ORIGINAL ARTICLE
Utility of p16 Expression for Distinction of Uterine Serous
Carcinomas From Endometrial Endometrioid and
Endocervical Adenocarcinomas
Immunohistochemical Analysis of 201 Cases
Anna Yemelyanova, MD,* Hongxiu Ji, MD, PhD,* Ie-Ming Shih, MD, PhD,*w z
Tian-Li Wang, PhD,w z Lee-Shu-Fune Wu, MHS,y and Brigitte M. Ronnett, MD*z
Abstract: Uterine serous carcinomas typically have a character-
istic morphology (papillary architecture, high-grade nuclei) and
immunoprofile (diffuse/strong p53 expression, loss of hormone
receptor expression) that distinguish them from most endome-
trial endometrioid carcinomas. However, glandular variants of
serous carcinoma can simulate Fédération Internationale de
Gynécologie et d’Obstétrique (FIGO) grade 2 endometrioid
carcinomas, and some serous carcinomas lack p53 expression
and retain hormone receptor expression, making classifi-
cation difficult. P16 expression patterns distinguish endo-
metrioid carcinomas (patchy) from human papillomavirus
(HPV)-related endocervical adenocarcinomas (diffuse/strong)
but utility for distinction of serous carcinomas from endome-
trioid carcinomas and endocervical adenocarcinomas has not
been evaluated in a large series. Immunohistochemical analysis
of p16 expression was performed on 201 uterine and endocervi-
cal adenocarcinomas in hysterectomy specimens, including 49
serous carcinomas, 101 endometrial endometrioid carcinomas
(44 FIGO grade 1, 40 FIGO grade 2, and 17 FIGO grade 3), and
51 HPV-related endocervical adenocarcinomas. All serous
carcinomas demonstrated diffuse/moderate-strong p16 expres-
sion, with percentage of positive tumor cells ranging from 90%
to 100% (mean/median: 95%/100%). In contrast, endometrial
endometrioid carcinomas exhibited less diffuse and less intense
expression, with percent of positive tumor cells ranging from
From the Departments of *Pathology; wOncology; zGynecology and
Obstetrics, The Johns Hopkins University School of Medicine; and
yDepartment of International Health, Bloomberg School of Public
Health, The Johns Hopkins University, Baltimore, MD.
Disclosure: Dr Ronnett has recently been a paid lecturer for mtm
laboratories. The terms of this arrangement are managed by the
Johns Hopkins University in accordance with its conflict of interest
policies. This study was completed before the establishment of this
relationship. The authors retain full control and authority over this
study, which is not subject to earlier approval by any commercial
interest. All opinions expressed and implied in this study are solely
those of the authors and do not represent or reflect the views of the
Johns Hopkins University or the Johns Hopkins Hospital and
Health System.
Dr Ji’s current affiliation is Eastside Pathology Inc, Bellevue, WA.
Correspondence: Anna Yemelyanova, MD, Department of Pathology,
The Johns Hopkins Hospital, Weinberg Building Room 2242, 401 N.
Broadway, Baltimore, MD 21231 (e-mail: ayemely1@jhmi.edu).
Copyright r 2009 by Lippincott Williams & Wilkins
1504 | www.ajsp.com
10% to 90% (mean/median: 38%/30%; staining intensity:
variable). Similar to serous carcinomas, all endocervical
adenocarcinomas exhibited diffuse/moderate-strong p16 expres-
sion, with percentage of positive tumor cells ranging from 90%
to 100% (mean/median: 94%/90%). P16 can serve as an
additional diagnostic marker, used as part of an immunohisto-
chemical panel, including p53 and hormone receptors, for
distinction of uterine serous carcinomas from endometrioid
carcinomas. Distinction of serous carcinomas from endocervical
adenocarcinomas (HPV-related type), both of which share
diffuse p16 expression and frequently lack hormone receptor
expression, relies on morphology and diffuse/strong p53
expression in the former and detection of HPV in the latter.
