus,
ependymitis, and meningitis, but this mechanism
remains in doubt [118]. On occasion, SE arises
concurrently with other primary neoplasms, such
as glioblastoma, meningioma, or choroid plexus
papilloma, but this seems to be coincidental
[119,120].
The incidence of SE at autopsy is 0.4%, and
the incidence of SE in intracranial surgical tumor
specimens ranges from 0.2% to 0.7% [121].
The most common site of presentation is the
fourth ventricle, but SEs may also occur in the
lateral ventricle and the aqueduct of Sylvius. SEs
may occur at extraventricular locations, which
include the septum pellucidum and the cervicothoracic spinal cord. In one series of 69 surgical
patients, 70% were male and the average age was
39 years [122]. In this report, two thirds of
patients had tumors in an infratentorial location,
one third had supratentorial tumors, and 2% had
tumors in the cervicothoracic spinal cord [122].
SEs are relatively rare in the septum pellucidum,
consisting of 5% of all reported cases [112].
Whether incidental or symptomatic, SEs occur
far more frequently in the adult population. There
do not seem to be reports of SE detected in infants.
Most SEs are asymptomatic during the life of the
patient and discovered at the time of autopsy,
despite the fact that some of these lesions are quite
sizable. Of those patients who present in the clinic,
there seems to be a bimodal age distribution that
depends on the mode of presentation. Symptomatic patients usually present with obstructive
hydrocephalus and tend to be younger (average
age of 40 years) as opposed to those asymptomatic
patients who present at an average age of 60 years
[123]. It is estimated that 40% of SEs become
symptomatic [124]. Most asymptomatic patients
have lesions that are discovered on imaging studies
performed for clinical indications that are unrelated to the neoplasm (Fig. 2). Nevertheless, there
579
is a report of a sudden death related to a previously

asymptomatic SE [125]. Symptomatic SEs usually
present either with increased pressure because of
CSF obstruction or hemorrhage [126]. Predictably,
tumors that arise from the septum pellucidum, the
region of the foramen of Monro, or the aqueduct of
Sylvius are the most likely to cause symptoms
because of obstruction of CSF ow. Symptomatic
tumors tend to be larger, and obstructive hydrocephalus is present in up to 88% of cases [122].
Fig. 2. Subependymoma. This patient originally presented with progressive ataxia. (A) Initially, the symptoms were
attributed to the mildly enhancing fourth ventricular mass (white arrow). (B) On closer inspection, the patient had
signicant spinal cord signal abnormality and myelomalacia (white arrow) related to basilar invagination and an
occipitalized atlas (black arrow). (C ) The patient underwent posterior fossa and C-1 decompression, total resection of the
subependymoma, and fusion.
580
Prominent symptoms and signs on presentation are

headache, visual disturbance, papilledema, gait
ataxia, memory disturbances, cranial nerve paresis,
nystagmus, spasticity, vertigo, and vomiting. Tumors arising in the area of the septum pellucidum
may produce personality disorder, memory impairment, loss of consciousness, or seizure disorder
[127]. Spinal cord SEs usually result in cord
compression and produce symptoms referable to
the spinal level of involvement.
Radiology
SEs most often arise from the region of the
lower medulla and project into the fourth ventricle.
Fourth ventricular tumors grow from the oor or
roof and may extend laterally via the foramina of
Luschka to occupy the subarachnoid space. In
fact, a tumor originating in the lateral recess of the
fourth ventricle may grow out the foramen of
Luschka, erode the petrous bone, and simulate
other CP angle tumors [128]. Another common
location is in the frontal horn of the lateral
ventricle, where they may attach to the septum
pellucidum or the lateral ventricular wall. A few
SEs are found along the midbody of the lateral
ventricle [123]. CT reveals a well-delineated mass,
which is hypodense, isodense, or even slightly
hyperdense to brain parenchyma. Contrast enhancement is often not seen on CT, but, when
present, enhancement tends to be homogeneous
but not intense (Fig. 3). Edema tends to be
uncommon. On CT, intense diuse enhancement
with edema should arouse suspicion of a mixed
ependymoma-SE, which has a signicantly more
aggressive natural history [129]. Cyst formation,
focal calcication, and hemorrhage can be seen
on CT. Dense calcications are more common for
tumors that arise in the fourth ventricular location.
MRI reveals homogeneous hypointense to
isointense masses on T1-weighted imaging. SEs
may be mildly hyperintense on T2-weighted and
gradient echo imaging. Signal heterogeneity can be
present as a result of cyst formation. Contrast
enhancement is typically absent, but gadolinium
enhancement may occur (Fig. 4). When contrast
enhancement is present, it is more likely to occur in
fourth ventricle sites [130,131]. The dierential
diagnosis of fourth ventricular SE with similar
imaging characteristics includes metastasis or ependymoma. Central neurocytoma is one of the major
dierential considerations for frontal horn SE.
Cerebral angiography may disclose arterial
displacement around the mass and stretched
subependymal veins indicating ventricular dilatation, but neovascularity is absent. Cerebral

angiography probably does not have signicant
utility as a diagnostic test because of lack of
signicant tumor vascularity.
Pathology
SEs are characterized by their intraventricular
location, circumscribed nature, lobulated appearance, infrequent multiplicity, sharp demarcation,
slow growth, and usually noninvasive nature.
Growth is typically by expansion rather than by
inltration. These lesions originate immediately
beneath the ependymal surface and tend to displace
the ependymal surface as they enlarge over time.
On gross inspection, SEs are rm, well-demarcated, grayish-white to tan, avascular, intraventricular
masses that are rmly attached to their site or
origin at the septum pellucidum, foramen of
Monro, lateral ventricle, or inferior fourth ventricle. In the fourth ventricle, the tumor may be
primarily attached to the oor, roof, or lateral
recesses via a vascular pedicle. Secondarily, SEs
may adhere to adjacent ependymal surfaces other
than the site or origin, particularly as the tumors
grow larger. The gross appearance may be modied
by calcication, hemorrhage, or cyst formation.
Once considered a variant of ependymoma,
SEs are now placed in a separate subcategory of
ependymal tumors. SEs are designated as World
Health Organization grade I. The cell of origin for
SE may be a bipotential subependymal cell with
capabilities to dierentiate into ependymal or
astrocytic cells [116,128]. Microscopic examination discloses a sparsely cellular neoplasm with
a prominent brillary background. The concept of
ependymal dierentiation in SE is supported by
its nuclear characteristics, ultrastructural features,
and occasional coexistence of a mixed ependymoma component. Supporting an astrocytic lineage is the abundance of long processes that are
rich in glial laments. This characteristic microscopic appearance led to the term subependymal
glomerate astrocytoma [132]. Microcystic changes
are common, but hypercellularity, true ependymal
rosette formation, neovascularity, and necrosis
are generally absent. Nuclear atypia and limited
mitotic activity do not seem to be of prognostic
signicance. Lateral ventricular SEs have occasional mitoses and hyalinized vessels, are more
infrequently calcied, and are more astrocytic
in appearance. SEs of the fourth ventricle are
usually more suggestive of ependymomas. Tumors
581
Fig. 3. Subependymoma. (A) A nonenhancing lesion in the right frontal horn of the lateral ventricle. (B) The mass has
signicant hyperintensity on gradient echo MRI. (C ) An unusual appearance of intense homogeneous enhancement after
administration of intravenous gadolinium.
at these locations usually lack microcystic change,

calcication is more common, and the nuclear
chromatin pattern is more reminiscent of ependymal cells.
Immunohistochemical analysis reveals frequent GFAP positivity and common vimentin
and S-100 positivity, but the Ki-67 labeling index
(MIB-1) is normally quite low [133]. Electron
microscopy shows closely packed cell processes
lled with intermediate glial laments. Ultrastructural ependymal features, such as microvilli and
cilia, are most frequent among tumors arising in
the fourth ventricle. Some histologic variants of
SE have been reported, including rhabdomyosarcomatous dierentiation, sarcomatous proliferation of the vasculature, and the presence of
melanin pigment [134136]. When mixed with
ependymoma, these areas reveal true or pseudorosettes, increased vascularity, increased mitosis,
necrosis, and increased cellularity [129].
Treatment and prognosis
Surgery
Advances in microsurgery have improved
surgical outcomes. One study compared the
582
Fig. 4. Subependymoma. (A) Isointense mass on T1-weighted images in patient with headaches and nystagmus. (B)
Minimal enhancement after intravenous gadolinium administration. (C ) Total resection was achieved despite broad
attachment to the oor of the ventricle.
outcome of patients who underwent resection

between 1950 and 1974 with the outcome of those
who underwent surgery between 1975 and 1989
[137]. Mortality decreased from 25% to 0%. The
plan for surgical treatment is usually radical
excision. This goal is often achievable, because
SE growth patterns tend to be expansive rather
than inltrative.
Tumor location and extent of resection are the
most important prognostic factors. SEs of the
lateral ventricle may be resected with relative ease.

Septum pellucidum and frontal horn lesions may
be approached either through a transcallosal or
transcortical transventricular route as previously
described. Total excision of fourth ventricular
tumors may be more dicult because of attachment to the oor of the fourth ventricle. Reports
have described sharp demarcation allowing complete removal in about half of the cases and
higher perioperative morbidity in fourth ventric-
ular locations [121]. Fourth ventricular SEs are

usually approached through a standard midline
posterior fossa approach with splitting of the
vermis, similar to the approach described for
ependymomas. For symptomatic lesions, surgical
debulking may re-establish CSF ow. Particularly
in older patients, debulking may be equivalent to
a cure in some instances, because it could take
many years for a symptomatic mass to have
a substantial recurrence. In a retrospective review,
no tumor recurrence was noted in a series of 12
patients who underwent total or subtotal resection [133].
Treatment of asymptomatic lesions that are
incidentally discovered is decidedly less clear. If
the lesions remain asymptomatic and do not
exhibit growth on serial imaging, expectant
management is usually undertaken. If ventricular
enlargement is proven, tumor growth is detected
radiographically, or there is considerable doubt
with regard to the diagnosis, surgical treatment
should be considered, however.
Radiation and chemotherapy
The literature is too sparse to draw conclusions
for a signicant benet from radiation therapy.
Thus, radiation therapy is generally not recommended for incompletely resected asymptomatic
tumors. Nevertheless, it is important to dierentiate SEs from ependymomas because of a much
better prognosis and diering treatment strategies.
Perhaps as many as one fourth of symptomatic
tumors may have an admixture of both SE and
ependymoma components and seem to have a less
favorable prognosis resembling that of pure
ependymoma [129]. In cases of mixed ependymoma-SE or where there is signicant nuclear
pleomorphism, radiation has been proposed [138].
CSF dissemination of SE has not been reported.
We are not aware of any study that advocates
chemotherapy for the treatment of SE.
Subependymal giant cell astrocytomas

SEGAs are a relatively rare form of astrocytoma that is characteristically associated with
tuberous sclerosis complex, an autosomal dominant phakomatosis. SEGAs may occur independent of tuberous sclerosis, although spontaneous
cases may occasionally represent a forme fruste
of the neurocutaneous syndrome. After tubers,
583
SEGAs are the second most common tumor

aecting patients with tuberous sclerosis [139].
Reports reveal that 3% to 14% of patients with
tuberous sclerosis have SEGAs [139,140]. In
a large series of 345 tuberous sclerosis patients,
6.1% had SEGAs [141]. Rarely is there an
association between malignant glial tumors and
tuberous sclerosis [142].
The peak age of incidence is between 5 years of
age and the midteens [139,140]. The earliest
reported occurrence is of that in a premature infant
[143]. Approximately 20% of SEGAs present in
adulthood, however [144]. There does not seem to
be a racial or gender predilection. In most cases,
the tumors are found associated with the ventricular wall near the foramen of Monro, resulting in
a presentation that is characteristic for obstructive
hydrocephalus. Symptoms in infants include enlarging head circumference, irritability, lethargy,
and vomiting. Older children may present with
headache, nausea, vomiting, or exacerbation of
seizure disorder. The diagnosis is usually made by
the location of the tumor and by the association of
other stigmata of tuberous sclerosis, including
adenoma sebaceum, mental retardation, and myoclonic seizures (Vogts triad for clinical diagnosis),
and the presence of subependymal or cortical
tubers and heterotopic gray matter. The dierential diagnosis for tumors of the lateral ventricle also
includes ependymoma, SE, neuroblastoma, astrocytoma, oligodendroglioma, meningioma, central
neurocytoma, and choroid plexus papilloma.
Radiology
On CT, SEGAs are well-circumscribed isodense or hyperdense masses that demonstrate
intense homogeneous contrast enhancement. The
tumors tend to protrude into the ventricle and
arise from the sulcus terminalis [145]. The tumors
may contain some calcication. SEGAs associated
with tuberous sclerosis are typically heavily
calcied and have strong but inhomogeneous
enhancement after contrast administration. Tubers, although they may calcify, do not enhance
after intravenous injection of contrast dye [146].
On MRI, SEGAs are isointense or hypointense
on T1-weighted imaging and hyperintense or
heterogeneous on T2-weighted imaging and may
enhance signicantly with contrast administration
(Fig. 5). Calcications may be seen as signal voids.
584
Fig. 5. Subependymal giant cell astrocytoma. Axial (A) and coronal (B) MRI after intravenous gadolinium
administration demonstrating an intensely enhancing mass expanding the atrium of the left lateral ventricle. This
patient did not have stigmata of tuberous sclerosis.
Most SEGAs inhomogeneously enhance after

gadolinium administration. In contradistinction,
tubers generally do not exhibit hyperintensity on
T2-weighted imaging and do not usually show
contrast enhancement [147]. The current diagnostic imaging modality of choice is MRI with T2weighted images, with and without gadolinium,
including coronal sections through the region of
the foramen of Monro.
SEGAs have variable vascularity on angiography; a prominent blush may be present on the late
arterial phase. Venous phase lms show stretched
and elongated subependymal veins when ventriculomegaly is present.
Pathology
SEGAs are usually well-demarcated lobulated
masses that often appear calcied. Even macroscopically, cysts are not uncommon. SEGAs may
be quite vascular and have a reddish or hemorrhagic appearance. An origin of SEGA from
subependymal tubers has been postulated and
supported by the observations that there are
transitional lesions between tubers and SEGA
and that serial imaging studies have demonstrated
transformation of subependymal nodules into
symptomatic tumors [148150]. Current molecular studies suggest that SEGAs are the neoplastic
counterpart in the spectrum of central hamartomatous tuberous sclerosis complex lesions [151].
Microscopically, SEGAs consist of bizarre

spindle cells, large swollen astrocytes packed with
glial laments (giant cells), prominent thick processes, and occasional ganglion cells. Angiocentric
arrangement of glial cells around vascular structures to form pseudorosettes is common. The
vasculature is typically devoid of endothelial proliferation. There can be an association with microscopic hemorrhage. Rare focal areas of atypia,
mitosis, or necrosis do not indicate aggressive
behavior [141,152]. Calcication is infrequently
noted. Mast cells may be present, and these are
detected only in SEGAs, hemangioblastomas, and
meningiomas in the CNS. Histologically, the
dierential diagnosis for SEGA includes gemistocytic astrocytoma and giant cell glioblastoma.
GFAP staining is variable. One report found
that approximately half of SEGAs stain for
GFAP but that 6 of 7 tumors were S-100 positive
and postulated that SEGAs may arise from cells
in the germinal matrix that have not yet fully
dierentiated along astrocytic or neuronal pathways [153]. Interestingly, another author reports
that tumors associated with tuberous sclerosis
were less likely to stain positively for GFAP [154],
but this is not a widely held opinion [151]. In an
analysis of 20 tuberous sclerosisassociated tumors, investigators found immunoreactivity for
both glial- and neuron-associated epitopes and
neuropeptides within tumor cells with the same
morphology, suggesting that SEGAs have the
ability to undergo divergent glioneuronal and
neuroendocrine dierentiation, perhaps to a greater extent than other mixed glial-neuronal neoplasms [151]. The Ki-67 labeling index (MIB-1) is
usually low [155].
Treatment and prognosis
Surgery
The surgical management of SEGAs is based
on the patients symptoms and the serial changes
on neuroimaging. An asymptomatic patient with
minimal changes in tumor size may be followed
expectantly with serial imaging. Patients who have
symptomatic obstructive hydrocephalus require
surgical intervention. In this instance, the treatment of choice is surgical resection, with gross
total resection as the surgical goal. In the past,
unilateral or bilateral shunting without tumor
resection has been advocated. With advances in
neuroanesthesia and microsurgery, however, these
lesions may be approached safely through transcallosal or transcortical transventricular approaches with acceptable morbidity as described
previously. Gross total tumor resection and
ventriculoventriculostomy (fenestration of the
septum pellucidum) may obviate the need for
shunting. It is recommended that neonates undergoing surgical treatment have preoperative
cardiac clearance because there is a reported
incidence of cardiac rhabdomyomas of tuberous
sclerosis complex, resulting in potential fatal
arrhythmias [156].
The frequency of tumor recurrence is low, with
a 10-year survival rate of nearly 80% after
surgical treatment [141]. Long-term survival is
possible, even after subtotal resection, because of
the limited growth potential of remaining tumor.
Because of the fact that rapid tumor regrowth is
reported [157], however, yearly follow-up MRI
should be obtained to monitor for tumor regrowth and hydrocephalus.
Radiation and chemotherapy
There is no signicant experience of the
treatment of SEGA with radiation therapy or
chemotherapy, although radiation can be considered in the rare setting of malignant degeneration.
The treatment of choice in the primary and
recurrent settings remains surgery.
Miscellaneous intraventricular glial tumors

Although the most common and important
forms of intraventricular gliomas have been
585
discussed, case reports of other types have

appeared in the literature. Nearly 30 cases of
chordoid gliomas of the third ventricle have
recently been reported [13]. Chordoid gliomas
refer to a slow-growing and rare neoplasm of the
third ventricle with an uncertain histogenesis and
chordoid appearance, occurring predominantly in
middle-aged women [13]. The clinical signs and
symptoms seem to be nonspecic and relate to
mass eect in the tumors vicinity (including visual
loss, hydrocephalus, endocrine disturbances,
memory changes, and psychiatric disorders). On
imaging studies, the lesions are typically well
circumscribed with an ovoid shape. They appear
hyperdense on CT imaging and isointense on T1weighted MRI with intense enhancement [13].
The pathology is quite consistent [158] and
involves clusters of oval-to-polygonal epithelioid
tumor cells with plentiful eosinophilic cytoplasm
with a mucinous, vacuolated, and periodic acid
Schi-positive matrix similar to chordomas. A
paucity of mitotic gures and anaplastic features
is noted. Additionally, a lymphoplasmacytic inltrate with Russell bodies without formation of
follicles and germinal centers is usually seen. The
tumors generally possess low growth potential,
and there are no physaliphorous cells, whorl
formations, psammoma bodies, or nuclear pseudoinclusions and no ependymal canals or rosettes.
Given the rarity of this tumor, optimal treatment
after histologic diagnosis is unclear. Surgery
followed by radiation therapy has been reported
[3]. Future reports of this unusual tumor should
help to dene its nature and appropriate treatment further. Other common tumors may also
present in an uncommon intraventricular location, including glioblastoma multiforme, oligodendrogliomas, and gangliogliomas [46,159].
Summary
Gliomas are the most common primary brain
tumor in adults, and those within or relating to
the ventricular surface represent a less common
but important subcategory. The most common
intraventricular gliomas include ependymomas,
SEs, and SEGAs. Other less common varieties
have been reported, including chordoid gliomas,
glioblastoma multiforme, and mixed glial-neuronal tumors. Each type of intraventricular glioma
is associated with its own unique constellation of
epidemiologic, clinical, radiologic, and pathologic
dening characteristics. Each tumor type has its
586
own management considerations and nuances

with unique prognostic indicators and outcomes.
The outcome for certain intraventricular gliomas (especially ependymomas) remains relatively
poor. Future advancements in surgical technique
are likely to have only a modest impact on
improvement of outcome. Translational research
aiming to advance the knowledge of tumor biology
into new targeted cellular and molecular therapies
holds tremendous promise to improve the overall
outcome. Additionally, more thorough delineation
of prognostic factors as well as modications and
renements to radiation and chemotherapy may
help to improve the still signicantly poor outcomes for patients harboring these lesions. Future
cooperative intra- and interinstitutional eorts
between scientists and clinicians will hopefully
culminate in an improved outlook and eventual
cure for patients with gliomas.
[9]
[10]
[11]
[12]
[13]
[14]
References
[1] Sato K, Kubota T, Ishida M, Yoshida K, Takeuchi H, Handa Y. Immunohistochemical and
ultrastructural study of chordoid glioma of the
third ventricle: its tanycytic dierentiation. Acta
[2] Grand S, Pasquier B, Gay E, Kremer S, Remy C,
Le Bas JF. Chordoid glioma of the third ventricle:
CT and MRI, including perfusion data. Neuroradiology 2002;44(10):8426.
[3] Pasquier B, Peoch M, Morrison AL, Gay E,
Pasquier D, Grand S, et al. Chordoid glioma of the
third ventricle: a report of two new cases, with
further evidence supporting an ependymal dierentiation, and review of the literature. Am J Surg
Pathol 2002;26(10):133042.
[4] Lee TT, Manzano GR. Third ventricular glioblastoma multiforme: case report. Neurosurg Rev
1997;20(4):2914.
[5] Guibaud L, Champion F, Buenerd A, Pelizzari M,
Bourgeois J, Pracros JP. Fetal intraventricular
glioblastoma: ultrasonographic, magnetic resonance imaging, and pathologic ndings. J Ultrasound Med 1997;16(4):2858.
[6] Jaeger M, Hussein S, Schuhmann MU, Brandis A,
Samii M, Blomer U. Intraventricular trigonal
ganglioglioma arising from the choroids plexus.
Acta Neurochir (Wien) 2001;143(9):9535.
[7] Dyer S, Prebble E, Davison V, Davies P, Ramani P,
Ellison D, et al. Genomic imbalances in pediatric
intracranial ependymomas dene clinically relevant
groups. Am J Pathol 2002;161:213341.
[8] Granzow M, Popp S, Weber S, Schoell B,
Holtgreve-Grez H, Senf L, et al. Isochromosome
1q as an early genetic event in a child with
intracranial ependymoma characterized by molec-
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
ular cytogenetics. Cancer Genet Cytogenet 2001;

130(1):7983.
Hirose Y, Aldape K, Bollen A, James CD, Brat D,
Lamborn K, et al. Chromosomal abnormalities
subdivide ependymal tumors into clinically relevant groups. Am J Pathol 2001;158:113743.
Singh PK, Gutmann DH, Fuller CE, Newsham IF,
Perry A. Dierential involvement of protein 4.1
family members DAL-1 and NF2 in intracranial
and intraspinal ependymomas. Mod Pathol 2002;
15(5):52631.
Virchow RLK. Cellular pathology as based upon
physiology and pathological histology. Philadelphia: JB Lippincott. 1971. [Chance F, Trans.;
original work published 1863.]
Fokes E, Earle K. Ependymomas: clinical and
pathological aspects. J Neurosurg 1969;30:58594.
Gurney JG, Severson RK, Davis S, Robison LL.
Incidence of cancer in children in the United
States. sex-, race-, and 1-year age-specic rates by
histologic type. Cancer 1995;75:218695.
Helseth A, Mork SJ. Neoplasms of the central
nervous system in Norway: III. Epidemiological
characteristics of intracranial gliomas according to
histology. APMIS 1989;97:54755.
Kuratsu J, Ushio Y. Epidemiological study of
primary intracranial tumors in childhood. Pediatr
Neurosurg 1996;25:2407.
Miller RW, Young JL, Novakovic B. Childhood
cancer. Cancer 1995;75:395405.
Svien H, Mabon R, Kernohan J, et al. Ependymoma of the brain: pathological aspects. Neurology 1953;3:115.
Birgisson S, Blondal H, Bjornsson J, et al.
Tumours in Iceland: 15. Ependymoma: a clinicopathological and immunohistological study.
APMIS 1992;100:294300.
Goldwein JW, Leahy JM, Packer RJ, et al.
Intracranial ependymomas in children. Int J
Radiat Oncol Biol Phys 1990;19:1497502.
Robertson PL, Zeltzer PM, Boyett JM, et al.
Survival and prognostic factors following radiation
therapy and chemotherapy for ependymomas in
children: a report of the Childrens Cancer Group.
J Neurosurg 1998;88:695703.
Gilles FH, Sobel EL, Tavare CJ, et al. Age-related
changes in diagnoses, histological features, and
survival in children with brain tumors: 19301979.
The Childhood Brain Tumor Consortium. Neurosurgery 1995;37:105668.
Nazar GB, Homan HJ, Becker LE, et al. Infratentorial ependymomas in childhood: prognostic
factors and treatment. J Neurosurg 1990;72:40817.
Polednak AP, Flannery JT. Brain, other central nervous system, and eye cancer. Cancer 1995;75:3307.
Rousseau P, Habrand JL, Sarrazin D, et al.
Treatment of intracranial ependymomas in children: review of a 15-year experience. Int J Radiat
Oncol Biol Phys 1994;28:3816.

[25] Vernet O, Farmer JP, Meagher-Villemure K, et al.
Supratentorial ectopic ependymoma. Can J Neurol
Sci 1995;22:3169.
[26] Ernestus RI, Schroder R, Klug N. Spontaneous
intracerebral hemorrhage from an unsuspected
ependymoma in early infancy. Childs Nerv Syst
1992;8:35760.
[27] Naidich T, Lin J, Leeds N, et al. Primary tumors
and other masses of the cerebellum and fourth
ventricle: dierential by computed tomography.
[28] Centeno RS, Lee AA, Winter J, et al. Supratentorial ependymomas: neuroimaging and clinicopathological correlation. J Neurosurg 1986;64:
20915.
[29] Swartz J, Zimmerman R, Bilaniuk L. Computed
tomography of intracranial ependymomas. Radiology 1982;143:97101.
[30] Armington WG, Osborn AG, Cubberley DA, et al.
Supratentorial ependymoma: CT appearance. Radiology 1985;157:36772.
[31] Spoto GP, Press GA, Hesselink JR, et al. Intracranial ependymoma and subependymoma: MR
manifestations. AJNR Am J Neuroradiol 1990;11:
8391.
[32] Choi JY, Chang KH, Yu IK, et al. Intracranial
and spinal ependymomas: review of MR images in
61 patients. Korean J Radiol 2002;3(4):21928.
[33] Kleihues P, Cavenee WK, editors. World Health
Organization classication of tumours. Pathology
and genetics of tumours of the nervous system.
Lyon: IARC Press; 2000.
[34] Ikezaki K, Matsushima T, Inoue T, et al.
Correlation of microanatomical localization with
postoperative survival in posterior fossa ependymomas. Neurosurgery 1993;32:3844.
[35] Lyons MK, Kelly PJ. Posterior fossa ependymomas; report of 30 cases and review of the literature.
[36] Paulino AC, Wen BC, Buatti JM, et al. Intracranial ependymomas. An analysis of prognostic factors and patterns of failure. Am J Clin Oncol
2002;25(2):11722.
[37] van Veelen-Vincent MC, Pierre-Kahn A, Kalifa C,
et al. Ependymoma in childhood: prognostic
factors, extent of surgery, and adjuvant therapy.
J Neurosurg 2002;97:82735.
[38] Ernestus RI, Wilcke O, Schroder R. Supratentorial
ependymomas in childhood: clinicopathological
ndings and prognosis. Acta Neurochir (Wien)
1991;111:96102.
[39] Sutton LN, Goldwein J, Perilongo G, et al.
Prognostic factors in childhood ependymomas.
Pediatr Neurosurg 1990;16:5765.
[40] Pierre-Kahn A, Hirsch J, Roux F, et al. Intracranial
ependymomas in childhood: survival and functional
results of 47 cases. Childs Brain 1983;10:14556.
[41] Undjian S, Marinov M. Intracranial ependymomas in children. Childs Nerv Syst 1990;6:1314.
587
[42] Kovalic JJ, Flaris N, Grigsby PW, et al. Intracranial ependymoma long term-outcome, patterns of failure. J Neurooncol 1993;15:12531.
[43] Palma L, Celli P, Cantore G. Supratentorial
ependymomas of the rst two decades of life:
long-term follow-up of 20 cases (including two
subependymomas). Neurosurgery 1993;32:16975.
[44] Fokes E, Earle K. Ependymomas: clinical and
pathological aspects. J Neurosurg 1969;30:58594.
[45] Mork S, Loken A. Ependymoma: a follow-up
study of 100 cases. Cancer 1977;40:90715.
[46] Perilongo G, Massimino M, Sotti G, et al.
Analyses of prognostic factors in a retrospective
review of 92 children with ependymoma: Italian
Pediatric Neurooncology Group. Med Pediatr
Oncol 1997;29:7985.
[47] Phillips TL, Sheline GE, Boldrey E. Therapeutic
consideration in tumors aecting the central
nervous system: ependymomas. Radiology 1964;
83:98105.
[48] Kim Y, Fayos JV. Intracranial ependymomas.
Therapeut Radiol 1977;124:8058.
[49] Garrett PG, Simpson WJK. Ependymomas: results
of radiation therapy. Int J Radiat Oncol Biol Phys
1983;9:11214.
[50] Kun LE, Kovnar EH, Sanford RA. Ependymomas
in children. Pediatr Neurosci 1988;14:5763.
[51] Leibel SA, Sheline GE. Radiation therapy for
neoplasms of the brain. J Neurosurg 1987;66:
122.
[52] Wallner KE, Wara WM, Sheline GE, et al.
Intracranial ependymomas: results of treatment
with partial or whole brain irradiation without
spinal irradiation. Int J Radiat Oncol Biol Phys
1986;12:93741.
[53] Staord SL, Pollock BE, Foote RL, et al.
Stereotactic radiosurgery for recurrent ependymoma. Cancer 2000;88:8705.
[54] Oya S, Shibamoto Y, Nagata Y, et al. Postoperative radiotherapy for intracranial ependymoma: analysis of prognostic factors and patterns
of failure. J Neurooncol 2002;56(1):8794.
[55] Salazar OM, Castro-Vita H, Van Houtte P, et al.
Improved survival in cases of intracranial ependymoma after radiation therapy. Late report and
recommendations. J Neurosurg 1983;59:6529.
[56] Hoppe-Hirsch E, Brunet L, Laroussinie F, et al.
Intellectual outcome in children with malignant
tumors of the posterior fossa: inuence of the eld
of irradiation and quality of surgery. Childs Nerv
Syst 1995;11:3405.
[57] Scheurlen W, Kuhl J. Current diagnostic and
therapeutic management of CNS metastasis in
childhood primitive neuroectodermal tumors and
ependymomas. J Neurooncol 1998;28:1815.
[58] Merchant TE, Haida T, Wang MH, et al.
Anaplastic ependymoma: treatment of pediatric
patients with or without craniospinal radiation
therapy. J Neurosurg 1997;86:9439.
588
[59] Goldwein JW, Corn BW, Finlay JL, et al. Is

craniospinal irradiation required to cure children
with malignant (anaplastic) ependymoma? Cancer
1991;67:276671.
[60] Heideman RL, Packer RJ, Reaman GH, et al. A
phase II evaluation of thiotepa in pediatric central
nervous system malignancies. Cancer 1993;72:
271275.
[61] Salazar OM. A better understanding of CNS
seeding and a brighter outlook for postoperatively
irradiated patients with ependymomas. Int J
Radiat Oncol Biol Phys 1983;9:12314.
[62] Palma L, Celli P, Mariottini A, et al. The
importance of surgery in supratentorial ependymomas. Long-term survival in a series of 23 cases.
Childs Nerv Syst 2000;16:1705.
[63] Douek E, Kingston JE, Malpas JS, et al. Platinumbased chemotherapy for recurrent CNS tumors in
young patients. J Neurol Neurosurg Psychiatry
1991;54:7225.
[64] Lesser GJ, Grossman SA. The chemotherapy of
adult primary brain tumors. Cancer Treat Rev
1993;19:26181.
[65] Strauss LC, Killmond TM, Carson BS, et al.
Ecacy of postoperative chemotherapy using
cisplatin plus etoposide in young children with
brain tumors. Med Pediatr Oncol 1991;19:1621.
[66] Tamura M, Ono N, Kurihara H, et al. Adjunctive
treatment for recurrent childhood ependymoma of
the IV ventricle: chemotherapy with CDDP and
MCNU. Childs Nerv Syst 1990;6:1869.
[67] van Eys J, Cangir A, Coody D, et al. MOPP
regimen as primary chemotherapy for brain
tumors in infants. J Neurooncol 1985;3:23743.
[68] Lefkowitz I, Evans A, Sposto R, et al. Adjuvant
chemotherapy of childhood posterior fossa ependymoma: craniospinal irradiation with or without
CCNU, vincristine, and prednisone. Proc Am Soc
Clin Oncol 1989;8:87.
[69] Ater JL, van Eys J, Woo SY, et al. MOPP
chemotherapy without irradiation as primary postsurgical therapy for brain tumors in infants and
young children. J Neurooncol 1997;32:24352.
[70] White L, Johnston H, Jones R, et al. Postoperative
chemotherapy without radiation in young children
with malignant non-astrocytic brain tumours. A
report from the Australia and New Zealand
Childhood Cancer Group (ANZCCSG). Cancer
Chemother Pharmacol 1993;32:4036.
[71] Geyer JR, Zeltzer PM, Boyett JM, et al. Survival
of infants with primitive neuroectodermal tumors
or malignant ependymomas of the CNS treated
with eight drugs in 1 day: a report from the
Childrens Cancer Group. J Clin Oncol 1994;
12:160715.
[72] Ayan I, Darendeliler E, Kebudi R, et al. Evaluation of response to postradiation eight in one
chemotherapy in childhood brain tumors. J Neurooncol 1995;26:6572.
[73] Evans AE, Anderson JR, Lefkowitz-Boudreaux

