See

discussions, stats, and author profiles for this publication at:

https://www.researchgate.net/publication/22713998

Risk Factor Analysis in Differentiated

Thyroid Cancer

Article in Cancer March 1979

DOI: 10.1002/1097-0142(197903)43:3<810::AID-CNCR2820430306>3.0.CO;2-B Source: PubMed

CITATIONS READS

232 476

6 authors, including:

Blake Cady
Harvard Medical School
361 PUBLICATIONS 12,728 CITATIONS

SEE PROFILE

All content following this page was uploaded by Blake Cady on 16 September 2014.

The user has requested enhancement of the downloaded file. All in-text references underlined in blue are added to the original document
and are linked to publications on ResearchGate, letting you access and read them immediately.
 RISK FACTOR ANALYSIS IN DIFFERENTIATED
THYROID CANCER
BLAKECADY,MD, CORNELIUSE. SEDGWICK,MD, WILLIAMA. MEISSNER,MD,
MARVINS. WOOL, MD, FERDINAND MD, A N D J O A N WERBER,A B
A. SALZMAN,

Six hundred patients with primary differentiated thyroid carcinoma had fol-
low-up studies for a minimum of 15 years and a maximum of 45 years. Re-
currence rate and death rate were significantly different in defined high-risk
and low-risk groups of patients. These basic risk groups were defined by
age and sex alone; low risk consisted of men 40 years of age and younger
and women 50 years of age and younger whereas the high-risk group were
older patients. Recurrence and death rates in patients at high risk were 33%
and 27% while respective figures for patients at low risk were 11% and 4%.
In more recent years these results have shown significant improvement. Basic
risk group definition outweighed the effect of pathologic type, local disease
extension, type of treatment, and site of recurrence or metastasis. For instance,
radioactive iodine cured 70% of patients at low risk with metastatic disease
but only 10% of patients at high risk. Less aggressive biologic behavior of
thyroid cancer before the age of menopause implies that an estrogen-rich
milieu may alter the effects of initiating and promoting factors in carcinogene-
sis. It also suggests that therapeutic trials of estrogen be undertaken in pro-
gressive metastatic differentiated thyroid cancer.
Cancer 43:810-820, 1979.

T HYROID CANCER HAS AROUSED more atten-

tion than its frequency would warrant
because of unique biologic behavior patterns
group definition in differentiated thyroid
cancer and methods and results of therapy
for recurrences and metastases. Risk group
ranging from extremely indolent, essentially analysis explains in part the biologic variabil-
nonlethal varieties to forms that are uniformly ity of thyroid cancer and indicates that at
and rapidly fatal, all arising from the same least two varieties of differentiated thyroid
follicular cell. The frequency of thyroid can- cancer occur in women, with the distinction
cer in young persons, its association with low- probably related to the hormonal milieu at
dose irradiation given in childhood, and the induction or promotion of carcinogenesis.
ability to treat metastatic disease with thyroid These unique induction or promotion char-
hormone suppression and specific tumor seek- acteristics seem to hold for the life of the
ing radioactive iodine have heightened inter- patient.
est to endocrinologists, oncologists, epidemi-
ologists, and surgeons. MATERIALS
A N D METHODS
In recent a n a l y s e ~ , ~we
~ ' ~have reported
changing patterns of clinical presentation, Between 1931 and 1960,631 differentiated
therapy, and results in patients with primary thyroid cancers were primarily treated at the
differentiated and undifferentiated thyroid Lahey Clinic Foundation. Pathology material
cancers. This report adds analysis of risk was available for review in 1975. Histologic
definitions conform entirely to the Armed
Presented at the 31st Annual Meeting of T h e Society Forces Institute of Pathology fascicle on thy-
of Surgical Oncology, San Diego, California, April 2 - 6 ,
1978.
roid disease,I3 the World Health Organiza-
From the Departments of Surgery, Internal Medicine, tion nomenclature,10and the American Joint
and Radiotherapy, Lahey Clinic Foundation and Ldbora- Committee for Cancer Staging and End-Re-
tory of Pathology, New England Deaconess Hospital, sults Reporting"; 98% had follow-up infor-
Boston, Massachusetts. mation for a minimum of 15 years. Cancer
Address for reprints: Blake Cady, MD, Department of
Surgery, Lahey Clinic Foundation, 605 Commonwealth was defined as follicular if the histologic pat-
Avenue, Boston, MA 02215. tern was entirely of a follicular pattern and
Accepted for publication November 6, 1978. included all subcategories, such as Hiirthle
0008-543X/79/0300/0810 $1.05 0 American Cancer Society
810
No. 3 T H Y R OCA
~DRISKFAcTok ANALYSIS Cady et al. 811

1. Risk of Recurrence and Death by Decade

TABLE

Death

As percent of As percent of
Total Recurrence No. Total Recurrence

Men* <31) 28 5 (18%) 3 11% 60%

3 1-40 34 4 (12%) 2 6% 50%
41-50 31 10 (32%) 5 16% 50 %
51-60 25 8 (32%) 6 24% 75%
>6 1 14 6 (43%) 4 29% 67%
Woment <30 115 17 (15%) 3 3% 18%
31-40 122 I 1 (9%) 5 4% 46%
41-50 85 5 (6%) 2 2% 40%
51-60 80 23 (28%) 21 24% 88%
>6 1 66 24 (36%) 22 33% 92%
rOTAL 600 113 73

