Comparative Analysis of Alternative and Traditional

Immunohistochemical Markers for the Distinction

of Ovarian Sertoli Cell Tumor From Endometrioid
Tumors and Carcinoid Tumor
A Study of 160 Cases
Chengquan Zhao, MD,* Gary L. Bratthauer, MS, MT(ASCP), CLS,* Ross Barner, MD,*
and Russell Vang, MDw
Abstract: The main neoplasms in the dierential diagnosis for
primary ovarian tumors with a tubule-rich pattern are pure
Sertoli cell tumor, endometrioid tumors (including borderline
tumor, well-dierentiated carcinoma, and the sertoliform
variant of endometrioid carcinoma), and carcinoid tumor.
Because traditional immunohistochemical markers [pan-cyto-
keratin (pan-CK), low molecular weight cytokeratin (CK8/18),
epithelial membrane antigen (EMA), inhibin, calretinin, CD99,
chromogranin, and synaptophysin] can occasionally have
diagnostic limitations, the goal of this study was to determine
whether or not any alternative markers [cytokeratin 7 (CK7),
estrogen receptor (ER), progesterone receptor (PR), CD10, and
CD56] have better diagnostic utility when compared with
traditional markers for this dierential diagnosis. Immunohis-
tochemical stains for alternative, as well as traditional, markers
were performed on the following primary ovarian tumors: pure
Sertoli cell tumor (n=40), endometrioid borderline tumor
(n=38), sertoliform endometrioid carcinoma (n=13), well-
dierentiated endometrioid carcinoma (n =27), and carcinoid
tumor (n=42). Extent and intensity of immunostaining were
semiquantitatively scored. In addition, immunohistochemical
composite scores (ICSs) in positive cases were calculated on the
basis of the combination of extent and intensity scores.
Cytokeratin 7 (CK7) was positive in 97% of endometrioid
tumors, 13% of Sertoli cell tumors, and 24% of carcinoid
tumors. The dierences in the mean ICSs for endometrioid
tumors versus Sertoli cell tumor or carcinoid tumor were
statistically signicant (P values ranging from <0.001 to 0.018).
ER and PR were positive in 87% and 86% of endometrioid
tumors, 8% and 13% of Sertoli cell tumors, and 2% each of
carcinoid tumors, respectively. The dierences in the mean ICSs
for endometrioid tumors versus Sertoli cell tumor were
statistically signicant (P values ranging from <0.001 to
0.012). Among the epithelial markers, EMA seemed to be the
most discriminatory but only slightly better than CK7, ER, or
PR. Pan-CK and CK8/18 were not helpful. CD10 showed
overlapping patterns of expression in all categories of tumors.
Among the sex cord markers, CD10 was markedly less useful
than inhibin or calretinin; CD99 was not discriminatory. CD56
showed overlapping patterns of expression in all categories of
tumors. Among the neuroendocrine markers, CD56 was less
useful than chromogranin or synaptophysin. When traditional
immunohistochemical markers are problematic for the dier-
ential diagnosis of ovarian Sertoli cell tumor versus endo-
metrioid tumors versus carcinoid tumor, adding CK7, ER, and/
or PR to a panel of markers can be helpful. Endometrioid
tumors more frequently express CK7, ER, and PR and show a
greater extent of immunostaining in contrast to Sertoli cell
tumor and carcinoid tumor. Compared with traditional
epithelial markers, CK7, ER, and PR are nearly as advanta-
geous as EMA. Inhibin is the most discriminatory sex cord
marker, and CD10 is not helpful in the dierential diagnosis.
Chromogranin and synaptophysin are excellent discriminatory
markers for carcinoid tumor, and CD56 is neither suciently
sensitive nor specic enough for this dierential diagnosis to
warrant its use in routine practice.
Key Words: immunohistochemistry, ovary, Sertoli cell tumor,
endometrioid carcinoma, carcinoid tumor
(Am J Surg Pathol 2007;31:255266)
T
he dierential diagnosis of primary ovarian tumors
with a tubule-rich pattern may sometimes be challen-
ging. Specically, Sertoli cell tumor may be histologically
mimicked by other tumors, such as endometrioid
tumors (including borderline tumor, well-dierentiated Copyright
r
2007 by Lippincott Williams & Wilkins
From the *Department of Gynecologic and Breast Pathology, Armed
Forces Institute of Pathology, Washington, DC; and wDivision of
Gynecologic Pathology, Johns Hopkins Hospital, Baltimore, MD.
The opinion and assertions contained herein are the private views of the
authors and are not to be construed as ocial or as representing the
views of the Department of the Army or the Department of Defense
Current Address: Chengquan Zhao, MD, Department of Pathology,
University of Pittsburgh, PA. Ross Barner, MD, Department of
Pathology, Walter Reed Army Medical Center, Washington, DC.
Reprints: Chengquan Zhao, MD, Department of Pathology, Magee
Womens Hospital, 300 Halket Street, Pittsburgh, PA 15213 (e-mail:
chengquanzhao@yahoo.com).
ORIGINAL ARTICLE
Am J Surg Pathol

Volume 31, Number 2, February 2007 255
carcinoma, and the sertoliform variant of endometrioid
carcinoma) and carcinoid tumor. Most cases can be
distinguished from one another by traditional clinical,
gross, and histologic features. Sertoli cell tumor is usually
conned to the ovary, unilateral, and shows distinctive
open and closed tubular patterns with bland round nuclei.
In elongated solid tubules, the nuclei are characteristically
arranged in a paired cell pattern. Endometrioid
tumors, especially carcinoma, may be bilateral or of
advanced stage. They may display other growth patterns
such as villoglandular forms, conuent glandular appear-
ances, and inltrative growth with stromal desmoplasia.
Their nuclei may sometimes be somewhat atypical or
pleomorphic, and glandular lumens may contain muci-
nous secretions. Other characteristic features include
squamous metaplasia, endometriosis, and a background
of borderline tumor or adenobroma. Primary carcinoid
tumor is usually unilateral and conned to the ovary.
Distinctive trabecular or strumal patterns may be
appreciated, and the periphery of cells may show ne,
granular, and eosinophilic granules. The nuclei are round
and stippled with a salt and pepper appearance.
Occasionally, a background of a teratoma may be
evident, and the carcinoid syndrome may be present in
some patients.
Even though the tumors in this dierential diagnosis
are typically distinguished from one another based on
traditional clinicopathologic features, immunohistochem-
istry can be helpful in problematic cases. In general,
traditional immunohistochemical markers used for the
distinction of epithelial, sex cord-stromal, and neuro-
endocrine tumors, such as pan-cytokeratin (pan-CK), low
molecular weight cytokeratin (CK8/18), epithelial mem-
brane antigen (EMA), inhibin, calretinin, CD99, synap-
tophysin, and chromogranin, can be useful; however, they
may sometimes show overlapping patterns of expression.
For example, Sertoli cell tumors can express epithelial
markers such as cytokeratin (AE1/AE3 and CK8/18) in
38% to 100% of cases
11,13,21,22,37,49
whereas endometrioid
carcinomas can express sex cord markers such as inhibin
and calretinin in 0% to 25% and 0% to 36% of cases,
respectively.