Key Words: uterus, adenocarcinoma, serous carcinoma, endome-
trioid carcinoma, endocervical adenocarcinoma, p16, immuno-
histochemistry
(Am J Surg Pathol 2009;33:1504–1514)
U terine adenocarcinomas include those of endometrial
and those of endocervical origin. The endometrial
carcinomas include the common endometrioid carcino-
mas (85% to 90% of tumors) and the less common serous
carcinomas (10% to 15% of tumors).3 These subtypes
are usually distinguished by characteristic morphologic
features. Most endometrioid carcinomas exhibit some
degree of gland formation with glands comprised of tall
columnar cells with smooth luminal borders and uniform
nuclei with generally mild-to-moderate, but occasionally
marked, nuclear atypia. In contrast, typical serous
carcinomas have papillary architecture, display high-
grade pleomorphic nuclei, and have hobnail-type cells
that create scalloped luminal borders. However, pure
glandular variants of uterine serous carcinoma, including
some with more columnar-appearing cells, can be difficult
to distinguish from endometrial endometrioid carcino-
mas, and can be misclassified as Fédération Internatio-
nale de Gynécologie et d’Obstétrique (FIGO) grade 2
based on the presence of notable (marked) atypia.14
Several studies have demonstrated the utility of immuno-
histochemical markers, including p53, Ki-67, and
hormone receptors [estrogen (ER) and progesterone
Am J Surg Pathol  Volume 33, Number 10, October 2009
Am J Surg Pathol  Volume 33, Number 10, October 2009
(PR) receptors], for distinguishing the common lower-
grade endometrioid carcinomas from serous carcino-
mas.2,15,23,24,40 The vast majority of uterine serous
carcinomas are characterized by diffuse/strong p53
expression, have limited or absent hormone receptor
expression, and have markedly elevated Ki-67 prolifera-
tion indices, whereas most endometrial endometrioid
carcinomas, particularly FIGO grade 1 and 2 tumors,
lack significant p53 expression, retain hormone receptor
expression, and have lower Ki-67 proliferation indices.23
However, the utility of these markers is limited by the
occurrence of some tumors with overlapping immuno-
profiles, including those serous carcinomas that retain
hormone receptor expression and lack p53 expression and
higher-grade endometrial endometrioid carcinomas that
lack hormone receptor expression and acquire significant
p53 expression.15,16,24,25,37,42 Distinction of these tumors
is important because both the therapy and prognosis of
these subtypes differ significantly.3,8,10–12 Uterine corpus
(endometrial) adenocarcinomas must also be distin-
guished from endocervical adenocarcinomas because
surgical management of these tumors often differs. Due
to their columnar glandular morphology and frequent
endometrioid or hybrid endometrioid and mucinous
(‘‘usual-type’’)49 differentiation, differential diagnosis
generally concerns distinction of endocervical adenocar-
cinomas from endometrial endometrioid carcinomas
rather than uterine serous carcinomas. The vast majority
of endocervical adenocarcinomas (approximately 90%)
are human papillomavirus (HPV)-related and exhibit
diffuse p16 expression because of complex molecular
mechanisms by which high-risk HPV transforming pro-
teins (E6, E7) interact with cell cycle regulatory proteins
(p53, pRb) to generate a futile feedback loop resulting in
p16 overexpression.4–6,17–19,21,26–30,32,34–36,38,44,49 Although
certain features, such as usual-type differentiation, nu-
merous mitotic figures, and apoptotic bodies, can suggest
an HPV-related endocervical adenocarcinoma, these
features can be shared by endometrial endometrioid
carcinomas, thereby often precluding distinction of these
tumor types on the basis of morphology alone. In
contrast, endometrial endometrioid carcinomas are con-
sidered etiologically unrelated to HPV. They have been
shown to exhibit patchy p16 expression of variable
intensity, which contrasts with the diffuse/moderate-
strong expression characteristic of HPV-related endocer-
vical adenocarcinomas.6,26,48 Interestingly, these HPV-
related endocervical adenocarcinomas also often lack
hormone receptor expression. Thus, a panel of immno-
histochemical markers comprised of ER, PR, and p16 has
been shown to be useful for distinguishing endometrial
endometrioid carcinomas from endocervical adenocarci-
nomas.6,38,41,48 HPV-related endocervical adenocarcino-
mas have wild-type TP53 in most cases and therefore lack
the diffuse/strong p53 expression pattern that charac-
terizes most uterine serous carcinomas.20,33,45 Although
endocervical adenocarcinomas and uterine serous carci-
nomas are generally readily distinguished by morphology
alone, their shared frequent lack of hormone receptor
r 2009 Lippincott Williams & Wilkins
P16 Expression in Uterine Adenocarcinomas
expression and increased proliferation indices can occa-
sionally lead to misclassification, particularly with some
of the same problematic subsets of uterine serous
carcinomas, namely, the glandular variants and those
lacking p53 expression.