IB, et al. Adjuvant chemotherapy of childhood
posterior fossa ependymoma: cranio-spinal irradiation with or without adjuvant CCNU, vincristine,
and prednisone: a Childrens Cancer Group study.
Med Pediatr Oncol 1996;27:814.
[74] Needle MN, Goldwein JW, Grass J, et al.
Adjuvant chemotherapy for the treatment of
intracranial ependymoma of childhood. Cancer
1997;80:3417.
[75] Duner PK, Krischer JP, Sanford RA, et al.
Prognostic factors in infants and very young
children with intracranial ependymomas. Pediatr
Neurosurg 1998;28:21522.
[76] Bouet E, Foreman N. Chemotherapy for intracranial ependymomas. Childs Nerv Syst 1999;
15:56370.
[77] Bertolone SJ, Baum ES, Krivit W, et al. A phase II
study of cisplatin therapy in recurrent childhood
brain tumors: a report from the Childrens Cancer
Study Group. J Neurooncol 1989;7:511.
[78] Corden BJ, Strauss LC, Killmond T, et al.
Cisplatin, ara-C and etoposide (PAE) in the
treatment of recurrent childhood brain tumors.
[79] Sexauer CL, Khan A, Burger PC, et al. Cisplatin in
recurrent pediatric brain tumors: A POG Phase II
studya Pediatric Oncology Group Study. Cancer
1985;56:1497501.
[80] Levin VA. Chemotherapy of primary brain tumors. Neurol Clin 1985;3:85566.
[81] Shapiro WR. Chemotherapy of primary malignant
brain tumors in children. 1975. Cancer 1975;
35(Suppl):96572.
[82] Gaynon PS, Ettinger LJ, Baum ES, et al. Carboplatin in childhood brain tumors. A Childrens
Cancer Study Group Phase II trial. Cancer 1990;
66:24659.
[83] Friedman HS, Krisher JP, Burger P, et al.
Treatment of children with progressive or recurrent brain tumors with carboplatin or iproplatin: a Pediatric Oncology Group randomized phase
II study. J Clin Oncol 1992;10:24956.
[84] Allen JC, Hancock C, Walker R, et al. PCNU and
recurrent childhood brain tumors. J Neurooncol
1987;5:2414.
[85] Ragab AH, Burger P, Badnitsky S, et al. BCNU in
the treatment of recurrent medulloblastoma and
ependymomaa POG study. J Neurooncol 1986;
2:3412.
[86] Chastagner P, Sommelet OD, Kalifa C, et al. Phase
II study of ifosfamide in childhood brain tumors:
a report by the French Society of Pediatric Oncology
(SFOP). Med Pediatr Oncol 1993;21:4953.
[87] Heideman RL, Douglass EC, Langston JA, et al.
Phase II study of every other day high-dose
ifosfamide in pediatric brain tumors: a Pediatric
Oncology Group study. J Neurooncol 1995;25:
7784.

[88] Heidman RL, Packer RJ, Allen JC. A phase II
study of thiopental in pediatric central nervous
system tumors. Pediatr Neurosci 1989;15:1467.
[89] Razzouk BI, Heidman RL, Friedman HS, et al. A
phase II evaluation of thiotepa followed by other
multiagent chemotherapy regimens in infants and
young children with malignant brain tumors.
Cancer 1995;75:27627.
[90] Schold SC Jr, Friedman HS, Bjornsson TD, et al.
Treatment of patients with recurrent primary brain
tumors with AZQ. Neurology 1984;34:6159.
[91] Ettinger LJ, Ru N, Krailo M, et al. A phase II
study of diaziquone in children with recurrent or
progressive primary brain tumors: a report from
the Childrens Cancer Study Group. J Neurooncol
1990;9:6976.
[92] Arndt C, Krailo MD, Steinherz L, et al. A phase II
clinical trial of idarubicin administered to children
with relapsed brain tumors. Cancer 1998;83:8136.
[93] Pendergrass TW, Milstein JM, Geyer JR, et al.
Eight drugs in one day chemotherapy for brain
tumors: experience in 107 children and rationale
for preradiation chemotherapy. J Clin Oncol 1993;
5:12213.
[94] Ettinger LJ, Sinniah D, Siegel SE. Combination
chemotherapy with cyclophosphamide, vincristine,
procarbazine, and prednisone (COPP) in children
with brain tumors. J Neurooncol 1985;3:2639.
[95] Mason WP, Goldman S, Yates AJ, et al. Survival
following intensive chemotherapy with bone
marrow reconstitution for children with recurrent intracranial ependymomaa report of the
Childrens Cancer Group. J Neurooncol 1998;
37:13543.
[96] Grill J, Kalifa C, Doz F, et al. A high-dose
busulfan-thiotepa combination followed by autologous bone marrow transplantation in childhood
recurrent ependymoma. A phase-II study. Pediatr
Neurosurg 1996;25:712.
[97] Duner PK, Cohen ME, Myers MH, et al.
Survival of children with brain tumors: SEER
Program, 19731980. Neurology 1986;36:597601.
[98] Papadopoulos DP, Giri S, Evans RG. Prognostic
factors and management of intracranial ependymomas. Anticancer Res 1990;10:68992.
[99] Vanuytsel LJ, Bessell EM, Ashley SE, et al.
Intracranial ependymoma: long-term results of
a policy of surgery and radiotherapy. Int J Radiat
[100] Pollack IF, Gerszten PC, Martinez AJ, et al.
Intracranial ependymomas of childhood: longterm outcome and prognostic factors. Neurosurgery 1995;37:65566.
[101] Healey EA, Barnes PD, Kupsky WJ, et al. The
prognostic signicance of postoperative residual
tumor in ependymoma. Neurosurgery 1991;28:
66671.
[102] Guyotat J, Signorelli F, Desme S, et al. Intracranial ependymomas in adult patients: analyses
[103]
[104]
[105]
[106]
[107]
[108]
[109]
[110]
[111]
[112]
[113]
[114]
[115]
[116]
[117]
[118]
589
of prognostic factors. J Neurooncol 2002;60(3):

25568.
Rorke LB. Relationship of morphology of ependymoma in children to prognosis. Prog Exp
Tumor Res 1987;30:1704.
Shaw EG, Evans RG, Scheithauer BW, et al.
Postoperative radiotherapy of intracranial ependymoma in pediatric and adult patients. Int J Radiat
Figarell-Branger D, Gambarelli D, et al. Infratentorial ependymomas of childhood. Correlation
between histological features, immunohistological
phenotype, silver nucleolar organizer region staining values and postoperative survival in 16 cases.
Schier D, Chio A, Cravioto H, et al. Ependymoma: internal correlations among pathological
signs: the anaplastic variant. Neurosurgery 1991;
29:20610.
Chiu JK, Woo SY, Ater J, et al. Intracranial
ependymoma in children: analysis of prognostic
factors. J Neurooncol 1992;13:28390.
Ross GW, Rubinstein LJ. Lack of histopathological correlation of malignant ependymomas with
postoperative survival. J Neurosurg 1989;70:316.
Marks JE, Adler SJ. A comparative study of
ependymomas by site of origin. Int J Radiat Oncol
Biol Phys 1982;8:3743.
Jayawickreme DP, Hayward RD, Harkness WF.
Intracranial ependymomas in childhood: a report
of 24 cases followed for 5 years. Childs Nerv Syst
1995;11:40913.
Pierre-Kahn A, Hirsch JF, Roux FX, et al.
Intracranial ependymomas in childhood. Survival
and functional results of 47 cases. Childs Brain
1983;10:14556.
Scheithauer BW. Symptomatic subependymoma:
report of 21 cases with review of the literature.
J Neurosurg 1978;49:68996.
Scheinker IM. Subependymoma: a newly recognized tumor of subependymal derivation. J Neurosurg 1945;2:23240.
Clarenbach P, Kleihues P, Metzel E, Dichgans J.
Simultaneous manifestation clinical manifestation
of subependymoma of the fourth ventricle in identical twins: case report. J Neurosurg 1979;50:6559.
Honan WP, Anderson M, Carey MP, et al.
Familial subependymomas. Br J Neurosurg 1987;
1:31721.
Ryken T, Robinson R, VanGilder J. Familial
occurrence of subependymoma. J Neurosurg
1994;80:110811.
Ho KL. Concurrence of subependymoma and
heterotopic leptomeningeal neuroglial tissue. Arch
Pathol Lab Med 1983;107:13640.
Russell D, Rubinstein L. Tumors of central neuroepithelial origin. In: Russell DS, Rubinstein LJ,
editors. Pathology of the central nervous system.
Baltimore: Williams & Wilkins; 1989. p. 192206.
590
[119] Chanson JL. Subependymal mixed gliomas.

J Neuropathol Exp Neurol 1956;15:46170.
[120] Hashimoto M, Tanaka H, Oguro K, Masuzawa T.
Subependymoma of the lateral ventricle: case
report. Neurol Med Chir (Tokyo) 1991;31:7325.
[121] Matsumura A, Ahyai A, Hori A, Schaake T.
Intracerebral subependymomas: clinical and neuropathological analyses with special reference to
the possible existence of a less benign variant. Acta
Neurochir (Wien) 1989;96:1525.
[122] Scheithauer BW, Bruner JM. Central nervous
system tumors. Clin Lab Med 1987;7:15779.
[123] Jelenik J, Smirniotopoulos JG, Parisi JE, Kanzer
M. Lateral ventricular neoplasms of the brain:
dierential diagnosis based on clinical, CT, and
MR ndings. AJNR Am J Neuroradiol 1990;
11:56774.
[124] Nishio S, Morioka T, Mihara F, Fukui M.
Subependymoma of the lateral ventricles. Neurosurg Rev 2000;23:98103.
[125] Ortiz-Reyes R, Dragovic L, Eriksson A. Sudden
unexpected death resulting from previously nonsymptomatic subependymoma. Am J Forensic
Med Pathol 2002;23:637.
[126] Changaris DG, Powers JM, Perot PL Jr, Hungerford GD, Neal GB. Subependymoma presenting as
subarachnoid hemorrhage: case report. J Neurosurg 1981;55:6435.
[127] French JD, Bucy PC. Tumors of the septum
pellucidum. J Neurosurg 1948;5:43349.
[128] Azarelli B, Rekate HL, Roessmann U. Subependymoma: a case report with ultrastructural study.
[129] Rengachary SS. Subependymomas. In: Wilkins
RH, Rengachary SS, editors. Neurosurgery. New
York: McGraw-Hill; 1996. p. 12013.
[130] Chiechi MV, Smirniotopoulos JG, Jones RV.
Intracranial subependymomas: CT and MR imaging features in 24 cases. AJR Am J Roentgenol
1995;165:124550.
[131] Hoeel C, Boukobza M, Polivka M, Lot G,
Guichard JP, Latte F, et al. MR manifestations
of subependymomas. AJNR Am J Neuroradiol
1995;16:21219.
[132] Godwin JT. Subependymal glomerate astrocytoma: report of two cases. J Neurosurg 1959;
16:3859.
[133] Prayson RA, Suh JH. Subependymomas: clinicopathologic study of 14 tumors, including comparative MIB-1 immunohistochemical analysis with
other ependymal neoplasms. Arch Pathol Lab Med
1999;123:3069.
[134] Louis DN, Hedley-Whyte ET, Martuza RL. Case
report: sarcomatous proliferation of the vasculature in subependymoma. Acta Neuropathol (Berl)
1989;78:3325.
[135] Rosenblum MK, Erlanson RA, Aleksic SN, Budsilovich GN. Melanotic ependymoma and subependymoma. Am J Surg Pathol 1990;14:72936.
[136] Tomlinson FH, Scheithauer BW, Kelly PJ, Gorbes

GS. Subependymoma with rhabdomyosarcomatous dierentiation: report of a case and literature
review. Neurosurgery 1991;28:7618.
[137] Artico M, Bardella L, Ciapetta P, Raco A. Surgical
treatment of subependymomas of the central
nervous system. Report of 8 cases and review of
the literature. Acta Neurochir (Wien) 1989;98:
2531.
[138] Marsh WR, Laws E Jr. Intracranial ependymomas. Prog Exp Tumor Res 1987;30:17580.
[139] Frerebeau P, Benezech J, Segnarbieux F, Harbi H,
Desy A, Marty-Double C. Intraventricular tumors
in tuberous sclerosis. Childs Nerv Syst 1985;
1:458.
[140] Sima AAF, Robertson DM. Subependymal giantcell astrocytoma: case report with ultrastructural
study. J Neurosurg 1979;50:2405.
[141] Shepherd CW, Scheithauer BW, Gomez MR,
Altermatt HJ, Katzmann JA. Subependymal
giant cell astrocytoma: a clinical, pathological,
and ow cytometric study. Neurosurgery 1991;
28:8648.
[142] Padmalatha C, Harru RC, Ganick D, Hafez GR.
Glioblastoma multiforme with tuberous sclerosis.
Arch Pathol Lab Med 1980;104:64950.
[143] Chou TM, Chou SM. Tuberous sclerosis in the
premature infant: a report of a case with immunohistochemistry on the CNS. Clin Neuropathol
1989;8:4552.
[144] Holanda FJ, Holanda GM. Tuberous sclerosis:
neurosurgical implications in intraventricular tumors. Neurosurg Rev 1980;3:13950.
[145] Winter J. Computed tomography in diagnosis of
intracranial tumors versus tubers in tuberous
sclerosis. Acta Radiol 1982;23:33744.
[146] Lee BCP, Gawler J. Tuberous sclerosis: comparison of computed tomography and conventional
neuroradiology. Radiology 1978;127:4037.
[147] Martin N, Debussche C, DeBroucker T,
Mompoint D, Marsault C, Nahum H. Gadolinium-DTPA enhanced MR imaging in tuberous
sclerosis. Neuroradiology 1990;31:4927.
[148] Russell DS, Rubinstein LJ. Pathology of tumors of
the central nervous system. 5th edition. Baltimore:
[149] Morimoto K, Mogami H. Sequential CT study of
subependymal giant-cell astrocytoma associated
with tuberous sclerosis. Case report. J Neurosurg
1986;65:8747.
[150] Nishio S, Morioka T, Suzuki S, Kira R, Mijhara F,
Fukui M. Subependymal giant cell astrocytoma:
clinical and neuroimaging features of four cases.
J Clin Neurosci 2001;8(1):8314.
[151] Lopes MBS, Altermatt HJ, Scheithauer BW,
Shepherd CW, VandenBerg SR. Immunohistochemical characterization of subependymal giant
cell astrocytomas. Acta Neuropathol (Berl) 1996;
91:36875.

[152] Scheithauer BW. The neuropathology of tuberous
sclerosis. J Dermatol 1992;19:897903.
[153] Nakamura Y, Becker LE. Subependymal giantcell tumor. Acta Neuropathol (Berl) 1983;60:
2717.
[154] Bonnin JM, Rubinstein LJ, Papasozomenos SC,
Marangos PJ. Subependymal giant cell astrocytoma. Signicance and possible cytogenetic implications of an immunohistochemical study. Acta
Neuropathol (Berl) 1984;62:18593.
[155] Gyure KA, Prayson RA. Subependymal giant cell
astrocytoma: a clinicopathologic study with
HMB45 and MIB-1 immunohistochemical analysis. Mod Pathol 1997;10:3137.
591
[156] Painter MJ, Pang D, Ahdab-Barmada M, Bergman

I. Connatal brain tumors in patients with tuberous
sclerosis. Neurosurgery 1984;14:5703.
[157] Yamamoto K, Yamada K, Nakahara T, Ishihara
A, Takaki S, Kochi M, et al. Rapid regrowth of
solitary subependymal giant cell astrocytoma: case
report. Neurol Med Chir (Tokyo) 2002;42:2247.
[158] Brat DJ, Scheithauer BR, Staugaitis SM, et al.
Third ventricular chordoid glioma: a distinct
clinicopathologic entity. J Neuropathol Exp
Neurol 1998;57:28390.
[159] Garza-Mercado R, Campa H, Grajeda J. Primary
oligodendroglioma of the septum pellucidum.
Surgical resection of metastatic

intraventricular tumors
Giacomo G. Vecil, MD, Frederick F. Lang, MD*
Department of Neurosurgery, The University of Texas, M.D. Anderson Cancer Center, 1515 Holcombe Boulevard,
Unit 442, Houston, TX 77030-4009, USA
Metastasis from a systemic (noncerebral)

primary cancer must be considered in the dierential diagnosis of an intraventricular mass.
Although less common than metastases that occur
within the brain parenchyma, metastatic brain
tumors located within the cerebral ventricles are
a unique challenge for neurosurgical oncologists.
Because of the deep location of these tumors and
their juxtaposition to critical brain structures, the
management of patients with intraventricular metastases is generally more complex than that of most
patients with intraparenchymal brain metastases.
Nevertheless, continued experience with intraventricular surgery has resulted in a progressive improvement in the surgical treatment of intraventricular
metastases. In addition, the development of
stereotactic radiosurgery oers a less invasive
nonsurgical alternative for small intraventricular
metastatic tumors.
Despite these complexities, there is a paucity of
information regarding the clinical features and
surgical treatment of intraventricular metastases.
Indeed, the literature on this topic is essentially
limited to case reports [121] or to a few examples
within larger series of intraventricular tumors in
general [2224]. Consequently, the purpose of this
article is to review the unique features and surgical
management strategies pertinent to intraventricular metastatic brain tumors based primarily on an
This work was supported by a grant from the

Anthony Bullock Foundation.
E-mail address: ang@mdanderson.org
(F.F. Lang).
analysis of our institutional experience with these

lesions.
Clinical Features
Classication
True intraventricular metastases are distinct
tumor masses that arise within the ventricle and
originate most commonly from the choroid plexus
or from focal attachment to the ependyma/
subependyma. Parenchymal tumors with secondary extension into the ventricle should be classied as intraparenchymal metastases. More
importantly, nodular deposits, even if large, that
develop in patients with meningeal carcinomatosis
(eg, Khoshyomn et al [25] and Bugiani et al [26])
should not be classied as intraventricular metastases, because the treatment and prognosis of
these lesions dier markedly from the treatment
and prognosis of true intraventricular metastatic
tumors [27].
Incidence
When only true metastatic focal tumors are
considered, brain metastases to the ventricles are
rare, and although the exact incidence is dicult
to determine, best estimates suggest that intraventricular metastases comprise about 6% of all
intraventricular tumors (Kohno et al [13]) and
occur in less than 5% of patients with cancer
[13,16,28,29]. Arendt and colleagues [29] found 7
cases of intraventricular metastasis (4.6%) among
150 autopsied cancer patients. Kohno and colleagues [13] reported that tumors metastatic to the
lateral ventricle comprised 0.9% of brain metastases based on a review of the Brain Tumor
doi:10.1016/S1042-3680(03)00056-1
594
G.G. Vecil, F.F. Lang / Neurosurg Clin N Am 14 (2003) 593606
Registry of Japan. Even less usual is the occurrence of a single intraventricular brain metastasis.
Schreiber and colleagues [28] reported that intraventricular metastases were found in 2.6% of
patients with cancer at autopsy but that single
intraventricular lesions were found in only 1
(0.14%) of the 737 cases analyzed.
Table 1 lists the cases of intraventricular
metastases reported in the literature. A total of 23
cases have been described, with the largest series
comprising 3 cases. Excluded from this list are cases
identied as being part of a larger series of intraventricular tumors [2224]; these cases are omitted
because the reports typically focus on the radiographic features or the surgical aspects of resecting
or biopsying intraventricular tumors in general and
not on the specics of metastatic lesions. For
example, Kelly and coworkers [30] reported on the
use of computer-assisted stereotactic surgery in the
removal of 58 intraventricular tumors, 3 of which
were intraventricular metastases; detailed descriptions of these cases were not given.
A search of the database of the Department of
Neurosurgery at the University of Texas M.D.
Anderson Cancer Center (M.D. Anderson) identied 35 patients (1.8%) with true intraventricular
metastases among a group of 1930 patients with
brain metastases treated during the 10-year period
between June 1993 and February 2003 (Table 2).
Thirty-one of the 35 patients had an intraventricular metastasis at the rst presentation of brain
metastasis, whereas 4 presented rst with an intraparenchymal metastasis and subsequently developed an intraventricular lesion. Of the 31 patients
with an intraventricular lesion at initial presentation, 12 (0.6% of all patients with metastases) had
single intraventricular metastatic tumors and 19
had multiple lesions (2 patients had 2 intraventricular lesions, 11 patients had 1 intraventricular
lesion and 1 intraparenchymal lesion, and 6 patients had 2 or more [range: 211] intraparenchymal lesions in addition to the intraventricular
lesion). This population undoubtedly represents
an underestimation of the total number of patients with intraventricular lesions seen at our
institution, because some patients with intraventricular metastases may have been treated by
medical or radiation oncologists without neurosurgical consultation.
Presentation and diagnosis
Patients with intraventricular metastases either
are known to have systemic cancer at the time of
presentation of the brain lesion (metachronous

presentation) or the brain lesion is part of the
initial presentation of the cancer (synchronous
presentation of brain metastasis). In both groups,
MRI of the brain is the critical diagnostic test.
In patients with known systemic cancer, the
nding of a mass showing gadolinium contrast
enhancement within the ventricle must be considered a metastasis until proven otherwise. The presence of multiple lesions virtually assures the
diagnosis; thus, high-quality MRI is critical. In
our series of 35 patients, 33 were known to have
cancer before the diagnosis of brain metastasis,
and the median interval from time of diagnosis
of the primary tumor to presentation with the
intraventricular brain metastases was 23 months
(range: 1.484 months). For the 18 cases reported
in the literature in which the presenting interval
was described (see Table 1), 11 arose in patients
with known cancer (metachronous presentation)
and the median interval to diagnosis of the brain
metastasis was 48 months. Both of these sets of
data suggest that the time from diagnosis of the
primary tumor to occurrence of intraventricular
metastasis is longer than that observed for intraparenchymal metastases, probably because of the
unique histologies of intraventricular metastases
[31]. It is noteworthy that 6 of the cases in our
series were identied during routine surveillance
for systemic metastases in asymptomatic patients, demonstrating the value of this screening
approach.
In patients without known systemic cancer
(synchronous presentation), the dierential diagnosis of an intraventricular lesion is obviously
broader than in patients with known cancer.
Distinctions between metastasis and other intraventricular tumors may be dicult based on
imaging. It has been suggested [7] that the presence of extensive peritumoral edema within the
brain parenchyma is more characteristic of brain
metastases than other intraventricular lesions (eg,
meningioma). Although this feature has not been
specically analyzed in a large study of intraventricular lesions to our knowledge, the presence of
marked peritumoral brain edema should suggest
the possibility of an intraventricular metastasis.
Indeed, of the 20 cases reported in the literature
that comment on this feature, brain edema was
documented in 70%. Patients with intraventricular tumors without a documented primary cancer,
particularly if in the fth to seventh decades of
life, should be evaluated for the possibility of
systemic cancer. This evaluation typically includes
a careful history, physical examination, routine

laboratory tests, and plain chest radiographs. A
more extensive search for a primary cancer (eg,
CT of the chest and abdomen) is probably not
warranted until the brain lesion is proven to be
a metastasis [32]. Although only two of our
patients did not have a history of cancer at the
time of diagnosis of the brain metastasis (see
Table 2), 37% of the cases in the literature were
found in patients without a diagnosis of cancer
(see Table 1). This dierence is undoubtedly
a result of the dierent referral patterns of our
tertiary cancer center and the hospitals reported in
the literature.
With regard to specic symptoms, patients
with intraventricular metastases present with the
same symptoms as patients with other intraventricular tumors [31]. In our series of patients, most
patients presented with symptoms of elevated
intracranial pressure, including headache (49%)
and nausea/vomiting (17%). In contrast, focal
symptoms were rare, and seizures occurred in
5.6% of patients. Headache (47%) was also the
most common symptom in the cases reported in
the literature (see Table 1). Although not often
recognized, hemorrhage as the cause of the
symptoms is not uncommon for patients with
intraventricular metastases [9,33]. Nakabayashi
and colleagues [9] were the rst to report
hemorrhage in an intraventricular metastasis.
Since then, several other reports have veried
such intratumoral hemorrhages (see Table 1)
[121]. In our series, four patients experienced
signicant intratumoral hemorrhages. Spetzger
and colleagues [12] reported a noteworthy case
of subarachnoid and intraventricular hemorrhage
in a patient with incidental bilateral posterior
communicating artery aneurysms, in whom the
cause of the bleeding was eventually and correctly
attributed to a fourth ventricular metastatic
carcinoma.
Histology of primary tumor
A striking feature of intraventricular metastases is the high frequency of renal cell carcinoma
(Fig. 1). Among the cases reported in the
literature, nine metastases (39%) were renal cell
carcinomas. In our series, 46% of intraventricular
metastases were renal cell carcinomas, followed by
melanoma (14%), breast cancer (14%), and lung
cancer (9%) in decreasing order of frequency (see
Table 2). This distribution is in contrast to the
distribution of intraparenchymal brain metasta-
595
ses, among which lung cancer typically predominates (30%60%), followed by breast cancer
(10%30%), melanoma (5%21%), and, less
commonly, renal cell carcinoma (1%2%) [31].
The high frequency of intraventricular renal
carcinomas suggests that there may be a tropism
of this tumor type for the ventricle, specically for
the choroid plexus, from which most of them
appear to arise. Matsumora and colleagues [15]
pointed out that renal cell carcinomas are divided
into two types: the rapidly progressive type,
which is often fatal within a few years of
diagnosis, and the slowly progressive type in
which patients survive for long periods with an
indolent tumor. The long time between diagnosis
of the brain metastases and the initial diagnosis of
renal cell carcinoma in the kidney that is common
in the cases reported in the literature suggests that
metastasis to the ventricle (choroid plexus) from
renal cell cancer is more prevalent in the slowly
progressive type. Taken together, these observations suggest that there may be a biologic basis for
the tendency of certain renal cell carcinomas to
metastasize to the ventricle. Notably, the kidney
and the choroid plexus function as plasma
ltering systems, and this suggests, albeit speculatively, that there may be a cellular and
molecular interaction between renal cell carcinoma and the cells comprising the choroid plexus
that is more important than simple mechanical
forces, such as blood ow. To our knowledge,
the nature of this interaction has not been
determined.
Location
In the literature (see Table 1) and our series
(see Table 2), intraventricular metastases were
observed to arise most commonly in the lateral
ventricle (68% of our cases), whereas metastases
to the fourth ventricle (21%) (Fig. 2) and the third
ventricle (11%) (Fig. 3) were less usual. Within the
lateral ventricle, the trigone was the most common site of metastasis (60% of our cases),
followed by the body (28%). In our series, 18 of
the 25 lateral ventricle tumors arose on the left
side, whereas there was less of a propensity for
localization in the dominant hemisphere in the
reports in the literature.
The distribution of intraparenchymal brain
metastases (ie, cerebral hemispheres [80%85%],
cerebellum [10%15%], and brain stem [3%5%])
roughly parallels the relative tissue and blood
volume of these regions of the brain. Using this
Reference (year)
[language if not
English]
596
Table 1
Published cases of intraventricular brain metastasesa
Time from
diagnosis
of primary Presenting
(months) symptoms
Bleeding/ No.
Location in
edema
metastases ventricle
Sex/age
(years)
Site of primary
tumor (pathology)
F/73
Lung (oat cell

carcinoma)
Decreased
consciousness
No/?
Bilateral trigones None (autopsy)
Died 1 day after