7,8,10,1416,19,20,25,2932,35,38,41,47,54
Although sex cord-stromal markers such as inhibin
and calretinin have been extensively studied in ovarian
sex cord-stromal tumors, most studies are predominantly
composed of granulosa cell tumors. The number of
Sertoli cell tumors that have been analyzed for various
immunohistochemical markers in any given study is very
small, and the majority of available immunohistochem-
ical data are based on a mixture of pure Sertoli cell
tumors, Sertoli-Leydig cell tumors, and mixed granulosa
cell-Sertoli cell tumors (gynandroblastoma) from
ovarian and testicular sites.
3,6,7,1113,17,19,2022,25,26,28,30,33,
3537,38,44,47,49,51,53
Thus, a formal immunohistochemical
comparison of ovarian pure Sertoli cell tumor, endo-
metrioid tumors, and carcinoid tumor is lacking from
the literature. Recently, a clinicopathologic study of
54 ovarian pure Sertoli cell tumors was reported, and
immunohistochemical results for various markers were
available in 23 cases; however, other tumors in the
dierential diagnosis were not evaluated because that was
not the aim of that study.
37
Neuroendocrine markers such
as chromogranin and synaptophysin have been exten-
sively evaluated in nonovarian carcinoid tumors, but the
number of reported ovarian carcinoid tumors studied for
these markers, and other immunohistochemical markers
for this dierential diagnosis, in any single study is also
quite small.
4,10,20,24,25,32,43,4547,53
Cytokeratin 7 (CK7), estrogen receptor (ER), and
progesterone receptor (PR) are often expressed in
endometrioid tumors, but these markers have not been
thoroughly evaluated in ovarian Sertoli cell tumor or
carcinoid tumor.
19,21,22,38,49
CD10 expression has been
described in Sertoli cell tumor, but sucient data on the
use of this marker for this dierential diagnosis is lacking
from the literature.
23,36,39,40,49
CD56 can be expressed in
nonovarian neuroendocrine tumors; although, it is not
known whether it provides any diagnostic advantage in
the setting of an ovarian carcinoid tumor. These
alternative markers (CK7, ER, PR, CD10, and CD56)
have not been studied as thoroughly as the traditional
markers in this dierential diagnosis and may have
potential diagnostic utility.
The goal of this study was to determine whether any
of these alternative markers for this dierential diagnosis
are diagnostically useful and if they have any advantages
over more traditional markers. Furthermore, the patterns
of expression of CK7, ER, PR, CD10, and CD56 were
formally characterized in ovarian pure Sertoli cell tumor,
endometrioid tumors, and carcinoid tumor.
MATERIALS AND METHODS
Case Selection
Primary ovarian Sertoli cell tumors, endometrioid
borderline tumors, endometrioid carcinomas (including
the sertoliform variant), and carcinoid tumors were
retrieved from the les of the Armed Forces Institute of
Pathology (AFIP) from 1970 to 2004 after Institutional
Review Board approval and were reviewed by the authors.
Immunohistochemical stains for pan-CK, CK7, CK8/18,
EMA, ER, PR, calretinin, CD10, CD99, inhibin, CD56,
chromogranin, synaptophysin, and TTF-1 were per-
formed in the AFIPs immunohistochemistry laboratory
on 40 pure Sertoli cell tumors, 38 endometrioid borderline
tumors, 13 invasive sertoliform endometrioid carcinomas,
27 invasive well-dierentiated endometrioid carcinomas,
and 42 carcinoid tumors (33 insular type, 8 trabecular
type, and 1 mixed insular-trabecular type).
Immunohistochemistry
The antibody clone names, sources, dilutions, and
antigen pretreatments are listed in Table 1. Briey,
formalin-xed, paran-embedded tissue sections were
deparanized, and immunohistochemical staining was
performed using protocols optimized for each anti-
body. Immunostaining for calretinin, CD10, chromo-
granin, CK7, CK8/18, EMA, ER, pan-cytokeratin, PR,
Zhao et al Am J Surg Pathol

Volume 31, Number 2, February 2007
256 r 2007 Lippincott Williams & Wilkins
synaptophysin, and TTF-1 was performed using the
Ventana Benchmark Autostainer (Ventana Medical
System, Inc, Tucson, AZ) at 421C for 32 minutes. The
detection system used was Ventanas Basic DAB Detec-
tion Kit. Immunostaining for CD56 and CD99 was
performed using the Dako Autostainer (DakoCytoma-
tion, Carpinteria, CA) at room temperature for 30
minutes. The detection system used was a polymer link
system method (Envision+; DakoCytomation). Immu-
nostaining for inhibin was performed manually for one
hour at room temperature and detected using the ABC
method with the VIP chromogenic substrate (Vector
Laboratories, Burlingame, CA).
Interpretation and Scoring of
Immunohistochemical Preparations
The following patterns of immunostaining were
considered positive: CD10, CD56, CD99, chromogranin,
CK7, CK8/18, EMA, inhibin, pan-CK, synaptophysin
cytoplasmic and/or membranous (staining for CK7 is
typically cytoplasmic; however, if a membrane-accentu-
ated pattern was present, it was considered positive);
calretinincytoplasmic and nuclear; and ER, PR, and
TTF-1nuclear. For overall positivity, immunostaining
in >5% of cells was considered positive, and r5%
positive cells was considered negative. Additionally, both
extent (on the basis of the percentage of positive cells) and
intensity of immunostaining were evaluated by a semi-
quantitative system. Extent was scored as: 0, r5%; 1+
(1 point), 6% to 25%; 2+ (2 points), 26% to 50%; 3+
(3 points), 51% to 75%; and 4+ (4 points), 76% to
100%. Intensity was arbitrarily scored as: weak (1 point),
moderate (2 points), or strong (3 points). Intensity was
designated as weak when immunostaining was present
but only barely detectable. In positive cases, these values
were converted into composite immunohistochemical
scores by multiplying the individual scores of extent by
intensity (possible range of values from 1 to 12). For
example, a case with 3+ extent (3 points) and moderate
intensity of immunostaining (2 points) would have an
immunohistochemical composite score of 3 2 =6. In
endometrioid tumors, only immunostaining in the gland-
ular component was scored (immunostaining within the
squamoid component was not recorded).
Statistical Analysis
The dierences in the mean immunohistochemical
composite scores for alternative markers between tumor
groups were statistically analyzed using the t test (2-
tailed) with P values <0.05 considered statistically
signicant.
RESULTS
Overall positivity for alternative and traditional
immunohistochemical markers is listed in Table 2. Extent
of staining for epithelial, sex cord, and neuroendocrine
markers is detailed in Tables 3 to 5, respectively. Immuno-
histochemical composite scores are listed in Table 6, and
patterns of expression of alternative markers (CK7, ER,
CD10, and CD56) are illustrated in Figures 1 to 4.