This study was designed to evaluate p16 expression
in a large series of uterine adenocarcinomas, with
emphasis on uterine serous carcinomas, to determine
the utility and limitations of this marker as part of an
immunohistochemical panel for distinction of the sub-
types of uterine adenocarcinomas.
MATERIALS AND METHODS
Case Selection
The study was approved by the Johns Hopkins
University School of Medicine Institutional Review
Board. Uterine adenocarcinomas in hysterectomy speci-
mens were retrieved from the surgical pathology files of
the Johns Hopkins Hospital. The 201 cases retrieved
included 101 endometrioid carcinomas (44 FIGO grade 1,
40 FIGO grade 2, and 17 FIGO grade 3), 49 serous
carcinomas with typical morphologic features, and 51
endocervical adenocarcinomas previously determined to
be HPV-positive by in situ hybridization or polymerase
chain reaction (PCR). All tumors were reviewed by 2
pathologists (A.Y. and B.M.R.) to confirm that their
morphologic features conformed to the established
diagnostic criteria for endometrioid and serous subtypes.
Mixed tumors and those with overlapping features not
readily classified by morphology alone were excluded.
The specific basis (solid growth, marked nuclear atypia)
for establishing endometrioid carcinomas as FIGO grades
2 and 3 was assessed as well.
Immunohistochemical Analysis
of p16 Expression
Formalin-fixed paraffin-embedded tissues were
used. Five-micron sections were deparaffinized and
rehydrated. Antigen retrieval was performed under
standard optimized conditions for each antibody. Im-
munoperoxidase labeling using the streptavidin-biotin
peroxidase complex technique and 30 ,30 -diamino-benzi-
dine as the chromagen was performed with the automated
BioTek-Tech Mate 1000 Staining System (Ventana/Bio-
tek Solutions, Inc) and the BenchMark XT IHC Staining
Module (Ventana Medical Systems) at room temperature.
Anti-p16 clone E6H4 (mtm laboratories AG, Germany)
was initially used with the first staining system in the
earlier phase of the study. Subsequently, when the
laboratory changed staining systems, anti-p16 clone
16p04, prediluted (Cell Marque) was used with the second
staining system. Then, when the laboratory changed
reagent suppliers, anti-p16 clone E6H4 (mtm laboratories
AG, Germany) was again used with the second staining
system. The Johns Hopkins Immunopathology Labora-
tory validated these reagents for diagnostic use on both
staining systems and appropriate positive and negative
controls were performed. Cases from each of the 3 groups
www.ajsp.com | 1505
Yemelyanova et al Am J Surg Pathol  Volume 33, Number 10, October 2009

of uterine adenocarcinoma subtypes were collected Research on Cancer, which were not considered as a
throughout the course of the study and thus tumors from mutation in this study.1
each of these 3 groups were stained with both systems and
both reagents. The consistent p16 staining patterns HPV DNA Detection
observed within the 3 groups of tumor subtypes over All endocervical adenocarcinomas were evaluated
time (see Results) provide evidence that the change in for the presence of HPV DNA by in situ hybridization or
staining system and antibody source did not influence the PCR analysis (the latter restricted to cases in which HPV
immunohistochemical analysis. Interpretations of the was not detected by in situ hybridization), as described
reactions were performed by 2 pathologists (A.Y. and earlier.38
B.M.R.). Percentage of positive tumor cells and staining
intensity (weak, moderate, strong) were assessed, with Statistical Analysis
both nuclear and cytoplasmic staining regarded as a Statistical analysis was performed in SAS version
positive reaction. Extent of staining was estimated to the 9.1 (SAS institute Inc, Carry, NC), and the significance
nearest 10%, with staining in 90% or greater of tumor level was set at 0.05. Ranges and frequency distributions
cells considered a diffuse pattern and staining of less than of all continuous and categorical variables were exam-
90% of tumor cells considered a patchy (nondiffuse) ined. The t test, w2 test, and analysis of variance
pattern. The p16 immunostaining results for some of (ANOVA) were used for comparison.