* p = 0.02 recurrence; p = NS deaths.

tp = 0.001 recurrence; p = 0.001 deaths.

cell carcinoma. All other forms of differen-
tiated thyroid cancer ranging from pure pap-
illary to predominantly follicular but with the
presence of some papillary features were
mixed papilla9 and follicular. No patients had
any other histology except for four patients
whb died with anaplastic carcinoma, two of
whom had small areas of anaplastic carcinoma
in the original mixed papillary and follicular
carcinoma.
First disease recurrence was categorized as
local recurrence if ih the thyroid bed, residual
thyroid tissue, trachea, esophagus, or midline
neck in the operative area, nodal metastasis if
in ipsilateral or contralateral cervical areas or
upper mediastinum, and distant metastases if
disease was defined outside the neck, thyroid,
or upper mediastinal area. If disease reap-
peared in more than one site simultaneously,
it was categorized as the location most remote
from the thyroid area. Second and third re-
currences of disease were defined as progres-
sion or new recrudescence after a period of
regression or successful therapy or disease ap-
pearing in a new site.
Thirty-one patients initially seen with either
metastatic or locally unresectable differenti-
ated thyroid carcinoma were eliminated from
consideration because this presentation is so
uncommon in current practice4 (f1%). Of

2. Risk Group Definition

TABLE

Total
Recurrence Death Percent of
Inci- recurrences that
No. dence No. Rate No. Rate die of disease

Highest risk group*

Follicular 83 14% 33 40% 30 36% 91%
Men >40
Women >50
Intermediate risk group*
Papillary 133 22% 38 29% 28 21% 74 %
Men >40
Women >50
Low risk group*
Follicular or papillary 384 64% 42 11% 15 4% 36%
Men <40
Women <50
TOTAL 600 113 19% 73 12% 65%

* High vs. intermediate: NS recurrence; p = O.b2 p = 0.001 death. High + intermediate vs. low: p = 0.001
death. High vs. low: p = 0.001 recurrence; p = 0.001 recurrence; p = 0.001 death.
death. Intermediate vs. low: p = 0.001 recurrence;
812 CANCER
March 1979 VOl. 43

TABLE 3. Changes in Sex Composition of Risk Groups by Time Period

193 1- 1950 1951- 1960 1961- 1970

No. Percent No. Percent No. Percent

Low risk*
Men 29 16% 33 17% 17 18%
Women 157 84% 165 83% 77 82%
High risk?
Men 29 26% 41 39% 37 55%
Women 82 74% 64 61% 30 45%
Women$
High risk 82 34% 64 28% 30 28%
Low risk 157 66% 165 72% 77 72%
Men5
High risk 29 50% 41 55% 37 69%
Low risk 29 50% 33 45% 17 31%
Proportion of men
in series 58 1b% 74 24% 54 34%

* Men vs. women: NS. $ High vs. low: NS.

t Men vs. women: p = 0.0005 P High vs. low: NS.