CK7 was much more frequently positive in en-
dometrioid tumors compared with Sertoli cell tumor and
carcinoid tumor (Table 2). The extent of immunostaining
in positive cases was frequently diuse in endometrioid
tumors and focal in Sertoli cell tumor and carcinoid
tumor (Table 3). The dierences in the mean immuno-
histochemical composite scores between Sertoli cell tumor
and the 3 endometrioid tumor categories were statistically
signicant (endometrioid borderline tumor, P<0.001;
sertoliform endometrioid carcinoma, P=0.018; and
well-dierentiated endometrioid carcinoma, P<0.001),
as was the dierence between endometrioid tumors and
carcinoid tumor (P<0.001).
TABLE 1. Antibodies and Pretreatment Conditions
Antigen Clone Dilution Source Pretreatment
Calretinin Rabbit polyclonal Prediluted Ventana Medical System, Inc, Tucson, AZ Heat*; EDTAw
CD10 56C6 Prediluted Ventana Medical System, Inc, Tucson, AZ None
CD56 123C3 1:200 Zymed, South San Francisco, CA None
CD99 12E7 1:150 DakoCytomation, Carpinteria, CA None
Chromogranin Rabbit polyclonal 1:200 DakoCytomation, Carpinteria, CA None
CK7 OV-TL 12/30 1:400 DakoCytomation, Carpinteria, CA Protease (8 min at 421C)
CK8/18 5D3 Prediluted Ventana Medical System, Inc, Tucson, AZ Protease (8 min at 421C)
EMA E29 Prediluted Ventana Medical System, Inc, Tucson, AZ Protease (8 min at 421C)
ER 6F11 Prediluted Ventana Medical System, Inc, Tucson, AZ Heat*; EDTAw
Inhibin R1 1:80 Oxford Bio-Innovation, Oxfordshire, UK Microwave, 10 mM citrate
(pH 6.0)
Pan-CK Cocktail of AE1/AE3
and LP34
1:400 Chemicon, Temecula, CA (AE1/AE3) and
DakoCytomation, Carpinteria, CA (LP34)
Protease (8 min at 421C)
PR 1A6 Prediluted Ventana Medical System, Inc, Tucson, AZ Heat*; EDTAw
Synaptophysin Rabbit polyclonal Prediluted Cell Marque, Hot Springs, AR Heat*; EDTAw
TTF-1 8G7G3/1 Prediluted Cell Marque, Hot Springs, AR Heat*; EDTAw
*Heat-induced antigen retrieval was performed within autostainer at 1001C for 30 min.
wCC1 antigen retrieval stock solution (Ventana Medical System, Inc, Tucson, AZ).
EDTA indicates ethylenediaminetetraacetic acid.
Am J Surg Pathol

Volume 31, Number 2, February 2007 Immunohistochemistry of Ovarian Sertoli Form Tumors
r
2007 Lippincott Williams & Wilkins 257
ER and PR were usually positive in endometrioid
tumors and negative in Sertoli cell tumor and carcinoid
tumor (Table 2). The extent of immunostaining in
positive cases was usually diuse in endometrioid tumors
and focal in Sertoli cell tumor and carcinoid tumor (Table
3). The dierences in the mean immunohistochemical
composite scores for ER and PR between Sertoli cell
tumor and all three endometrioid tumor categories were
statistically signicant (ER: endometrioid borderline
tumor, P<0.001; sertoliform endometrioid carcinoma,
P=0.012; and well-dierentiated endometrioid carcino-
ma, P<0.001; PR: endometrioid borderline tumor,
P=0.001; sertoliform endometrioid carcinoma,
P<0.001; and well-dierentiated endometrioid carcino-
ma, P<0.001). Statistical analysis for carcinoid tumors
was not performed because only 1 carcinoid tumor was
positive for ER/PR.
Among the traditional epithelial markers, EMA was
positive in all endometrioid tumors but infrequently
positive in carcinoid tumors and negative in all Sertoli
cell tumors, respectively (Table 2). In positive cases,
immunostaining in endometrioid tumors tended to be
diuse whereas usually focal in positive carcinoid tumors
(Table 3). Pan-CK and CK8/18 were frequently positive
in all categories of tumors (Table 2). The extent of
immunostaining for pan-CK and CK8/18 was frequently
diuse in endometrioid tumors and carcinoid tumor and
variable in Sertoli cell tumor (Table 3).
A subset of all categories of tumors were positive for
CD10 (Table 2). Variable extents of immunostaining were
seen in all positive cases regardless of the type of tumor
(Table 4). The dierence in the mean immunohistochem-
ical composite scores between Sertoli cell tumor and well-
dierentiated endometrioid carcinoma was statistically
signicant (P=0.047), but it was not signicant for
Sertoli cell tumor versus the other tumors (endometrioid
borderline tumor, P=0.327; sertoliform endometrioid
carcinoma, P=0.944; and carcinoid tumor P=0.868).
Among the traditional sex cord markers, inhibin
and calretinin were frequently positive in Sertoli cell
tumor and negative in most endometrioid tumors and
carcinoid tumors (Table 2). In particular, endometrioid
borderline tumors and sertoliform endometrioid carcino-
mas were never positive for inhibin, and sertoliform
endometrioid carcinomas and carcinoid tumors were
never positive for calretinin. The extent of immunostain-
ing for inhibin and calretinin in positive cases tended to
be diuse in Sertoli cell tumor whereas it was mostly focal
in the other categories of tumors (Table 4). The frequency
of positivity for CD99 was variable in all categories of
tumors (Table 2). The extent of immunostaining in
positive cases tended to be diuse in Sertoli cell tumor
and carcinoid tumor and focal in endometrioid tumors
(Table 4).
The frequency of positivity for CD56 was variable
in all categories of tumors (Table 2). The extent of
immunostaining in positive cases was diuse in carcinoid
tumor and focal to variable in endometrioid tumors and
Sertoli cell tumor, respectively (Table 5). No statistically
signicant dierence in the mean immunohistochemical
composite scores for CD56 was found for carcinoid
tumor versus Sertoli cell tumor (P=0.289); however, a
signicant dierence was found for carcinoid tumor
versus endometrioid tumors (P=0.001).
Among the traditional neuroendocrine markers,
chromogranin and synaptophysin were positive in almost
all cases of carcinoid tumor and positive in a small subset
of endometrioid tumors and Sertoli cell tumors (Table 2).
TABLE 2. Overall Positivity for Alternative and Traditional Immunohistochemical Markers
Antigen Sertoli Cell Tumor
Endometrioid
Borderline Tumor
Sertoliform
Endometrioid
Carcinoma
Well-dierentiated
Endometrioid
Carcinoma Carcinoid Tumor
Epithelial markers
CK7* 13% 100% 85% 100% 24%
ER* 8% 87% 85% 89% 2%
PR* 13% 84% 77% 93% 2%
Pan-CKw 65% 100% 100% 100% 98%
CK8/18w 43% 100% 93% 96% 74%
EMAw 0% 100% 100% 100% 12%
Sex cord markers
CD10* 25% 39% 31% 37% 9%
Inhibinw 98% 0% 0% 4% 2%
Calretininw 60% 11% 0% 18% 0%
CD99w 68% 16% 23% 33% 40%
Neuroendocrine markers
CD56* 48% 16% 16% 30% 57%
Synaptophysinw 35% 8% 8% 22% 98%
Chromograninw 13% 3% 15% 11% 100%
*For purposes of this study, CK7, ER, PR, CD10, and CD56 were considered alternative markers for the dierential diagnosis of ovarian Sertoli cell tumor vs.
endometrioid tumors vs. carcinoid tumor.
wFor purposes of this study, pan-CK, CK8/18, EMA, inhibin, calretinin, CD99, synaptophysin, and chromogranin were considered traditional markers for the
dierential diagnosis of ovarian Sertoli cell tumor vs. endometrioid tumors vs. carcinoid tumor.