CT was done
F/?
Breast (?)
No/+
Right trigone
?/?
Skin (melanoma)
Yes/+
Fourth ventricle
Kidney (renal cell

carcinoma)
156
No/+
Left trigone
Approach not
specied
Parietal
craniectomy
F/55
Decreased
consciousness
Headache
Kart et al (1986) M/61
60
Asymptomatic
Yes/+
Mertens et al
(1987)
[German]b
Shigemori et al
(1987)
[Japanese]b
Tanimoto et al
(1991)
Lung (poorly
dierentiated
epithelial
carcinoma)
?/newborn ? (neuroblastoma)
?/?
M/58
Kidney (renal cell

carcinoma)
32
No/
M/64
Lung (large cell

carcinoma)
Headache,
motor
weakness
Headache
No/+
F/59
Kidney (renal cell

carcinoma)
48
Headache
No/+
Gastric (adenocarcinoma)
Decreased
Yes/
consciousness,
? seizure
Killebrew et al
(1983)
Mizuno et al
(1992)
Nakabayshi et al M/64
(1994)
Outcome
GTR, postoperative
bleed, reoperation;
required shunting;
alive at 4 years with
mild hemiparesis
Left trigone
Stereotactic
Radiation therapy to
biopsy
primary lung lesion;
died 2 months after
primary diagnosis
Right trigone
?
Received multiagent
chemotherapy; no
relapse at 2 years
Right body
Transventricular
Postoperative subdural
(near foramen)
hematoma; died on
postoperative day 3
Right trigone
Transcortical
Rapid regrowth
(middle
of tumor, progressive
temporal
hemiplegia; died
gyrus)
because of acute MI
before reoperation
Left lateral body Posterior
Tumor
interhemispheric
recurrence at 19
transcallosal
months, bedridden
Left center body Anterior
Died 2 months after
interhemispheric
surgery from
(transcallosal)
respiratory
complications
Healy and
Rosenkrantz
(1980)
Kendall et al
(1983)
Surgical
procedure
Suetake et al
(1994)
[Japanese]b
Berkow and
Kelly (1995)
M/78
Kidney (renal cell

carcinoma)
Decreased
consciousness
Yes/+
Right trigone
Stereotactic biopsy Vegetative state
F/13
Mediastinum (mixed 12
months
malignant germ cell
tumormature
teratoma)
Decreased
consciousness
No/?
Autopsy
Spetzger et al
(1995)
F/60
Kidney (renal cell

carcinoma)
48
Headache,
nausea,
decreased
consciousness
Yes/
Fourth ventricle
(oor)
Suboccipital
craniotomy
Kohno et al
(1996)
M/45
Colon
(adenocarcinoma)
36
Hemiparesis,
aphasia,
hemianopsia
No/+
Left trigone
Transcortical
(parietal)
M/66
Kidney (renal cell

carcinoma)
84
No/+
Right trigone
Interhemispheric
(transcallosal)
M/66
Lung (adenocarcinoma)
Hemiparesis,
disorientation
and memory
disturbance
Seizures
No/+
Right inferior
horn
Transcortical
(temporal)
Brandicourt et al ?/64
(1997)
[French]b
Matsumura
M/68
et al (1997)
Colon
(adenocarcinoma)
No/+
Third ventricle
Kidney (renal cell

carcinoma)
84
Headache
No/
Right lateral
body
Raila et al
(1998)
F/47
Kidney (renal cell

carcinoma)
Headache and
somnolence
Yes/+
Right trigone
Transventricular
via small
corticotomy
(frontal)
Approach not
specied
Arbelaez et al
(1999)
F/48
? (melanoma)
Headache
No/
Left trigone
Disease-free at
30 months
Death at 3 weeks
from secondary
complications
Partial resection,
no primary
found; stable
at 3 months
597
Transcortical
(temporal)
Received chemotherapy,
then presented with
brain herniation
from intraventricular
metastasis
Local rebleed on
postoperative day 2,
meningitis 1 week
after surgery,
recurrence at 18
months
Improved hemiparesis
and speech after gross
total resection,
followed by
postoperative
radiation therapy
Gross total resection,
postoperative
radiation therapy
Gross total resection,
postoperative
radiation therapy
Died 7 months later
(continued on next page)
598
1
No/
Headache
?
M/32
Iwatsuki et al
(1999)
[Japanese]b
Qasho et al
(1999)
Escott (2001)
Abbreviations: F, female; GTR, gross total resection; M, male; MI, myocardial infarction; , absent; +, present.
a
References were excluded for the reasons indicated: [42] article written in German, unable to conrm; [26] article written in Italian, unable to rule out meningeal
carcinomatosis; [43] clearly meningeal carcinomatosis; and [25] clearly meningeal carcinomatosis.
b
English abstract only.
Lesion decreased in size

after radiosurgery
Gross total resection
Transcortical
(temporal)
Gamma Knife
radiosurgery
1
Bladder
(adenocarcinoma)
? (melanoma)
M/40
No/+
Right lateral
ventricle
Left trigone
Improved condition?
Approach not
specied
Left trigone
1
Yes/+
Headache,
decreased
consciousness
Seizure
Kidney (renal cell
carcinoma)
F/75
Bleeding/
edema
Site of primary
tumor (pathology)
Sex/age
(years)
Reference (year)
[language if not
English]
Table 1 (continued)
Time from
diagnosis
of primary
(months)
Presenting
symptoms
No.
metastases
Location in
ventricle
Surgical
procedure
Outcome
same logic, the high proportion of atrial lesions

seems to correlate with the large amount of
choroid plexus within the trigone (often referred
to as the glomus of the choroid plexus). Because
metastases are generally believed to reach the
brain via the blood stream (hematogenous
spread), this higher proportion of choroid plexus
may account for their frequent location at the
trigone. If one considers the production of
cerebrospinal uid (CSF) to be the ltered
product of blood (or plasma) passing through
the choroid plexus, it is logical to hypothesize that
tumor emboli may be trapped in this ltering
system. On review of the operative notes of our
cases, an attachment of the metastasis to the
choroid plexus was conrmed in 11 of 24 cases
operated on (46%), which supports this concept.
Likewise, choroid plexus involvement was documented in all the cases reported in the literature.
In most of these cases, cerebral angiography
conrmed that the primary blood supply of the
lesion arose from the choroidal vessels. Nevertheless, this simple trapping hypothesis cannot
entirely explain the occurrence of tumors in the
ventricle, because the choroid is aected less
commonly than would be predicted by its
signicant vasculature [34]. Moreover, the low
rate of metastasis from lung cancer (the most
common systemic tumor) and the high rate of
metastasis from renal cell cancer (a more rare
cancer) suggest that other biologic factors must
govern the occurrence of intraventricular metastases.
Surgical management strategies

Similar to its role in the treatment of the more
common intraparenchymal metastases [32], surgical resection is an important component in the
management of patients with intraventricular
metastases. The overall goal of management is
to cure the brain lesion(s) while avoiding
signicant morbidity to the patient. Although
surgery is a critical treatment modality, the
surgeon must also consider the role of other
options, particularly whole-brain radiotherapy
(WBRT) and stereotactic radiosurgery, when
recommending the most appropriate treatment.
Deciding which patients are appropriate for
surgery requires that clinicians weigh the immediate risk of surgical intervention against the longterm benets of quality survival. Thus, surgical
decision making for patients with intraventricular
599
Table 2
Summary characteristics of patients at the University of Texas M.D. Anderson Cancer Center with intraventricular
metastases
Single metastasis
Number of patientsa
Age (years)
Sex
Intraventricular metastases
Presentation (with respect
to primary)
Time from diagnosis of
primary (months)
Presenting symptomsb
Hemorrhage
Location in ventriclec
Side of lateral ventricle

Primary (histology)
12
64
9
3
12
0
1
Median (range)
Male
Female
At presentation
Developed subsequently
Synchronous
Metachronous
Median (range)
(34%)
(3669)
(75%)
(25%)
(100%)
(0%)
(8%)
11 (92%)
30.5 (084)
Headache
Nausea and vomiting
Decreased consciousness
Ataxia
Seizure
Motor weakness
Other
Asymptomatic
Lateral
Trigone
Body
Anterior horn
Temporal horn
Third
Fourth
Left
Right
Breast (adenocarcinoma)
Colon (adenocarcinoma)
Esophagus (adenocarcinoma)
Kidney (renal cell carcinoma)
Lung (adenocarcinoma)
(sarcoma)
Thyroid (papillary carcinoma)
Skin (melanoma)
6
2
1
2
0
1
3
2
2
10
(50%)
(17%)
(8%)
(17%)
(0%)
(8%)
(25%)
(17%)
(17%)
(27%)
8 (21%)
1 (3%)
0 (0%)
1 (3%)
1 (3%)
1 (3%)
8 (32%)
2 (8%)
1 (8%)
2 (17%)
0 (0%)
7 (58%)
1 (8%)
0 (0%)
0 (0%)
1 (8%)
Multiple metastases
23
50
12
11
19
4
1
(66%)a
(2069)
(52%)
(48%)
(83%)
(17%)
(4%)
22 (96%)
14.8 (022.8)
11
4
2
4
2
4
2
4
2
15c
(48%)
(17%)
(9%)
(17%)
(9%)
(17%)
(9%)
(17%)
(9%)
(40%)
7 (18%)
6 (16%)
1 (3%)
1 (3%)
3 (8%)
7 (19%)
10 (50%)
5 (10%)
4 (17%)
0 (0%)
1 (4%)
9 (39%)
2 (9%)
1 (4%)
2 (9%)
4 (17%)
Nineteen of the 23 patients with multiple metastases had an intraventricular metastasis at the initial diagnosis of
brain metastases, whereas 4 patients presented with an intraparenchymal metastasis and subsequently developed
intraventricular metastases.
b
Patient may have had more than one symptom.
c
Two patients had 2 intraventricular tumors (total number of tumors = 37).
metastases must take into account the fact that

intraventricular lesions are deep within the brain;
thus, surgical resection of these lesions may be
associated with a slightly higher risk than for
supercially located intraparenchymal lesions.
Ultimately, the appropriate candidates for surgery
are selected by carefully analyzing the clinical
status of the patient, pertinent radiographic
studies, and the histology of the tumor. Each of
these areas has been analyzed in detail elsewhere

[32], but they are reviewed here as they specically
relate to intraventricular metastases.
Clinical assessment
The status of the systemic disease (the
primary tumor and noncerebral metastases) is
the most important determinant of overall
600
Fig. 1. (A) Preoperative axial (left) and sagittal (right) contrast-enhanced MRI scans of a 50-year-old man with known
renal cell carcinoma and a left atrium/trigone lesion. (B) Corresponding immediate postoperative MRI scans after a gross
total resection via a transcortical superior lobule approach. The trajectory of this approach is evident.
survival in patients with cerebral metastases. Up

to 70% of patients undergoing surgery for single
metastases die from progression of the systemic
disease rather than from neurologic causes [35].
Consequently, surgery is generally considered
only in patients with absent, controlled, or
limited systemic cancer. Although the denition of controlled or limited systemic cancer is
clearly subjective, patients who are expected to
survive more than 4 months are generally
considered appropriate candidates for surgical
resection. This consideration is particularly important for patients with deep lesions, such as
those in the ventricle, because their recovery time
from such surgery may be somewhat longer than
for patients with supercially located lesions,
which may reduce the benet of surgical intervention.
Radiographic assessment
Radiographic assessments are used to determine the number of lesions and the size of each
lesion.
Single intraventricular metastases
Patients with single brain metastases are the
most appropriate surgical candidates. Indeed, two
independent prospective randomized studies demonstrated that surgical resection is superior to
WBRT in the treatment of single brain metastases
in terms of morbidity, recurrence, and overall
survival [35,36]. Although these studies did not
specically focus on intraventricular metastases, it
is probably appropriate to conclude that surgical
resection of intraventricular metastases is likely to
be superior to WBRT as long as the lesion can be
601
Fig. 2. (A) Preoperative axial (left) and sagittal (right) contrast-enhanced MRI scans of a 65-year-old woman with
known breast cancer and a fourth ventricular metastasis. (B) Corresponding immediate postoperative MRI scans after
a gross total resection via a midline suboccipital approach.
resected completely and with minimal morbidity

(see Fig. 1).
More recently, however, it has been suggested
that stereotactic radiosurgery may be an eective
alternative to surgery in the treatment of single
brain metastases [37]. Indeed, the deep location of
intraventricular metastases makes radiosurgery
particularly attractive because it is less invasive.
To date, however, there has been no randomized
trial comparing conventional surgery with stereotactic surgery in the treatment of brain metastases.
Until such a trial is completed, it remains our
contention that surgery oers signicant advantages over stereotactic radiosurgery even for deepseated intraventricular lesions. First, surgical
resection provides histologic conrmation that
the lesion is a metastasis, whereas tissue is not
obtained with radiosurgery. Second, surgery is
eective regardless of tumor size, whereas radiosurgery is not indicated for lesions greater than
3 cm in maximum diameter, and recent studies
suggest that good tumor control is achieved
primarily for lesions less than 1 cm in diameter
[38]. Thus, radiosurgery may be best applied to
relatively small lesions. Third, because surgery
rapidly removes the lesion, it reverses symptoms
more eciently than radiosurgery and reduces the
complications associated with long-term corticosteroid use. Fourth, surgery is eective regardless
of tumor histology, whereas the eects of radiosurgery are less predictable, particularly for
lesions that commonly produce ventricular metastases, such as renal cell carcinoma. Lastly, the
risk of delayed radiation injury that can occur
with radiosurgery is avoided with surgical intervention. The consequence of this injury may be
602
Fig. 3. Contrast-enhancing axial (A) and sagittal (B) MRI scans demonstrating a third ventricular tumor in a 42-yearold man with a history of thyroid cancer. The patient initially presented with an intraventricular hemorrhage that was
treated with ventricular drainage and a ventriculoperitoneal shunt. The images in this gure were obtained 2 months
after the initial presentation when the hemorrhage had resolved. The patient was treated with stereotactic radiosurgery
because of the deep location and small size of the lesion.
more pronounced for intraventricular lesions that

are near vital structures, such as the thalamus and
fornices.
Thus, given our present knowledge, our preferred treatment of single intraventricular brain
metastases is surgical resection (see Figs. 1 and 2).
We reserve stereotactic radiosurgery for patients
with uncontrolled or widespread systemic disease
or with medical conditions that preclude surgery.
In addition, we often recommend radiosurgery
for asymptomatic patients with intraventricular
lesions less than 1 cm in maximum diameter that
are found incidentally during screening. Consistent with this philosophy, of the 12 patients with
single intraventricular metastases treated at our
institution (see Table 2), 11 underwent conventional surgery as the rst line of therapy, whereas
1 was treated with stereotactic radiosurgery
because of the small tumor size (see Fig. 3).
Multiple intraventricular metastases
Although the presence of more than one brain
metastasis has traditionally been considered a contraindication to surgery, more recently, surgical
resection has been used in the treatment of
patients with multiple (ie, 24) brain metastases. This approach is based on a retrospective
study from M.D. Anderson demonstrating that
removal of multiple metastatic lesions was as
eective as resection of single metastases provided
that all lesions were removed [39]. Resection of an

intraventricular metastasis in the presence of
multiple brain metastases requires careful consideration, however. In fact, in our series of 35 cases,
19 patients presented with multiple brain lesions.
Two of these patients had two intraventricular
lesions; both lesions in 1 patient were treated
initially with WBRT, and both lesions in the other
patient received stereotactic radiosurgery. Of 11
patients who had both an intraventricular lesion
and an intraparenchymal lesion, only 3 were
treated with resection of the intraventricular
lesion, whereas 8 had an intraventricular lesion
treated with radiosurgery and 1 received WBRT.
Lastly, 7 patients presented with an intraventricular metastasis plus two or more intraparenchymal lesions; 4 patients underwent surgical
resection of their symptomatic intraventricular
lesions. Thus, of the 19 patients with multiple
lesions that included an intraventricular tumor,
only 7 (37%) were treated initially with surgery.
Indeed, given the complexity of resecting multiple
lesions, less invasive strategies, such as radiosurgery, are an attractive alternative.
Histologic assessment
It is important to consider the radiosensitivity
and chemosensitivity of the primary tumor before
proceeding to surgery. Metastases from small cell
lung cancer and germ cell tumors are particularly

sensitive to radiation and chemotherapy and are
probably best treated with these modalities. These
tumor types, however, rarely occur as intraventricular lesions. Melanoma, renal cell carcinoma,
and most sarcomas are essentially resistant to
fractionated radiation and are best treated with
surgery. The most common type of cancer producing intraventricular metastases, renal cell
carcinoma, is usually resistant to fractionated
radiation therapy and to most chemotherapeutic
agents; thus, surgery is the preferred treatment
modality. Tumors like non-small cell lung cancer
and breast cancer are intermediately sensitive to
radiation, and surgery should be considered as
part of a multidisciplinary scheme. In addition, it
must be remembered that the responsiveness of
tumors to single high-dose radiosurgery may
dier from their responsiveness to fractionated
radiotherapy. Indeed, radiosurgery may be an
eective treatment for many of the more resistant tumors, including renal cell carcinoma.
Surgical Techniques
Successful removal of intraventricular cerebral
metastases is safely accomplished in the modern
era because of advances in surgical technique.
Better understanding of the surgical anatomy of
these lesions has led to safer operative approaches, and accurate localization of lesions on
MRI has been translated into better intraoperative localization using image-guided methods.
Although the approaches to intraventricular
tumors have been discussed elsewhere in this issue
and certainly apply to metastatic tumors, certain
technical aspects unique to metastatic tumors are
worthy of note.
Surgical anatomy
Several anatomic features unique to intraventricular metastases must be recognized for safe
and eective resection. Intraventricular metastases are typically well-dened lesions that are
invariably attached to the choroid plexus. This
attachment may be quite broad, and inltration of
the choroid is common. Consequently, removal of
the choroid back to a zone that appears normal
may be required to prevent recurrence of the
lesion. In addition, attachment or inltration of
the wall of the ventricle may occur because of the
malignant nature of these tumors. Metastases
typically have a gliotic pseudocapsule, however,
603
and circumferential dissection in this gliotic plane

generally ensures gross total resection, because
there are typically no tumor cells in this zone.
Lastly, intraventricular metastases almost invariably receive their blood supply from the choroidal
arteries and drain via the choroidal veins. These
vessels are usually at the base of the lesion at
a position farthest from the initial surgical
exposure. Thus, entering an intraventricular
metastasis at the beginning of the resection
without interrupting the blood supply results in
signicant hemorrhage (especially if a renal cell
carcinoma) and obscures vision. Consequently, we
generally prefer to perform a circumferential
dissection and to identify the attachment to the
choroidal arteries before entering the lesion. En
bloc removal is also preferred to an inside-out
approach because it reduces the possibility of
shedding tumor cells into the ventricular system.
Nevertheless, the size of the lesion and the small
deep surgical corridor preclude en bloc removal in
many cases. In these situations, circumferential
dissection is performed as widely as possible and
internal debulking is done. Progressive circumferential dissection and internal debulking are
undertaken until the supplying blood vessels can
be isolated, coagulated, and cut. Bleeding from
the tumor stops when complete resection is
achieved and the choroidal vessels are secured.
The presence of bleeding invariably signals residual tumor. Cottonoids are used to prevent
blood from spilling throughout the ventricles.
Surgical approaches
The surgical approach to metastatic intraventricular tumors depends on the location of the
lesion within that ventricular compartment. A
variety of surgical approaches have been described for dierent locations in the ventricles,
and each is discussed elsewhere in this issue. All
approaches to the ventricle require transgressing
normal brain structures at some point. Therefore,
our general philosophy is to use an approach that
limits neurologic injury, even if the approach
requires a longer surgical corridor.
Because most intraventricular metastatic tumors arise within the atrium/trigone of the lateral
ventricle, special attention to this location is in
order. For lesions in the dominant hemisphere,
the atrium of the lateral ventricle is one of
the more complex areas to access, because the surrounding cortical structures are critical to language and visual functions. Interhemispheric
604
approaches usually do not provide enough lateral

exposure; thus, transcortical approaches are invariably required. Although transtemporal and
lateral transparietal approaches have been suggested for dominant hemisphere atrial lesions
[40,41], these approaches usually result in significant aphasia and visual eld cuts because of the
important role of the posterior temporal gyrus
and the underlying white matter bers in these
functions. Approaches through the inferior temporal gyrus are generally too low to provide access
to the atrium. Consequently, we generally attack
these lesions through the superior parietal lobule,
typically splitting the adjacent sulcus before
entering the brain. Computer-assisted stereotactic
guidance is used to plan the craniotomy and to
maintain the corridor during surgery. The use
of a computer-driven robotic microscope (eg,
SurgiScope, Intelligent Surgical Instruments and
Systems, Saint Martin DHeres, France) is particularly applicable in these cases. Functional
mapping of the position of the sensory cortex
ensures passage through the superior parietal lobule
and prevents injury to important motor and sensory
ber tracts. A corridor of 2 cm 2 cm usually
provides the necessary exposure to the lesion. This
approach typically avoids aphasia and spares the
visual projection bers. In contrast, for lesions in
the nondominant (right) hemisphere, violation of
the superior parietal lobule results in visual-spatial
dysfunction that manifests at least partly by
dressing apraxia. Consequently, for right-sided
atrial lesions, we typically use a middle temporal
gyrus or superior temporal sulcal exposure.
Surgical complications and outcome
Because of the relatively short lifespan of
patients with systemic cancer, it is important that
surgical resection of brain metastases, including
intraventricular metastases, be undertaken with
minimal morbidity. Review of the cases in the
literature (see Table 1) demonstrates that postoperative complications from intraventricular
surgery in the past have been signicant. Nevertheless, in the modern era, resection of intraventricular tumors can be achieved with low mortality
(<5%) and morbidity (<10%) [22]. In our M.D.
Anderson series of 24 intraventricular tumors
operated on by several surgeons, there was no
operative mortality and three patients (12%)
experienced signicant neurologic complications.
The outcome of patients with intraventricular
metastases has not been specically analyzed
heretofore. In our series of 11 patients with single

intraventricular metastases who were treated with
surgical resection, the median survival time was
13.6 months. Thus, the survival of these patients
seems to be similar to that of the more general
population of surgically treated patients with
single intraparenchymal metastases [31,32]. Of
course, our patients with intraventricular tumors
had a predominance of renal cell carcinomas,
which may be prognostic for a longer survival
interval [32]. Interestingly, despite having a tumor
within the ventricle, only 2 of 11 patients with
single brain metastases eventually developed
meningeal carcinomatosis. In all, resection of
deep-seated intraventricular tumors seems to be
as eective as resecting more supercially located
brain metastases.
Summary
The ventricle is a rare site of brain metastases.
Renal cell carcinoma has a higher propensity to
metastasize to the ventricle compared with more
common metastatic tumors (eg, lung cancer). The
trigone is the predominant location for intraventricular metastases, presumably because of the
high concentration of choroid plexus in this region.
Surgical resection is an important component of
the management of these lesions, particularly if
there is only a single intraventricular lesion.
Despite the deep location of these tumors within
the ventricle, survival in patients undergoing
surgery for them is comparable to that in patients
receiving surgery for intraparenchymal metastases.
Acknowledgment
The authors thank Weiming Shi for his
excellent assistance with the gures and Sandra
Flores for her excellent assistance with the
preparation of this manuscript.
References
[1] Healy JF, Rosenkrantz H. Intraventricular metastases demonstrated by cranial computed tomography. Radiology 1980;136(1):124.
[2] Kendall B, Reider-Grosswasser I, Valentine A.
Diagnosis of masses presenting within the ventricles
on computed tomography. Neuroradiology 1983;
25(1):1122.
[3] Killebrew K, Krigman M, Mahaley MS Jr, Scatli
JH. Metastatic renal cell carcinoma mimicking
a meningioma. Neurosurgery 1983;13(4):4304.