these 31 patients, 21 (68%) died of disease.
Three of these deaths occurred postopera-
tively after vigorous attempts to remove lo-
cally unresectable disease and were the only
postoperative deaths in the entire group. Pa-
tients with such advanced disease would pro-
duce a markedly unfavorable bias in analyzing
basic risk group definitions that should be of
wide application.
Six hundred patients, therefore, constituted
the study group and had surgically resectable
and potentially curable differentiated thyroid
carcinoma. Of these 600 patients, disease re-
curred or metastases developed in 113, and
73 died of disease.
Patients were considered dead of disease if
they had active disease when last seen. Thy-
roid cancer was the usual but not invariable
direct cause of death. All other patients were
documented as living disease free or lost to
follow-up study. Statistical methods were
standard chi square or Fisher exact test un-
less otherwise noted.
RESULTS
Clear correlations between age, sex, path-
ologic type, and results of therapy were ap-
parent. The results of treatment by age dec-
ade at initial operation were analyzed for both
men and women (Table 1). Two striking fea-
tures of this analysis, which ignore the particu-
lar pathologic type, are the sharp increase in
the risk of recurrence and death and the
lethality of recurrences that occur at age 50
in women who constitute 78% of the cases.
Thus, among women less than 50 years of
age, the overall risk of recurrence was 10%
and risk of death but 3%. Of patients in whom
disease recurred or metastases developed,
only 30% died of disease. T h e respective fig-
ures for women over 50 were 32%, 30%, and
89%. These changes with age were highly
significant statistically for women and of bor-
derline significance in terms of recurrence for
men. Changes in death rate were not signifi-
cant in men.
When the two biologically distinctive histo-
logic groups (follicular and mixed papillary
and follicular) were combined with age and
sex, three basic risk groups became apparent
(Table 2). Those at highest risk consisted of
men over 40 years of age and women over
50 years of age with follicular carcinoma.
These 83 patients constituted 14% of the se-
ries and had a 40% recurrence rate. The inter-
mediate risk group (133 patients or 22% of
series) were men over 40 years of age and
w,omen over 50 years of age with mixed pap-
illary and follicular carcinoma. The low-risk
group, consisting of all men 40 years of age
and under and all women 50 years of age
and under, regardless of pathologic type of
tumor, constituted 64% of the series and had
only an 11% recurrence rate.
The variation in risk is emphasized even
more by the overall death rates of 36%, 21%,
and 4% for the high-risk, intermediate-risk,
and low-risk groups, respectively. Further-
more, the mortality rates of patients with re-
current disease for the three risk groups are
91%, 74%, and 36%, respectively. No differ-
No. 3 THYROID ANALYSIS. Cady et al.
CA RISKFACTOR
ence in recurrence or death rates was seen
within the low-risk group based on histologic
type of tumor, although the particular pattern
of recurrence of disease in terms of location
and incidence of metastases was distinctive for
the two histologic varieties.*
When analyzed statistically by means of chi
square with Yates continuity correction, no
significant differences in results could be
demonstrated between the highest and inter-
mediate risk groups. However, highly signifi-
cant differences were present in terms of re-
currence and death rates when highest or
intermediate group or these groups combined
were compared with the low-risk group.
Therefore, because of basic similarities in re-
currence rates, death rates, and lethality of
recurrences or metastases, high-risk and low-
risk groups defined by age and sex alone, re-
gardless of pathologic type of tumor, are logi-
cal and pertinent. In some of the analyses that
follow, the three defined risk groups will be
utilized while for others only the two major
risk groups will be compared.
Table 3 displays an interesting change in
differentiated thyroid cancer over four dec-
ades. For this analysis, 161 patients encoun-
tered during the years 1961 to 1970 were
added. The proportion of men rose from 19%
in the years before 1950 to 34% in the 1960s.
Within the low-risk group, the proportion of
men and women remained exactly the same
over the 40 years. In the high-risk group, how-
ever, the proportion of men rose dramatically
and significantly from 26% to 55%. Essen-
tially the same proportion of women occurred
at high risk and at low risk throughout the
four decades, but the proportion of men who
were older, and therefore at high risk, in-
creased progressively in time so that now more
4. Chronology of Recurrence and Death
TABLE
5 10
Recurrence*
Highest risk group 56%
Intermediate risk group 55%
Low risk group 48%
Deathst
Highest risk group 43%
Intermediate risk group 39%
Low risk group 27%
~~~
* NS.
t NS except high + intermediate vs. low p < 0.001.
813
than two-thirds of men fall in the high-risk
group.
Differentiated thyroid cancer frequently
displays a protracted clinical course even
when recurrent or metastatic. Table 4 docu-
ments the percentages of all recurrences and
deaths that have occurred by five-year follow-
up intervals. A slowing of the biologic pat-
tern is apparent in the low-risk group so that
the few deaths that do occur evolve over a
more protracted period.
Table 5 illustrates the recurrence and death
rates for the three risk groups by extent of
primary cancer. lntraglandular cancers had
no evidence of penetration of the thyroid
capsule by either macroscopic or pathologic
estimations; extraglandular disease had both
pathologic and clinical evidence of carcinoma-
tous involvement of skeletal muscles, nerves,