Zhao et al Am J Surg Pathol

Volume 31, Number 2, February 2007
258 r 2007 Lippincott Williams & Wilkins
TABLE 3. Extent of Immunohistochemical Expression of Alternative and Traditional Epithelial Markers*
Sertoli Cell
Tumor
Endometrioid Borderline
Tumor
Sertoliform Endometrioid
Carcinoma
Well-dierentiated Endometrioid
Carcinoma
Carcinoid
Tumor
Antigen 0 1+ 2+ 3+ 4+ 0 1+ 2+ 3+ 4+ 0 1+ 2+ 3+ 4+ 0 1+ 2+ 3+ 4+ 0 1+ 2+ 3+ 4+
CK7w 87% 8% 3% 3% 0% 0% 3% 3% 24% 71% 15% 8% 15% 39% 23% 0% 4% 4% 37% 55% 76% 10% 10% 0% 5%
ERw 92% 5% 3% 0% 0% 13% 5% 32% 18% 32% 15% 23% 15% 15% 31% 11% 0% 11% 37% 41% 98% 2% 0% 0% 0%
PRw 87% 5% 8% 0% 0% 16% 13% 11% 24% 37% 23% 0% 0% 23% 54% 7% 0% 15% 41% 37% 98% 0% 2% 0% 0%
Pan-CKz 35% 15% 18% 20% 12% 0% 0% 3% 5% 92% 0% 0% 0% 8% 92% 0% 0% 4% 0% 96% 2% 14% 29% 14% 41%
CK8/18z 57% 18% 7% 5% 13% 0% 8% 13% 26% 53% 8% 0% 15% 23% 54% 4% 4% 11% 22% 59% 26% 7% 21% 22% 24%
EMAz 100% 0% 0% 0% 0% 0% 3% 5% 13% 79% 0% 15% 15% 15% 55% 0% 0% 8% 8% 84% 88% 5% 2% 5% 0%
*All rows for a given marker in each tumor category may not add up to 100% because the numbers in each column were rounded up to the next whole number.
wFor purposes of this study, CK7, ER, and PR were considered alternative markers for the dierential diagnosis of ovarian Sertoli cell tumor vs. endometrioid tumors vs. carcinoid tumor.
zFor purposes of this study, pan-CK, CK8/18, and EMA were considered traditional markers for the dierential diagnosis of ovarian Sertoli cell tumor vs. endometrioid tumors vs. carcinoid tumor.
TABLE 4. Extent of Immunohistochemical Expression of Alternative and Traditional Sex Cord Markers*
Sertoli Cell
Tumor
Endometrioid Borderline
Tumor
Sertoliform Endometrioid
Carcinoma
Well-dierentiated Endometrioid
Carcinoma
Carcinoid
Tumor
Antigen 0 1+ 2+ 3+ 4+ 0 1+ 2+ 3+ 4+ 0 1+ 2+ 3+ 4+ 0 1+ 2+ 3+ 4+ 0 1+ 2+ 3+ 4+
CD10w 75% 8% 8% 8% 3% 61% 5% 13% 8% 13% 69% 15% 15% 0% 0% 63% 0% 11% 15% 11% 90% 2% 2% 2% 2%
Inhibinz 2% 7% 13% 18% 60% 100% 0% 0% 0% 0% 100% 0% 0% 0% 0% 96% 4% 0% 0% 0% 98% 0% 0% 2% 0%
Calretininz 40% 15% 10% 15% 20% 88% 3% 3% 3% 3% 100% 0% 0% 0% 0% 82% 4% 11% 0% 3% 100% 0% 0% 0% 0%
CD99z 32% 10% 13% 25% 20% 84% 8% 8% 0% 0% 77% 8% 15% 0% 0% 67% 11% 18% 4% 0% 60% 5% 10% 13% 12%
*All rows for a given marker in each tumor category may not add up to 100% because the numbers in each column were rounded up to the next whole number.
wFor purposes of this study, CD10 was considered an alternative marker for the dierential diagnosis of ovarian Sertoli cell tumor vs. endometrioid tumors vs. carcinoid tumor.
zFor purposes of this study, inhibin, calretinin, and CD99 were considered traditional markers for the dierential diagnosis of ovarian Sertoli cell tumor vs. endometrioid tumors vs. carcinoid tumor.
A
m
J
S
u
r
g
P
a
t
h
o
l

V
o
l
u
m
e
3
1
,
N
u
m
b
e
r
2
,
F
e
b
r
u
a
r
y
2
0
0
7
I
m
m
u
n
o
h
i
s
t
o
c
h
e
m
i
s
t
r
y
o
f
O
v
a
r
i
a
n
S
e
r
t
o
l
i
F
o
r
m
T
u
m
o
r
s
r
2
0
0
7
L
i
p
p
i
n
c
o
t
t
W
i
l
l
i
a
m
s
&
W
i
l
k
i
n
s
2
5
9
In positive cases, expression of chromogranin and
synaptophysin tended to be diuse in carcinoid tumor
and focal in endometrioid tumors and Sertoli cell tumor
(Table 5). TTF-1 was assessed only in carcinoid tumors;
all cases were negative.
DISCUSSION
The main neoplasms in the dierential diagnosis for
primary ovarian tumors with a tubule-rich pattern are
pure Sertoli cell tumor, endometrioid tumors (including
borderline tumor, well-dierentiated carcinoma, and the
sertoliform variant of endometrioid carcinoma), and
carcinoid tumor. These tumors are usually distinguished
from one another by traditional clinicopathologic fea-
tures. However, not all cases can be reliably distinguished
from one another by clinical, gross, and histologic
features alone. Immunohistochemistry may be helpful,
but even this ancillary study may at times yield conicting
results, making distinction of these tumors from one
another somewhat dicult. Hence, accurate data regard-
ing the characteristic immunophenotypes of each of these
types of tumors are necessary if one is going to use
immunohistochemistry as a reliable diagnostic tool. This
is particularly important because the number of ovarian
pure Sertoli cell tumors and carcinoid tumors that have
been immunohistochemically studied in the literature is
small. In this study, we compared the utility of alternative
(CK7, ER, PR, CD10, and CD56) and traditional [pan-
CK (AE1/AE3), low molecular weight cytokeratin (CK8/
18), EMA, inhibin, calretinin, CD99, chromogranin, and
synaptophysin] immunohistochemical markers for the
dierential diagnosis of primary ovarian pure Sertoli cell
tumor, endometrioid tumors, and carcinoid tumor to
determine whether any of these alternative markers may
be more useful than traditional ones.