these tumors have been reported earlier.6,38 The p16
expression was also assessed in endometrial intraepithelial
carcinoma (EIC) when present in the section subjected to RESULTS
immunohistochemical analysis. Immunohistochemical Analysis
of p16 Expression
Ancillary Analysis of Uterine Serous Carcinomas
Immunohistochemical analysis of p16 expression is
To assure that analysis of p16 expression was being
summarized in Table 1. All serous carcinomas demon-
performed on a rigorously classified set of uterine serous
strated diffuse p16 expression, with percentage of positive
carcinomas, the serous carcinomas were also evaluated
tumor cells ranging from 90% to 100% (mean/median:
for p53 expression to provide additional support for their
95%/100%) (Figs. 1, 2); staining intensity was usually
morphologic classification as serous (as these tumors have
strong but occasionally moderate. EIC was identified in
a known high frequency of diffuse/strong p53 expression,
41 cases. In 29 cases, EIC was present in the section
which is highly associated with the presence of TP53
evaluated immunohistochemically and displayed diffuse/
mutations).8,13,15,24,25,31,40 Immunoperoxidase labeling
moderate-strong p16 expression, similar to the adjacent
was performed with the automated XT iVIEW DAB
invasive serous carcinoma.
V.1 procedure on the BenchMark XT IHC/ISH Staining
All endometrial endometrioid carcinomas demon-
Module, Ventana with anti-p53 (clone Bp53-11, Venta-
strated some degree of p16 expression, with percentage of
na). In addition, this analysis allowed for identification of
positive tumor cells ranging from 10% to 90% (mean/
that subset of serous carcinomas lacking p53 expression,
median: 38%/30%). All tumors displayed a patchy
which might be misclassified as endometrial endometrioid
pattern of expression and staining intensity varying from
or endocervical adenocarcinoma when subjected to an
predominantly weak-to-moderate but occasionally mod-
immunohistochemical marker panel. Although not the
erate-to-strong (Figs. 3, 4), with the exception of 3 tumors
focus of this study, this subset of serous carcinomas
(2 FIGO grade 3 and 1 FIGO grade 1), which exhibited
(n = 7) that were completely lacking immunohistochem-
diffuse/moderate-strong expression (90% of tumor cells
ical expression of p53 was further analyzed for the
positive). For FIGO grade 1 endometrioid carcinomas,
presence of TP53 mutations (as earlier studies have
the percentage of positive tumor cells ranged from 10% to
shown that some p53-negative serous carcinomas can
harbor a TP53 mutation leading to lack of immunohis-
tochemically recognized protein).24,42 This was pursued to TABLE 1. The p16 Expression in Uterine Adenocarcinomas
allow for discussion of outlier results among the serous Age (y) p16 Expression
carcinomas (those lacking p53 expression) when assessing No. (Mean/ (% Positive Cells)
both the value and limitations of p16 as part of a marker Tumor Type Patients Median) Mean/Median Range
panel that includes p53 for distinguishing subtypes of Serous carcinoma 49 68/67 95/100 90-100
uterine adenocarcinomas. Genomic DNA was isolated Endometrial 101 61/62 38/30 10-90
from the paraffin blocks using Agencourt FormaPure Kit endometrioid
(Beckman Coulter). Codons 4 through 9 were analyzed. carcinoma
FIGO grade 1 44 62/64 37/30 10-90
Primers used for PCR amplification were described FIGO grade 2 40 61/58 36/30 10-80
earlier.39,42,46 Mutational analysis was performed by FIGO grade 3 17 61/61 42/40 10-90
nucleotide sequencing at the Core Facility of the Johns Endocervical 51 44/44 94/90 90-100
Hopkins Medical Institutions. All the mutations were adenocarcinoma
screened for known polymorphisms of the TP53 gene FIGO indicates Fédération Internationale de Gynécologie et d’Obstétrique.
using the database from the International Agency for

1506 | www.ajsp.com r 2009 Lippincott Williams & Wilkins

Am J Surg Pathol  Volume 33, Number 10, October 2009
FIGURE 1. Serous carcinoma of endometrium. The tumor displays classical (typical) papillary architecture (A) and high-grade
nuclear features (B). Tumor exhibits diffuse/strong expression of p16 (C) and p53 (D) (100% of tumor cells positive for both
markers).