trachea, esophagus, or surrounding soft tis-
sue. The extent of the primary cancer had no
relationship to recurrence or death rates in
the low-risk group, whereas in the intermedi-
ate-risk or high-risk groups, both the recur-
rence rates and death rates were significantly
worse with extrathyroid primary disease.
Table 6 illustrates the relationship between
lymph node metastases removed at the pri-
mary operation and later recurrence of dis-
ease and death as a function of risk group;
47% of patients in the low-risk category and
33% of patients in the high-risk category had
neck dissections performed. Of these patients,
5 1% of patients in the high-risk group but 74%
of patients in the low-risk group actually had
nodal metastases histologically. In the patients
in the low-risk group who did not have a neck
dissection performed, disease recurred in 22
patients, and in 8 of these patients (36%) re-
current disease included cervical node metas-
Years after primary curative surgery
15 20 25 30
88% 97% 97% 100% 100%
79% 100% 100% 100% 100%
83% 93 % 95% 100% 100%
77% 90% 97% 100% 100%
61% 82% 96% 100% 100%
47% 67% 73 % 93 % 100%
814 CANCERMarch 1979 Vol. 43

TABLE5. Recurrence and Death as a Function of Disease Extension

Low risk group* Intermediate risk groupt High risk groupt

Recur- Recur- Recur-

Total rence Death Total rence Death Total rence Death

Intraglandular 334 36 ( 1 1 % ) 12 (4%) 103 18 (17%) 12 (12%) 68 23 (34%) 21 (31%)

Percent recur-
rences dying 33% 67% 91%
Extraglandular 50 6(12%) 2(4%) 30 20(67%) 16(53%) 15 10(67%) 9(60%)
Percent recur-
rences dying 33% 80% 90%
TOTAL 384 133 83

* NS. t Combined intermediate + high. Intraglandular vs.

extraglandular p < 0.001 recurrence; p < 0.001 death.

tases. Only 1 of these 8 patients (13%) died
of disease whereas 10 of the 14 patients (7 1%)
with recurrent disease not including lymph
node metastases died of their disease-a sig-
nificant difference statistically. Overall only
1.5% of patients with lymph node metastases
detected in the primary neck dissection died
of disease whereas 4% of patients without
nodal metastases initially died of disease.
In the high-risk group recorded in Table 6,
no relationship between node metastases and
survival was detectable; 17 of 25 patients
(68%) who had recurrent disease with nodal
metastases died whereas 41 of 46 patients
(89%) who had recurrent disease without
nodal metastases died, a difference that was
not significant.
Other interesting aspects of the relation-
ship between nodal metastases, survival, and
risk groups abound, all emphasizing the bet-
ter prognosis of patients with nodal metastases
in the low-risk group. For instance, in the low-
risk group, 49 of 196 patients (25%) seen in
the years 1951 to 1960 initially presented with
palpable nodal metastases. Only 5 of these
49 patients (10%) who had therapeutic neck
dissections had recurrent disease, and none of
these involved neck nodes. Only one of these
five patients died of disease. Thus, 48 of 49
patients (98%) in the low-risk group who had
palpable nodal metastases at initial presenta-
tion were cured. In the high-risk group, by
contrast, only 9 of 103 patients (9%) initially
presented with palpable nodal metastases and
therefore underwent therapeutic nodal dis-
section; 4 of these 9 patients (44%) had re-
current disease, 2 in cervical nodes. Only one
of these four patients survived. Overall, three
of nine patients (33%) presenting with palpa-
ble node metastases died of disease, a death
rate no worse than the group as a whole.
Table 7 illustrates the location and lethality
of first recurrences in the three risk groups.
In low-risk patients, 4 1% had recurrent dis-
ease in the form of lymph node metastases
either alone or with accompanying local re-
currence; 12% of these patients eventually
died of disease. Only 63% of patients in the
low-risk group with distant metastases even-
tually died of disease in contrast to the nearly
uniform deaths of patients in the intermedi-
ate-risk or high-risk group. In the highest
risk group, node metastases were uncommon
as a first recurrence (9%) and conveyed a
much worse prognosis as two of three such
patients died of disease. Although the pattern

6. Effect of Lymph Node Metastases

TABLE

Low risk High risk

Recur- Recur-
Total rence Death Total rence Death

Node dissection
Negative nodes 47 2 (4%) 2 (4%) 35 11(31%) 9(26%)
Positive nodes 134 18(13%) 2(1.5%) 37 13(35%) 12(32%)
N o node dissection
All cases 203 22(11%) 1 1 (5%) 144 47(33%) 37(26%)
No. 3 CA RISKFACTORANALYSIS *
THYROID Cady et al. 815

7. Survival Related to Risk Group

TABLE
~ ~ ~ ~~~~

Local Node Distant

Total recurrence* metastases* metastasest

Highest risk group

Recurrence 33 3 (9%) 3 (9%) 27 (82%)
Deaths 2 (67%) 2 (67%) 26 (96%)
Intermediate risk group
Recurrence 38 12 (32%) 12 (32%) 14 (37%)
Deaths 9 (75%) 6 (50%) 13 (93%)
Low risk group
Recurrence 42 9 (21%) 17 (41%) 16 (38%)
Deaths 3 (33%) 2 (12%) 10 (63%)
TOTAL 113 24 (21%) 32 (28%) 57 (50%)

* NS.
t p = 0.006.

of recrudescence (mixed papillary and folli- Thirty-four of 42 patients in the low-risk

cular) in the intermediate-risk group was simi- group (81%) but only 48 of 71 patients in
lar to that in the low-risk group, the lethality the high-risk group (68%)were actually treated
of each form of recurrence was similar to that by surgery or radiotherapy for recurrences.
found in the high-risk group; thus, 93% of Patients not classified as treated were given
patients with distant metastases, 50% of pa- thyroid medication after testing for radio-
tients with nodal metastases, and 75% of pa- active iodine uptake by scans. One patient in
tients with local recurrences died. These fig- the low-risk group not treated survived for
ures reemphasize the overriding importance 25 years with stable proven pulmonary metas-