TABLE 5. Extent of Immunohistochemical Expression of Alternative and Traditional Neuroendocrine Markers*
Sertoli Cell
Tumor
Endometrioid
Borderline
Tumor
Sertoliform
Endometrioid
Carcinoma
Well-dierentiated
Endometrioid
Carcinoma
Carcinoid
Tumor
Antigen 0 1+ 2+ 3+ 4+ 0 1+ 2+ 3+ 4+ 0 1+ 2+ 3+ 4+ 0 1+ 2+ 3+ 4+ 0 1+ 2+ 3+ 4+
CD56w 52% 13% 15% 8% 13% 84% 8% 8% 0% 0% 84% 8% 8% 0% 0% 70% 19% 11% 0% 0% 43% 7% 5% 21% 24%
Synaptophysinz 65% 20% 13% 2% 0% 92% 0% 3% 3% 2% 92% 0% 0% 8% 0% 78% 0% 15% 4% 3% 2% 0% 5% 14% 79%
Chromograninz 87% 8% 5% 0% 0% 97% 0% 3% 0% 0% 85% 15% 0% 0% 0% 89% 7% 4% 0% 0% 0% 5% 0% 0% 95%
*All rows for a given marker in each tumor category may not add up to 100% because the numbers in each column were rounded up to the next whole number.
wFor purposes of this study, CD56 was considered an alternative marker for the dierential diagnosis of ovarian Sertoli cell tumor vs. endometrioid tumors vs.
carcinoid tumor.
zFor purposes of this study, synaptophysin and chromogranin were considered traditional markers for the dierential diagnosis of ovarian Sertoli cell tumor vs.
endometrioid tumors vs. carcinoid tumor.
TABLE 6. Immunohistochemical Composite Scores for Alternative and Traditional Immunohistochemical Markers*
Antigen Sertoli Cell Tumor
Endometrioid
Borderline Tumor
Sertoliform Endometrioid
Carcinoma
Well-dierentiated
Endometrioid
Carcinoma
Carcinoid
Tumor
Epithelial markers
CK7w 3.8 (2-9) 10.7 (2-12) 8.3 (3-12) 10 (3-12) 5.9 (2-12)
ERw 1.3 (1-2) 7 (2-12) 4.6 (1-12) 9.3 (4-12) 1y
PRw 2.4 (1-6) 8.5 (2-12) 11.1 (9-12) 9.2 (2-12) 4y
Pan-CKz 6.6 (2-12) 11.7 (6-12) 11.2 (8-12) 11.8 (6-12) 8.1 (2-12)
CK8/18z 5.8 (2-12) 8.9 (1-12) 9.1 (4-12) 9.6 (3-12) 7.3 (2-12)
EMAz 0 10.6 (3-12) 8.7 (3-12) 11.3 (6-12) 4.8 (2-9)
Sex cord markers
CD10w 5.6 (1-12) 7.2 (1-12) 5.8 (4-9) 8.8 (4-12) 5.2 (3-6)
Inhibinz 7.9 (1-12) 0 0 2y 6y
Calretininz 5.2 (1-12) 3.8 (1-8) 0 3 (1-4) 0
CD99z 5.8 (2-12) 2.8 (1-4) 2.3 (1-4) 3.5 (1-6) 6.4 (2-12)
Neuroendocrine markers
CD56w 6 (2-12) 3.3 (1-6) 2.5 (1-4) 3.4 (2-6) 7.3 (1-12)
Synaptophysinz 2.1 (1-6) 4.3 (2-8) 3y 4.3 (2-9) 9.7 (2-12)
Chromograninz 3 (1-6) 2y 2.5 (2-3) 2 (1-3) 11.2 (2-12)
*Results are listed as mean value with range in parentheses.
wFor purposes of this study, CK7, ER, PR, CD10, and CD56 were considered alternative markers for the dierential diagnosis of ovarian Sertoli cell tumor vs.
endometrioid tumors vs. carcinoid tumor.
zFor purposes of this study, pan-CK, CK8/18, EMA, inhibin, calretinin, CD99, synaptophysin, and chromogranin were considered traditional markers for the
dierential diagnosis of ovarian Sertoli cell tumor vs. endometrioid tumors vs. carcinoid tumor.
yOnly 1 case was positive.
Zhao et al Am J Surg Pathol

Volume 31, Number 2, February 2007
260 r 2007 Lippincott Williams & Wilkins
In this study, CK7 was expressed in most endo-
metrioid tumors whereas only a subset of Sertoli cell
tumors and carcinoid tumors were positive. It is also
important to note that when Sertoli cell tumors or
carcinoid tumors were positive, they usually exhibited a
focal pattern (r50% positive cells) of reactivity as
opposed to the diuse pattern (>50% positive cells) seen
in endometrioid tumors. The classic diagnostic use of
CK7 in regards to ovarian tumors is for the distinction of
primary ovarian tumors of surface epithelial-stromal
origin from metastatic carcinomas involving the ovary
(mainly those of lower gastrointestinal tract origin).
However, CK7 is also useful for the distinction of ovarian
endometrioid tumors from Sertoli cell tumor and
carcinoid tumor. The frequent incidence of CK7 expres-
sion in the endometrioid tumors in this study is similar to
that in the literature, including the sertoliform variant of
endometrioid carcinoma.
5,9,14,18,19,42,50
Only a small
number of ovarian Sertoli cell tumors has been assessed
for CK7 in the literature. In one study, all 3 Sertoli-
Leydig cell tumors were reported as negative for CK7
although expression was noted in retiform areas in 2 of
those cases.
19
Our study assessed only pure Sertoli cell
tumors, and 13% of these were positive for this marker.
Whereas CK7 has been assessed in a number of studies of
nonovarian carcinoid tumors, the frequency of CK7
expression in ovarian carcinoid tumors in the literature is
essentially unknown. In our study, this marker seems to
be useful and was much more discriminative for
distinguishing endometrioid tumors from Sertoli cell
tumor and carcinoid tumor than pan-CK or CK8/18
because the latter 2 tumors can express pan-CK or CK8/
18 with a diuse pattern similar to endometrioid tumors.
Pan-CK and CK8/18, therefore, are not helpful because
it has also been noted in the literature that Sertoli cell
tumor can express pan-CK or CK8/18 in 38% to 100% of
cases.
11,13,21,22,37,49
Additionally, expression of pan-CK
has been reported in 5 of 5 carcinoid tumors of the ovary
in one study.
43
ER and PR were helpful for distinction of Sertoli
cell tumor and carcinoid tumor from endometrioid
tumors. They were frequently positive in endometrioid
tumors but only seldom expressed in Sertoli cell tumor
and carcinoid tumor. The low incidence of positivity
FIGURE 1. Expression of CK7. Well-differentiated endometrioid carcinoma: (A) hematoxylin and eosin. (B) Diffuse expression of
CK7 (4+ overall). Sertoli cell tumor: (C) hematoxylin and eosin. (D) Focal expression of CK7 (1+ overall).
Am J Surg Pathol

Volume 31, Number 2, February 2007 Immunohistochemistry of Ovarian Sertoli Form Tumors
r
2007 Lippincott Williams & Wilkins 261
for ER in Sertoli cell tumor in this study is similar to
that reported in the literature.