[4] Kart BH, Reddy SC, Rao GR, Poveda H. Choroid
plexus metastasis: CT appearance. J Comput Assist
Tomogr 1986;10(3):53740.
[5] Mertens R, Muhler E, Heimann G. Disseminated
neuroblastoma with intracerebral metastasis in
a newborn infant. Klin Padiatr 1987;199(6):4248.
[6] Shigemori M, Shimamoto H, Noguchi S, Yoshitake
Y, Sugita Y, Kuramoto S. Choroid plexus metastasis of renal cell carcinoma. CT Kenkyu 1987;9(5):
6036.
[7] Tanimoto M, Tatsumi S, Tominaga S, Kamikawa
S, Nagao T, Tamaki N, et al. Choroid plexus
metastasis of lung carcinomacase report. Neurol
Med Chir (Tokyo) 1991;31(3):1525.
[8] Mizuno M, Asakura K, Nakajima S, Sampei T,
Sayama I, Kawamura S, et al. Renal cell carcinoma
metastasizing to choroid plexus of lateral ventricle;
a case report. No Shinkei Geka 1992;20(4):
4694.
[9] Nakabayashi H, Murata K, Sakaguchi M, Nakajima K, Katsuyama J. Choroid plexus metastasis
from gastric cancercase report. Neurol Med Chir
(Tokyo) 1994;34(3):1836.
[10] Suetake K, Shinya T, Takeda M. A choroid plexus
metastasis of renal cell carcinoma: a case report. No
Shinkei Geka 1994;3:43641.
[11] Berkow RL, Kelly DR. Isolated CNS metastasis as
the rst site of recurrence in a child with germ cell
tumor of the mediastinum. Med Pediatr Oncol
1995;24(1):369.
[12] Spetzger U, Mull M, Sure U, Gilsbach J. Subarachnoid and intraventricular hemorrhage caused
by hypernephroma metastasis, accompanied by
innocent bilateral posterior communicating artery
aneurysms. Surg Neurol 1995;44(3):2758.
[13] Kohno M, Matsutani M, Sasaki T, Takakura K.
Solitary metastasis to the choroid plexus of the
lateral ventricle. Report of three cases and a review
of the literature. J Neurooncol 1996;27(1):4752.
[14] Brandicourt P, Toussaint P, Le Gars D, Depriester
D, Deramond H. Metastasis to the 3rd ventricle of
colonic adenocarcinoma. Presse Med 1997;26(2):75.
[15] Matsumura H, Yoshimine T, Yamamoto S, Maruno M, Hayakawa T, Ono Y, et al. Single solitary
metastasis of the slowly progressive type of renal
cell carcinoma to the choroid plexuscase report.
Neurol Med Chir (Tokyo) 1997;37(12):9169.
[16] Raila FA, Bottoms WT Jr, Fratkin JD. Solitary
choroid plexus metastasis from a renal cell carcinoma. South Med J 1998;91(12):115962.
[17] Gaab MR, Schroeder HW. Neuroendoscopic approach to intraventricular lesions. J Neurosurg 1998;
88(3):496505.
[18] Arbelaez A, Castillo M, Armao DM. Imaging
features of intraventricular melanoma. AJNR Am
J Neuroradiol 1999;20(4):6913.
[19] Iwatsuki K, Sato M, Taguchi J, Fukui T, Kiyohara
H, Yoshimine T, et al. Choroid plexus metastasis
of renal cell carcinoma causing intraventricular
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
605
hemorrhage: a case report. No Shinkei Geka 1999;

27(4):35963.
Qasho R, Tommaso V, Rocchi G, Simi U, Delni
R. Choroid plexus metastasis from carcinoma of the
bladder: case report and review of the literature.
J Neurooncol 1999;45(3):23740.
Escott EJ. A variety of appearances of malignant
melanoma in the head: a review. Radiographics 2001;
21(3):62539.
Morita A, Kelly PJ. Resection of intraventricular
tumors via a computer-assisted volumetric stereotactic approach. Neurosurgery 1993;32(6):9206.
Chou MS, Tsai TC, Yang CW, Liu GC, Howng SL.
Ventricular tumorsCT evaluations of 28 cases.
Gaoxiong Yi Xue Ke Xue Za Zhi 1993;9(3):14352.
Majos C, Coll S, Aguilera C, Acebes JJ, Pons LC.
Intraventricular mass lesions of the brain. Eur
Radiol 2000;10(6):95161.
Khoshyomn S, Bra SP, McKenzie MA, Florman
JE, Pendlebury WW, Penar PL. Metastatic intraventricular melanoma. Case illustration. J Neurosurg 2002;97(3):726.
Bugiani O, Bava GL. Metastasis of bronchial
carcinoma to the choroid plexus and the roots of
the cranial and spinal nerves. Anatomo-clinical
observation. Sist Nerv 1968;20(6):4037.
Aparicio A, Chamberlain MC. Neoplastic meningitis. Curr Neurol Neurosci Rep 2002;2(3):22535.
Schreiber D, Bernstein K, Schneider J. Metastases
of the central nervous system: a prospective study.
3rd communication: metastases in the pituitary
gland, pineal gland, and choroid plexus. Zentralbl
Allg Pathol 1982;126(12):6473.
Arendt A, Dieck I, Fuchs U. Frequency of
metastases of extracranial tumours in the choroid
plexus. Zentralbl Allg Pathol 1980;124(5):3957.
Kelly PJ. Resection of intraventricular tumors via
a computer-assisted volumetric stereotactic approach. Neurosurgery 1993;33(4):7712.
Sawaya R, Bindal RK. Metastatic brain tumors.
In: Kay AH, Laws ER, editors. Brain tumors:
an encyclopedic approach. Edinburgh: Churchill
Lang FF, Sawaya R. Surgical management of
cerebral metastases. Neurosurg Clin North Am
1996;7(3):45984.
Taheri SA, Wani MA, Lewko J. Uncommon causes
of intraventricular hemorrhage. Clin Neurol Neurosurg 1990;92(3):195202.
Takakura K, Sano K, Hojo S, Hirano A. Metastatic
tumors of the central nervous system. Tokyo: Igaku
Shoin; 1982.
Patchell RA, Tibbs PA, Walsh JW, Dempsey RJ,
Maruyama Y, Kryscio RJ, et al. A randomized trial
of surgery in the treatment of single metastases to
the brain. N Engl J Med 1990;322(8):494500.
Vecht CJ, Haaxma-Reiche H, Noordijk EM,
Padberg GW, Voormolen JH, Hoekstra FH, et al.
Treatment of single brain metastasis: radiotherapy
606
alone or combined with neurosurgery? Ann Neurol

1993;33(6):58390.
[37] Kondziolka D, Lunsford LD. Brain metastases. In:
Apuzzo MLJ, editor. Brain surgery: complication
avoidance and management. New York: Churchill
[38] Chang EL, Hassenbusch SJ, Shiu AS, Lang FF,
Allen PK, Sawaya R, et al. The role of tumor size
in the radiosurgical management of patients with
equivocal brain metastasis. Neurosurgery 2003;
53:27281.
[39] Bindal RK, Sawaya R, Leavens ME, Lee JJ.
Surgical treatment of multiple brain metastases.
[40] Santoro A, Salvati M, Frati A, Polli FM, Delni R,

Cantore G. Surgical approaches to tumours of the
lateral ventricles in the dominant hemisphere.
J Neurosurg Sci 2002;46(2):605.
[41] Andoh T, Shinoda J, Miwa Y, Hirata T, Sakai N,
Yamada H, et al. Tumors at the trigone of
the lateral ventricleclinical analysis of eight
cases. Neurol Med Chir (Tokyo) 1990;30(9):
67684.
[42] Faber V. Fall von carcinomatos entartetem papillom des seit. Frankf Z Pathol 1934;47:16872.
[43] Al-Anazi A, Shannon P, Guha A. Solitary metastasis to the choroid plexus. Case illustration.
J Neurosurg 2000;92(3):506.
Intraventricular congenital lesions and colloid cysts

Aurelia Peraud, MDa, Anna Illner, MDb,
James T. Rutka, MD, PhD, FRCSCa,*
a
Division of Neurosurgery, Suite 1502, The Hospital for Sick Children, The University of Toronto,
555 University Avenue, Toronto, Ontario, Canada M5G 1X8
b
Division of Neuroradiology, The Hospital for Sick Children, The University of Toronto,
555 University Avenue, Toronto, Ontario, Canada M5G 1X8
Intraventricular congenital lesions and colloid

cysts comprise a rather large spectrum of dierent
pathologic conditions, which are outlined in the
following article. Treatment is not warranted in
most cases unless there is progressive ventricular
obstruction with hydrocephalus or growth of the
lesion itself, making tissue biopsy and histopathologic diagnosis necessary. Accordingly, a precise
neuroradiologic evaluation is of the utmost
importance, because most lesions, if not symptomatic, only require clinical and radiologic
follow-up.
Developmental midline intracranial cysts constitute a separate entity and are usually associated
with malformation disorders of the brain. They
can be intraventricular, paraventricular, or intraarachnoid in location. Detailed evaluation with
neuroimaging is benecial because it directly
aects subsequent management. The presence of
communication with the ventricular system, associated hydrocephalus, and mass eect are features
that help the neurosurgeon to plan appropriate
treatment.
Neurosurgical techniques, such as endoscopy,
allow a more direct approach for fenestration of
either the cyst or the ventricular system if there is
associated hydrocephalus. The move toward these
less invasive methods of treatment demands
Dr. Aurelia Peraud was supported by the Deutsche

Forschungsgemeinschaft, grant Pe 758/2-1.
E-mail address: james.rutka@sickkids.ca
(J.T. Rutka).
detailed anatomic roadmaps that can be provided

by appropriate neuroimaging.
Before we can fully appreciate congenital intraventricular lesions, it is important to recognize
cavities that exist as normal variants, such as the
cavum septum pellucidum (CSP), cavum vergae,
and cavum veli interpositi (CVI).
Embryology
The neural crest begins to form at 3 weeks of
gestation of the trilaminar embryo (containing
ectoderm, mesoderm, and endoderm) by folding
of the neural plate into the neural tube. The
anterior neuropore closes at 18 days of gestation;
by the end of the fourth week, the primary brain
vesicle exists and the hindbrain exures outline
the primary divisions of the brain. The forebrain
divides into two secondary vesicles during the fth
week, forming the telencephalon and the diencephalon. With further invagination of the
telencephalon and growth of the cerebral hemispheres in a dorsorostral direction, the frontal,
temporal, and occipital poles form around the
diencephalon during the seventh week. The lateral
ventricles follow accordingly and are drawn from
the frontal pole to the temporal pole in a Cshaped manner.
A vascular layer of the pia mater fuses with the
ependyma to form the tela choroidea, which subsequently invaginates into the ventricles through
the choroidal ssure as the choroid plexus is
developing. Cerebrospinal uid (CSF) production starts, and, nally, in the second trimester,
the thin roof of the fourth ventricle bulges and
doi:10.1016/S1042-3680(03)00057-3
608
A. Peraud et al / Neurosurg Clin N Am 14 (2003) 607619
ruptures in three locations, giving rise to the

foramina of Luschka and Magendie and allowing
CSF access into the subarachnoid space.
Normal variants
Cavum septum pellucidum
As a single midline structure, the septum
pellucidum separates the two anterior horns of
the lateral ventricles (Fig. 1). The CSP is demarcated by the genu of the corpus callosum
anteriorly , by the columns and body of the fornix
posteriorly, by the body of the corpus callosum
superiorly, and by the rostrum of the corpus
callosum inferiorly. It consists of an ependymal
lining toward the ventricles and contains neuronal
and glial cell elements. These cell elements have
connections to the hypothalamus and the hippocampus. At birth, the two layers of the septum
pellucidum are separate and enclose a cavum.
Later in life, these two layers typically fuse into
a single septum. Autopsy and imaging studies
have shown that all premature infants and 97% of
term infants have a CSP, with the incidence
dropping to 41% by 3 months of age and to 15%

by 6 months of age [1]. The CSP may remain
dilated in the context of some congenital disorders
with arrest of normal brain development or may
secondarily enlarge with repetitive brain trauma,
such as in boxers. In such instances, obstructive
hydrocephalus results from compression at the
foramina of Monro and may require neurosurgical treatment.
Cavum vergae
If the layers of the septum pellucidum posterior to the columns of the fornix do not merge,
they leave a cavum vergae, which is commonly
seen in combination with a CSP (see Fig. 1) [2]. It
is not clear whether the cavum vergae is the
posterior portion of the CSP or whether it
develops independently and communicates with
the CSP. The cavum vergae is bordered by the
body of the corpus callosum superiorly, by the
hippocampal ssure inferiorly, by the crus of
the fornices laterally, and by the splenium of the
corpus callosum posteriorly. This anatomic variant is present in about one third of newborns and
persists only rarely until adulthood. Interestingly,
Fig. 1. Cavum septum pellucidum (CVP) and cavum velum interpositum (CVI). (A) Axial T2-weighted image
demonstrates the CVP anteriorly (arrow) and CVI posteriorly (open arrow). (B) Sagittal T1-weighted image demonstrates
anterior and superior displacement of the fornix (open arrow) distinguishing the CVI from the cavum vergae. Note
characteristic inferior displacement of the internal cerebral veins (arrow).
the cavum vergae disappears before the CSP.

Cystic enlargement of the cavum vergae may
cause hydrocephalus by obstruction of either the
foramen of Monro or the body of the lateral
ventricle.
Cavum veli interpositi
This rare variant develops through an anterior
extension of the pia-arachnoid membrane that
arises from the quadrigeminal plate cistern. The
CVI is situated between the crus of the fornices
and lies inferior to the hippocampal commissure
and the corpus callosum and superior to the roof
of the third ventricle [3]. The CVI may extend as
far as the columns of the fornix. It is formed from
a double layer of pia mater, the tela choroidea,
which covers the ependymal roof of the third
ventricle, and results in uid accumulation within
the potential space of these two layers when the
posterior end of the tela choroidea remains open.
The internal cerebral veins and the medial posterior choroidal artery lie within the two layers
and can be displaced by cystic expansion of the
CVI inferolaterally. Cystic enlargement of the
CVI requiring treatment is exceptional, with only
a few case reports in the literature [4,5].
Intraventricular cysts
The development of intracranial cysts, especially those in the midline, can occur in isolation
or may be associated with further anomalies of the
brain, such as dysgenesis of the corpus callosum,
and incomplete forebrain cleavage within the
spectrum of holoprosencephaly [6,7]. Midline
cysts may be composed of ependymal, glioependymal, epithelial, neuroepithelial, or true arachnoid features.
There are several classication systems of
interhemispheric cysts, with most being based on
the communication with the ventricular system
and location in relation to brain parenchyma.
More recently, interhemispheric cysts have been
classied according to a combination of criteria,
including the presence of further developmental
anomalies or clinical characteristics [8,9].
Colloid cysts
Third ventricular colloid cysts constitute 0.5%
to 1% of all intracranial mass lesions and arise
from the anterior roof of the third ventricle. There
is a slight male predominance [10] and a pre-
609
dilection for the third and fourth decades of life.

The clinical symptomatology may vary considerably. Sudden death as the result of acute
obstruction of the foramen of Monro has been
reported [11]. Most patients present with headaches that have a sudden onset with change in
head position. An explanation for this phenomenon is acute ventricular obstruction from cyst
occlusion of the foramen of Monro. The bobblehead doll syndrome typically described in the
context of colloid cysts has the same underlying
pathophysiology [12]. Because these cysts are
located between the columns of the fornix, an
association with short-term memory decits is not
uncommon [10].
Colloid cysts are lled with thick gelatinous
uid that is rich in cholesterol and contains calcications. They are typically hyperdense on CT
and hyperintense on T1-weighted images (Fig. 2)
[13]. The cyst appearance on MRI can vary
considerably, and the signal intensity is directly
related to the viscosity of the cyst, which can be
helpful in deciding whether stereotactic aspiration
or open craniotomy for removal should be performed. It is not unusual to nd colloid cysts at
autopsy as an incidental nding; however, those
individuals who become clinically symptomatic
require surgical intervention. Denitive cure can
be achieved only by microsurgical resection either
by an endoscopic route or via open transcallosal
or transcortical resection [14]. Cyst aspiration
alone carries the risk of reaccumulation of cyst
material with subsequent re-expansion of the cyst
and recurrence of symptoms. Memory impairment as a consequence of endoscopic or open
surgery is not an uncommon postoperative
phenomenon caused by trauma to the fornices.
The callosotomy should not be longer than 2.5 cm
and should remain in the anterior third to avoid
disconnection syndromes. Seizures can occur after
a transcortical route, and sagittal sinus thrombosis is a feared complication, albeit rare, after
a transcallosal approach [10].
Arachnoid cysts
Arachnoid cysts occur most frequently in the
middle cranial fossa (30%50%) and less commonly over the convexity (10%), in the suprasellar
region (10%), in the quadrigeminal plate (10%),
in the cerebellopontine angle (10%), or in the
midline posterior fossa (10%). In rare cases, these
cysts may occur unrelated to the cerebral cisterns,
such as in the ventricles or the diploe of the
610
Fig. 2. Colloid cyst. (A) Unenhanced CT in an adult demonstrates a homogeneous hyperdense lesion in the region of the
foramen of Monro. The ventricles are mildly enlarged. (B, C ) Coronal T1-weighted postcontrast and axial T2-weighted
images in another adult patient show a nonenhancing cyst with medium signal intensity in the roof of the third ventricle
causing mild ventricular enlargement. The unusual signal characteristics and location of the lesion are typical of a colloid
cyst.
skull. Intraventricular arachnoid cysts are mainly

located in the lateral ventricles and only rarely in
the third or fourth ventricle [15,16]. Those in the
lateral ventricle develop by invagination of
arachnoid through the choroidal ssure into the
choroid plexus [17] and appear as CSF-lled
thin-walled lesions in the atrium of the ventricle

(Fig. 3) [18]. They can be classied as either primary or secondary. Primary intraventricular arachnoid cysts arise in the lateral or fourth ventricle,
whereas secondary cysts originate extra-axially
and extend secondarily into the ventricular
611
Fig. 3. Intraventricular arachnoid cyst. (A) Coronal T2-weighted image demonstrates a large trigonal arachnoid cyst.
(B) Axial T2-weighted image demonstrates involvement of the perimesencephalic cistern via the choroidal ssure
(arrow), indicating a secondary rather than primary arachnoid cyst.
system as with suprasellar arachnoid cysts. Primary intraventricular arachnoid cysts are rare, with
less than 20 cases reported in the literature. Patients
typically present with symptoms of a spaceoccupying lesion and with headaches, delayed
psychomotor development, macrocephaly, hydrocephalus, and seizures [16,1923].
Symptomatic cysts are usually treated with
fenestration by endoscopic or open techniques
[1922,24]. During surgery, they are easy to differentiate from ependymal cysts because of their
sole adherence to the choroid plexus, whereas the
latter show attachment to the ventricular ependyma [25].
Choroid plexus cysts
Choroid plexus cysts are of neuroepithelial
origin and are most often found in the latera ventricles according to autopsy studies (Fig. 4). They
are a common nding on prenatal ultrasound
during the second trimester, but most spontaneously resolve by birth or early infancy. Ventricular
asymmetry may be attributed to a transient choroid plexus cyst that resolves later in life [26]. Only
a few reports exist of choroid plexus cysts that
Fig. 4. Choroid plexus cyst. Axial T2-weighted image

demonstrates an oval, hyperintense, intraventricular cyst
in the region of the glomus of the choroid plexus (arrow),
characteristic of a choroid plexus cyst.
612
persist into adulthood. An increased association

with aneuploidy or trisomy 18 and 21 has been
reported [27]. Shangshotti and Metsky [28] have
described four stages of embryologic development
of the choroid plexus and attribute the occurrence
of these cysts to lobulation of the choroidal
neuroepithelium in stage 2. They present as
sharply marginated round lesions without septations and can be mobile on a pedicle. They may
cause intermittent obstruction of CSF, particularly at the foramen of Monro. Depending on the
cyst size, the symptomatic spectrum may include
seizures, motor or sensory impairment, papilledema, and an altered level of consciousness [29].
Treatment is indicated in symptomatic cysts that
persist past infancy.
contain cilia. They are most often found in the

atrium of the lateral ventricle. Occasionally, they
occur within the central white matter of the frontal
or temporoparietal lobes adjacent to the CSF space
and may clinically present with signs of increased
intracranial pressure or seizures [3133]. Ependymal cysts are thought to develop as a sequestration
of parts of the primitive ependymal lining into
the cortical mantle or the perimedullary mesh,
and they never communicate with the ventricular
system. They are incidental ndings on imaging
studies of the brain most of the time, but if they
become large enough to entrap the ventricle and
cause obstructive hydrocephalus, treatment becomes necessary and can be accomplished either by
draining of the cyst or resection [34].
Ependymal cysts
Pineal cysts
Ependymal cysts resemble arachnoid cysts, but

the protein content of the cyst uid is generally
higher and they appear as isointense or slightly
hyperintense to CSF on T1- and T2-weighted
images (Fig. 5) [30]. Their wall is composed of
columnar or cuboidal cells that may or may not
Pineal cysts are considered a normal anatomic

variant and can be found incidentally on MR
studies in 2% to 4% of the general population. An
even higher incidence of 40% has been reported at
autopsy [35]. There is a certain predilection for
women in the third decade of life. These cysts
Fig. 5. Ependymal cyst. (A) Axial T2-weighted image demonstrates asymmetric lateral ventricles with deviation of the
septum pellucidum. (B) Axial T1-weighted postcontrast image conrms the presence of an intraventricular cyst with
cerebrospinal uid intensity with visualization of the posterior margin of the cyst wall (arrow). Imaging ndings are
consistent with an ependymal or arachnoid cyst.
contain CSF-like uid. They rarely become

symptomatic. If they do, it is from enlargement
with obstruction of the aqueduct of Sylvius.
Typical clinical symptoms include headache, gaze
palsy, and papilledema [36]. Headaches may be
either chronic or intermittent, with intermittent
obstruction of the aqueduct responsible for the
latter. As described by Miyatake and colleagues
[37], compression of the veins in the pineal region,
including the precentral cerebellar vein, the internal cerebral veins, and the great vein of Galen,
may also contribute to headaches. A sudden onset
of headache can be caused by hemorrhage into the
cyst [36,38]. Parinaud syndrome with upward gaze
palsy is caused by focal compression of the superior colliculus of the tectum mesencephali. On
MRI scans, pineal cysts appear hyperintense on
T2-weighted images and can have a higher signal
than CSF on T1-weighted images because of their
higher protein content (Fig. 6). Typically, there is
no or mild enhancement of the cyst wall [39,40].
The cyst wall can be calcied, and this is best visualized on CT scans (see Fig. 6).
The current opinion about the ideal treatment
of symptomatic pineal cysts is controversial. Even
after adequate decompression, symptoms like headache may persist. The aqueduct may not reopen if
secondary adhesions have occurred. Operative
indications should be restricted to large cysts with
compression of nervous structures causing Parinaud syndrome, visual eld defects, or disturbances of motor or sensory long-tract bers. Open
surgery via either the infratentorial-supracerebellar or interhemispheric-transcallosal approach has
been advocated by some authors [4146]. Surgery
has the advantage of direct visualization of the
surrounding veins. Stereotactic cyst aspiration
was proposed by others [4750]. The associated
risk of hemorrhagic complications after stereotactic aspiration seems to be lower than formerly estimated. Nevertheless, improper tissue sampling as
well as the high probability of cyst re-expansion
makes the stereotactic approach less favorable.
Michielsen and colleagues [51] have clearly stated
that the endoscopic or endoscopic-assisted microneurosurgical approach in experienced hands is
the best form of treatment for pineal cysts.
Neurenteric cysts
Abnormal separation of germ cell layers in the
third week of gestation is thought to be responsible for persistence of the endodermal tissue
elements that can form neurenteric cysts in the
613
central nervous system (CNS). In most reported

cases, the cysts are spinal in location [52,53] often
with associated vertebral anomalies caused by
abnormal development of the adjacent mesoderm.
Fewer than 20 cases of intracranial neurenteric
cysts have been reported thus far, with most of
these located anterior to the brain stem [54,55] or
at the craniovertebral junction [56]. Histologically, neurenteric cysts are composed of a thin epithelial wall, which may contain cuboidal or
columnar cells with or without ciliation that lie
on a basement membrane and connective tissue of
varying vascularity [54,57]. If other germ cell
elements are present within the cyst, the diagnostic trend is toward a teratoma. Afshar and
Scholtz [58] reported on an enterogenous cyst of
the fourth ventricle. Because of their expanding
nature, surgical cyst excision is the treatment of
choice. Aspiration of the cyst content alone is
considered insucient, because the cyst may reexpand by an osmotic gradient or by active mucin
secretion of the epithelium. These cysts are known
to be attached rmly to surrounding structures,
however, making them dicult to remove completely without further neurologic decits. If cyst
removal is not safely possible, opening of the cyst
and partial removal may be eective.
Epidermoid cysts
These intracranial cysts are rare and are
thought to arise from ectopic ectoderm that was
misplaced during the time of neural tube closure.
Favorite sites are the cerebellopontine angle and
the chiasmal region, but they also occur within the
cerebral hemispheres and the ventricles [59]. They
may lie within the brain parenchyma adjacent to
or in communication with the ventricles [59,60].
There are reports of intrathecal seeding along the
CSF pathways as a result of rupture of the cyst
into the ventricle and subsequent scattering of oily
contents into the CSF space [61,62].
Dandy-Walker malformation
The typical triad of Dandy-Walker syndrome
is agenesis or hypoplasia of the cerebellar vermis,
cystic dilatation of the fourth ventricle, and
supratentorial hydrocephalus (Fig. 7). Most cases
are sporadic, but there is evidence of a chromosomal disorder in some reports involving
chromosomes 9p and 12p as well as the X-chromosome [63,64] Associated ndings are cranial
and anterior rotation of the cerebellar vermis as
614
Fig. 6. Pineal cyst. (A) Unenhanced CT scan shows a low-density lesion with thick rim calcications in the region of the
pineal gland. Coronal T2-weighted image (B) and sagittal T1-weighted postcontrast image (C) conrm the presence of
a nonenhancing cyst within the pineal gland. Note normal appearance of the third ventricle and lack of calcication
visualization on the MRI scan.
well as cranial displacement of the tentorium

cerebelli, falx cerebelli, torcular herophili, and
transverse sinuses. There is no communication of
the cystic CSF collection with the basal subarachnoid space.
Most pediatric patients present early within the
rst year of life with increased head circumference
or symptoms caused by hydrocephalus. The
incidence of cerebellar signs or cranial nerve

dysfunction is unexpectedly low. Delayed psychomotor development may be present when the
syndrome is diagnosed after the rst year of life.
In up to 68% of patients, other CNS abnormalities can be found, including agenesis of the
corpus callosum, occipital encephalocele or meningocele, aqueductal stenosis, CSP, choroid
Fig. 7. Dandy-Walker malformation. Sagittal T2weighted image demonstrates the classic features of
a Dandy-Walker malformation. The vermis is severely
hypogenetic (arrow). The fourth ventricle is open,
communicating with a large retrocerebellar cyst. The
posterior fossa is enlarged with torcular-lambdoidal
inversion. Note the small occipital meningocele, a common association with Dandy-Walker malformations.
plexus cysts, infundibular hamartomas, macrogyria, microcephaly, syringomyelia, posterior fossa

lipoma, and tuberous sclerosis. In addition, systemic abnormalities are found in one fourth of the
cases, especially cardiac defects (eg, atrial and
ventricular septal defects, patent ductus arteriosus,
tetralogy of Fallot), but facial, skeletal, gastrointestinal, and urogenital defects can also be present.
The terminology of posterior fossa cystic lesions
has been confused by the application of dierent
terms and classications. The Dandy-Walker complex includes the Dandy-Walker malformation, the
Dandy-Walker variant, and the mega-cisterna
magna. The persistent Blakes pouch cyst has been
added by Tortori-Donati and colleagues [65].
In the Dandy-Walker variant, the anterior
rotation of the hypoplastic cerebellum is not as
prominent and the roof as well as the lateral walls
of the fourth ventricle can be recognized. The
Dandy-Walker syndrome and the Dandy-Walker
variant are thought to originate from a failure of
assimilation of the area membranacea anterior
within the tela choroidea of the fourth ventricle.
The foramen of Magendie is usually patent;
however, it may open in a delayed fashion. The
persistent Blakes pouch is dened by failed
regression of the embryonal pouch secondary
615
to nonperforation of the foramen of Magendie.

Physiologically, the lateral foramina of Luschka
open later than the foramen of Magendie, and the
fourth ventricle tends to enlarge during the rst
stage of development when the latter is not patent.
The cerebellar hemispheres and vermis appear
hypoplastic but re-expand after shunting, thus
indicating mainly compression of the cerebellar
structures and not hypoplasia [66]. The aqueduct
is not aected.
The mega-cisterna magna is characterized by
an enlargement of the cisterna magna with free
communication with the basal subarachnoid space
and the fourth ventricle. There has been a considerable debate in the past about the optimal
treatment of posterior fossa cystic lesions in cases
associated with supratentorial hydrocephalus. The
question, whether a ventriculo-peritoneal (VP)
shunt or a cysto-peritoneal (CP) shunt, or
a combination of both, would be superior to the
other techniques remains unanswered. There is
evidence that after shunting of the posterior fossa
cyst, the normal CSF pathway is not always
restored, most likely because of secondary aqueductal stenosis. A study performed by Asai [67]
and colleagues demonstrated that 9 of 21 patients
with initial VP shunts subsequently required CP
shunts, whereas only 1 of 10 patients with a CP
shunt as the rst treatment needed a VP shunt
thereafter. Conversely, CP shunts carry a higher
risk for secondary brain stem injury, and the
occlusion/malfunction rate is higher [68].
Intraventricular congenital lesions
Lipoma of the corpus callosum
Intracranial lipomas are considered to be
benign developmental anomalies; they mainly
arise in the interhemispheric ssure and only rarely
in the interpeduncular, chiasmatic, Sylvian, quadrigeminal, or ambient cistern. Interhemispheric
lipomas can be associated with lipomas of the
choroid plexus and aplasia of the corpus callosum,
anomalies of the cingulate gyrus and the septum
pellucidum, and other dysraphic states (Fig. 8)
[6972]. These lesions rarely become symptomatic.
Conservative therapy is recommended.
Tuberous sclerosis
The tuberous sclerosis complex is an autosomal dominant inherited disease, but most cases
are sporadic. Lesions occurring in this context
involve all three primary germ cell layers and are
616
Fig. 8. Interhemispheric lipoma. (A) Sagittal T1-weighted image shows a hyperintense lesion surrounding a minimally
hypogenetic corpus callosum (absent splenium). (B) Coronal T1-weighted postcontrast image demonstrates
intraventricular extension (arrow), a common nding.
classied as hamartias, hamartomas, or benign

tumors. Typical skin lesions are small hypopigmented areas known as ash leaf spots, facial
angiobromas, subungual bromas, and shagreen
patches. Other features include cardiac rhabdomyomas, pulmonary lymphangiomatosis, renal
angiomyolipomas, renal cysts, and retinal hamartomas. The most common lesions in the CNS are
cortical tubers with seizures, either partial or
infantile spasms, as the main presenting symptom.
Subependymal nodules along the surface of the
lateral ventricles sometimes resembling candle
gutterings are the lesions that relate to the topic
reviewed here (Fig. 9). They are most commonly
situated in the caudothalamic groove, have a predilection for the foramen of Monro, and may
cause obstructive hydrocephalus. They are often
calcied, and their number increases gradually
with age until the age of 10 years. The study of
Hosoya and colleagues [73] clearly demonstrates
that patients with ve or more subependymal
nodules have a signicantly greater number of
cortical tubers and white matter lesions as well as
a signicantly higher percentage of infantile
spasms or mental retardation. If these nodules
show a tendency to grow, they likely represent
subependymal giant cell astrocytomas (SEGAs)
(see Fig. 9).
Two distinct genetic loci have been identied in
association with the tuberous sclerosis complex.
TSC1 is located on chromosome 9q, and TSC2
resides on chromosome 16p. Both loci seem to
Fig. 9. Tuberous sclerosis. Axial T1-weighted postcontrast image demonstrates multiple enhancing and nonenhancing subependymal tubers. The large enhancing
lesion in the vicinity of the foramen of Monro (black
arrow) is consistent with a giant cell tumor. Mild
enlargement of the left lateral ventricle may be related
to intermittent obstruction of the foramen of Monro.
Note multiple periventricular parenchymal cysts (white
arrows), not uncommonly identied in tuberous sclerosis.
possess tumor suppressor gene activity. Their