of age and sex definition of patients in risk tases; another did well for a few years but
groups rather than the pathologic type of eventually required radioactive iodine for
tumor. further disease extension and was cured. Of
A statistically significant difference in death the 34 patients in the low-risk group actually
rates of patients with distant metastases was treated by surgery or radiotherapy for their
found when the highest and intermediate first recurrence, 24 (7 1%) were cured. Twenty-
groups were compared to the low-risk group. three of the patients in the high-risk category
However, the variations did not achieve sta- were not treated for a variety of reasons, and
tistical significance when analyzed by local re- all died of disease. Of the 48 patients at high
currence or nodal recurrence because of the risk actually treated, 13 (27%) were cured.
small size of the groups involved. The variation in cure rate of patients treated
Table 8 indicates that the number of re- for recurrent or metastatic disease in the two
currences does not alter the poor prognosis risk groups was highly significant (p = 0.002).
of patients in the high-risk group as about Surgery only as therapy was utilized par-
80% of patients died whether one, two, or
more recurrences appear. In patients in the TABLE
8. Death Rate as a Function of Number
low-risk group, however, a single recurrence and Time of Recurrences
was associated with a low mortality, and even Low risk High risk
two episodes of recurrent disease were still
associated with a fair prognosis. Prognosis was No. Deaths No. Deaths
poor if more than two separate episodes of Number of
disease recrudescence occur, however. In recurrences*
neither group was prognosis altered by the One 32 9(28%) 45 37(82%)
time of appearances of the first episode of Two 5 2(40%) 12 10(83%)
More than two 5 4(80%) 14 1 1 (79%)
recurrent disease. Time to first
All types of therapy for patients with re- recurrence
current disease in the low-risk group were Up to five years 21 7(33%) 44 37(84%)
relatively effective whereas therapy of any More than five
years 21 8 (38%) 27 21 (78%)
kind was far less useful in terms of cure in
patients in the high-risk group (Table 9). * p < 0.001
816 March 1979
CANCER VOl. 43

9. Results of Treatment of Recurrences

TABLE However, by applying the results of Table 4,
the maximum recurrence rate would be 8%
Low risk High risk
and the death rate 1 to 2% in the low-risk
No. Deaths No. Deaths group; the respective figures in the high-risk
group would be approximately 15 and 13%.
First therapy of Considerable reduction in both recurrence
recurrence and death rates is apparent, although this im-
Surgeryalone 11 1(9%) 8 4(50%)
External radio- proving trend began about 1950 in the low-
therapy 7 3(43%) 24 19(79%) risk group but not until 1960 in the high-risk
Surgery + ra- group. Indeed, a recent increased recurrence
diotherapy 10 4 (40%) 8 5 (63%) rate is seen in patients at low risk undoubt-
Radioactive
iodine 6 2(33%) 8 7(88%) edly related to the fewer, and less extensive,
TOTAL 34 lO(29%) 48 35(73%) nodal resections performed in the 1960s. Be-
All therapy cause of the good prognosis in patients with
Radioactive node metastases, no impact on death rate of
iodine* 10 3(30%) 10 9(90%) this increased node recurrence is expected.
~

* p = 0.002. The reason for the improving results of ther-

ticularly for isolated local recurrences and
nodal metastases, and thus produced a better
outcome than combinations of therapy or ex-
ternal radiotherapy. In general, radioactive
iodine was utilized initially for distant metas-
tases and resulted in cure in 67% of patients
in the low-risk group but in only 1 of 8 pa-
tients (13%)in the high-risk group.
The last line of Table 9 is the summary of
all radioactive iodine therapy utilized at any
time in these patients. Again, the effectiveness
of radioactive iodine is noted in the low-risk
group and ineffectiveness in the high-risk
group in terms of cure. Of course, periods
of palliation may be achieved, but these are
not analyzed here. The difference is not at
the level of statistical significance (p = 0.02),
but this is related to the small size of the
group.
The changing incidences of recurrence and
death in time by risk group is displayed in
Table 10. The most recent decade, 1961 to
1970, is not completely comparable to the pre-
vious decades as the median follow-up period
is but 10 years and ranges from 5 to 15 years.
apy are clearly related to improved clinical
material in the high-risk group. In the low-
risk group, however, where results are not
related to extension of disease, the cause of
the improved results is obscure.
The incidence of other carcinomas in pa-
tients in the two risk groups is noted in Table
1 1 . Despite the younger age, better prognosis,
and therefore greater overall years at risk, the
risk of second cancers is identical in the two
groups, 17%;37% of all second primary can-
cers in the low-risk group are carcinoma of
the breast and 14% are carcinoma of the ovary.
These respective figures are 22 and 5% in
the high-risk group. Carcinoma of the breast
developed in 7.5% of women at low risk and
in 5.5% of women at high risk, and ovarian
carcinoma developed in 2.8% of patients at
low risk and in 1.4% of patients at high risk.
DISCUSSION
Young patients with differentiated thyroid
carcinoma do very well with surgical and
sometimes other therapy; conversely, older
patients displaying the same histologic features
have a considerably worse prognosis. How-

10. Recurrence and Death Rates in Time

TABLE

Low risk High risk

Time
period Total Recurrence* Deatht Total Recurrence$ Death8

193 1 - 1950 186 (63%) 33 (18%) 13 (7%) 1 1 1 (37%) 38 (34%) 30 (27%)

195 1 - 1960 198 (65%) 9 (5%) 2 (1%) 105 (35%) 33 (31%) 28 (27%)
196 1- 1970" 94 (58%) 6 (6%) 0 67 (42%) 8 (12%) 5 (8%)

* p < 0.001. 8p = 0.003.