90% (mean/median: 37%/30%). For FIGO grade 2
tumors, the percentage of positive tumor cells ranged
from 10% to 80% (mean/median: 36%/30%). For those
endometrioid carcinomas classified as FIGO grade 2
based on the presence of solid growth (n = 26), the
percentage of positive tumor cells ranged from 10% to
80% (mean/median: 35%/30%). For those endometrioid
carcinomas classified as FIGO grade 2 based on the
presence of notable (marked) nuclear atypia (n = 14), the
percentage of positive tumor cells ranged from 10% to
60% (mean/median: 39%/40%). For FIGO grade 3
endometrioid carcinomas, the percentage of positive
tumor cells ranged from 10% to 90% (mean/median:
42%/40%). Rereview of the 3 outliers with 90% staining
did not reveal any morphologic features to suggest that
these tumors should be reclassified.
All endocervical adenocarcinomas exhibited diffuse/
moderate-strong p16 expression, with percentage of
positive tumor cells ranging from 90% to 100%
r 2009 Lippincott Williams & Wilkins
P16 Expression in Uterine Adenocarcinomas
(mean/median: 94%/90%) (Fig. 5); HPV DNA was
detected in all of these [43 cases by in situ hybridization,
8 cases by PCR; HPV 16 in 35 cases, HPV 18 in 8,
both HPV 16 and 18 in 1, HPV 45 in 3, HPV 31 in 1, and
HPV type unknown in 3 (positive with wide spectrum
probe by in situ hybridization but PCR was not
performed)].
There were statistically significant differences in p16
expression among serous, endometrioid, and endocervical
adenocarcinomas (P<0.0001, ANOVA). Using a value
of >90% positive tumor cells as a diffuse positive
result, this pattern of p16 expression distinguished serous
carcinomas from endometrioid carcinomas (P<0.0001,
w2 test), and endometrial endometrioid carcinomas
from endocervical adenocarcinomas (P<0.0001, w2 test)
(Fig. 6). There were no significant differences in p16
expression among the different FIGO grades of endome-
trial endometrioid adenocarcinoma (P = 0.5, ANOVA)
(Fig. 7).
www.ajsp.com | 1507
Yemelyanova et al
FIGURE 2. Serous carcinoma of endometrium. Tumor has the typical microscopic features, with irregular, scalloped luminal
borders (A), and marked nuclear atypia (B). This example also exhibits diffuse/strong expression of p16 (C) but completely lacks
expression of p53 (D).
Immunohistochemical and Mutational Analysis
of TP53 in Serous Carcinomas
P53 expression was strong and diffuse (>90%
tumor cells positive) in 42 serous carcinomas and nega-
tive in 7 (Fig. 2). These 7 tumors that were lacking
immunohistochemical expression of p53 were subjected to
mutational analysis. Exons 4 through 9 were sequenced.
The alterations described as known TP53 polymorphisms
were not considered mutations in this analysis. Mutations
were identified in 3 cases. These included a nonsense
mutation in 1 (exon 6, codon 637, CGA-TGA), an
intronic mutation in 1 (nucleotide substitution in the
second base pair of the 30 flanking codon of exon 4), and
a silent mutation in 1 (exon 4, codon 213, CCT-CCC);
the latter was not considered a functional mutation
(ie, not a somatic change in nucleotide sequence lead-
ing to structural change of the encoded p53 protein).
The remaining 4 cases lacked mutations in the regions
analyzed.
1508 | www.ajsp.com
Am J Surg Pathol  Volume 33, Number 10, October 2009
DISCUSSION
Classifying endometrial adenocarcinomas into en-
dometrioid and serous types and determining the site of
origin of uterine adenocarcinomas (endometrial vs.
endocervical) are important for guiding surgical manage-
ment and assessing prognosis. Earlier studies have
characterized immunophenotypic differences between
serous and endometrioid carcinomas of the endometrium,
which can be exploited by using an immunohistochemical
marker panel, including p53, hormone receptors (ER and
PR), and Ki-67, to subtype these endometrial adenocar-
cinomas.9,14,25,31,40 In addition, recent studies have
demonstrated the value of ancillary techniques, including
immunohistochemical analysis of p16, ER, and PR
expression, and HPV DNA detection, for distinguishing
endometrial and endocervical adenocarcinomas (those of
endometrioid and mucinous types).6,26,41,48 In most situa-
tions, the differential diagnosis concerns either subtyping
of endometrial adenocarcinoma as endometrioid versus
r 2009 Lippincott Williams & Wilkins
Am J Surg Pathol  Volume 33, Number 10, October 2009 P16 Expression in Uterine Adenocarcinomas

FIGURE 3. P16 expression in endometrial endometrioid carcinomas. FIGO grade 1 endometrioid carcinoma (A) exhibits patchy
expression of p16 (B; approximately 20% positive tumor cells overall). FIGO grade 2 endometrioid carcinoma (C) exhibits patchy
expression of p16 (D; foci of rather diffuse staining admixed with areas that are negative, with approximately 60% positive tumor
cells overall). FIGO grade 3 endometrioid carcinoma (E) exhibits patchy expression of p16 (F; approximately 30% positive tumor
cells overall). FIGO indicates Fédération Internationale de Gynécologie et d’Obstétrique.