38,49
Although the incidence
of positivity for PR varies widely in other studies, only
13% of our cases expressed this marker.
22,38,49
Despite
the fact that some Sertoli cell tumors may express
hormonal markers in some studies, the patterns of
expression in Sertoli cell tumor and carcinoid tumor
dier from those of endometrioid tumors. In our study,
endometrioid tumors characteristically expressed
ER/PR in a diuse pattern (>50% positive cells) in
contrast to the focal pattern (r50% positive cells)
that is usually seen in Sertoli cell tumor and carcinoid
tumor. Distinction from Sertoli cell tumor should be
straightforward when the immunostaining for ER/PR is
evaluated in the context of other helpful markers, such
as EMA, CK7, inhibin, and chromogranin. Though
the incidence of positivity for ER/PR in carcinoid tumor
in our study is low, these markers have not been
previously assessed in other studies of ovarian carcinoid
tumor.
Of all the epithelial markers evaluated in this study,
one of the best discriminatory markers for this dierential
diagnosis was EMA based on its frequency and extent.
Although EMA showed slightly better discrimination
between these tumors than did CK7, ER, and PR, these
3 alternative markers were comparable to EMA.
In this study, CD10 was infrequently expressed in
Sertoli cell tumors (25%), and it was also expressed in
37% of endometrioid tumors and 9% of carcinoid
tumors. CD10 was previously thought to be specic for
mesonephric/wolan origin and frequently positive in
mesonephric remnants/hyperplasia/adenocarcinoma of
the cervix, rete ovarii, and mesonephric adenocarcinoma
of the uterine corpus.
39,40
This marker initially seemed to
be useful in the distinction from endometrioid carcinomas
of the uterus, but subsequent studies showed that
nonmesonephric cervical adenocarcinoma and endo-
metrial endometrioid carcinoma could in fact express
this marker, diminishing the diagnostic use of CD10 in
those anatomic locations.
34,52
Although this marker may
be nonspecic in the uterus, data for ovarian endome-
trioid carcinomas was previously limited to just a few
studies, and expression of CD10 has been demonstrated
in Sertoli cell tumor.
23,36,39,40,49
Therefore, we attempted
FIGURE 2. Expression of ER. Sertoliform endometrioid carcinoma: (A) hematoxylin and eosin. B, Diffuse expression of ER
(4+ overall). Sertoli cell tumor: (C) hematoxylin and eosin. D, Focal expression of ER (1+ overall).
Zhao et al Am J Surg Pathol

Volume 31, Number 2, February 2007
262 r 2007 Lippincott Williams & Wilkins
to determine if this marker could have potential
diagnostic value in the ovary. In the current study,
CD10 was neither a sensitive nor specic marker for
Sertoli cell tumor.
Among the sex cord markers, inhibin and calretinin
were far superior to CD10 in terms of being able to
discriminate Sertoli cell tumor from endometrioid tumor
and carcinoid tumor. The prior literature suggests that
CD99 may be helpful in distinguishing Sertoli cell tumor
from endometrioid carcinoma (including variants that
resemble sex cord-stromal tumors), each expressing this
marker in 14% to 100% and 0% to 14% of cases,
respectively.
8,17,26,28,30,37,43,49
In contrast, in our study
CD99 did not seem to be sensitive or specic for sex cord
lineage, similar to CD10; however, the frequency of
positivity and extent of staining for CD99 were greater in
Sertoli cell tumor and carcinoid tumor compared with
endometrioid tumors.
CD56 was positive in variable proportions of
carcinoid tumors, Sertoli cell tumors, endometrioid
borderline tumors, well-dierentiated endometrioid
carcinomas, and sertoliform endometrioid carcinomas.
Expression of CD56 has been described in nonovarian
carcinoid tumor; however, the utility of this neuroendo-
crine marker has not been tested for the dierential
diagnosis of ovarian carcinoid tumor versus endometrioid
tumor or Sertoli cell tumor.
1,27,48
Although endometrioid
tumors could express CD56 with a frequency only slightly
less than that of carcinoid tumor in this study (21% vs.
57%, respectively), the pattern of staining in endome-
trioid tumors was typically focal (r50% positive cells) as
opposed to the frequently diuse pattern (>50% positive
cells) in carcinoid tumor. However, CD56 did not seem to
be substantially useful for the discrimination of carcinoid
tumor from Sertoli cell tumor.
Of the neuroendocrine markers in this study,
chromogranin and synaptophysin demonstrated a much
better ability to distinguish carcinoid tumor from
endometrioid tumors and Sertoli cell tumor than did
CD56. Although chromogranin and synaptophysin have
only been studied in small numbers of ovarian carcinoid
tumors in the literature, our ndings are similar to those
reported by others in ovarian sites.
4,24,45,46
Neuroendo-
crine markers have been insuciently studied in ovarian
FIGURE 3. Expression of CD10. Sertoli cell tumor: (A) hematoxylin and eosin. B, Diffuse expression of CD10 (3+ overall). Well-
differentiated endometrioid carcinoma: (C) hematoxylin and eosin. D, Diffuse expression of CD10 (4+ overall).
Am J Surg Pathol

Volume 31, Number 2, February 2007 Immunohistochemistry of Ovarian Sertoli Form Tumors
r
2007 Lippincott Williams & Wilkins 263
Sertoli cell tumor. However, in one study, none of the
tumors were positive for chromogranin.
37
Although
the results for neuroendocrine markers in our study are
slightly higher (13% and 35% of cases positive for
chromogranin and synaptophysin, respectively), these
incidences are still substantially lower than for carcinoid
tumor (Table 2). Moreover, the extent of immunostaining
for chromogranin and synaptophysin in Sertoli cell tumor
FIGURE 4. Expression of CD56. Carcinoid tumor: (A) hematoxylin and eosin. B, Diffuse expression of CD56 (4+ overall). Sertoli
cell tumor: (C) hematoxylin and eosin. D, Diffuse expression of CD56 (4+ overall). Well-differentiated endometrioid carcinoma:
(E) hematoxylin and eosin. F, Focal expression of CD56 (1+ overall).
Zhao et al Am J Surg Pathol

Volume 31, Number 2, February 2007
264 r 2007 Lippincott Williams & Wilkins
is substantially dierent from that in carcinoid tumor
(Table 5). Chromogranin and synaptophysin expression
have not been previously assessed in ovarian endome-
trioid tumors; however, in one study of endometrioid
carcinomas of the endometrium, synaptophysin was
expressed in 21% of cases.
2
That nding is similar to
the low incidence of positivity for the neuroendocrine
markers chromogranin and synaptophysin in ovarian
endometrioid tumors in our study (Table 2). Given that
the incidence of positivity for chromogranin and synap-
tophysin in ovarian endometrioid tumors is not 0%,
positivity for either of these markers (without considera-
tion of the extent of staining) poses a potential pitfall
resulting in the misdiagnosis of carcinoid tumor. Regard-
less, our results show that chromogranin and synapto-
physin are good diagnostic markers that are helpful in
this dierential diagnosis when used together as part of a
panel of antibodies.