exact role in tuber development is not completely
understood, but it is speculated that progenitor
cells on the germinal layer of the ventricular zone,
which are destined for the cortex, undergo
inactivation of the TSC1 or TSC2 gene. Those
immature neuroepithelial cells that carry the
double-hit mutations on either locus are believed
to proliferate, migrate, and dierentiate abnormally [74]. It is speculated that abnormal cerebral
function results from improper synaptic delivery
of a neurotransmitter in the mutated cell. There
are single case reports of intraventricular hamartomas that required surgical removal or CSF
shunting because they caused obstructive hydrocephalus [75,76]. Hamartomas on the oor of the
third ventricle are known to cause gelastic
epilepsy.
[7]
[8]
[9]
[10]
[11]
Summary
Intraventricular congenital lesions and colloid
cysts comprise a rather large spectrum of dierent
pathologic conditions. In most cases, treatment in
not warranted unless there is progressive ventricular obstruction with hydrocephalus or growth of
the lesion itself, making tissue biopsy and
histopathologic diagnosis necessary. Accordingly,
a precise neuroradiologic evaluation is of the
utmost importance, because most lesions, if not
symptomatic, only require clinical and radiologic
follow-up.
References
[1] Shaw CM, Alvord EC Jr. Cava septi pellucidi et
vergae: their normal and pathological states. Brain
1969;92(1):21323.
[2] Nakajima Y, Yano S, Kuramatsu T, Ichihashi K,
Miyao M, Yanagisawa M, et al. Ultrasonographic
evaluation of cavum septi pellucidi and cavum
vergae. Brain Dev 1986;8(5):5058.
[3] Kier LE. The evolutionary and embryologic basis
for the development and anatomy of the cavum veli
interpositi. AJNR Am J Neuroradiol 2000;21(3):
61214.
[4] Gangemi M, Donati P, Maiuri F, Sigona L. Cyst of
the velum interpositum treated by endoscopic fenestration. Surg Neurol 1997;47(2):1346.
[5] Gangemi M, Maiuri F, Godano U, Mascari C,
Longatti PL, Marzucco M. Endoscopic treatment
of para- and intraventricular cerebrospinal uid
cysts. Minim Invasive Neurosurg 2000;43(3):1538.
[6] Barth PG, Uylings HB, Stam FC. Interhemispheral
neuroepithelial (glio-ependymal) cysts, associated
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
617
with agenesis of the corpus callosum and neocortical maldevelopment. A case study. Childs Brain
1984;11(5):3129.
Mori K. Giant interhemispheric cysts associated
with agenesis of the corpus callosum. J Neurosurg
1992;76(2):22430.
Raybaud C, Girard N. Anatomic MRI study of
commissural agenesis and dysplasia of the telencephalon (agenesis of the corpus callosum and related
anomalies). Clinical correlations and morphogenetic interpretation. Neurochirurgie 1998;44(1 Suppl):
3860.
Barkovich AJ, Simon EM, Walsh CA. Callosal
agenesis with cyst: a better understanding and new
classication. Neurology 2001;56(2):2207.
Desai KI, Nadkarni TD, Muzumdar DP, Goel AH.
Surgical management of colloid cyst of the third
ventriclea study of 105 cases. Surg Neurol 2002;
57(5):295302.
Buttner A, Winkler PA, Eisenmenger W, Weis S.
Colloid cysts of the third ventricle with fatal outcome: a report of two cases and review of the
literature. Int J Legal Med 1997;110(5):2606.
Wiese JA, Gentry LR, Menezes AH. Bobble-head
doll syndrome: review of the pathophysiology and
CSF dynamics. Pediatr Neurol 1985;1(6):3616.
Maeder PP, Holtas SL, Basibuyuk LN, Salford LG,
Tapper UA, Brun A. Colloid cysts of the third
ventricle: correlation of MR and CT ndings with
histology and chemical analysis. AJR Am J Roentgenol 1990;155(1):13541.
Hopf NJ, Perneczky A. Endoscopic neurosurgery
and endoscope-assisted microneurosurgery for the
treatment of intracranial cysts. Neurosurgery 1998;
43(6):13306.
Di Rocco C, Di Trapani G, Iannelli A. Arachnoid
cyst of the fourth ventricle and arrested hydrocephalus. Surg Neurol 1979;12(6):46771.
Korosue K, Tamaki N, Fujiwara K, Matsumoto S.
Arachnoid cyst of the fourth ventricle manifesting
normal pressure hydrocephalus. Neurosurgery 1983;
12(1):10810.
Nakase H, Hisanaga M, Hashimoto S, Imanishi M,
Utsumi S. Intraventricular arachnoid cyst. Report
of two cases. J Neurosurg 1988;68(3):4826.
Yeates A, Enzmann D. An intraventricular arachnoid cyst. J Comput Assist Tomogr 1979;3(5):697
700.
Kurokawa Y, Sohma T, Tsuchita H, Kitami K,
Suzuki S, Ishikawa A. A case of intraventricular
arachnoid cyst. How should it be treated? Childs
Nerv Syst 1990;6(6):3657.
Martinez-Lage JF, Poza M, Sola J, Puche A.
Congenital arachnoid cyst of the lateral ventricles
in children. Childs Nerv Syst 1992;8(4):2036.
Nakase H, Ishida Y, Tada T, Sakaki T, Goda K,
Tunoda S, et al. Neuroepithelial cyst of the lateral
ventricle. Clinical features and treatment. Surg
Neurol 1992;37(2):94100.
618
[22] Pelletier J, Milandre L, Peragut JC, Cronqvist S.

Intraventricular choroid plexus arachnoid cyst.
MRI ndings. Neuroradiology 1990;32(6):5235.
[23] Turgut M, Ozcan OE, Onol B. Case report and
review of the literature: arachnoid cyst of the fourth
ventricle presenting as a syndrome of normal pressure hydrocephalus. J Neurosurg Sci 1992;36(1):
557.
[24] Okamura K, Watanabe M, Inoue N, Kanoh M,
Ohno T, Mitsui Y, et al. Intraventricular arachnoid
cyston the origin of intraventricular arachnoid
cysts. No To Shinkei 1996;48(11):101521.
[25] Maiuri F, Iaconetta G, Gangemi M. Arachnoid cyst
of the lateral ventricle. Surg Neurol 1997;48(4):
4014.
[26] Heilman CB. Transient choroid plexus cysts and
benign asymmetrical ventricles: a case suggesting
a possible link: case report. Neurosurgery 2003;
52(1):2135.
[27] Geary M, Patel S, Lamont R. Isolated choroid
plexus cysts and association with fetal aneuploidy
in an unselected population. Ultrasound Obstet
Gynecol 1997;10(3):1713.
[28] Shangshotti S, Metsky M. Neuroepithelial cysts of
the nervous system. Am J Anat 1966;16:887903.
[29] Odake G, Tenjin H, Murakami N. Cyst of the
choroid plexus in the lateral ventricle: case report
and review of the literature. Neurosurgery 1990;
27(3):4706.
[30] Ismail A, Tampieri D, Melanson D, Pokrupa R,
Villemure JG, Bertrand G. Glioependymal cysts:
CT and MR ndings. J Comput Assist Tomogr
1992;16(6):8604.
[31] Pawar SJ, Sharma RR, Mahapatra AK, Dev EJ.
Giant ependymal cyst of the temporal hornan
unusual presentation. Case report with review of the
literature. Pediatr Neurosurg 2001;34(6):30610.
[32] Friede RL, Yasargil MG. Supratentorial intracerebral epithelial (ependymal) cysts: review, case
reports, and ne structure. J Neurol Neurosurg
Psychiatry 1977;40(2):12737.
[33] Haddad FS, Abla A, Allam C. Ependymal brain
cyst. Surg Neurol 1982;18(4):2469.
[34] Pant B, Uozumi T, Hirohata T, Arita K, Kurisu K,
Nakahara T, et al. Endoscopic resection of intraventricular ependymal cyst presenting with psychosis. Surg Neurol 1996;46(6):5736.
[35] Hasegawa A, Ohtsubo K, Mori W. Pineal gland in
old age; quantitative and qualitative morphological
study of 168 human autopsy cases. Brain Res
1987;409(2):3439.
[36] Mena H, Armonda RA, Ribas JL, Ondra SL,
Rushing EJ. Nonneoplastic pineal cysts: a clinicopathologic study of twenty-one cases. Ann Diagn
Pathol 1997;1(1):118.
[37] Miyatake S, Kikuchi H, Yamasaki T, Terashima T,
Asahi M, Asato R, et al. Glial cyst of the pineal
gland with characteristic computed tomography,
magnetic resonance imaging, and pathological
[38]
[39]
[40]
[41]
[42]
[43]
[44]
[45]
[46]
[47]
[48]
[49]
[50]
[51]
[52]
[53]
ndings: report of two cases. Surg Neurol 1992;

37(4):2939.
Mukherjee KK, Banerji D, Sharma R. Pineal cyst
presenting with intracystic and subarachnoid haemorrhage: report of a case and review of the literature. Br J Neurosurg 1999;13(2):18992.
Barboriak DP, Lee L, Provenzale JM. Serial MR
imaging of pineal cysts: implications for natural
history and follow-up. AJR Am J Roentgenol 2001;
176(3):73743.
Petitcolin V, Garcier JM, Mohammedi R, Ravel A,
Mod R, Viallet JF, et al. Prevalence and morphology of pineal cysts discovered at pituitary MRI:
review of 1844 examinations. J Radiol 2002;83
(2 Part 1):1415.
Konovalov AN, Pitskhelauri DI. Infratentorial
supracerebellar approach to the colloid cysts of the
third ventricle. Neurosurgery 2001;49(5):111622.
Hancq S, De Witte O, Brotchi J. Pineal region
surgery. Experience in 22 patients. Neurochirurgie
2002;48(1):1424.
Fain JS, Tomlinson FH, Scheithauer BW, Parisi JE,
Fletcher GP, Kelly PJ, et al. Symptomatic glial cysts
of the pineal gland. J Neurosurg 1994;80(3):45460.
Oeckler R, Feiden W. Benign symptomatic lesions of
the pineal gland. Report of seven cases treated surgically. Acta Neurochir (Wien) 1991;108(12):404.
Stein BM. The infratentorial supracerebellar approach to pineal lesions. J Neurosurg 1971;35(2):
197202.
Wiso JH, Epstein F. Surgical management of
symptomatic pineal cysts. J Neurosurg 1992;77(6):
896900.
Kreth FW, Schatz CR, Pagenstecher A, Faist M,
Volk B, Ostertag CB. Stereotactic management of
lesions of the pineal region. Neurosurgery 1996;
39(2):2809.
Stern JD, Ross DA. Stereotactic management of
benign pineal region cysts: report of two cases.
Vallee B, Meriot P, Dam HP, Person H. Benign
hemorrhagic cyst of the pineal region. Treatment by
stereotactic puncture. 4-year follow-up. Neurochirurgie 1997;43(5):3037.
Vallee B, Meriot P, Person H, Rodriguez V,
Sidhamed S, Dam HP. Benign glial cysts of the
pineal region. Neurochirurgie 1997;43(5):299302.
Michielsen G, Benoit Y, Baert E, Meire F,
Caemaert J. Symptomatic pineal cysts: clinical manifestations and management. Acta Neurochir (Wien)
2002;144(3):23342.
Agrawal D, Suri A, Mahapatra AK, Sharma MC.
Intramedullary neurenteric cyst presenting as infantile paraplegia: a case and review. Pediatr
Neurosurg 2002;37(2):936.
Rao MB, Rout D, Misra BK, Radhakrishnan VV.
Craniospinal and spinal enterogenous cystsreport
of three cases. Clin Neurol Neurosurg 1996;98(1):
326.

[54] Hirai O, Kawamura J, Fukumitsu T. Prepontine
epithelium-lined cyst. Case report. J Neurosurg 1981;
55(2):3127.
[55] Malcolm GP, Symon L, Kendall B, Pires M.
Intracranial neurenteric cysts. Report of two cases.
J Neurosurg 1991;75(1):11520.
[56] Fuse T, Yamada K, Kamiya K, Inagaki H.
Neurenteric cyst at the craniovertebral junction:
report of two cases. Surg Neurol 1998;50(5):4316.
[57] Inoue T, Matsushima T, Fukui M, Iwaki T,
Takeshita I, Kuromatsu C. Immunohistochemical
study of intracranial cysts. Neurosurgery 1988;23(5):
57681.
[58] Afshar F, Scholtz CL. Enterogenous cyst of the
fourth ventricle: case report. J Neurosurg 1981;54(6):
8368.
[59] Gelabert-Gonzalez M, Garcia-Allut A, GonzalezGarcia J, Martinez-Rumbo R. Epidermoid cyst of
the third ventricle. Neurocirugia 2002;13(5):38992.
[60] Tosaka M, Oya N, Kobayashi S, Kamagata M,
Kohga H, Sasaki T. Pineal epidermoid cyst
visualized by diusion-weighted magnetic resonance imaging. Acta Neurochir (Wien) 2001;143(2):
2056.
[61] Miyagi Y, Suzuki SO, Iwaki T, Ishido K, Araki T,
Kamikaseda K. Magnetic resonance appearance of
multiple intracranial epidermoid cysts: intrathecal
seeding of the cysts? Case report. J Neurosurg
2000;92(4):7114.
[62] el Quessar A, Chakir N, Bouyaakoub F, el Hassani
MR, Jiddane M, Boukhrissi N. Spontaneous
rupture of an intracerebral dermoid cyst. Ann
Radiol (Paris) 1996;39(6):2536.
[63] Chen CP, Chang TY, Shih JC, Lin SP, Lin CJ,
Wang W, et al. Prenatal diagnosis of the DandyWalker malformation and ventriculomegaly associated with partial trisomy 9p and distal 12p deletion.
Prenat Diagn 2002;22(12):10636.
[64] Wakeling EL, Jolly M, Fisk NM, Gannon C,
Holder SE. X-linked inheritance of Dandy-Walker
variant. Clin Dysmorphol 2002;11(1):158.
[65] Tortori-Donati P, Fondelli MP, Rossi A, Carini S.
Cystic malformations of the posterior cranial fossa
originating from a defect of the posterior membranous area. Mega cisterna magna and persisting
[66]
[67]
[68]
[69]
[70]
[71]
[72]
[73]
[74]
[75]
[76]
619
Blakes pouch: two separate entities. Childs Nerv

Syst 1996;12(6):3038.
Calabro F, Arcuri T, Jinkins JR. Blakes pouch
cyst: an entity within the Dandy-Walker continuum. Neuroradiology 2000;42(4):2905.
Asai A, Homan HJ, Hendrick EB, Humphreys
RP. Dandy-Walker Syndrome: experience at the
Hospital for Sick Children, Toronto. Pediatr
Neurosci 1989;15:6673.
Gupta PK, Mann KS. Shunt migration into
a Dandy Walker cyst. Neurol India 2000;48(1):
8990.
Kieslich M, Ehlers S, Bollinger M, Jacobi G.
Midline developmental anomalies with lipomas in
the corpus callosum region. J Child Neurol
2000;15(2):859.
Ickowitz V, Eurin D, Rypens F, Sonigo P, Simon I,
David P, et al. Prenatal diagnosis and postnatal
follow-up of pericallosal lipoma: report of seven
new cases. AJNR Am J Neuroradiol 2001;22(4):
76772.
Kawamura T, Nishio S, Morioka T, Fukui K.
Callosal anomalies in patients with spinal dysraphism: correlation of clinical and neuroimaging
features with hemispheric abnormalities. Neurol
Res 2002;24(5):4637.
Kim TH, Joh JH, Kim MY, Kim YM, Han KS.
Fetal pericallosal lipoma: US and MR ndings.
Korean J Radiol 2002;3(2):1403.
Hosoya M, Naito H, Nihei K. Neurological
prognosis correlated with variations over time in
the number of subependymal nodules in tuberous
sclerosis. Brain Dev 1999;21(8):5447.
Yeung RS. Tuberous sclerosis as an underlying
basis for infantile spasm. Int Rev Neurobiol
2002;49:31532.
Sasaki T, Matsuno A, Inoh Y, Asai A, Kirino T. A
rare case of hamartoma in the lateral ventricle: case
report. Surg Neurol 1997;47(1):237.
Rossiter JP, Khalifa MM, Nag S. Diencephalic
neuronal hamartoma associated with congenital
obstructive hydrocephalus, anophthalmia, cleft lip
and palate and severe mental retardation: a possible
new syndrome. Acta Neuropathol (Berl) 2000;
99(6):68590.
Choroid plexus tumors in children

Nalin Gupta, MD, PhD
Division of Pediatric Neurosurgery, Department of Neurological Surgery, University of California at
San Francisco, 505 Parnassus Avenue, M-779, San Francisco, CA 94143, USA
The choroid plexus is anatomically located

within each of the four ventricles at the parenchymal/ventricular junction and is derived from the
specialization of ventricular ependymal cells along
certain segments of the neural tube. The roof of
the neural tube in areas of the forebrain and hindbrain expands into thin sheets of epithelium with
an underlying loose mesenchymal stroma. The
stroma consists of leptomeningeal cells, blood
vessels, and connective tissue. The epithelial layer
consists of cuboidal cells linked by tight junctions,
thereby creating a bloodcerebrospinal uid
(CSF) barrier. Although the evolutionary origin
of the choroid plexus is unknown, it is the source
of most CSF. In lower mammals, there is a welldeveloped lymphatic system and the volume of
CSF produced is small. The loss of a lymphatic
system in the brain may have led to the corresponding increase in the volume of CSF produced
in primates.
As with other tissues of the central nervous
system, benign and malignant tumors can arise
from the choroid plexus. It is assumed that
malignant tumors arise de novo, with conversion
from a benign to malignant phenotype occurring
rarely [1]. Guerard was the rst to describe
a choroid plexus tumor in 1833. The rst surgical
resection was reported by Bielschowsky and
Unger in 1906. Thereafter, both Davis and
Cushing [2] and Dandy [3] reported their experiences with this unusual tumor.
E-mail address: guptan@neurosurg.ucsf.edu

(N. Gupta).
Epidemiology
Choroid plexus papillomas (CPPs) and their
malignant counterpart, choroid plexus carcinomas
(CPCs), are uncommon, comprising only 0.5% to
0.6% of all brain tumors. They are encountered in
all age groups but are mainly seen in childhood.
There is a particular preponderance in infants and
extremely young children. In his review of all
published cases before 1974, Laurence [4] reported
that 45% presented in the rst year of life, whereas
74% occurred in the rst decade. Wol and
colleagues [5] studied all available cases reported
in the literature until 1998 and performed a detailed
meta-analysis. This report noted a male-to-female
ratio of 1.2:1, a median age at diagnosis of 3.5.
years, and a striking dierence in tumor location
and age. Supratentorial tumors (lateral and third
ventricles) occurred mostly in infants, with a median age of 1.5 years at diagnosis in this group. By
comparison, the median ages at diagnosis of
tumors in the fourth ventricle and cerebellopontine
angle (CPA) were 22.5 and 35.5 years, respectively.
Laurence [4] did note that 50% of cases reviewed
were situated in the lateral ventricles, 37% in the
fourth ventricle, 9% in the third ventricle, and the
remainder in other locations.
As expected, reviews from pediatric centers
report that a higher percentage (1.8%2.9%) of
their cases are choroid plexus tumors [68]. Two
reviews of tumors in the rst year of life by
Galassi et al [9] and Haddad et al [10] found that
choroid plexus tumors comprised 14% and 12.8%
of all cases, respectively. Most series noted
previously have not reported any predilection
for right or left ventricle or sex. CPCs comprise
29% to 39% of all choroid neoplasms [8,11,12].
doi:10.1016/S1042-3680(03)00058-5
622
N. Gupta / Neurosurg Clin N Am 14 (2003) 621631
Pathology
Gross appearance and histopathology
Typically, most astrocytic tumors are soft and
poorly cohesive masses. In contrast, CPPs recapitulate the normal appearance of the choroid
plexus with well-dierentiated secondary structures, resulting in a cauliower-like appearance
(Fig. 1). The surface of the tumor is similar to the
soft fronds of normal choroid, with the overall
shape being roughly globular. The stroma can
be brous and tough in consistency. Evidence of
previous hemorrhage is sometimes apparent.
Because papillomas are benign, they tend to
expand the ventricle rather than invade adjacent
brain. Nevertheless, the proximity of these tumors
to deep-seated structures, such as the internal
cerebral veins and limbic structures, can make their
removal dicult.
The microscopic appearance of papilloma is
also similar to that of the normal choroid plexus.
There are many papillae covered with a simple
cuboidal or columnar epithelia. The stroma of
these brovascular structures is composed of
connective tissue and small blood vessels. The
presence of the connective tissue stroma is notable
mainly because it allows one to distinguish
between CPP and papillary forms of ependymoma
(whose stroma is composed of brillary neuroglia).
Fig. 1. A large choroid plexus papilloma located within

the lateral ventricle. A sagittal, precontrast, T1-weighted
image illustrates the gross architecture of the tumor. The
mass consists of many small fronds of tissue surrounding
a central mesenchymal stroma.
In addition, choroid epithelial cells do not contain

cilia or blepharoplasts as do ependymal cells.
Mitotic gures are rare. Unusual pathologic
features, such as mucinous degeneration, tubular
dierentiation, and formation of structures like
bone, are sometimes observed [13].
Grossly, CPCs tend to be softer and more
friable than papillomas, but it can be dicult to
dierentiate the grade of the neoplasm based only
on the gross appearance. Although carcinomas rarely metastasize from the intracranial or
intraspinal compartment, they can disseminate
throughout the CSF pathways [14]. The denitive
diagnosis is usually based on routine histopathology. There are three major features that have been
associated with malignancy. First are cytologic
criteria of malignancy, such as nuclear atypia,
increased nuclear-to-cytoplasmic ratio, and prominent mitotic gures. Second is the loss of the
normal papillary architecture. Third is the presence of brain invasion by the tumor, which usually
involves extension through the ependymal lining.
The signicance of stromal invasion has been
questioned in a recent report, where three of four
patients with CPP and invasion after gross total
resection (GTR) were cured without the need for
adjunctive therapy [15].
Occasionally, diagnostic confusion can arise in
the setting of a tumor with epithelial features. The
epithelial nature of the frank malignancy can
create confusion, because other tumors, such as
metastatic adenocarcinoma and medulloepithelioma, are histologically similar. If the tumor
arises in a young patient, the chance of the tumor
being metastatic is extremely low. Electron
microscopy can reveal details, such as cilia, that
are not normally present in choroid plexus
tumors. Rarely, if a tumor demonstrates some
atypical features without evidence of invasion,
it can be classied as an atypical papilloma.
Villous hypertrophy of the choroid plexus is
a poorly dened entity. Characteristically, the
choroid plexus of both lateral ventricles is enlarged and is associated with hydrocephalus
from birth. The hydrocephalus is related to
hyperactivity of the choroid, whereas the cytologic appearance of the tissue is normal. Villous
hypertrophy has been used synonymously with
bilateral CPP, but this is not accurate in the
strictest sense if histologic evidence of neoplastic
growth is not present and expansion of the
choroid plexus is occurring diusely [16]. Finally,
a recent report highlighted the presence of
rhabdoid cells within CPCs [17]. These cells are
more typically encountered in atypical teratoid

rhabdoid tumors (ATRTs), with electron microscopy required to distinguish between the two
diagnoses.
Usually, the diagnosis is clear from standard
hematoxylin-eosin stains, whereas immunohistochemistry is often inconclusive. The calciumbinding protein S-100 is positive in most choroid
tumors [18,19]. This is of limited diagnostic value,
because glial tissues and normal choroid express
S-100 in a parallel fashion with glial brillary acid
protein (GFAP). Other markers, such as vimentin,
GFAP, and cytokeratins, can be positive but also
lack specicity [20,21]. Prealbumin or transthyretin (TTR) was initially believed to be a specic
marker, but another report noted that 20% of
choroid tumors were TTR-negative [22,23]. These
investigators did nd that prognostic information could be gleaned from immunohistochemical
data. A poor prognosis was found in those tumors
when less than 50% of the cells in a given tumor
were heavily stained for S-100. In addition, the
absence of TTR-positive cells also correlated with
a poor prognosis. Cellular proliferation as measured by Ki67/MIB1 labeling is low in papillomas
and signicantly higher for carcinomas [17,24].
Genetics and molecular biology
Choroid plexus tumors arise sporadically without a known cause. Aside from one report that
described two cases occurring in one family,
a genetic predisposition is not present for most
cases [25]. Some experimental evidence links SV40,
a primate DNA virus, with choroid plexus tumor
formation. As with most viruses, a portion of the
virus genome expresses proteins used to subvert
normal cell functions and to interfere with growth
regulatory pathways. When expressed in mice,
large T antigen, the major regulator of late viral
gene products of the SV40 virus, induces the
formation of choroid plexus neoplasms [26]. The
large T antigen is expressed only in the choroid
plexus and seems to interact with the product of
the p53 gene [27]. T antigen interferes with the
function of both p53 and retinoblastoma (RB)
tumor suppressor proteins. This function is likely
required for viral replication and growth but may
also have the eect of increasing tumor formation.
Using polymerase chain reaction (PCR), SV40
DNA sequences were demonstrated in 50% of
choroid plexus tumors and most ependymomas
623
[28]. Active T-antigen and p53 complexes have

also been demonstrated in brain tumors [29]. More
recently, the expression of transgenes of the viral
oncoproteins E6 and E7 from human papilloma
virus has also been shown to produce tumors in
71% of ospring, of which 26% of the tumors
were choroid plexus tumors [30].
A subset of central primitive neuroectodermal
tumors (PNETs), CPCs, and medulloblastoma
were recently shown to have frequent mutations
in the hSNF2/INI1 gene, which encodes for a
component of the adenosine triphosphate (ATP)
dependent chromatin remodeling complex [31].
The same authors proposed that constitutional
mutations in this gene lead to a greater incidence
of renal and extrarenal malignant rhabdoid
tumors, CPCs, central PNETs, and medulloblastoma, a complex they have coined as the rhabdoid
predisposition syndrome [32]. The penetrance of
the disease is high, with many probands developing
malignant tumors before 3 years of age. Some
pathologic data also suggest a connection between
malignant rhabdoid tumors and CPCs [17].
A number of chromosomal abnormalities have
been identied in CPP and CPC [33]. Surprisingly,
even benign CPP (32 of 34 cases) demonstrated
chromosomal aberrations. The patterns of aberrations in CPP dier from those observed in CPC.
Wyatt-Ashmead and colleagues [17] reported the
presence of deletions involving either an entire copy
(3 of 4 cases) or a portion of chromosome 22 (1 of 4
cases) among a group of 5 cases of CPC studied.
Clinical features
For most patients with a choroid plexus tumor,
hydrocephalus is responsible for the presenting
symptoms. It is caused by overproduction of CSF
and, in certain cases, the obstruction of CSF
pathways, although it seems that overproduction
is the major contributing factor (Fig. 2) [34].
Resolution of hydrocephalus has been reported
after complete tumor removal, suggesting that
CSF hypersecretion was responsible for ventriculomegaly [3537]. Variations are likely to exist,
because a normal rate of CSF production has been
reported in a patient harboring a papilloma [38].
The most common presentation of choroid
plexus neoplasms is related to increased intracranial hypertension secondary to obstructive hydrocephalus or CSF overproduction [4,8,39]. Because
most cases occur in infants and young children,
there are characteristic features of raised
624
Fig. 2. An axial T2-weighted image demonstrates severe

ventriculomegaly in a patient with a lateral ventricle
choroid plexus papilloma. The entire ventricular system
is enlarged, suggesting that hydrocephalus is primarily
caused by cerebrospinal uid (CSF) overproduction.
Transependymal absorption of CSF is visible at the
margins of the ventricles.
intracranial pressure (ICP). Ellenbogen et al [8]

described the usual presenting signs and symptoms. The most common symptoms were nausea/
vomiting, irritability, headache, visual diculty,
and seizure. As expected, the most common signs
were craniomegaly, papilledema, and decreased
level of consciousness. The duration of symptoms
reported in this series varied from 2 months in
those patients younger than 2 years of age to 6
months on average in those patients older than 2
years. Although choroid plexus neoplasms are
viewed as slow-growing tumors, the presence of
stupor or coma as the presenting sign in 25% of
children suggests a more acute clinical course in
some patients. Rapid decompensation can occur
either from massive hydrocephalus or from tumoral hemorrhage. Of 21 patients who had CSF
examined, 2 were found to have grossly bloody
uid. Lateralizing signs are found in a few patients
and are usually related to asymmetric ventricular
dilatation. Hydrocephalus was present in 78% of
cases at the Hospital for Sick Children and in 95%
of cases at the Childrens Hospital in Boston [8,39].
Neuroimaging
The CT scan appearance of CPP is often
characteristic for this tumor. The mass is well
demarcated from the brain, lobulated, and often

has punctate calcication. These tumors enhance
homogeneously and intensely after contrast,
reecting a luxuriant blood supply [40]. An
enlarged choroidal artery leading into the tumor
mass can sometimes be seen in postcontrast
images. At times, the massive size of these lesions
may obscure the exact site of origin. Some
carcinomas display a diuse border between
tumor and normal brain, which may reect areas
of brain invasion. On the basis of CT, certain
features distinguish a suspected choroid plexus
tumor from other possibilities. Cerebellar astrocytomas tend to be less homogeneously staining
and often have cystic areas. Medulloblastomas
are characterized by a more heterogeneous appearance, although they also stain vividly with
contrast and may cause confusion with a fourth
ventricular choroid papilloma. Ependymomas
arise physically in similar locations but tend to
enhance inhomogeneously. Finally, meningiomas
can occur in an intraventricular location, but
these tumors are rarely encountered in children.
Papillomas are isodense to brain on T1weighted images (see Fig. 1). Areas of high signal
indicate hemorrhage necrosis. After gadolinium
administration, the tumor enhances brightly
(Fig. 3A), although this can be patchy in nature,
reecting areas of high ow. T2-weighted images
demonstrate an intermediate to high signal intensity with areas of heterogeneous internal signal
(see Fig. 3B) [41]. MR spectroscopy of CPP and
CPC is characterized by a prominent choline peak
and absence of N-acetyl aspartate [42]. As with
CT, an enlarged choroidal artery is often noted,
especially with larger tumors. The vascularity of
these tumors is easily demonstrated with specic
perfusion sequences. MR angiography can sometimes help with dening the location of feeding
vessels (see Fig. 3C). With CPC, the boundary
between the tumor and surrounding brain can be
indistinct in areas, but this is not a universal
nding.
Treatment
Overall objectives
GTR is associated with the most favorable
outcome in patients with choroid plexus tumors
[5,43]. Therefore, any treatment plan should be
tailored to reach this objective. The following steps are likely to be encountered as part of
the surgical plan: (1) temporary or permanent
625
Fig. 3. An 18-month-old girl with a large left ventricular choroid plexus papilloma. (A) T1-weighted image after
injection of gadolinium. (B) T2-weighted axial image. (C) MR angiogram demonstrating a choroidal artery arising from
the posterior cerebral artery leading to the tumor. (D) A postoperative, coronal, T1-weighted image demonstrating the
surgical route clearly shows an enlarged choroidal artery leading into the tumor. Postoperative coronal MRI
demonstrates the route through the temporal lobe used to access the feeding artery initially and then the actual tumor
itself.
resolution of hydrocephalus, (2) denition of the