r 2009 Lippincott Williams & Wilkins www.ajsp.com | 1509

Yemelyanova et al Am J Surg Pathol  Volume 33, Number 10, October 2009

FIGURE 4. Patterns of p16 expression in endometrial endometrioid carcinomas. FIGO grade 1 endometrioid carcinoma (A)
displays extensive yet patchy expression of p16, with intervening negative cells (B). FIGO grade 3 endometrioid carcinoma (C)
displays a region of diffuse expression (D, left) and a region with only focal expression of p16 (D, right). A small sample from the
diffuse region could cause misclassification as either endocervical adenocarcinoma or serous carcinoma, depending on
morphology and p53 results. FIGO indicates Fédération Internationale de Gynécologie et d’Obstétrique.

serous or distinction of endometrial endometrioid carci- HPV-related endocervical adenocarcinomas, which we

noma from HPV-related endocervical adenocarcinoma.
A marker panel comprised of p53/ER/PR usually suffices
for the former and one comprised of p16/ER/PR usually
resolves the latter.
In the course of evaluating many uterine adenocar-
cinomas in routine and consultation practice cases, we
have occasionally encountered situations in which all 3
types of carcinomas were being considered in the
differential diagnosis of individual cases for a variety of
reasons (age of patient, location of tumor in a fractional
curettage specimen, morphologic features such as marked
nuclear atypia or papillary architecture, conflicting
immunophenotypic data using limited or traditional
markers such as carcinoembryonic antigen and vimentin).
As a result, we noticed anecdotally that uterine serous
carcinomas had striking p16 expression similar to
reported earlier in abstract form.47 Only a few studies
have recently described similar findings.13,37 This obser-
vation led us to systematically evaluate p16 expression in
a large series of uterine adenocarcinomas to explore the
value and limitations of this marker for the differential
diagnoses cited above.
This study establishes that uterine serous carcino-
mas are uniformly characterized by diffuse/moderate-
strong p16 expression identical to that encountered in
HPV-related endocervical adenocarcinomas. The exact
molecular mechanism underlying this diffuse expression
of p16 in serous carcinomas is not clear. Some studies
have suggested that it reflects abnormalities in the pRb
pathway.7,22 As HPV has not been implicated in the
pathogenesis of uterine serous carcinomas, it stands to
reason that other molecular alterations that interfere with

1510 | www.ajsp.com r 2009 Lippincott Williams & Wilkins

Am J Surg Pathol  Volume 33, Number 10, October 2009
FIGURE 5. Endocervical adenocarcinoma. Tumor has an
endometrioid appearance at low-magnification (A) and dis-
plays typical cytologic features (B; hybrid endometrioid and
mucinous features with mitotic figures and occasional apop-
totic bodies). Tumor exhibits diffuse/strong expression of p16
(C). In situ hybridization for HPV demonstrated the presence
of HPV 16 (not shown). HPV indicates human papillomavirus.
r 2009 Lippincott Williams & Wilkins
P16 Expression in Uterine Adenocarcinomas
FIGURE 6. Box plot illustrating p16 expression in uterine
serous carcinomas (USC), uterine endometrial endometrioid
carcinomas (UEC), and endocervical adenocarcinomas (ECC).
There were statistically significant differences between these
tumor types (Pr0.001, analysis of variance). For each group,
the box plot provides data as statistical values: the horizontal
line within the shaded box denotes the median value, the
lower and upper limits of the shaded box indicate the 25th
and 75th percentile values, the error bars (‘‘whiskers’’) below
and above the shaded boxes indicate the 10th and 90th
percentile values, and the black circles represent outliers.
the pRb pathway in such a way as to stimulate a futile
attempt to regulate the cell cycle would lead to p16
overexpression similar to that induced by high-risk HPV-
related oncoproteins. The exact nature of such alterations
remains to be determined by molecular analyses.