On the basis of the ndings in this study and the
literature, we make the following conclusions for the
immunohistochemical distinction of ovarian pure Sertoli
cell tumor, endometrioid tumors, and carcinoid tumor.
Despite the histologic similarity of some endometrioid
carcinomas to sex cord-stromal tumors (ie, the sertoli-
form variant of endometrioid carcinoma), these histologic
variants show immunohistochemical proles similar to
the classic forms of endometrioid tumors (well-dieren-
tiated carcinoma and borderline tumor). Among the
epithelial markers, the most discriminative marker for the
distinction of ovarian endometrioid tumors from Sertoli
cell tumor and carcinoid tumor is EMA. CK7, ER, and
PR are also useful in this regard, and their use is
comparable to that of EMA. Pan-CK and CK8/18 should
not be used for this dierential diagnosis because of
overlap in expression between the dierent types of
tumors. Among the sex cord markers, inhibin is the most
useful. CD10 is neither sensitive nor specic for sex cord
lineage and not useful for the distinction of Sertoli cell
tumor from endometrioid tumors or carcinoid tumor.
Among the neuroendocrine markers, chromogranin and
synaptophysin are the most helpful. CD56 is neither
highly sensitive nor specic for neuroendocrine lineage
and not very useful for the distinction of carcinoid tumor
from endometrioid tumors or Sertoli cell tumor. Of the
alternative markers studied for the dierential diagnosis
in this study, the only ones that have practical diagnostic
value are CK7, ER, and PR. Traditional markers such as
EMA, inhibin, calretinin, chromogranin, and synapto-
physin should still be used as part of an immunohisto-
chemical panel. CK7, ER, and PR can be added to
supplement this panel; however, immunohistochemical
ndings should always be interpreted in the context of
traditional clinicopathologic features.
ACKNOWLEDGMENTS
The authors thank the Department of Scientic
Laboratories at the AFIP for the voluminous task of
cutting unstained sections and performance of immunohis-
tochemical stains.
REFERENCES
1. Al-Khafaji B, Nosinger AE, Miller MA, et al. Immunohistologic
analysis of gastrointestinal and pulmonary carcinoid tumors.
Hum Pathol. 1998;29:992999.
2. Alkushi A, Irving J, Hsu F, et al. Immunoprole of cervical and
endometrial adenocarcinomas using a tissue microarray. Virchows
Arch. 2003;442:271277.
3. Augusto D, Leteurtre E, De La Taille A, et al. Calretinin: a valuable
marker of normal and neoplastic Leydig cells of the testis.
Appl Immunohistochem Mol Morphol. 2002;10:159162.
4. Baker PM, Oliva E, Young RH, et al. Ovarian mucinous carcinoids
including some with a carcinomatous component: a report of 17
cases. Am J Surg Pathol. 2001;25:557568.
5. Berezowski K, Stastny JF, Kornstein MJ. Cytokeratins 7 and 20
and carcinoembryonic antigen in ovarian and colonic carcinoma.
Mod Pathol. 1996;9:426429.
6. Cao QJ, Jones JG, Li M. Expression of calretinin in human ovary,
testis, and ovarian sex cord-stromal tumors. Int J Gynecol Pathol.
2001;20:346352.
7. Cathro HP, Stoler MH. The utility of calretinin, inhibin, and WT1
immunohistochemical staining in the dierential diagnosis of
ovarian tumors. Hum Pathol. 2005;36:195201.
8. Choi YL, Kim HS, Ahn G. Immunoexpression of inhibin alpha
subunit, inhibin/activin betaA subunit and CD99 in ovarian tumors.
Arch Pathol Lab Med. 2000;124:563569.
9. Chu P, Wu E, Weiss LM. Cytokeratin 7 and cytokeratin 20
expression in epithelial neoplasms: a survey of 435 cases. Mod
Pathol. 2000;13:962972.
10. Ciris M, Erhan Y, Zekioglu O, et al. Inhibin alpha and beta
expression in ovarian stromal tumors and their histological
equivalences. Acta Obstet Gynecol Scand. 2004;83:491496.
11. Costa MJ, Morris RJ, Wilson R, et al. Utility of immunohisto-
chemistry in distinguishing ovarian sertoli-stromal cell tumors from
carcinosarcomas. Hum Pathol. 1992;23:787797.
12. Costa MJ, Ames PF, Walls J, et al. Inhibin immunohistochemistry
applied to ovarian neoplasms: a novel, eective, diagnostic tool.
Hum Pathol. 1997;28:12471254.
13. Deavers MT, Malpica A, Liu J, et al. Ovarian sex cord-stromal
tumors: an immunohistochemical study including a comparison of
calretinin and inhibin. Mod Pathol. 2003;16:584590.
14. Dionigi A, Facco C, Tibiletti MG, et al. Ovarian metastases from
colorectal carcinoma. Clinicopathologic prole, immunophenotype,
and karyotype analysis. Am J Clin Pathol. 2000;114:111122.
15. Doglioni C, Tos AP, Laurino L, et al. Calretinin: a novel
immunocytochemical marker for mesothelioma. Am J Surg Pathol.
1996;20:10371046.
16. Flemming P, Grothe W, Maschek H, et al. The site of inhibin
production in ovarian neoplasms. Histopathology. 1996;29:465468.
17. Gordon MD, Corless C, Renshaw AA, et al. CD99, keratin, and
vimentin staining of sex cord-stromal tumors, normal ovary, and
testis. Mod Pathol. 1998;11:769773.
18. Groisman GM, Meir A, Sabo E. The value of Cdx2 immunostaining
in dierentiating primary ovarian carcinomas from colonic carci-
nomas metastatic to the ovaries. Int J Gynecol Pathol. 2004;23:
5257.
19. Guerrieri C, Franlund B, Malmstrom H, et al. Ovarian endome-
trioid carcinomas simulating sex cord-stromal tumors: a study using
inhibin and cytokeratin 7. Int J Gynecol Pathol. 1998;17:266271.
20. Hildebrandt RH, Rouse RV, Longacre TA. Value of inhibin in the
identication of granulosa cell tumors of the ovary. Hum Pathol.
1997;28:13871395.
21. Iczkowski KA, Bostwick DG, Roche PC, et al. Inhibin A is a
sensitive and specic marker for testicular sex cord-stromal tumors.
Mod Pathol. 1998;11:774779.
22. Kato N, Fukase M, Ono I, et al. Sertoli-stromal cell tumor of the
ovary: immunohistochemical, ultrastructural, and genetic studies.
Hum Pathol. 2001;32:796802.
23. Khin EE, Kikkawa F, Ino K, et al. Neutral endopeptidase/CD10
expression in the stroma of epithelial ovarian carcinoma. Int J
Gynecol Pathol. 2003;22:175180.
24. Kimura N, Pilichowska M, Okamoto H, et al. Immunohistochem-
ical expression of chromogranins A and B, prohormone convertases
Am J Surg Pathol

Volume 31, Number 2, February 2007 Immunohistochemistry of Ovarian Sertoli Form Tumors
r
2007 Lippincott Williams & Wilkins 265
2 and 3, and amidating enzyme in carcinoid tumors and pancreatic
endocrine tumors. Mod Pathol. 2000;13:140146.