anatomic location and blood supply, and (3)
planning of the surgical approach to allow maximal exposure and resection.
Management of hydrocephalus
Because most patients present with symptoms
of intracranial hypertension secondary to hydrocephalus, the initial step is to decide whether CSF
drainage is required. In infants less than 8 to 10
months of age, the presence of open sutures allows

compensation by slow enlargement of the head,
and treatment of hydrocephalus can often wait
until the planned surgical date. If rapid neurologic
deterioration is occurring, an external ventricular
drain should be inserted immediately, regardless
of the age of the patient. Placement of a ventriculo-peritoneal (VP) shunt, although an acceptable alternative, does not allow external CSF
drainage and ICP monitoring in the intraoperative
626
and postoperative settings. Furthermore, the

presence of intraventricular blood and debris after
resection may lead to blockage of the shunt. If it is
necessary to delay denitive surgical treatment
beyond a few days, a VP shunt may allow time for
the procedure to be performed electively.
A substantial number of patients have resolution of hydrocephalus after tumor resection.
Matson and Crofton [36] and others have reported that the successful removal of a tumor
obviates the need for shunting. It is likely that
other factors, such as ventricular blood, postoperative changes, or meningitis, can also render
a patient shunt dependent, however. Ellenbogen
et als series [8] noted that 37% of surviving patients required shunting. Two other series
reported much higher rates of shunt dependency,
ranging from 57% to 78% [39,44]. Raimondi
and Gutierrez [45] have recommended that third
and fourth ventricular tumors require immediate
shunt placement, followed by a delay of 7 to
14 days before surgery. At our institution, a VP
shunt is only placed in the postoperative period if
there is clear evidence of hydrocephalus.
Preoperative imaging
The primary preoperative imaging study is
a high-quality MRI scan with postgadolinium
sequences performed in all three planes. Choroid
plexus tumors enhance brightly and can be
distinguished from normal brain tissue easily.
Obtaining multiple planes assists with surgical
planning and determining the relation of the
tumor to various structures within the ventricles.
Conventional catheter angiography is not required
for diagnosis. Rather, its primary role is as a
preoperative adjunct to reduce tumor vascularity
[46]. Angiography has demonstrated that the
vascular supply of papillomas is from normal
choroidal vessels, which often enlarge as the tumor
grows. Tumors of the lateral ventricle or third
ventricle are generally supplied by branches of the
anterior or posterior choroidal arteries. Mass
eect tends to displace the internal occipital artery
and the basal vein of Rosenthal in an inferior
direction. A fourth ventricular tumor receives its
blood supply from medullary or vermian branches
of the posterior inferior cerebellar artery.
Operative treatment
As with most intracranial tumors, the exact
approach is determined by avoiding eloquent
tissue (eg, primary motor or sensory cortex,
speech centers, visual cortex). The two features

of choroid plexus tumors that can make resection
exceedingly dicult are profuse vascularity and
large size. Within the tumor, arterial vessels
arborize rapidly; thus, control of hemorrhage
within the tumor requires slow and tedious
dissection. The most eective strategy focuses
on initial exposure of the feeding artery and its
ligation. In certain cases, approaching a tumor
from the preferred angle to facilitate tumor
resection does not allow early visualization of
the feeding artery. In this case, a separate initial
approach to allow coagulation of the feeding
artery before tumor resection should be considered. In general, en bloc excision is recommended
[45], although this may not be feasible with large
tumors. Detailed descriptions of surgical approaches to the lateral and third ventricles are
provided elsewhere in this issue.
For lateral ventricle papillomas, a direct cerebrotomy posterior to the angular gyrus allows
access to the entire trigone and permits the pedicle
of the tumor to be identied and coagulated. For
more anteriorly located tumors, an incision can be
made in the frontal convolutions and the lateral
ventricle approached from an anterolateral direction. Lateral ventricular tumors can also be
approached through a cerebrotomy through the
superior or middle temporal gyri. Tumors located
within the temporal horn are easily approached
through the middle or inferior temporal gyri (see
Fig. 3D). Extension of the cortical incision
anteriorly or posteriorly allows access to the
anterior or posterior choroidal vessels. Obviously,
this approach must be tailored to avoid injury to
eloquent areas in the dominant hemisphere.
Third ventricular tumors are rare and are
approached via a midline transcallosal route. The
anterior aspect of the ventricle is entered through
a generous opening in the corpus callosum
extending from the rostrum to the supraoptic
recess. In this way, the tumor can be separated
from the choroid of the tela choroidea, where it is
usually attached, and the accompanying bridging
vessels can be identied and divided.
Fourth ventricular tumors almost always produce triventricular obstructive hydrocephalus
and may require preoperative CSF drainage and
stabilization as noted previously. Tumors in this
location arise from the caudal part of the roof of
the fourth ventricle (Fig. 4). The tumor may
extend into the lateral recesses or through the
foramen of Magendie. The approach is via
a standard midline posterior fossa craniectomy
627
A 10-kg child has a circulating blood volume

of 800 to 1000 mL (7080 mL/kg). The entire
blood volume can easily be lost during the
resection of a particularly vascular tumor. Therefore, preoperative planning in the small child must
anticipate the need for intraoperative transfusion,
potentially several blood volumes, and the accompanying coagulopathy. Specically, several
units of blood and blood products should be
available before starting the procedure. Adequate
intravascular access is crucial, and a central line
should be considered if peripheral access is
limited. During surgery, continual communication with the anesthesiologist allows transfusion
to be started before the development of adverse
events, such as hypotension and hypoxia. Infants
from 2 to 6 months of age reach a physiologic
hemoglobin and hematocrit nadir, which may
need to be considered if a signicant degree of
blood loss occurs. Prevention of a coagulopathy is
critical, because hemostasis is exceedingly dicult
once a coagulation defect is present. Finally, if
several blood volumes have been replaced and
tumor resection is not complete, serious consideration should be given to staging the procedure
(Fig. 5).
Fig. 4. A typical choroid plexus papilloma arising in the
fourth ventricle. T1-weighted images before (top) and
after (bottom) gadolinium enhancement are shown.
or craniotomy exposing the vermis and tonsils. The

blood supply from branches of the posterior
inferior cerebellar artery (PICA) are visualized
from a medial vantage point. If vessels arise from
the region of the superior vermis, it may be necessary to split the vermis to reach these vessels. In
certain situations, the tumor may extend through
the foramina of Luschka; in such cases, the tonsil
can be reected laterally to allow access to this area.
Alternatively, at that sitting or during a staged
procedure, a retrosigmoid approach can be used to
reach the space anterolateral to the brain stem.
Special considerations in the pediatric population
The primary perioperative concerns are those
related to blood loss, resuscitation, and anesthesia. It cannot be emphasized enough that the
anesthesia team should be completely familiar
with managing uid replacement and blood loss
in a pediatric patient. Although self-evident, it
cannot be assumed that experience with adult
perioperative problems guarantees expertise with
similar problems in a child.
Treatment of choroid plexus carcinomas

Overall, reported results in the literature
conrm that GTR has a favorable impact on
survival for carcinomas (see section on outcome).
For this reason, as with papillomas, aggressive
surgical treatment with GTR should be the
primary objective. Nevertheless, GTR with carcinoma is achieved in less than 50% of cases.
Combined with adjunctive therapy, radiation, or
chemotherapy, survival after GTR ranges from
67% to 91% [47]. Technical considerations with
CPC include the expected increased tumor vascularity, lack of a well-developed plane between the
brain and tumor, and excessive friability of the
tumor tissue. The rate of recurrence associated
with GTR alone suggests that adjunctive therapy
is useful, although denitive guidelines are not
available [47].
Most chemotherapy regimens rely on cyclophosphamide, etoposide, vincristine, and a platinum agent [12,43,48]. Wol and colleagues [5]
noted that only 8 of 22 carcinomas responded to
chemotherapy, a disappointing observation. Use
of combination chemotherapy (ifosfamide, carboplatin, and etoposide) after an initial surgical
procedure was found to reduce tumor volume and
628
Fig. 5. A 2-month-old child with a right atrial choroid plexus papilloma is shown. (A) The preoperative T1-weighted
image demonstrates bright enhancement after injection of gadolinium. (B) During the initial attempted resection, the
deeper vascular pedicle could not be secured; after two blood volumes had been lost, the procedure was stopped. The
postoperative CT scan demonstrates a large intraventricular and intraparenchymal hematoma. (C) After 2 weeks,
the patient was taken back to the operating room, where the residual tumor and hematoma were resected. Postoperative
CT shows resolution of the hydrocephalus with large extra-axial cerebrospinal uid collections bilaterally. The patient
did not require placement of ventriculo-peritoneal (VP) shunt.
allow a more complete resection during the

second-stage operation [12,50]. Importantly, the
vascularity of the tumor seemed to be greatly
reduced, because measured blood loss during the
second procedure was on average 15% of blood
volume compared with an average of 64% of
blood volume during the rst procedure.
Postoperative radiation is usually recommended if the child is older than 3 years of age,
although this therapy has not been subjected to
a clinical trial. Radiation is also used in the
presence of leptomeningeal dissemination, subtotal resection, and drop metastases. In one series,
10 patients with CPC were treated with either
chemotherapy or craniospinal radiation [51].
Some of these patients demonstrated no evidence
of disease after chemotherapy alone, but some
required radiation to achieve disease control. The
authors do suggest that radiation can be used as
salvage therapy, but whether radiation for all
patients with carcinoma would reduce relapse
remains unclear. Fitzpatrick and colleagues [47]
noted that after subtotal resection, radiation
therapy, alone or in combination with chemotherapy, oered a survival advantage. The question of whether to use adjunctive therapy after
GTR remains unclear, although the presence of
relapse despite chemotherapy and radiation in
a small group of patients suggests that surgery
alone is not sucient for CPC. Wol and
colleagues [49] support this view and state that

GTR alone is insucient for carcinoma and
should be supplemented with radiation. The role
of conformal radiation and radiosurgery is unknown; neither is the role of intrathecal chemotherapy. The experience reported by Packer et al
[48] suggests that disease relapse confers a poor
prognosis.
Outcome
Most patients with CPPs can expect excellent
long-term survival. The survival for CPC, however, is much worse. In a recent meta-analysis, the
1-, 5-, and 10-year survival rates for papilloma
were 90%, 81%, and 77% compared with only
71%, 41%, and 35% for carcinoma [5]. The extent
of surgery is the most important treatment variable inuencing long-term survival for papilloma
and carcinoma [5,8,48]. The degree of surgical
resection varied from 96% for papillomas to one
of six cases in a series of carcinomas [8,12]. The
overall crude survival rate in Ellenbogen et als
series [8] was 88% for patients with papillomas
and 50% for those with carcinomas.
Packer et al [48] reported that GTR for
carcinoma without adjunctive therapy oers the
highest likelihood of success. Four of ve patients
with GTR remained disease-free at a median of 45
months after diagnosis. Five of six patients who
629
As noted previously, the treatment of hydrocephalus goes hand in hand with the treatment of
choroid plexus tumors. One treatment-related
complication is the development of large subdural
collections. These can result from a transcortical
approach or by craniocerebral disproportion. In
the former, the cause is believed to be a persistent
ventriculosubdural stula. Boyd and Steinbok [53]
report that placement of pial sutures at the
conclusion of the procedure can prevent the development of subdural collections. In the latter,
reduction in the size of the ventricles and removal
of a large mass can lead to an enlarged calvarium
relative to the size of the brain (see Fig. 5C;
Fig. 6). At times, these subdural collections may
require treatment by the placement of a subduralperitoneal shunt.
Fig. 6. Persistent subdural collections in a child several
months after removal of a lateral ventricle choroid
plexus papilloma. The patient remained asymptomatic
with a gradual reduction in the size of the collections;
thus, surgical drainage was not required.
had a subtotal resection relapsed. Although the

authors of this study were pessimistic regarding
the value of adjunctive therapy, the timing of this
therapy may prove to be important with regard to
survival after multimodality treatment. Two other
reports noted that 5-year survival after GTR of
carcinomas ranged from 26% to 40% [43,52].
Berger et al [43] also noted that surgery was the
most important prognostic factor for CPC. The
role of radiation also remains unclear. A brief
report recently noted that the 5-year survival for
patients with carcinoma and GTR followed by
radiation was 68% compared with 16% for those
not irradiated [49]. The two groups were not
exactly comparable, but the clear suggestion is
that surgery alone is insucient to prevent recurrence of carcinomas.
Although papillomas are histologically benign
and potentially curable, morbidity and mortality
are signicant concerns. With respect to operative
mortality, modern series provide gures of 8% to
9.5% [39,44]. In the series from the Hospital for
Sick Children, the cumulative mortality was 36%,
most of which (six of eight cases) occurred in
patients less than 12 months of age. Morbidity
remains an important problem. In one series, 33%
of patients with papillomas had persisting motor
sequelae and psychomotor retardation [44]. In
another series, 26% of patients were classied as
having a fair or poor recovery [8].
Summary
Choroid plexus tumors represent a well-dened
subset of brain tumors that occur mainly in young
children. Surgical resection for papilloma is
usually curative, although careful surgical planning is required to minimize the potential risks.
Although adjunctive therapy for carcinoma includes chemotherapy or radiation, the long-term
survival for carcinoma remains poor.
References
[1] Chow E, Jenkins JJ, Burger PC, Reardon DA,
Langston JW, Sanford RA, et al. Malignant
evolution of choroid plexus papilloma. Pediatr
Neurosurg 1999;31:12730.
[2] Davis LE, Cushing H. Papillomas of the choroid
plexus with a report of six cases. Arch Neurol
Psychiatry 1925;13:681710.
[3] Dandy W. Diagnosis, localization, and removal of
tumours of the third ventricle. Bull Johns Hopkins
Hosp 1922;33:1889.
[4] Laurence KM. The biology of choroid plexus
papilloma and carcinoma of the lateral ventricle.
In: Vinken PJ, Bruyn GW, editors. Handbook
of clinical neurology. New York: Elsevier; 1974.
p. 55595.
[5] Wol JE, Sajedi M, Brant R, Coppes MJ, Egeler
RM. Choroid plexus tumours. Br J Cancer 2002;87:
1086191.
[6] Sarkar C, Sharma MC, Gaikwad S, Sharma C, Singh
VP. Choroid plexus papilloma: a clinicopathological
study of 23 cases. Surg Neurol 1999;52:379.
[7] Asai A, Homan HJ, Hendrick EB, Humphreys
RP, Becker LE. Primary intracranial neoplasms
in the rst year of life. Childs Nerv Syst 1989;5:
2303.
630
[8] Ellenbogen RG, Winston KR, Kupsky WJ. Tumors

of the choroid plexus in children. Neurosurgery
1989;25:32735.
[9] Galassi E, Godano U, Cavallo M, Donati R, Nasi
MT. Intracranial tumors during the 1st year of life.
Childs Nerv Syst 1989;5:28898.
[10] Haddad SF, Menezes AH, Bell WE, Godersky JC,
A AK, Bale JF. Brain tumors occurring before
1 year of age: retrospective reviews of 22 cases in
an 11-year period (19971987). Neurosurgery 1991;
29:813.
[11] Johnson DL. Management of choroid plexus tumors
in children. Pediatr Neurosci 1989;15:195206.
[12] St. Clair SK, Humphreys RP, Pillay PK, Homan
HJ, Blaser SI, Becker LE. Current management of
choroid plexus carcinoma in children. Pediatr
Neurosurg 1991;17:22533.
[13] Doran SE, Blaivas M, Dauser RC. Bone formation
within a choroid plexus papilloma. Pediatr Neurosurg 1995;23:2168.
[14] McComb RD, Burger PC. Choroid plexus carcinoma. Report of a case with immunohistochemical
and ultrastructural observations. Cancer 1983;51:
4705.
[15] Levy ML, Goldfarb A, Hyder DJ, Gonzales-Gomez
I, Nelson M, Gilles FH, et al. Choroid plexus
tumors in children: signicance of stromal invasion.
[16] Hirano H, Hirahara K, Asakura T, Shimozuru T,
Kadota K, Kasamo S, et al. Hydrocephalus due to
villous hypertrophy of the choroid plexus in the
lateral ventricles. Case report. J Neurosurg 1994;
80:3213.
[17] Wyatt-Ashmead J, Kleinschmidt-DeMasters B,
Mierau GW, Malkin D, Orsini E, McGavran L,
et al. Choroid plexus carcinomas and rhabdoid
tumors: phenotypic and genotypic overlap. Pediatr
Dev Pathol 2001;4:5459.
[18] Ho DM, Wong TT, Liu HC. Choroid plexus
tumors in childhood. Histopathologic study and
clinico-pathological correlation. Childs Nerv Syst
1991;7:43741.
[19] Paulus W, Janisch W. Clinicopathologic correlations
in epithelial choroid plexus neoplasms: a study of 52
cases. Acta Neuropathol (Berl) 1990;80:63541.
[20] Cruz-Sanchez FF, Rossi ML, Hughes JT, Coakham
HB, Figols J, Eynaud PM. Choroid plexus papillomas: an immunohistological study of 16 cases.
Histopathology 1989;15:619.
[21] Mannoji H, Becker LE. Ependymal and choroid
plexus tumors. Cytokeratin and GFAP expression.
Cancer 1988;61:137785.
[22] Paulus W, Baur I, Schuppan D, Roggendorf W.
Characterization of integrin receptors in normal
and neoplastic human brain. Am J Pathol 1993;
143:15463.
[23] Herbert J, Cavallaro T, Dwork AJ. A marker for
primary choroid plexus neoplasms. Am J Pathol
1990;136:131725.
[24] Vajtai I, Varga Z, Aguzzi A. MIB-1 immunoreactivity reveals dierent labelling in low grade and in
malignant epithelial neoplasms of the choroid
plexus. Histopathology 1996;29:14751.
[25] Zwetsloot CP, Kros JM, Paz y Gueze HD. Familial
occurrence of tumours of the choroid plexus. J Med
Genet 1991;28:4924.
[26] Brinster RL, Chen HY, Messing A, van Dyke T,
Levine AJ, Palmiter RD. Transgenic mice harboring SV40 T-antigen genes develop characteristic
brain tumors. Cell 1984;37:36779.
[27] Marks JR, Lin J, Hinds P, Miller D, Levine AJ.
Cellular gene expression in papillomas of the
choroid plexus from transgenic mice that express
the simian virus 40 large T antigen. J Virol 1989;
63:7907.
[28] Bergsagel DJ, Finegold MJ, Butel JS, Kupsky WJ,
Garcea RL. DNA sequences similar to those of
simian virus 40 in ependymomas and choroid plexus
tumors of childhood. N Engl J Med 1992;326:
98893.
[29] Zhen HN, Zhang X, Bu XY, Zhang ZW, Huang
WJ, Zhang P, et al. Expression of the simian virus
40 large tumor antigen (Tag) and formation of Tagp53 and Tag-pRb complexes in human brain
tumors. Cancer 1999;86:212432.
[30] Arbeit JM, Munger K, Howley PM, Hanahan D.
Neuroepithelial carcinomas in mice transgenic with
human papillomavirus type 16 E6/E7 ORFs. Am J
Pathol 1993;142:118797.
[31] Sevenet N, Lellouch-Tubiana A, Schoeld D,
Hoang-Xuan K, Gessler M, Birnbaum D, et al.
Spectrum of hS.N.F5/INI1 somatic mutations in
human cancer and genotype-phenotype correlations. Hum Mol Genet 1999;8:235968.
[32] Sevenet N, Sheridan E, Amram D, Schneider P,
Handgretinger R, Delattre O. Constitutional mutations of the hS.N.F5/INI1 gene predispose to
a variety of cancers. Am J Hum Genet 1999;65:
13428.
[33] Rickert CH, Wiestler OD, Paulus W. Chromosomal
imbalances in choroid plexus tumors. Am J Pathol
2002;160:110513.
[34] Eisenberg HM, McComb JG, Lorenzo AV. Cerebrospinal uid overproduction and hydrocephalus
associated with choroid plexus papilloma. J Neurosurg 1974;40:3815.
[35] Wilkins RH, Rutledge BJ. Papillomas of the
choroid plexus. J Neurosurg 1961;18:148.
[36] Matson DD, Crofton FD. Papilloma of choroid
plexus in childhood. J Neurosurg 1960;17:100227.
[37] Gudeman SK, Sullivan HG, Rosner MJ, Becker
DP. Surgical removal of bilateral papillomas of the
choroid plexus of the lateral ventricles with
resolution of hydrocephalus. Case report. J Neurosurg 1979;50:67781.
[38] Sahar A, Feinsod M, Beller AJ. Choroid plexus
papilloma: hydrocephalus and cerebrospinal uid
dynamics. Surg Neurol 1980;13:4768.

[39] Humphreys RP, Nemoto S, Hendrick EB, Homan
HJ. Childhood choroid plexus tumors. Concepts
Pediatr Neurosurg 1987;7:118.
[40] Hopper KD, Foley LC, Nieves NL, Smirniotopoulos JG. The interventricular extension of choroid
plexus papillomas. AJNR Am J Neuroradiol 1987;
8:46972.
[41] Coates TL, Hinshaw DB Jr, Peckman N, Thompson JR, Hasso AN, Holshouser BA, et al. Pediatric
choroid plexus neoplasms: MR, CT, and pathologic
correlation. Radiology 1989;173:818.
[42] Horska A, Ulug AM, Melhem ER, Filippi CG,
Burger PC, Edgar MA, et al. Proton magnetic
resonance spectroscopy of choroid plexus tumors in
children. J Magn Reson Imaging 2001;14:7882.
[43] Berger C, Thiesse P, Lellouch-Tubiana A, Kalifa C,
Pierre-Kahn A, Bouet E. Choroid plexus carcinomas in childhood: clinical features and prognostic
factors. Neurosurgery 1998;42:4705.
[44] Lena G, Genitori L, Molina J, Legatte JR, Choux
M. Choroid plexus tumours in children. Review of
24 cases. Acta Neurochir (Wien) 1990;106:6872.
[45] Raimondi AJ, Gutierrez FA. Diagnosis and surgical
treatment of choroid plexus papillomas. Childs
Brain 1975;1:81115.
[46] Do HM, Marx WF, Khanam H, Jensen ME.
Choroid plexus papilloma of the third ventricle:
[47]
[48]
[49]
[50]
[51]
[52]
[53]
631
angiography, preoperative embolization, and histology. Neuroradiology 2001;43:5036.

Fitzpatrick LK, Aronson LJ, Cohen KJ. Is there
a requirement for adjuvant therapy for choroid
plexus carcinoma that has been completely resected?
Packer RJ, Perilongo G, Johnson D, Sutton LN,
Vezina G, Zimmerman RA, et al. Choroid
plexus carcinoma of childhood. Cancer 1992;69:
5805.
Wol JE, Sajedi M, Coppes MJ, Anderson RA,
Egeler RM. Radiation therapy and survival in choroid plexus carcinoma [letter]. Lancet 1999;353:2126.
Razzaq AA, Cohen AR. Neoadjuvant chemotherapy for hypervascular malignant brain tumors of
childhood. Pediatr Neurosurg 1997;27:296303.
Chow E, Reardon DA, Shah AB, Jenkins JJ,
Langston J, Heideman RL, et al. Pediatric choroid
plexus neoplasms. Int J Radiat Oncol Biol Phys
1999;44:24954.
Pencalet P, Sainte-Rose C, Lellouch-Tubiana A,
Kalifa C, Brunelle F, Sgouros S, et al. Papillomas
and carcinomas of the choroid plexus in children.
Boyd MC, Steinbok P. Choroid plexus tumors:
problems in diagnosis and management. J Neurosurg 1987;66:8005.
Neurosurg Clin N Am 14 (2003 ) 633655
Cumulative Index 2003

Volume 14
January
PITUITARY SURGERY, pages 1180
April
SURGERY FOR PSYCHIATRIC DISORDERS, pages 181326
July
NEUROAUGMENTATION FOR CHRONIC PAIN, pages 327468
October
INTRAVENTRICULAR TUMORS, pages 469655
Note: Page numbers of article titles are in boldface type.
A
Abdominal muscle strength, evaluation of, for
pain management, 342
Abscess, pituitary, imaging of, 71
N-Acetyl aspartate, in magnetic resonance
spectroscopy, in obsessive-compulsive
disorder, 214215
Acetylcholine modulators, intrathecal delivery of,
361
Acidophilic stem cell adenoma, 27
Acromegaly, in somatotroph adenoma, 4345
after pituitary surgery, 134, 139141
treatment of, 8283, 139141
Adenoassociated virus, for gene transfer, for pain
management, 423424
Adenohypophysis
anatomy of, 5557
histogenesis of, 13
histology of, 25
hormones of, 1720
vascular anatomy of, 1415
Adenohypophysitis, lymphocytic, imaging in,
7273, 75
Adenoma
acidophilic stem cell, 27
corticotropinsecreting. See Corticotroph
adenoma.
gonadotropinsecreting. See Gonadotroph
adenoma.
growthhormone secreting. See Somatotroph
adenoma.
mammosomatotroph, 2728
null cell, 3435

plurihormonal, 2829, 36
prolactinsecreting. See Prolactinoma.
somatotroph. See Somatotroph adenoma.
somatotroph-lactotroph, 27
thyrotroph. See Thyrotroph adenoma.
Adenovirus, for gene transfer, for pain
management, 422423, 428, 430
Adrenalectomy, for Cushings disease, 84
Aicardi syndrome, choroid plexus tumors in, 472
Alpha-fetoprotein, in third ventricular tumors,
528
a-Amino-3-hydroxy-5-methyl-4isoxazolepropionate (AMPA) receptors, in
pain transmission, 451452
Amygdala, connections to, in depression, 216218
Analgesia
intrathecal delivery of. See Intrathecal pain
management.
pre-emptive, 452
Anaplasia, in ependymoma, 473, 573
Anderson-Fabry disease, pain in, gene therapy
for, 431
Anesthesia dolorosa, deep brain stimulation for,
333
Aneurysm, suprasellar, imaging of, 7576
Angiography
in choroid plexus tumors, 626
in intraventricular meningioma, 561
in neurocytoma, 484
in subependymal giant cell astrocytoma, 584
in subependymoma, 579
doi:10.1016/S1042-3680(03)00089-5
634
Cumulative Index / Hand Clin 14 (2003 ) 633655
Angiostatin gene transfer, for pain management,