Serous carcinomas and endocervical adenocarcino-
mas are usually readily distinguished by morphology
alone, aided by the significant difference in mean/median
ages of the patients with these tumor types. On occasion,
however, papillary/villoglandular architecture and nota-
ble nuclear atypia can lead to consideration of these
tumor types in the differential diagnosis for an individual
case. P53 expression usually readily resolves this differ-
ential diagnosis. Serous carcinomas have a high frequency
of diffuse/strong p53 overexpression, which is highly
correlated with the presence of p53 mutations.24,43 In
contrast, HPV-related (usual type) endocervical adeno-
carcinomas, which represent approximately 90% of
cervical adenocarcinomas, usually do not harbor p53
mutations; thus, they are not expected to overexpress
p53.20,33,45 However, as demonstrated in this study and
earlier studies,24,43 a subset of serous carcinomas com-
pletely lacks expression of p53. Some of these tumors
have been shown to harbor p53 mutations, resulting in a
‘‘truncated’’ nonimmunoreactive protein. The reasons for
lack of expression in those tumors lacking mutations
include the presence of mutations in regions of the gene
not examined and other molecular mechanisms, including
epigenetic changes. Therefore, distinction of these tumor
types is generally only problematic for the subset of
serous carcinomas completely lacking p53 expression.
In this situation, as both serous carcinomas and
www.ajsp.com | 1511
Yemelyanova et al
FIGURE 7. Box plot illustrating p16 expression in different
Fédération Internationale de Gynécologie et d’Obstétrique
grades of endometrial endometrioid carcinomas. There were
no statistically significant differences among the 3 groups
(P = 0.5, analysis of variance). For each group, the box plot
provides data as statistical values: the horizontal line within the
shaded box denotes the median value, the lower and upper
limits of the shaded box indicate the 25th and 75th percentile
values, the error bars (‘‘whiskers’’) below and above the
shaded boxes indicate the 10th and 90th percentile values,
and the black circles represent outliers.
HPV-related endocervical adenocarcinomas share diffuse/
strong p16 expression and most often also lack hormone
receptor expression, a serous carcinoma lacking p53
expression could be confused with an HPV-related
endocervical adenocarcinoma. Consequently, awareness
of the shared p16 expression patterns of these tumor types
is important.
The more common problem in differential diagnosis
of uterine adenocarcinomas concerns subclassification of
endometrial adenocarcinomas as endometrioid versus
serous. Many cases can be distinguished by morphology
alone or with the assistance of a marker panel comprised
of p53, ER, and PR. However, these tools do not resolve
all cases. Both serous and endometrioid carcinomas can
exhibit either glandular or papillary architecture and some
endometrioid carcinomas can acquire notable nuclear
atypia. Consequently, glandular variants of serous carci-
noma can be confused with those endometrioid carcino-
mas that are assigned FIGO grade 2 on the basis of
nuclear atypia (these have at least 95% glandular
architecture with notable nuclear atypia). In addition,
papillary variants of endometrioid carcinoma with atypia
or metaplastic (papillary eosinophilic or hobnail) changes
can be confused with serous carcinoma. Furthermore,
immunoprofiles of these tumors occasionally overlap.
Some serous carcinomas do retain expression of hormone
receptors and a subset lacks p53 expression, as discussed
above. Some endometrioid carcinomas can lose hormone
receptor expression and a subset of the higher-grade
tumors can overexpress p53 to the extent encountered in
1512 | www.ajsp.com
Am J Surg Pathol  Volume 33, Number 10, October 2009
serous carcinomas. Therefore, additional markers for the
distinction of these tumor types would be useful.
The diffuse/strong p16 expression characteristic of
serous carcinomas can be exploited in the subtyping of
endometrial adenocarcinomas as endometrioid versus
serous. Earlier studies and this study have demonstrated
that, with few exceptions, endometrial endometrioid
carcinomas do not exhibit the diffuse/strong p16 expres-
sion pattern that characterizes serous carcinomas.13,37,47
Expression in endometrioid carcinomas is variable and
can occasionally be extensive, but most of tumors have
some degree of patchy staining, with negative areas
scattered throughout the tumor. The importance of the
staining distribution pattern in these tumors cannot be
overemphasized. Limited specimens, such as small biop-
sies or the punch specimens used in tissue microarrays,
have the potential to prevent recognition of the staining
pattern in a given tumor. Therefore, the use of p16 in this
distinction is best accomplished with specimens having
sufficient material to reveal the staining pattern.