t p < 0.001. Follow-up averages only 10 years.
$ p = 0.008.
No. 3 THYROID
CA RISKFACTORANALYSIS* Cady et al. 817

TABLE
11. Occurrence of Other Cancers

Percent of all Percent of all

Total Low risk carcinomas High risk carcinomas
Breast 32 24 37% 8 22%
Colon 12 6 9% 6 16%
Ovary 11 9 14% 2 5%
Head and neck 7 3 5% 4 11%
Lung 7 3 5% 4 11%
Brain 4 3 5% 1
Leukemia 4 3 5% 1
Stomach 4 0 4 11%
Liver 3 3 5% 0
Melanoma 3 3 5% 0
Sarcoma 3 3 5% 0
Abdominal 2 1 1
Cervix (in situ) 2 2 0
Lymphoma 2 1 1
Prostate 2 0 2
Endometrium 1 0 1
Parotid 1 0 1
Renal 1 0 1
Vulva 1 1 0
TOTAL. 102 65 37
651384 = 17% 371216 = 17%

ever, the exact features of these different demonstrated a distinctive biologic pattern,
groups have not been outlined clearly nor the which has been well recognized, that sets it
reasons for the varied prognosis understood. apart from differentiated thyroid cancers that
Indeed, most reports of thyroid carcinoma have any papillary features on histologic ex-
completely ignore these aspects of variable amination and also implied a worse prognosis.
prognosis and consider all cases t ~ g e t h e r . ~ *Larger
~ ~ ~primary cancers had a worse prognosis
Some exceptions3 still emphasize aspects of if they were follicular carcinoma but not if
therapy. Occasional reports16 have indicated they were mixed papillary and follicular car-
that patients seldom, if ever, die of disease, cinomas. Histologic features were found to
and therefore therapy should be restrained relate to prognosis so that extensive, in con-
while other s e r i e ~ ,recommend
~*~ radical sur- trast to minor, breaching of the tumor capsule
gical approaches that carry significant risks of by cancer cells implied a worse prognosis in
iatrogenic morbidity that may be unwarranted follicular carcinoma but not in papillary car-
if dealing with cancers that are associated with cinoma; blood vessel invasion carried a worse
an excellent prognosis. Standard staging sys- prognosis in follicular carcinoma but not in
tems for carcinoma of the thyroid fail to con- mixed papillary and follicular carcinoma.
sider nonanatomic features,l2 A large group Thyroid hormone suppression of thyroid-
of patients was analyzed to define more closely stimulating hormone secretion improved
distinctive prognostic groups of patients and prognosis in mixed papillary and follicular
thereby make surgical and adjuvant therapy carcinomas but had no effect in pure follicular
more rational. In our previous study4 several carcinomas. Lymph node metastases had an
features were influential in prognosis. In de- ameliorating influence on prognosis, and the
creasing order of importance they seemed to be greater the number of lymph node metasta-
age, sex, extension of disease outside the thy- ses, the better the prognosis. This surprising
roid gland, histologic type, histologic features finding had been noted in several other re-
(capsular invasion), thyroid hormone sup- p o r t ~ ~ ,and ~ , thus appears to be yet another
pression after operation, and absence of unique feature of human thyroid cancer. Such
lymph node metastases. However, not all of a relationship has been suggested through the
these features affected prognosis in ways usu- years by the benign behavior of the lateral
ally predicted. aberrant thyroid.6
The previous report4 noted that extra- Risk groups defined by age and sex alone
glandular extension of carcinoma implied a clearly separate two prognostic categories,
poor prognosis. Pure follicular carcinoma and the influence of these basic risk groups
818 CANCER
March 1979
supersedes the effect of local disease exten-
sion, pathologic type, histologic features, type
of therapy, thyroid feeding, and lymph node
metastases. Thus, in young patients, no dif-
ferences between follicular and mixed papil-
lary and follicular types in terms of recurrence
and death rates could be detected despite the
fact that differing patterns of clinical disease
presentations were apparent. In older pa-
tients, there appeared to be some differences
based on pathologic type, but these proved
to be not statistically significant while each
was significantly different from the low-risk
group.
Extension of disease outside the thyroid
gland does not alter the good prognosis in
young patients but dramatically worsens the
prognosis of older patients in a highly signifi-
cant fashion and in both pathologic types. The
protective effects of lymph node metastases
are more apparent in patients in the low-risk
group. The presence of nodal metastases as
part of the recurrent disease presentation im-
plied a relatively good prognosis even if other
recurrent or metastatic sites were present, but
the prognosis was worse if nodal metastases
were not a part of the presentation of recru-
descent disease.
Although the sex composition of the low-
risk group remained the same throughout the
four decades, a male preponderance has de-
veloped in the high-risk group. Men com-
prised about 17% of the low-risk group over
the entire four decades, while in the high-risk
group, during the 40 years, the percent of
men has increased from 26% to 55%-a sig-
nificant change. These facts again point to
something unique about the age and sex cate-
gories.
The value of risk group definition can be
demonstrated equally with the treatment and
results of recurrent or metastatic disease as
for prognosis of primary therapy. Risk group
definition at the time of primary therapy gov-
erned the therapeutic results achieved years
later at the time of treatment of recurrent
disease. This preeminence of basic risk group
influence implies that the features that initiate
and promote the cellular events leading to
cancer are fixed, so that biologic behavior pat-
terns of the particular low-risk thyroid cancer
are not altered by a later change in the tumor
milieu or aging of the patient.