25. Kommoss F, Oliva E, Bhan AK, et al. Inhibin expression in ovarian
tumors and tumor-like lesions: an immunohistochemical study. Mod
Pathol. 1998;11:656664.
26. Kommoss F, Oliva E, Bittinger F, et al. Inhibin-a CD99, HEA125,
PLAP, and chromogranin immunoreactivity in testicular neoplasms
and the androgen insensitivity syndrome. Hum Pathol. 2000;31:
10551061.
27. Lantuejoul S, Moro D, Michalides RJ, et al. Neural cell adhesion
molecules (NCAM) and NCAM-PSA expression in neuroendocrine
lung tumors. Am J Surg Pathol. 1998;22:12671276.
28. Loo KT, Leung AK, Chan JK. Immunohistochemical staining of
ovarian granulosa cell tumours with MIC2 antibody. Histopatho-
logy. 1995;27:388390.
29. Lugli A, Forster Y, Haas P, et al. Calretinin expression in human
normal and neoplastic tissues: a tissue microarray analysis on 5233
tissue samples. Hum Pathol. 2003;34:9941000.
30. Matias-Guiu X, Pons C, Prat J. Mullerian inhibiting substance,
alpha-inhibin, and CD99 expression in sex cord-stromal tumors and
endometrioid ovarian carcinomas resembling sex cord-stromal
tumors. Hum Pathol. 1998;29:840845.
31. McCluggage WG, Maxwell P. Adenocarcinomas of various sites
may exhibit immunoreactivity with anti-inhibin antibodies. Histo-
pathology. 1999;35:216220.
32. McCluggage WG, Maxwell P. Immunohistochemical staining for
calretinin is useful in the diagnosis of ovarian sex cord-stromal
tumours. Histopathology. 2001;38:403408.
33. McCluggage WG, Maxwell P, Sloan JM. Immunohistochemical
staining of ovarian granulosa cell tumors with monoclonal antibody
against inhibin. Hum Pathol. 1997;28:10341038.
34. McCluggage WG, Oliva E, Herrington CS, et al. CD10 and
calretinin staining of endocervical glandular lesions, endocervical
stroma and endometrioid adenocarcinomas of the uterine corpus:
CD10 positivity is characteristic of, but not specic for, mesonephric
lesions and is not specic for endometrial stroma. Histopathology.
2003;43:144150.
35. Movahedi-Lankarani S, Kurman RJ. Calretinin, a more sensitive
but less specic marker than alpha-inhibin for ovarian sex cord-
stromal neoplasms: an immunohistochemical study of 215 cases.
Am J Surg Pathol. 2002;26:14771483.
36. Oliva E, Vu Q, Young RH. CD10 expression in sex cord-stromal
tumors (SCSTs) and steroid cell tumors (StCTs) of the ovary.
[Abstract]. Mod Pathol. 2002;15:204A.
37. Oliva E, Alvarez T, Young RH. Sertoli cell tumors of the ovary: a
clinicopathologic and immunohistochemical study of 54 cases. Am J
Surg Pathol. 2005;29:143156.
38. Ordi J, Schammel DP, Rasekh L, et al. Sertoliform endometrioid
carcinomas of the ovary: a clinicopathologic and immunohisto-
chemical study of 13 cases. Mod Pathol. 1999;12:933940.
39. Ordi J, Nogales FF, Palacin A, et al. Mesonephric adenocarcinoma
of the uterine corpus: CD10 expression as evidence of mesonephric
dierentiation. Am J Surg Pathol. 2001;25:15401545.
40. Ordi J, Romagosa C, Tavassoli FA, et al. CD10 expression in
epithelial tissues and tumors of the gynecologic tract: a useful
marker in the diagnosis of mesonephric, trophoblastic, and clear cell
tumors. Am J Surg Pathol. 2003;27:178186.
41. Pelkey TJ, Frierson HF Jr, Mills SE, et al. The diagnostic utility of
inhibin staining in ovarian neoplasms. Int J Gynecol Pathol. 1998;
17:97105.
42. Ramalingam P, Malpica A, Silva EG, et al. The use of cytokeratin 7
and EMA in dierentiating ovarian yolk sac tumors from
endometrioid and clear cell carcinomas. Am J Surg Pathol.
2004;28:14991505.
43. Riopel MA, Perlman EJ, Seidman JD, et al. Inhibin and epithelial
membrane antigen immunohistochemistry assist in the diagnosis of
sex cord-stromal tumors and provide clues to the histogenesis of
hypercalcemic small cell carcinomas. Int J Gynecol Pathol. 1998;
17:4653.
44. Rishi M, Howard LN, Bratthauer GL, et al. Use of monoclonal
antibody against human inhibin as a marker for sex cord-stromal
tumors of the ovary. Am J Surg Pathol. 1997;21:583589.
45. Sidhu J, Sanchez RL. Prostatic acid phosphatase in strumal
carcinoids of the ovary. An immunohistochemical study. Cancer.
1993;72:16731678.
46. Stagno PA, Petras RE, Hart WR. Strumal carcinoids of the ovary.
An immunohistologic and ultrastructural study. Arch Pathol Lab
Med. 1987;111:440446.
47. Stewart CJ, Jeers MD, Kennedy A. Diagnostic value of inhibin
immunoreactivity in ovarian gonadal stromal tumours and their
histological mimics. Histopathology. 1997;31:6774.
48. Van Eeden S, Quaedvlieg PF, Taal BG, et al. Classication of low-
grade neuroendocrine tumors of midgut and unknown origin. Hum
Pathol. 2002;33:11261132.
49. Vang R, Herrmann ME, Tavassoli FA. Comparative immunohis-
tochemical analysis of granulosa and sertoli components in ovarian
sex cord-stromal tumors with mixed dierentiation: potential
implications for derivation of sertoli dierentiation in ovarian
tumors. Int J Gynecol Pathol. 2004;23:151161.
50. Wang NP, Zee S, Zarbo RJ, et al. Coordinate expression of
cytokeratins 7 and 20 denes unique subsets of carcinomas. Appl
Immunohistochem. 1995;3:99107.
51. Yao DX, Soslow RA, Hedvat CV, et al. Melan-A (A103) and
inhibin expression in ovarian neoplasms. Appl Immunohistochem
Mol Morphol. 2003;11:244249.
52. Zamecnik M. Correspondence re: Mikami Y, Hata S, Kiyokawa T,
Manabe T. Expression of CD10 in malignant mullerian mixed
tumors and adenosarcomas: an immunohistochemical study.
Mod pathol. 2002;15:923930 [Letter]. Mod Pathol. 2003;16:
618619.
53. Zheng W, Sung CJ, Hanna I, et al. Alpha and beta subunits of
inhibin/activin as sex cord-stromal dierentiation markers. Int J
Gynecol Pathol. 1997;16:263271.
54. Zheng W, Luo MP, Welt C, et al. Imbalanced expression of inhibin
and activin subunits in primary epithelial ovarian cancer. Gynecol
Oncol. 1998;69:2331.
Zhao et al Am J Surg Pathol

Volume 31, Number 2, February 2007
266 r 2007 Lippincott Williams & Wilkins