430431
Annas, George, on psychiatric neurosurgery
regulations, 314
Bipolar mood disorder, thalamocortical

dysrhythmia in. See Thalamocortical
dysrhythmia.
Bleeding. See Hemorrhage.
Anticonvulsant action, of vagus nerve

stimulation, 276277, 279280
Body dysmorphic disorder, neuroimaging and

neurocircuitry in, 214
Antidepressants, for depression, 201. See also

Transcranial magnetic stimulation.
Bone disorders, in acromegaly, 4344
Antidiuretic hormone (vasopressin), 2122

deciency of, after pituitary surgery,
124126
inappropriate secretion of, after pituitary
surgery, 126129
regulation of, 123124
replacement of, for diabetes insipidus, 126
Antinociception, physiology of, 452454
Anxiety disorders, transcranial magnetic
stimulation for, 293294
Arachnoid cyst
intraventricular, 609611
parasellar, 6364
Arachnoid membrane, variants of, 609
Aspartylacyclase deciency, gene therapy for,
431
Astrocytoma
hypothalamic, imaging of, 6566
subependymal giant cell, 470471, 475476,
584585
third ventricular. See Third ventricular tumors.
Ataxia, in third ventricular tumor surgery,
542
B
Baclofen, intrathecal delivery of, 360361
Ballantine, H. Thomas, Jr., on psychiatric
neurosurgery, 308
Barber, Jesse, on psychiatric neurosurgery, 311
Behavioral therapy, for obsessive-compulsive
disorder, 200201
Benabid, Alim, deep brain stimulation studies of,
315
Bifrontal transbasal approach, to parasellar and
suprasellar lesions, 114119
Biopsy
in third ventricular tumors, 528
stereotactic, in neurocytoma, 494
Bone morphogenic protein gene transfer, for pain

management, 430
Brain
deep stimulation of. See Deep brain
stimulation.
intraventricular tumors of. See
Intraventricular tumors.
precentral stimulation of, 334, 437443
sensitization in, physiology of, 450452
Brain-derived neurotrophic factor, in
antinociception, 454
Broca, Paul, localization philosophy of, 181
Bromocriptine
for acromegaly, 83
for hyperprolactinemia, for prolactinoma, 30,
8182, 8990, 143
Burckhardt, Gottlieb, frontal lobotomy procedure
of, 181182
C
Cabergoline, for hyperprolactinemia, in
prolactinoma, 8182, 90, 143
Calcication
in ependymoma, 473, 572
Calcitonin gene-related peptide, in pain
transmission, 446, 452
Calcium, in pain transmission, 451452
Calcium channel blockers, intrathecal delivery of,
361
Califano, Joseph, psychiatric neurosurgery
regulations approved by, 314
Callostomy, for colloid cyst, 609
Camera, for endoscopic intraventricular surgery,
549550
Canavan disease, gene therapy for, 431
Cancer, metastasis from. See Metastasis.
Cumulative Index / Hand Clin N Am 14 (2003 ) 633655
Cancer pain
denition of, 330
intrathecal management of, 355, 359, 361,
383385
Capsule, internal, anterior limbs of, electrical
stimulation of, in obsessive-compulsive
disorder, 267274
Capsulotomy, for psychiatric disorders
depression, 219, 232233
history of, 191192
obsessive-compulsive disorder, 204208, 219,
232233
Carcinoma
choroid plexus
clinical features of, 623624
epidemiology of, 469, 621
imaging of, 624, 626
outcome of, 628629
pathology of, 469472, 622623
treatment of, 627628
pituitary, 3637
Cardiovascular disorders, in acromegaly, 44
Catheters, for intrathecal pain management,
356358, 382383
Cationic polymers, for gene transfer, 421
Cauda equina syndrome, in intrathecal
medication delivery, 358
Caudate nucleus
anatomy of, 511
connections to
in depression, 216219
in obsessive-compulsive disorder, 214,
218219
Cavernous sinuses
anatomy of, for pituitary surgery, 1617
exposure of, in pituitary surgery, 102
Cavum septum pellucidum, 608
Cavum veli interpositi, 609
Cavum vergae, 608609
Central neurocytoma. See Neurocytoma,
intraventricular.
Central sensitization, physiology of, 450452
Cerebellum, hypoplasia of, in Dandy Walker
malformation, 613615
Cerebral salt wasting, after pituitary surgery,
128
635
Cerebrospinal uid
analysis of, in depression, in vagus nerve
stimulation, 280
leakage of in intrathecal medication delivery,
358
Chemotherapy
for choroid plexus carcinoma, 627628
for ependymoma, 576577
for neurocytoma, 487, 501503
for subependymoma, 583584
Chlorpromazine, for psychiatric disorders, history
of, 190
Chlorpropamide, for diabetes insipidus, 132
Chondrosarcoma, clival, imaging of, 77
Chordoma, clival, imaging of, 7677
Choroid plexus
anatomy of, 512513
cysts of, 611612
tumors of
dierential diagnosis of, 472
genetic features of, 472, 623
imaging in, 624, 626
in children. See Pediatric patients, choroid
plexus tumors in.
in third ventricle. See Third ventricular
tumors.
metastatic, 598
outcome of, 628629
pathology of, 469472, 621623
subependymal giant cell astrocytoma,
585586
Chroman cells, intrathecal delivery of, 361
Cilia, in choroid plexus tumors, 470471
Cingulate cortex
anterior, connections to
218219
in nociception, 448
Cingulotomy, for psychiatric disorders, 191192,
225235
anatomic considerations in, 225227
complications of, 230
depression, 204206, 208209, 218, 225228
history of, 225
obsessive-compulsive disorder, 204206,
208209, 216, 225231
636
Cingulotomy (continued )
SPECT in, 238249
patient selection for, 227228, 231
physiologic considerations in, 225227
results of, 228231, 241243
technique for, 228, 239
versus alternative surgical methods, 231233
pineal cyst, 612613

tuberous sclerosis, 615617
versus normal variants, 608609
Conus medullaris compression, in intrathecal
medication delivery, 358
Cordotomy, for pain, 328
Cisterna magna, enlargement of, 615
Corpus callosum, lipoma of, 615
Clear cell ependymoma, 473, 478, 573
Corticosteroids, for hypopituitarism, after

pituitary surgery, 129
Clivus
exposure of, in pituitary surgery, 101102
neoplasms of, imaging of, 7678
Coccydynia, spinal cord stimulation for, 378
Corticostriatothalamocortical circuitry
218219
Colloid cyst, 552554, 609
Corticotroph(s), 25
Colorectal cancer
in acromegaly, 45
intraventricular metastasis from, 596597
Corticotroph adenoma
Cushings disease in, 4548
diagnosis of, 169
long-term management of, 169
outcome of, 169
medical treatment of, 142
pathology of, 3132
persistent, 141142
radiotherapy for, 142
silent, 32
surgical treatment of, 169
persistent, 141142
postoperative endocrine management in,
130131, 133134
preoperative management in, 94
Clonidine, intrathecal delivery of, 360, 384
Compensation systems, pain management and,

341, 349350
Complex regional pain syndrome, spinal cord
Computed tomography
in colloid cyst, 609
in craniopharyngioma, 6263
in empty sella syndrome, 6465
in ependymoma, 572
in hypothalamic astrocytoma, 6566
in intraventricular meningioma, 561
in lateral ventricle tumors, 514
in meningioma, 67
in neurocytoma, 484
in pituitary abscess, 71
in pituitary adenoma, 59
in schwannoma, 70
in suprasellar aneurysm, 75
Congenital intraventricular lesions, 607619
arachnoid cyst, 6364, 609611
choroid plexus cyst, 611612
colloid cyst, 552554, 609
corpus callosum lipoma, 615
Dandy-Walker malformation, 613615
embryology of, 607608
ependymal cyst, 612
epidermoid cyst, 613
neurenteric cyst, 613
Corticotropin, 1819, 25
adenomas secreting. See Corticotroph
adenoma.
deciency of, after pituitary surgery, 132133
measurement of, in corticotroph adenoma, 47
Corticotropin-releasing hormone stimulation test,
in corticotroph adenoma, 47
Cortisol
excess of, in corticotroph adenoma, 3132,
4548
persistent, 141142
preoperative management of, in pituitary
tumors, 94
Cranial nerve dysfunction, after pituitary
radiotherapy, 162
Craniopharyngioma. See also Third ventricular
tumors.
imaging of, 6163
637
Cre-lox system, for gene transfer, for pain

management, 422
Delgado, Jose M.R., on psychiatric neurosurgery,

309
Cushing, Harvey, transsphenoidal approach of,

25, 7
Deniker, Pierre, psychoactive drug therapy

discovery by, 190
Cushings disease
in corticotroph adenoma, 3132, 4548
outcome of, 169
persistent, 141142
Depression, 199212
epidemiology of, 201
evaluation of, before pain management,
411412, 414415
neurocircuitry models for, 216219
neuroimaging in, 216219
neurosurgery for
capsulotomy, 219, 232233
cingulotomy, 204206, 208209, 225228
current awareness of, 201202
eectiveness of, 207209
history of, 202203
limbic leukotomy, 205207, 209, 232
modern procedures for, 203205
neuroimaging and neurocircuitry related to,
218219
safety of, 205207
subcaudate tractotomy, 204205, 207, 218,
231
thalamocortical dysrhythmia in. See
Thalamocortical dysrhythmia.
transcranial magnetic stimulation for, 288293
maintenance, 292
meta-analysis of, 289292
with intentional seizure generation, 294295
treatment of, before pain management, 348
vagus nerve stimulation for, 275282
CyberKnife, for neurocytoma, 501

Cyst(s)
arachnoid
parasellar, 6364
choroid plexus, 611612
colloid, 552554, 609
ependymal, 612
epidermoid
parasellar, 64
neurenteric, 613
parasellar, imaging of, 6364
pineal, 612613
Rathkes cleft, 63
Cystitis, interstitial, peripheral nerve stimulation
for, 377
Cytokeratins, in choroid plexus tumors, 470
Cytomegalovirus, for gene transfer, for pain
management, 422
Desmopressin (DDAVP), for diabetes insipidus,

D
Dandy Walker malformation, 613615
Deep brain stimulation, 389399
for neuropathic pain, 390391
for nociceptive pain, 391
history of, 332334, 389390
mechanism of action for, 390391
o-label use of, 396
patient selection for, 391392
physiology of, 454455
preoperative evaluation for, 391392
rationale for, 390391
results of, 394396
target site selection for, 392
technique for, 392394
tolerance to, 395
Dexamethasone suppression test, in corticotroph

adenoma, 47
Diabetes insipidus, after pituitary surgery,
124126, 132
Diabetic neuropathy, pain in
gene therapy for, 430
spinal cord stimulation for, 378
Diaphragma sellae, anatomy of, 16
Diethelm, Oskar, on ethics of neuropsychiatric
surgery, 305307
Dizziness, in neurocytoma, 483
DNA, in gene therapy, for pain, 420421, 427430
Dopamine, deciency of, gene therapy for, 426
638
Dopamine agonists
for gonadotroph adenoma, 84
for prolactinoma, 3031, 8182, 8990, 143
Dott, Norman, transsphenoidal approach of, 57
Dynorphin, in antinociception, 453
E
Edema, in intrathecal morphine delivery, 385386
Electrical stimulation
deep brain, 332334, 389399
implantable stimulators for, 331334, 372375,
403406
internal capsule anterior limb, for obsessivecompulsive disorder, 267274
motor cortex, 334
peripheral nerve, 331, 401408
precentral, 334, 437443
spinal cord, 331332, 365380
vagus nerve, for depression, 275282
Electroconvulsive therapy, for psychiatric
disorders, 184, 201
versus transcranial magnetic stimulation,
287289, 291292, 294
Electroencephalography, in depression, in vagus
nerve stimulation, 280
Electropermeabilization, in gene therapy, for
pain, 420
Eloquent cortex, localization of, transcranial
magnetic stimulation for, 287
Empty sella syndrome, 16, 6465
Endocrinopathies, in pituitary tumors. See
Pituitary tumors, endocrinopathies in.
Endometriosis, pain in, gene therapy for, 430431
Endonasal transsphenoidal approach, to pituitary
tumors, 99100
lateral (Hirsch), 12
Endoscopy
in intraventricular surgery, 547557
colloid cyst resection, 552554
equipment for, 549551
history of, 547548
pineal tumor management, 554555
two-portal, 551
ventriculostomy, 551552
in transsphenoid approach, to pituitary
tumors, 104105
Enkephalins, in antinociception, 453
Ependymal cyst, 612

Ependymoma
classication of, 572573
anaplastic, 573
cellular, 573
clear cell, 473, 478, 573
clinical presentation of, 572
genetic features of, 474
imaging of, 572
infratentorial, 574576
papillary, versus choroid plexus tumors,
472
pathology of, 470474, 572574
prognosis of, 577578
supratentorial, 574576
tanycytic, 573
third ventricular. See Third ventricular
tumors.
versus central neurocytoma, 478
versus subependymoma, 475
Epidermoid cyst
parasellar, 64
Epidermoid tumors, in third ventricle. See Third
ventricular tumors.
Epilepsy
transcranial magnetic stimulation for,
286287, 293
vagus nerve stimulation for, 276277
Epithelial membrane antigen, in choroid plexus
tumors, 470
Equine infectious anemia virus, for gene transfer,
for pain management, 425
Ervin, Frank R., on psychiatric neurosurgery,
309
Estrogen, for persistent prolactinoma, 142
Ethical issues, in neuropsychiatric surgery,
303319
historical aspects of, 303308
regulatory aspects of, 312315
sociopolitical aspects of, 308312
F
Federal Employees Compensation Act, 349
Federal Employers Liability Act, 349
Feline immunodeciency virus, for gene transfer,

Gamma knife capsulotomy, for psychiatric

disorders, 204208
Fiamberti, Amarro, transorbital lobotomy

procedure of, 189
Gamma knife radiosurgery

for pituitary tumors, 156161
for somatotroph adenoma, 83
Fluid restriction, in antidiuretic hormone

inappropriate secretion, 128129
Follicle stimulating hormone, 19, 2526
excess of, in gonadotroph adenoma, 34, 4243,
84
Foramen of Monro, anatomy of, 510511
Foreign body reactions, in deep brain stimulation,
395
Fornices, anatomy of
in endoscopic surgery, 548
in third ventricular surgery, 531
Fourth ventricle, dilation of, in Dandy Walker
malformation, 613615
Freeman, Walter, psychiatric neurosurgical
procedures of, 182183, 185190, 303306
Frontal horn, anatomy of, 510511
Frontal lobotomy, history of, 182190, 203,
303308
Frontal transcortical (middle frontal gyrus)
approach, to lateral ventricle tumors, 517518
Frontotemporal transcavernous approaches, to
parasellar and suprasellar lesions, 109121
bifrontal transbasal, 114119
orbitozygomatic, 109114
pterional, 109111
supraorbital, 119121
transzygomatic, 109111
Fulton, John F., neurophysiologic studies of, 182,
185187
Functional neuroimaging, in
obsessive-compulsive disorder, 215
Furosemide, for antidiuretic hormone
639
Ganglioneurocytoma, 477
Gasserian ganglion, in nociception, 446447
Gate theory of pain, 328331, 451
Gemistocyte-like cells, in subependymal giant cell
astrocytoma, 475476
Gene-gun method, for gene transfer, 428
Gene therapy, for pain, 419435
advantages of, 419, 422
cell lines for, 429430
ex vivo, 429430
neuromodulation, 425429
nonviral, 420421
root causes, 429430
viral vectors in, 421425
Germinoma, parasellar, imaging of, 71
Giant cell astrocytoma, subependymal, 470471,
475476, 584585
Glial-derived neurotrophic factor, in nociception,
446
Glial brillary acidic protein
for gene transfer, for pain management, 422
in ependymoma, 473
Glioma. See also Ependymoma;
Subependymoma.
intraventricular
choroid, 585586
subependymal giant cell astrocytoma,
470471, 475476, 584585
types of, 571
parasellar, after pituitary radiotherapy, 162
Glutamate, in pain transmission, 451452
G
a-Galactosidase A deciency, pain in, gene
therapy for, 431
Gall, Franz Joseph, on phrenology, 181
Gamma-aminobutyric acid
deciency of, gene therapy for, 426
receptors for, in antinociception, 453454
Glutamate decarboxylase deciency, gene therapy

for, 426
Glycine, intrathecal delivery of, 360
Goltz, Friedrich, temporal lobotomy procedure
of, 181
Gonadotroph(s), 2526
640
Gonadotroph adenoma
diagnosis of, 4243
endocrinopathy in, 4243
imaging of, 43
pathology of, 34
Herpes simplex virus 1, for gene transfer, for pain

management, 424425, 428429
Gonadotropins, 19, 2526

adenoma secreting. See Gonadotroph
adenoma.
Histiocytosis X, of pituitary, 72, 74
Hippocampus, connections to, in depression,

216218
Hirsch, Oscar, transsphenoidal approach of, 12
Grant, Francis, lobotomies by, 186
Hormones, pituitary, 1722, 2526. See also

specic hormones.
abnormalities of. See Pituitary tumors,
endocrinopathies in.
Granuloma, at catheter tip, in intrathecal drug

delivery, 386
Human chorionic gonadotropin, in third

ventricular tumors, 528
Graves disease, in thyrotroph adenoma, 49
Human immunodeciency virus, for gene transfer,

Grinker, Roy, on ethics of neuropsychiatric

surgery, 306307
Growth hormone, 1920, 25
pituitary tumors secreting. See Somatotroph
adenoma.
Growth hormone-releasing hormone, 20
Guiot, Gerard, transsphenoidal approach of,
6, 8
H
Hallucinations, thalamocortical dysrhythmia in.
See Thalamocortical dysrhythmia.
Hardy, Jules, transsphenoidal approach of, 810
Headache
in deep brain stimulation, 396
in intraventricular metastasis, 595
in neurocytoma, 483
peripheral nerve stimulation for, 404405
Hemangiopericytoma, parasellar, imaging of,
6870
Hemianopsia, bitemporal, in prolactinoma, 29
Hemiparesis, in lateral ventricle tumor surgery,
523
Hemorrhage, intracranial
in lateral ventricle tumor surgery, 523
in third ventricular tumor surgery, 542
Herniorrhaphy, postoperative pain in, peripheral
nerve stimulation for, 406
Hydrocephalus
after lateral ventricle tumor removal, 523
after third ventricular tumor removal, 542
endoscopic ventriculostomy for, 551552
in Dandy Walker malformation, 613615
in ependymoma, 572, 574
Hydrocortisone, for hypopituitarism, after
pituitary surgery, 129
Hyperalgesia
central mechanisms of, 450451
in gene therapy for pain, 431
Hypercortisolemia, in corticotroph adenoma,
3132, 4548
persistent, 141142
Hypernatremia, after pituitary surgery, 124126
Hyperprolactinemia
in acromegaly, 44
in prolactinoma, 2931, 4142
persistent, 142143
Hyperthyroidism, in thyrotroph adenoma, 4850
Hypocortisolism, after pituitary surgery, 129130
Hypogonadism
in gonadotroph adenoma, 42
in prolactinoma, 8182
Hyponatremia, after pituitary surgery, 126129
Hypophyseal arteries, anatomy of, 1415

Hypophysectomy, in persistent acromegaly, 140
Hypophysis. See Pituitary gland; Pituitary
tumors.
Hypopituitarism
after pituitary surgery, 129130
in gonadotroph adenoma, 42
in somatotroph adenoma, 44
Hypothalamic-pituitary-adrenal axis, connections
to, in depression, 216
641
Inguinal pain, peripheral nerve stimulation for,

406
Inositol triphosphate, in pain transmission, 452
Institutional review boards, for psychiatric
neurosurgery research, 313
Insula, in nociception, 448
Insulin-like growth factor-1, in somatotroph
adenoma, 139141
Insulin resistance, in acromegaly, 44
Hypothalamus
anatomy of, in third ventricular surgery, 529
astrocytoma of, imaging of, 6566
dysfunction of, after pituitary radiotherapy,
161162
Interhemispheric approaches
to intraventricular meningioma, 565566
to lateral ventricle tumors, 520522
Hypoxanthine-guanine phosphoribosyltransferase
gene transfer, 422423
Interleukin-12 gene transfer, for pain, 427428
Intermediate lobe, pituitary, physiology of, 2021
in choroid plexus tumors, 470, 623
in ependymoma, 473
in neurocytoma, 485
Intracranial pressure, increased

in ependymoma, 572
in neurocytoma, 483
Implantable devices, for pain management

electrical stimulation, 331334, 372375
peripheral nerve, 403406
spinal cord, 372375
intrathecal drug delivery, 331334, 372375
Impulse control disorder, thalamocortical
dysrhythmia in. See Thalamocortical
dysrhythmia.
Incisural space, posterior, anatomy of, in third
ventricular surgery, 531
Infections
in intrathecal medication delivery, 357
in spinal cord stimulation, 375376
Inferior petrosal sinus sampling, in corticotroph
adenoma, 4748
Inammation, at catheter tip, in intrathecal drug
delivery, 386
Infratentorial supracerebellar approach
to third ventricular tumors, 532538
Infundibulum, in pituitary embryogenesis, 1112
Interhemispheric-transcallosal approach, to third

ventricular tumors, 532, 534, 538539
Intermediate laments, in subependymal giant cell

astrocytoma, 476
Intrathecal pain management, 353363

contraindications to, 355
delivery of, 381387
clinical studies of, 383385
inammatory mass interference with, 386
pharmacodynamics of, 381383
side eects of, 385386
sites for, 354
devices for
evaluation of, 358
failure of, 358
implantation of, 357358
management of, 359
replacement of, 359
selection of, 355357
dose changes in, 359
drugs used in, 360361
future of, 360362
history of, 353354
intraventricular route of, 354, 356
outcomes of, 359360
psychologic variables in, 410412
642
Intrathecal (continued )
trial of, 355
versus epidural delivery, 354
versus medical treatment, 383
Intraventricular tumors
choroid plexus, 469472, 621631
congenital, 607619
ependymoma. See Ependymoma.
epidemiology of, 469482. See also specic
tumor types.
glioma, 571591
lateral, 509525
meningioma, 470471, 559569
of third ventricle. See Third ventricular
tumors.
metastatic, 593606
neurocytoma, 470471, 476478, 483508
pathology of, 469482
posterior, 527545
subependymal giant cell astrocytoma, 470471,
475476, 584585
subependymoma. See Subependymoma.
surgical treatment of, endoscopic adjuncts to,
547557
Irrigation, in endoscopic intraventricular surgery,
550551
Lanreotide
for somatotroph adenoma, 83, 140
for thyrotroph adenoma, 84
Lateral ventricles, tumors of, 509525
blood supply of, 513514
clinical presentations of, 509510
intraoperative considerations in, 514516
outcome of, 522524
postoperative considerations in, 516, 522524
preoperative considerations in, 514
surgical approaches to, 516522
frontal transcortical (middle frontal gyrus),
517518
interhemispheric, 520522
parietal transcortical (superior parietal
lobule), 519520
temporal transcortical (middle temporal
gyrus), 518519
types of, 509
Leksell, Lars, capsulotomy procedure of, 191
Lentiviruses, for gene transfer, for pain
management, 425
Kainate receptors, in pain transmission, 451
Leukotomy
limbic, for psychiatric disorders, 193
depression, 205207, 209, 232
obsessive-compulsive disorders, 205207,
209, 218219, 232
versus cingulotomy, 232
orbitomedial, for obsessive-compulsive
disorder, 218
Kelly, Desmond, limbic leukotomy procedure of,

193
Li-Fraumeni syndrome, choroid plexus tumors in,

472
Ketamine, intrathecal delivery of, 360
Lima, Almeida, early frontal lobotomies by, 182
Ketoconazole, for Cushings disease, 8384, 142
Limbic leukotomy, for psychiatric disorders, 193

depression, 205207, 209, 232
obsessive-compulsive disorders, 205207, 209,
218219, 232
J
Jacobsen, Carlyle, neurophysiologic studies of,
182
Ki-67 antigen
in ependymoma, 473
Knight subcaudate tractotomy, 193
L
Laborit, Henri, psychoactive drug therapy
discovery by, 190
Lactotroph(s), 25. See also Prolactin;
Prolactinoma.
Laminotomy, for spinal cord stimulation implant,
375
Limbic system, in psychiatric disorders

dysfunction of, 226
pathway disruption in. See Cingulotomy;
Limbic leukotomy.
Linear accelerator-based radiosurgery
for neurocytoma, 501
Lipoma, of corpus callosum, 615
Liposome-mediated gene transfer, for pain, 421
Lobotomy
frontal, history of, 182190, 203, 303308
prefrontal, history of, 182190
transorbital, 189
Long Shore and Harbor Workers Act, 349
Lung carcinoma, intraventricular metastasis from,
596597
Luteinizing hormone, 19, 2526
excess of, in gonadotroph adenoma, 34, 4243,
84
Lyerly, James, psychiatric neurosurgical
procedures of, 186187
Lymphocytic adenohypophysitis, imaging in,
7273, 75
Lymphoma, clival, imaging of, 78
corticotroph, 47
for radiotherapy planning, 151152
for stereotactic radiosurgery, 160
gonadotroph, 43
preoperative, 94
prolactinoma, 42
in precentral stimulation, 438
in Rathkes cleft cyst, 63
in schwannoma, 70
in stereotactic cingulotomy, 228, 238
in suprasellar aneurysm, 76
morphometric
in depression, 217
in obsessive-compulsive disorder, 214
Lysosomes, in subependymal giant cell

astrocytoma, 476
Magnetic resonance spectroscopy

in neurocytoma, 484
Magnetic stimulation, transcranial. See

Transcranial magnetic stimulation.
Macroadenoma, pituitary, imaging of, 5960

Macrostimulation, in deep brain stimulation, 393
Magnetic resonance imaging
blood oxygenation level-dependent
in depression, in vagus nerve stimulation,
279
in transcranial magnetic stimulation
studies, 288
in arachnoid cyst, 64
in choroid plexus tumors, 624, 626
in colloid cyst, 609
in craniopharyngioma, 63
in empty sella syndrome, 6465
in ependymal cyst, 612613
in ependymoma, 572
in epidermoid cyst, 64
in germinoma, 71
in hemangiopericytoma, 6970
in histiocytosis, 74
in hypothalamic astrocytoma, 66
in lymphocytic adenohypophysitis, 73, 75
in meningioma, 6768, 561, 564
in metastatic intraventricular tumors, 594
in neurocytoma, 484
in pituitary abscess, 71
in pituitary histiocytosis X, 72
in pituitary tuberculosis, 71
in pituitary tumors, 5761, 94
643
Magnetoencephalography, in thalamocortical
dysrhythmia, 253256, 260
Mammillary bodies, anatomy of, in endoscopic
surgery, 548549
Mammosomatotroph adenoma, 2728
Manganese superoxide dismutase gene transfer,
Mania, transcranial magnetic stimulation for, 292
Mark, Vernon H., on psychiatric neurosurgery,
309310, 315
Mearns, Edward, on psychiatric neurosurgery,
311313
Medical history, in pain management, review of,
340342
Melanotropins, 2021
Melzack-Wall gate theory of pain, 328331, 451
Memory loss, in lateral ventricle tumor surgery,
523
Meningioma
intraventricular, 478, 559569
incidence of, 559
origin of, 559560
pathology of, 560
644
Meningioma (continued )
radiography in, 561562
observation in, 561, 563
options for, 561, 563
radiosurgery in, 563
radiotherapy in, 563
surgical, 563568
parasellar, imaging of, 6768
Meningitis, in intrathecal medication delivery,
357358
Mental status alterations, in neurocytoma, 483
Metastasis
intraventricular, 478, 593606
assessment of, 599600
diagnosis of, 594595
histology of, 595598, 602603
incidence of, 593594
location of, 595598
metachronous presentation of, 594595
multiple, 602
outcome of, 604
radiography in, 600602
single, 600602
synchronous presentation of, 594595
to clivus, imaging of, 78
to pituitary, imaging of, 71
N-Methyl-D-aspartate (NMDA) receptors, in
pain transmission, 451453
antagonists of, 360
MK-801, intrathecal delivery of, 360

Moniz, Egas, psychiatric neurosurgical
procedures of, 182183, 185186, 190, 305306
Morphine, intrathecal delivery of, 354, 358, 360
clinical studies of, 383385
pharmacodynamics of, 381383
side eects of, 385386
Morphine-naloxone test, for deep brain
stimulation preoperative evaluation, 391392
Mortality, in lateral ventricle tumor surgery, 523
Motivation, evaluation of, before pain
management, 415
Motor cortex, stimulation of, 334, 437443
indications for, 438
literature review on, 439441
mechanism of, 437438
Motor disorders, in lateral ventricle tumors, 510
Movement disorders, transcranial magnetic
Myoclonus, in intrathecal morphine delivery,
385386
N
National Commission for the Protection of
Human Subjects of Biomedical and Behavioral
Research, psychiatric neurosurgery and,
312314, 316
Metyrapone, for Cushings disease, 8384
National Psychosurgery Advisory Board, 313315
Microadenoma, pituitary
imaging of, 59
Nausea and vomiting, in lateral ventricle tumors,

510
Microlumina, in ependymoma, 473
Nelsons syndrome, in corticotroph adenoma,

3132
Microscopy, in neurocytoma, 485
Nerve(s), injury of, peripheral sensitization in, 449
Microvilli, in choroid plexus tumors, 470471
Nerve bers, sensory, physiology of, 445
Middle temporal gyrus approach, to

intraventricular meningioma, 564565
Neural crest, congenital lesions developing from,

607608
Midline suboccipital approach, to intraventricular

meningioma, 566, 568
Neuralgia, occipital, peripheral nerve stimulation

for, 377378
Midnight plasma cortisol measurement, in

corticotroph adenoma, 47
Neurenteric cyst, 613
Mitotane, for Cushings disease, 142

Mixter, W. Jason, lobotomies by, 187
Neuroaugmentation. See also specic techniques.

denition of, 327
evaluation for, 339352
history of, 327337
deep brain stimulation, 332334

early theories on, 327328
gate theory, 328331
motor cortex stimulation, 334
peripheral nerve stimulation, 331
spinal cord stimulation, 331332
psychologic evaluation for, 409417
Neurocircuitry models, for psychiatric disorders,
213223
depression, 216219
218219
Neurocytoma, intraventricular, 483508
atypical, 486487
imaging of, 484
pathology of, 477478, 484487
recurrent, treatment of, 494, 497498
stereotactic biopsy in, 494
acute, 492
chemotherapy in, 487, 501503
follow-up in, 503504
observation, 492
overview of, 487492
paradigm for, 504
radiation therapy in, 487491, 493501
recurrent, 494, 497498
surgical, 487494
versus ependymoma, 472
645
spinal cord stimulation for, 378

Neuropeptide FF, in antinociception, 453
Neuropeptide Y, in antinociception, 453454
Neurotrophic growth factors, in peripheral
sensitization, 449450
Neville, Robert, on psychiatric neurosurgery, 312
Nitric oxide, in peripheral sensitization, 450
NMDA (N-methyl-D-aspartate) receptors, in pain
transmission, 451453
antagonists of, intrathecal delivery of, 360
Nociception
anatomy of, 445448
receptors in, 448449
Nociceptive pain, deep brain stimulation for,
391392, 395
Nonfunctioning or nonsecreting pituitary tumors.
See Pituitary tumors, nonfunctioning.
Nonsteroidal anti-inammatory drugs, for
diabetes insipidus, 132
Norepinephrine, in nociception, 455
Nose, pituitary surgical approach through, 12,
99100
Nucleus raphe magnus, in nociception, 454455
Nucleus ventralis posteriomedialis, electrical
stimulation of, 389, 391393
Null cell adenoma, pathology of, 3435
Neuroendoscopy. See Endoscopy.