Table 7 described the therapeutic results in
the three patterns of recurrent disease by risk
group. Although the intermediate-risk group
Vol. 43
has the presenting pattern of mixed papillary
and follicular carcinoma that also predomi-
nates in the low-risk group, the prognosis al-
most exactly mimics that of the older patients
with follicular carcinoma. Thus, again is noted
the relatively decreased importance of histo-
logic type in contrast to the age and sex de-
lineation. Surgery and radioactive iodine are
both highly effective treatments for patients
in the low-risk group in terms of cure but
much less effective in older patients.
It is encouraging to see improved results
in both high-risk and low-risk groups in more
recent years,7such that the current death rate
of patients in the low-risk group initially
found to have potentially curable disease at
surgery is only 1% and the death rate of pa-
tients in the high-risk group will probably be
no more than 15%. These figures were
achieved in later decades with a basically con-
servative surgical philosophy which removed
gross disease if at all possible but avoided the
use of total thyroidectomy and routine neck
dissections. Thyroid suppression by hormone
administration was routinely utilized, but lit-
tle if any routine radiotherapy was given
either by external beam or with radioactive
iodine. A generally conservative surgical phi-
losophy for patients at low risk would seem to
be most appropriate and will produce the least
iatrogenic disease but does not necessarily im-
ply that other than surgical therapy should be
practiced initially.
Although breast cancers are not uncommon
in patients with thyroid cancer, the rate of
7.5%in women at low risk and 5.57'o in ' women
at high risk seems not out of line with natural
incidence figures, considering the older age
and higher mortality of the latter group re-
sulting from their primary thyroid carcinoma
and thus the decreased opportunity for long-
term follow-up. Similarly, the risk of ovarian
cancer is within the expected range.
An explanation of the unique effects of age-
sex determined risk groups should incorpo-
rate the data presented. T h e sharp break in
the prognosis at about the age of menopause
in a disease that largely affects women coupled
with the fact that the changes described are
much less prominent in men strongly suggests
a relationship to estrogen levels. Although al-
most no therapeutic trials of estrogen admin-
istration have ever been r e p ~ r t e d , it' ~would
seem important to explore this relationship
both experimentally and clinically.
Of even greater theoretical interest is the
No. 3 THYROID
CA RISK FACTOR
ANALYSIS
implication of this age-sex risk group defini-
tion on cancer induction and promotion. Cur-
rent theory suggests at least a two-step initiat-
ing and promoting process with many cancers.
If the various thyroid cancers arising from
the follicular cell had separate and distinctive
initiating or promoting events, it would serve
to explain the great discrepancies between
low-risk and high-risk differentiated cancer,
the possible differences between men and
women with differentiated cancer, and the
distinctive behavior of the highly lethal ana-
plastic cancer of older patients with its unique
sex ratio, survival pattern, and frequent ac-
companying microscopic foci of differenti-
ated thyroid carcinoma.15 Conversion of a
clinical differentiated thyroid cancer to an
anaplastic carcinoma is a rare event even in
patients with long-standing differentiated
~ a n c e rFurthermore,
.~ small cell carcinoma of
the thyroid, also derived from the follicular
cell, is associated with a distinctive sex ratio
and clinical behavior pattern although it is
never accompanied by evidence of micro-
scopic mixed papillary and follicular car-
cinoma.
Thyroid cancers of the differentiated type
that arise in the estrogen-rich milieu of pre-
menopausal women would permanently be im-
printed with a distinctive biologic behavior
pattern. In men, the recurrence and mortal-
ity rates before age 40 are higher than in
women, and the rates after age 40 are lower
than in women as if the ameliorating effect of
early age were blunted and there was more of
a continuity in biologic behavior across the age
decades in contrast to women where the more
distinctive sharp change in the biologic be-
havior occurs.
The fact that nodal metastases are far com-
moner in the low-risk group and may even be
associated with a protective effect, particularly
in those patients with more than 10 lymph
nodes involved, again points up the distinctive
biology of that group of patients.
The treatment of patients with recurrent
differentiated thyroid cancer should be sur-
gical for neck nodes that are resectable and
for local recurrences or recurrences in the
opposite thyroid lobe. Neck dissections when
performed need not be formal radical neck
dissections, since differentiated thyroid can-
cers show little tendency toward implanta-
tion despite violating the traditional en bloc
concept. Indeed, recurrent disease in nodes
indicates that the patient may be displaying
Cady et al. 819
some host resistance or protective biologic ef-
fect of the cancer, particularly in the low-risk
group. Our current technique of neck dissec-
tion for differentiated thyroid cancer pre-
serves the spinal accessory nerve, the jugular
vein, and the sternocleidomastoid muscle and
spares the submandibular space. In patients
at low risk neck dissection alone for nodal
recurrence is sufficient, as patients seldom die
of disease and further neck recurrences are
uncommon; such patients need continued ob-
servation for the possibility of distant metas-
tases, however. In patients at high risk the
implication of nodal metastases is more worri-
some, but what lethality is implied occurs as a
result not of neck nodes but as a result of
other disease.
If the original thyroid surgery constituted a
lobectomy only and local recurrence is noted,
reexploration of the thyroid area should be
performed to assess resectability, to remove
bulk disease, and to ablate the residual thy-
roid lobe for ease in use of radioactive iodine.
If the original surgical procedure consisted of
total or near total thyroidectomy, reexplora-
tion of the local thyroid bed is less certainly
useful but should be attempted if the local
recurrence suggests resectability. Considera-
tion of tracheotomy is essential if the local
recurrence in any way compromises the air-
way. In low-risk groups when local recurrences
cannot be excised completely, every attempt
to induce radioactive iodine uptake should be
utilized regardless of pathologic type,. as
therapeutic results are so successful. In high-
risk groups such radioactive iodine adminis-
tration is much less successful.
Distant metastases are ominous in patients
in the high-risk group and seldom will therapy
of such a disease presentation result in cure;
only 2 of 47 such patients survived. In the
low-risk group, remission of distant metasta-
ses can be induced by thyroid feeding in many
patients if they were not taking thyroid hor-
mone previously; when this is not successful,
patients can frequently be successfully treated
by radioactive iodine sometimes even if the
lesion is largely papillary in nature.
For successful utilization of radioactive
iodine, the vast majority of normal thyroid
tissue must be ablated. Normal thyroid abla-
tion may be performed surgically rather than
with radioactive iodine to simplify manage-
ment and conserve the potential for radio-
active iodine usage for later therapy of metas-
tases. Thus, when metastatic disease appears
 820 CANCER
March 1979 Vol. 43

that requires radioactive iodine therapy, sur- one patient had a spinal cord tumor, one had
gical thyroidectomy should be performed if a brain tumor, and one had an isolated bone
not previously completed at the time of pri: lesion. Good palliation and long-term survival
mary resection. Technical details of radio- were achieved in each instance.
active iodine administration have been well Risk group definition will enable more
described in the 1 i t e r a t ~ r e . l ~ rational and predictable therapy of both pri-
Although isolated distant metastases, such mary and recurrent differentiated thyroid
as a brain metastasis or isolated pulmonary carcinoma to be achieved. It may also offer
metastases, are uncommon, when they cannot clues as to etiology by enabling the separation
be induced to regress with radioactive iodine, of distinctive biologic patterns that by implica-
surgical excision may be justified. Several ex- tion may have distinctive initiating and pro-
amples of such surgery occurred in this series; moting factors.

REFERENCES
1. Attie, J. N., Khafif, R. A., and Steckler, R. M.: Elec- Tumors # 11. Geneva, World Health Organization,
tive neck dissection in papillary carcinoma of the thyroid. 1974.
Am. J . Surg. 122:464-471, 1971. 11. Hirabayashi, R. N., and Lindsay, S.: Carcinoma of
2. Block, M . A,: Management of carcinoma of the the thyroid gland: a statistical study of 390 patients.].
thyroid. Ann. Surg. 185:133-144, 1977. Clzn. Endocnnol. Metah. 21:1596-1610, 1961.
3. Buckwalter, J . A., and Thomas, C. G., Jr.: Selec- 12. Manual for Staging of Cancer 1977. Chicago, Illi-
tion of surgical treatment for well differentiated thyroid nois, American Joint Committee for Cancer Staging and
carcinomas. Ann. Surg. 176:565-577, 1972. End-Results Reporting, 1977, 174 pp.
4. Cady, B., Sedgwick, C. E., Meissner, W. A,, Book- 13. Meissner, W. A., and Warren, S.: Tumors of the
walter, J. R., Romagosa, V., and Werber, J.: Changing thyroid gland. In Atlas of T u m o r Pathology. 2nd series,
clinical, pathologic, therapeutic, and survival patterns in fascicle 4, Washington, D. C . . Armed Forces Institute
differentiated thyroid carcinoma. Ann. Surg. 184:541- of Pathology, 1969.
553, 1976.
14. Rawson, R. W., and Leeper, R.: Factors influencing
5. Clark, R. L., Jr.. White, E. C., and Russell, W. 0.:
benignancy versus malignancy of thyroid neoplasms. In
Total thyroidectomy for cancer of the thyroid: Significancr
Thyroid Neoplasia, S. Young and D. R. Inman, Eds.
of intraglandular dissemination. ,4nn.Surg. 149:858-866,
New York, Academic Press, 1968; pp. 159-177.
1959.
6. Clark, R. I>., Hickey, R. C., Butler, J . J., Ibanez, 15. Rossi, R., Cady, B., Meissner, W. A., Sedgwick,
M. L., and Ballantyne, A . J.: Thyroid cancer discovered C. E., and Werber, J.: Prognosis in undifferentiated
incidently during treatment of an unrelated head and carcinoma and lymphoma of the thyroid. Am. J . Surg.
neck cancer: Review of 16 cases. Ann. Surg. 163:665- 135:589-595, 1978.
671, 1966. 16. Veith, F. J.. Brooks, J . R., Grigsby, U. P., and
7. Crile, G., Jr.: Changing end results in patients with Selenkow, H. A.: T h e nodular thyroid gland and cancer:
papillary carcinoma of the thyroid. Surg. Gynrcol. Oh.\tPt. A practical approach to the problem. N . Engl. J . M e d .
132~460-468, 1971. 270:43 1-436, 1964.
8. Fransilla, K. 0.: Prognosis in thyroid carcinoma. 17. Wool, M. S.: Thyroid Cancer. Cleveland, Ohio,
Canrrr 36:1138-1146, 1975. CKC Press, Inc., 1978 (In press).
9. Geissinger, W. T., Horsley, J. S., Parker, F. P., and 18. Woolner, L. B., Beahrs, 0. H., Black, B. M.,
Alrich, E. M.: Carcinoma of the thyroid. Ann. Surg. 179: McConahey, W. M., and Keating, F. R., Jr.: Thyroid car-
734-739, 1974. cinoma: general considerations and follow-up data on
10. Hedinger, C., Ed.: Histologic l y p i n g of Thyroid 1,181 cases. I n Thyroid Neoplasia, S. Young and D. R.
Tumors. International Histological Classification of Inman, Eds.: New York, Academic Press, 1968; pp. 51-77.

View publication stats