Neurohypophysis
anatomy of, 5556
histology of, 21
hormones of, 2122
vascular anatomy of, 15
Neuroimaging. See also specic techniques.
for psychiatric disorders, 213223
depression, 216219
218219
Neuroma, peripheral sensitization in, 449450
Neuronal specic enolase
for gene transfer, for pain management, 422
in central neurocytoma, 477
in neurocytoma, 485
Neuropathic pain
deep brain stimulation for, 390392, 394
O
Obesity, pain management and, 342
Obsessive-compulsive disorder, 199212
behavioral therapy for, 200201
diagnosis of, SPECT in, 238, 241242
neurocircuitry models for, 214216, 218219
neuroimaging in, 214216, 218219
neurosurgery for
capsule anterior limb electrical stimulation,
267274
capsulotomy, 204208, 219, 232233
cingulotomy, 204206, 208209, 216,
225231, 238249
current awareness of, 201202
eectiveness of, 207209
history of, 202203
646
Obsessive-compulsive (continued )
limbic leukotomy, 205207, 209, 218219,
232
modern procedures for, 203205
neuroimaging and neurocircuitry related to,
218219
safety of, 205207
subcaudate tractotomy, 204205, 207, 218,
231232
pathophysiology of, 226
positron emission tomography in, 226,
246248
SPECT in, 237250
diagnostic, 238, 241242
postoperative, 238241, 245249
thalamocortical dysrhythmia in. See
Thalamocortical dysrhythmia.
Occipital headache, peripheral nerve stimulation
for, 377378, 404405
Occipital transtentorial approach
to third ventricular tumors, 532, 539541
OCD. See Obsessive-compulsive disorder.
Octreotide
for gonadotroph adenoma, 84
intrathecal delivery of, 361
Oligodendroglioma
versus central neurocytoma, 478
Oncocytoma, pituitary, pathology of, 35
Opioids. See also Morphine.
endogenous, in antinociception, 453
intrathecal delivery of, 354, 358
Optic chiasm, anatomy of, pituitary surgery and,
17
Optic nerve, anatomy of, pituitary surgery and, 17
Orbitofrontal-subcortical connections
in obsessive-compulsive disorder, 214, 218219
interruption of, in subcaudate tractotomy, 193,
204205, 207
Orbitomedial leukotomy, for obsessivecompulsive disorder, neuroimaging in, 218
Orbitozygomatic approach, to parasellar and
Oxytocin, 2122
P
Pain, chronic, anatomy and physiology of,
445462
central sensitization, 450452
descending systems, 454455
nociception, 445448
peripheral sensitization, 449450
supraspinal plasticity, 455
supraspinal systems, 454454
Pain evaluation, 339352
compensation systems and, 341, 349350
for patient selection, 339340
medical history review in, 340342
physical ndings in, 342343
prior treatment evaluation, 345347
psychologic report review in, 343345
psychosocial issues in, 347348
secondary gain and, 349
Pain management
anatomic and physiologic considerations in,
445462
compliance with, 343, 345
decision not to treat, 340
deep brain stimulation for, 389399
failure of, 339340, 345
in spinal cord stimulation, 377, 410
intrathecal medications for. See Intrathecal
pain management.
neuroaugmentation for
history of, 327337
psychologic evaluation for, 409417
pain classication for, 329330
peripheral nerve stimulation for. See
Peripheral nerve stimulation.
precentral stimulation for, 334, 437443
regression of benet in, 339340, 345
spinal cord stimulation in. See Spinal cord
stimulation.
terminology of, 329330
Paleospinothalamic tract, in nociception, 447
Palliative care, for psychiatric disorders,
neurosurgery in, 316
Pallidotomy, anterior medial, for thalamocortical
dysrhythmia, 252263
magnetoencephalography with, 253256, 260
results of, 255, 259263
strategy for, 252253, 259

Papilledema, in lateral ventricle tumors, 510
Papilloma, choroid plexus
imaging of, 624, 626
outcome of, 628629
pathology of, 469472, 621623
647
carcinoma, 627628
hydrocephalus in, 625626
imaging in, 624, 626
outcome of, 628629
pathology of, 621623
Pegvisomant, for acromegaly, 83, 140
Paraplegia, in intrathecal drug delivery, 386
Pelvic pain, spinal cord stimulation for, 378
Parasellar and suprasellar lesions. See also

Pituitary tumors.
anatomic considerations in, 1617, 57
imaging of
aneurysm, 7576
arachnoid cyst, 6364
clivus neoplasms, 7678
craniopharyngioma, 6163
empty sella, 6465
epidermoid cyst, 64
germinoma, 71
hemangiopericytoma, 6870
hypothalamic astrocytoma, 6566
meningioma, 6768
Rathkes cleft cyst, 63
schwannoma, 70
surgical treatment of
bifrontal transbasal approach to, 114119
frontotemporal transcavernous approaches
to, 109121
orbitozygomatic approach to, 109114
pterional approach to, 109111
supraorbital approach to, 119121
transcranial approaches to, 109122
transzygomatic approach to, 109111
Percutaneous implantation, for peripheral nerve

stimulation, 403
Paresthesia, in spinal cord stimulation, 372, 377
Periventricular gray matter, electrical stimulation

of. See Deep brain stimulation.
Parietal occipital approach, superior, to

Parietal transcortical (superior parietal lobule)
Parinauds syndrome, in intraventricular
meningioma, 561
Parkinsons disease
deep brain stimulation for, 315
Pediatric patients, choroid plexus tumors in,
621631
Pergolide, for prolactinoma, 8182, 143

Periaqueductal gray matter. See Deep brain
stimulation.
Perineurioma, intraventricular, 478
Peripheral nerve stimulation, 401408
for occipital headache, 404405
for postoperative inguinal pain, 406
for supraorbital pain, 405
history of, 331, 401
mechanism of action of, 406407
open nerve dissection technique for, 402403
outcomes of, 407
percutaneous implant technique for, 403
spinal cord stimulation equipment for,
377378
subcutaneous, 403406
techniques for, 402
Peripheral sensitization, physiology of,
449450
Phantom limb pain, physiology of, 452

Phospholipase C, in pain transmission, 452
Phrenology, 182
Physical therapy, for pain, 340
Pia mater, variants of, 609
Pineal cyst, 612613
Pineal gland tumors. See Third ventricular
tumors.
Pituitary fossa, anatomy of, 1516
648
Pituitary gland
abscess of, imaging in, 71
anatomy of, 5557
dysfunction of, after pituitary radiotherapy,
161162
histiocytosis of, imaging in, 72, 74
histogenesis of, 13
histology of, 2526
hormones of, 1722, 2526. See also specic
hormones.
abnormalities of. See Pituitary tumors,
endocrinopathies in.
hyperplasia of, imaging in, 61
imaging of, 5557. See also Pituitary tumors,
imaging of.
abscess, 71
histiocytosis X, 72, 74
lymphocytic adenohypophysitis, 7273, 75
sarcoidosis, 71
techniques for, 57
tuberculosis, 71
versus hyperplasia, 61
lymphocytic adenohypophysitis of, 7273, 75
sarcoidosis of, imaging in, 71
tuberculosis of, imaging in, 71
tumors of. See Pituitary tumors.
Pituitary tumors
anatomic considerations in, 1123, 5557
embryologic, 1113
gross, 1314
surgical, 1517
vascular, 1415
carcinoma, pathology of, 3637
corticotropinsecreting. See Corticotroph
adenoma.
endocrinopathies in, 4154
acromegaly, 4345, 8283, 134, 139141
Cushings disease. See Cushings disease, in
corticotroph adenoma.
gonadotrophin abnormalities, 4243
hyperprolactinemia, 2931, 4142, 8182,
142143
hyperthyroidism, 4850
postoperative management of, 123138
acromegaly, 131, 133
antidiuretic hormone abnormalities,
123129, 132
Cushings disease, 130131, 133134
hypopituitarism, 129130
in nonfunctioning tumors, 132133
in prolactinoma, 131, 134
in somatotroph adenoma, 131, 133
in thyrotroph adenoma, 131132, 134

gonadotropinsecreting. See Gonadotroph
adenoma.
growthhormone secreting. See Somatotroph
adenoma.
imaging of
adenoma, 5761
corticotroph, 47
gonadotroph, 43
metastasis, 71
preoperative, 94
prolactinoma, 42
macroadenoma, 5960
malignant, after radiotherapy, 162
medical treatment of, 8187, 140144
in long-term management, 167171
in persistent endocrine activity, 140144
metastatic, imaging of, 71
microadenoma, 59, 8992
nonfunctioning
pathology of, 3435
132133
null cell adenoma, 3435
oncocytoma, 3435
outcome of, 167171
pathology of, 2529
carcinoma, 3637
gonadotroph adenoma, 34
nonfunctioning, 3435
plurihormonal, 36
prolactinoma, 2931
somatotroph adenoma, 2629
thyrotroph adenoma, 3233
versus normal pituitary, 2526
persistent functional, management of,
139145
physiologic considerations in, 1722
anterior lobe, 1720
intermediate lobe, 2021
posterior, 2122
plurihormonal, 2829, 36
prolactinsecreting. See Prolactinoma.
radiation therapy for. See Radiation therapy,
for pituitary tumors; Stereotactic
radiosurgery, for pituitary tumors.
bifrontal transbasal approach to, 114119
frontotemporal transcavernous approaches
to, 109121

in suprasellar and parasellar region,
109122
lateral endonasal approach (Hirsch) to, 12
microprolactinoma, 8992
nonfunctioning, 170
orbitozygomatic approach to, 109114
123145
preoperative considerations in, 94
pterional approach to, 109111
stereotactic radiosurgery. See Stereotactic
radiosurgery.
sublabial transsphenoidal approach
(Cushing) to, 24, 7
supraorbital approach to, 119121
transcranial approaches to, 45, 103104,
109122
transsphenoidal approach to. See
Transsphenoidal approach.
transzygomatic approach to, 109111
thyrotropinsecreting. See Thyrotroph
adenoma.
treatment of
long-term, 167171
medical. See Pituitary tumors, medical
treatment of.
microprolactinoma, 8992
persistent functional, 139145
radiation therapy. See Radiation therapy,
for pituitary tumors.
surgical. See Pituitary tumors, surgical
treatment of.
transsphenoidal approach to. See
Transsphenoidal approach.
versus hyperplasia, 61
Plasmacytoma, clival, imaging of, 78
Plasticity, supraspinal, in nociception, 455
Pleomorphism, in central neurocytoma, 486487
Plurihormonal adenoma, 2829, 36
Polyethyleneimine, for gene transfer, 421
Polyneuropathy, in acromegaly, 44
Polyuria, in diabetes insipidus, after pituitary
surgery, 125
Positron emission tomography
in neurocytoma, 484
246248
649
Posterior commissure, anatomy of, in endoscopic

surgery, 549
Precentral stimulation, 334, 437443
literature review on, 439441
mechanism of, 437438
trial stimulation in, 439
Prefrontal lobotomy, history of, 182190
Pregnancy, in acromegaly, 45
Preproenkephalin gene transfer, for pain
management, 428429
Preprogalanin gene transfer, for pain
management, 429
Primary aerent depolarization, in
Prolactin, 20, 25
excess of. See Hyperprolactinemia.
in acidophilic stem cell adenoma, 27
in mammosomatotroph adenoma, 2728
in somatotroph-lactotroph adenoma, 27
Prolactinoma
diagnosis of, 4243
endocrinopathy in, 42
hyperprolactinemia in, 2931, 4142
persistent, 142143
treatment of, 8182
imaging of, 42, 58
medical treatment of, 8182, 8990, 143, 167
outcome of, 167168
pathology of, 2931
persistent, 142143
surgical treatment of, 90, 143
micro-, 8992
outcome of, 167168
131, 134
Proopiomelanocortin, 17, 20
Proopiomelanocortin gene transfer, for pain,
427428, 430
Prostate hyperplasia, in acromegaly, 44
Protein kinase C, in pain transmission, 452
Proton beam radiosurgery, for pituitary tumors,
154155
650
Pseudorosettes, in ependymoma, 472473, 573

Psychiatric disorders
evaluation of, before pain management,
411412, 414415
in acromegaly, 45
Psychiatric neurosurgery. See also specic
procedures, e.g., Cingulotomy.
for depression. See Depression, neurosurgery
for.
for obsessive-compulsive disorder. See
Obsessive-compulsive disorder,
neurosurgery for.
for thalamocortical dysrhythmia, 251265
historical perspective of, 181197, 202203
capsulotomy, 191192
cingulotomy, 191192
early years, 181190
ethical aspects of, 303308
future techniques derived from, 193196
limbic leukotomy, 193
stereotactic methods, 190193
subcaudate tractotomy, 193
neuroimaging and neurocircuitry models for,
213223
palliative, 316
research on
ethical considerations in, 303319
regulation of, 312315
single-photon emission computed tomography
in, 237250
sociopolitical aspects of, 308312
transcranial magnetic stimulation, 283301
Psychologic evaluation, for neuroaugmentation,
409417
contraindications identied in, 413414
drug administration, 410412
factors in, 414416
instruments for, 415416
need for, 409410
personnel for, 412
procedure for, 412413
spinal cord stimulation, 410412
Psychological issues, in pain management
deep brain stimulation with, 391
depression, 348, 411412, 414415
intrathecal, 354
personal relationships, 348
preoperative psychologic evaluation,
409417
psychologist report review, 343345
return to work, 347348
substance abuse, 348
Pterional approach
to parasellar and suprasellar lesions, 109111
to pituitary tumors, 104
Pumps, for intrathecal pain management,
356359, 383384
Q
Quadrigeminal cistern, anatomy of, in third
Quinagolide, for hyperprolactinemia, 8182
R
Rabies-G pseudotyped lentivirus, for gene
transfer, for pain management, 425
Radiation therapy
for choroid plexus carcinoma, 627628
for ependymoma, 576
for intraventricular metastasis, 600602
for meningioma, 563
conformal techniques for, 151153
conventional techniques for, 149151
ecacy of, 147149
external beam, 147153, 161162
nonfunctioning, 170
planning for, 149
prolactinoma, 143
somatotroph adenoma, 140141, 168
stereotactic. See Stereotactic radiosurgery.
thyrotroph adenoma, 144
toxicity of, 148, 161163
for subependymoma, 583584
Radiosurgery
for intraventricular meningioma, 563
gamma knife
for somatotroph adenoma, 83
linear accelerator-based
proton beam, for pituitary tumors, 154155
stereotactic. See Stereotactic radiosurgery.
Raspberry, William, on psychiatric neurosurgery
regulations, 314
Rathkes cleft cyst, imaging of, 63
Rathkes pouch, in pituitary embryogenesis,

1112
Regulation, of neuropsychiatric surgery research,
312315
Renal cell carcinoma, intraventricular metastasis
from, 595598
Respiratory depression, in intrathecal pain
management, 358
Respiratory disorders, in acromegaly, 44
Retroviruses, for gene transfer, for pain
management, 425
651
Sensitization
central, 450452
peripheral, 449450
Sensory disorders
Septum pellucidum
anatomy of, 511
variants of, 608609
Seroma, in intrathecal medication delivery, 358
Shock therapies, for psychiatric disorders, 184
Rous sarcoma virus, for gene transfer, for pain

management, 422
Single-photon emission computed tomography

in depression, in vagus nerve stimulation, 279
in neurocytoma, 484
diagnostic, 238, 241242
postoperative, 238241, 245249
Ryan, J. Kenneth, on psychiatric neurosurgery, 313
Skin disorders, in acromegaly, 44
Richardson, Alan, limbic leukotomy procedure

of, 193
Rosettes, in ependymoma, 473, 573
Skull base, anterior, exposure of, in pituitary

surgery, 100101
Sacral pain, spinal cord stimulation for, 378
Smoking, pain management and, 343
Saline solution, for antidiuretic hormone

Social Security programs, for chronic pain,

349350
Salt wasting, cerebral, after pituitary surgery, 128
Somatosensory cortex, in nociception, 448
Sandostatin
Somatosensory evoked potentials, in precentral

stimulation, 439
Sarcoidosis, of pituitary, 71
Somatostatin analogues
for thyrotroph adenoma, 84, 144
Schizophrenia
lobotomy for, 308
Schwannoma, parasellar, imaging of, 70
Scintigraphy, in somatotroph adenoma, 45
Secondary gain, in pain
chronic, 349
evaluation of, before management, 415
Seizures. See also Epilepsy.
in precentral stimulation, 438, 441
induction of, in transcranial magnetic
stimulation, 294295
Selective serotonin reuptake inhibitors, for
obsessive-compulsive disorder, 200
Sellar tumors, anatomic considerations in,
1516, 57
Somatostatin, intrathecal delivery of, 361
Somatotroph(s), 25
Somatotroph adenoma
acromegaly in, 4345
persistent, 139141
densely granulated, 2627
diagnosis of, 45
mixed cell, 2729
outcome of, 168169
pathology of, 2629
persistent, 139141
sparsely granulated, 27
652
Somatotroph (continued )
persistent, 139140
131, 134
prolactinoma, 9091
proton beam, 154155
somatotroph adenoma, 141
toxicity of, 162163
SPECT. See Single-photon emission computed

tomography.
Stereotactic surgery
cingulotomy. See Cingulotomy.
for pain, 328, 438439
history of, 190193
Speech disorders, in lateral ventricle tumor

surgery, 523524
Stereotactic systems, in lateral ventricle tumors,

514
Spinal cord
compression of, in intrathecal medication
delivery, 358
dorsal horn of, reorganization of, in
sensitization, 451
dorsal root ganglia of, in nociception, 445446
laminar organization of, nociception and,
446447
Stimulation, electrical. See Electrical stimulation.
Somatotroph-lactotroph adenoma, 27
Spinal cord stimulation, 365380

advantages of, 366
applications of, 366367, 377378
equipment for, 371372
in neuropathic pain, 378
in peripheral nerve stimulation, 377378
failure of, 377, 410
history of, 331332, 365
lead implants in, 372374
outcomes of, 367, 369370, 409412
psychologic variables in, 409412
screening trial in, 371
terminology of, 366
Spinal nucleus, in nociception, 447
Spindle cells, in subependymal giant cell
astrocytoma, 475476, 584
Spinothalamic tract
interruption of, for pain, 328
lateral, in nociception, 446447
Stroke, after pituitary radiotherapy, 162163

Subcaudate tractotomy, for psychiatric disorders,
193
depression, 204205, 207, 218
218, 231232
Subcutaneous implantation, for peripheral nerve
stimulation, 403406
Subdural hematoma, in third ventricular tumor
surgery, 542
Subependymal giant cell astrocytoma, 470471,
475476, 584585
Subependymal glomerate astrocytoma, 581582
Subependymoma, intraventricular, 578584
asymptomatic, 579
pathology of, 474475, 579582
prognosis for, 583584
radiography in, 579
Subfrontal approach, anterior, to pituitary
tumors, 104
Stereotactic biopsy, in third ventricular tumors,

528
Sublabial transsphenoidal approach, to pituitary

tumors, 24, 7, 9499
Stereotactic radiosurgery
for intraventricular metastasis, 601602
gamma knife, 83, 156161
historical view of, 6, 8
linear accelerator-based, 155156
nonfunctioning, 170
Subnucleus caudalis, in nociception, 447

Suboccipital approach, midline, to
intraventricular meningioma, 566, 568
Substance abuse
evaluation of, before pain management, 415
treatment of, before pain management, 348
Substance P, in nociception, 446, 452, 454
Substantia gelatinosa, in nociception, 446

Superior parietal occipital approach, to
Superoxide dismutase gene transfer, for pain
management, 430
Supraorbital approach, to parasellar and
Supraorbital pain, peripheral nerve stimulation
for, 405
Suprasellar approach, to third ventricular tumors,
532538
Suprasellar tumors. See Parasellar and suprasellar
lesions.
Sympathetic nervous system, peripheral
sensitization in, 450
653
magnetoencephalography with, 253256,

260
results of, 255, 259263
dorsomedial, for psychiatric disorders, 191
Thalamus
anatomy of, 511, 529
connections to
218219
electrical stimulation of. See Deep brain
stimulation.
nuclei of, in nociception, 447448
ventrobasal, electrical stimulation of, 332333
Thermal injury, central sensitization in, 450451
Synaptophysin
in neurocytoma, 485
Thermocapsulotomy, for psychiatric disorders,

204208
Syndrome of inappropriate antidiuretic hormone

secretion, after pituitary surgery, 126129
Third ventricular tumors, 527545

endoscopic resection of, 551552, 554555
hydrocephalus in, treatment of, 528
pathology of, 527
radiography in, 528
surgical approaches to
aims of, 527528
infratentorial supracerebellar, 532538
interhemispheric transcallosal, 532, 534,
538539
intraoperative considerations in, 541
occipital transtentorial, 532, 539541
positioning for, 533534
postoperative care in, 542
selection of, 531532
T
Talairach capsulotomy procedure, 191
Tegmentum, anatomy of, in endoscopic surgery,
549
Tela choroidea, anatomy of, in third ventricular
surgery, 531
Temporal gyrus, middle, approach to, to
Temporal horn, anatomy of, 512
Temporal transcortical (middle temporal gyrus)
Tetracycline, for gene transfer, for pain
management, 422
Thalamocortical dysrhythmia, 251265
pathophysiology of, 251, 259, 262
surgery for, 252263
magnetoencephalography with, 253256,
260
results of, 255, 259263
Thalamostriate vein, anatomy of, 512513
Thalamotomy
central lateral, for thalamocortical
dysrhythmia, 252263
Thiazide diuretics, for diabetes insipidus, 132
Thirst, in diabetes insipidus, after pituitary

surgery, 125
Thyroid stimulating hormone, 19, 25
pituitary tumors secreting. See Thyrotroph
adenoma.
preoperative management of, in pituitary
tumors, 94
Thyrotroph(s), 25
Thyrotroph adenoma
diagnosis of, 4950
hyperthyroidism in, 4850
654
Thyrotroph (continued )
pathology of, 3233
persistent, 143
131132, 134
preoperative management in, 94
Tics, transcranial magnetic stimulation for, 293
Tourette syndrome
neuroimaging and neurocircuitry in, 214
Tractotomy, subcaudate, for psychiatric
disorders, 193
depression, 204205, 207, 218
218, 231232
Transcallosal approaches
to lateral ventricle tumors
anterior, 520521
posterior, 521522
to third ventricular tumors, 532, 534, 538539
Transcortical approaches, to lateral ventricle
tumors, 517520
Transcranial approaches
to parasellar and suprasellar lesions, 109122
to pituitary tumors, 45, 9394, 103104
Transcranial magnetic stimulation, 283301
animal models of, 287
drug use with, 293
electrical requirements of, 285
for anxiety disorders, 293294
for depression, 288293
maintenance, 292
meta-analysis of, 289292
with intentional seizure generation, 294295
for eloquent cortex localization, 287
for epilepsy, 286287, 293
for mania, 292
for movement disorders, 293
for schizophrenia, 293
functional imaging with, 287288
high-frequency, 285286
mechanisms of action of, 284286
paired-pulse, 286287
repetitive, 284
research uses of, 286288
safety of, 286
versus electroconvulsive therapy, 287289,
291292, 294
Transmaxillary transsphenoidal approach,

102103
Transorbital lobotomy, 189
Transposons, for gene transfer, 421
Transsphenoidal approach, to pituitary tumors,
93103
anterior skull base exposure in, 100101
cavernous sinus exposure in, 102
clivus exposure in, 101102
combined transmaxillary, 102103
Cushing technique for, 25, 7
Dott technique for, 57
endonasal, 12, 99100
endoscope-assisted, 104105
extended, 100102
Guiot technique for, 6, 8
Hardy technique for, 810
Hirsch technique for, 12
historical overview of, 110
sublabial, 9499
Transtentorial approach, to third ventricular
tumors, 532, 539541
Transzygomatic approach, to parasellar and
Trichotillomania, neuroimaging and
neurocircuitry in, 214
Trigeminal mesencephalic nucleus, in nociception,
447
Trigeminal nerve, in nociception, 446447
Trigeminothalamic tract, in nociception, 447
Tuber cinereum, anatomy of, in endoscopic
surgery, 549
Tuberculosis, of pituitary, 71
Tuberin, in subependymal giant cell astrocytoma,
476
Tuberous sclerosis complex, 615617
subependymal giant cell astrocytoma in,
475476, 584585
Tumor(s)
intraventricular. See Intraventricular tumors.
pituitary. See Pituitary tumors.
U
Ultrasonography, in lateral ventricle tumors, 514
Urinary free cortisol test, in corticotroph
adenoma, 46
V
Vagus nerve, electrical stimulation of
clinical applications of, 280
for depression, 275282
clinical trials of, 276
long-term results of, 278279
mechanism of action of, 277, 279280
open-label study of, 277278
placebo-controlled study of, 278279
practical considerations in, 280281
preclinical investigations of, 275276
Vesiculostomatitis virus, for gene transfer, for

pain management, 425
Vimentin
in ependymoma, 473
Viral vectors, for gene transfer, for pain
management, 421425
Valentine, Elliot S., on psychiatric neurosurgery,

312
Visual disorders
after pituitary radiotherapy, 162163
in neurocytoma, 483
in prolactinoma, 29
Vascular proliferation, in central neurocytoma,

487
Vasopressin. See Antidiuretic hormone.
Watts, James, psychiatric neurosurgical

procedures of, 185187, 189
Vein of Galen, anatomy of, in third ventricular

surgery, 531
Velum interpositum, anatomy of, in third
Ventricles, tumors in. See Intraventricular
tumors.
Ventriculoscopy. See Endoscopy, in
intraventricular surgery.
Ventriculostomy, third, endoscopic, 551552
Ventroposterior nucleus, in nociception, 447
Weinberger, Lawrence M., on ethics of

neuropsychiatric surgery, 306
Wide dynamic range cells, in nociception,
446447
Withdrawal reactions, in intrathecal pain
management, 359
Work, return to, after pain management,
psychosocial issues in, 347348
Workers compensation systems, pain
management and, 341, 349350
655

03.04 - Intraventricular Tumours

Uploaded by

Document Information

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

03.04 - Intraventricular Tumours

Uploaded by

Copyright:

Available Formats

INTRAVENTRICULAR TUMORS

Intraventricular tumors present a diagnostic challenge to the clinician because of a broad

NUMBER 4 OCTOBER 2003

Surgical Approaches to Tumors of the Lateral Ventricle

Surgical Approaches to Posterior Third Ventricular Tumors

Endoscopic Adjuncts to Intraventricular Surgery

Recently, endoscopic intraventricular surgery has been performed successfully in several

Surgical Resection of Metastatic Intraventricular Tumors

Intraventricular metastases are a unique challenge for neurosurgical oncologists. This

Intraventricular Congenital Lesions and Colloid Cysts

Choroid Plexus Tumors in Children

Cumulative Index 2003

THE CLINICS ARE NOW AVAILABLE ONLINE!

Neurosurg Clin N Am 14 (2003) ix

Andrew T. Parsa, MD, PhD

twenty-rst century. These lesions have a complex

Neurosurg Clin N Am 14 (2003) 469482

Epidemiology and pathology of intraventricular tumors

Intraventricular tumors present a diagnostic

dominantly occurs in childhood, although it can

J.S. Waldron, T. Tihan / Neurosurg Clin N Am 14 (2003) 469482

taneously positive for cytokeratin [17]. S-100

J.S. Waldron, T. Tihan / Neurosurg Clin N Am 14 (2003) 469482

J.S. Waldron, T. Tihan / Neurosurg Clin N Am 14 (2003) 469482

Molecular and genetic features

of overlapping histologic, ultrastructural, and

J.S. Waldron, T. Tihan / Neurosurg Clin N Am 14 (2003) 469482

cellularity, increased mitotic gures, and vascular

J.S. Waldron, T. Tihan / Neurosurg Clin N Am 14 (2003) 469482

hyaloplasmic lipid vacuoles can also contribute to

issue. Another diagnostic consideration is the

Pathologic dierential diagnosis

Formulation of the dierential diagnosis for

Subependymomas are slow-growing, benign

J.S. Waldron, T. Tihan / Neurosurg Clin N Am 14 (2003) 469482

histologic features in larger and symptomatic

seen in ventricular ependymomas. In general,

J.S. Waldron, T. Tihan / Neurosurg Clin N Am 14 (2003) 469482

clusters and occasional pseudorosettes [62]. The

J.S. Waldron, T. Tihan / Neurosurg Clin N Am 14 (2003) 469482

Rare cases have been reported in the third and

J.S. Waldron, T. Tihan / Neurosurg Clin N Am 14 (2003) 469482

role in their pathogenesis. In addition, N-myc and

tricular clear cell meningioma [93] and malignant

Other tumors and tumor-like lesions within the

J.S. Waldron, T. Tihan / Neurosurg Clin N Am 14 (2003) 469482

[1] Janisch W, Staneczek W. Primary tumors of the

choroid plexus papilloma. Acta Neuropathol

J.S. Waldron, T. Tihan / Neurosurg Clin N Am 14 (2003) 469482

J.S. Waldron, T. Tihan / Neurosurg Clin N Am 14 (2003) 469482

[87] Homan JC, Jr, Bufkin WJ, Richardson HD.

J.S. Waldron, T. Tihan / Neurosurg Clin N Am 14 (2003) 469482

[106] Gratton RJ, Hogge WA, Aston CE. Choroid plexus

the foramen of Monro. Br J Neurosurg 2002;16(1):

Neurosurg Clin N Am 14 (2003) 483508

J. Lee was supported in part by a grant from the

of neurocytoma have previously been reported as

J. Lee et al / Neurosurg Clin N Am 14 (2003) 483508

Abbreviations: HA, headache or reported as signs and

attachment site, MRI is preferred [15,46,97,106].

J. Lee et al / Neurosurg Clin N Am 14 (2003) 483508

ependymoma, and CNC. A denitive diagnosis

tochemistry or electron microscopy is required to

J. Lee et al / Neurosurg Clin N Am 14 (2003) 483508