On the basis of this study, we recommend addition
of p16 to a marker panel including p53, ER, and PR for
subtyping uterine adenocarcinomas. In our experience
and that in the literature, typical serous carcinomas are
p53 diffuse+/p16 diffuse+/ER  /PR  , with some
exceptions as noted above. Typical endometrial endome-
trioid carcinomas are p53  or patchy+/p16 patchy+/
ER+/PR+, again with some exceptions as noted above.
Typical endocervical adenocarcinomas are p16 diffuse+/
p53  /ER  /PR  , with some retaining focal/limited
expression of hormone receptors. We specifically recom-
mend the use of PR in conjunction with ER based on our
published38,41,48 and unpublished experiences indicating
that the subsets of both endocervical and serous
carcinomas that retain some expression of hormone
receptors most often retain ER expression to some degree
(often focal/weak-to-moderate) but often entirely lose PR
expression; thus, PR is the more discriminatory marker of
the two when trying to distinguish these two tumor
subtypes from endometrial endometrioid carcinomas.
HPV DNA detection is recommended for resolution of
tumors (those not confined to the cervix) with the p16
diffuse+/p53  /ER  /PR  immunoprofile, which is the
profile of HPV-related endocervical adenocarcinomas
shared by a subset of uterine serous carcinomas. We
have been using this marker panel for several years to
classify uterine adenocarcinomas in biopsy and curettage
specimens and thus guide surgical management. In our
experience, most cases are readily classified based on the
expected profiles described above. Formal analysis of a
prospective series correlated with hysterectomy findings is
warranted. Occasionally, problematic high-grade endo-
metrial adenocarcinomas with morphologic and immuno-
phenotypic features intermediate between typical serous
and endometrioid types are encountered. These tumors
were excluded from this study and are intended for
separate subsequent analysis. Such cases raise interesting
questions regarding the parameters and modalities used
to define and classify tumors, namely, whether specific
r 2009 Lippincott Williams & Wilkins
Am J Surg Pathol  Volume 33, Number 10, October 2009
tumor categories should be defined by a combination of
morphologic, immunophenotypic, and molecular fea-
tures.
REFERENCES
1. IARC TP53 Mutation Database [internet database]. International
Agency for Research on Cancer. Available at: http://www-p53.iarc.fr/
PolymorphismView.asp. Accessed on: December 11, 2008.
2. Alkushi A, Lim P, Coldman A, et al. Interpretation of p53
immunoreactivity in endometrial carcinoma: establishing a clinically
relevant cut-off level. Int J Gynecol Pathol. 2004;23:129–137.
3. Amant F, Moerman P, Neven P, et al. Endometrial cancer. Lancet.
2005;366:491–505.
4. Andersson S, Larson B, Hjerpe A, et al. Adenocarcinoma of the
uterine cervix: the presence of human papillomavirus and the
method of detection. Acta Obstet Gynecol Scand. 2003;82:960–965.
5. Andersson S, Rylander E, Larson B, et al. Types of human
papillomavirus revealed in cervical adenocarcinomas after DNA
sequencing. Oncol Rep. 2003;10:175–179.
6. Ansari-Lari MA, Staebler A, Zaino RJ, et al. Distinction of
endocervical and endometrial adenocarcinomas: immunohistochem-
ical p16 expression correlated with human papillomavirus (HPV)
DNA detection. Am J Surg Pathol. 2004;28:160–167.
7. Armes JE, Lourie R, de SM, et al. Abnormalities of the RB1
pathway in ovarian serous papillary carcinoma as determined by
overexpression of the p16(INK4A) protein. Int J Gynecol Pathol.
2005;24:363–368.
8. Baergen RN, Warren CD, Isacson C, et al. Early uterine serous
carcinoma: clonal origin of extrauterine disease. Int J Gynecol
Pathol. 2001;20:214–219.
9. Bonfitto VL, de Angelo Andrade LA. p53, estrogen and progester-
one receptors in diagnostic curettage for endometrial adenocarci-
noma and their correlation with morphological data and disease
stage at hysterectomy. Sao Paulo Med J. 2003;121:163–166.
10. Carcangiu ML, Chambers JT. Uterine papillary serous carcinoma:
a study on 108 cases with emphasis on the prognostic significance
of associated endometrioid carcinoma, absence of invasion, and
concomitant ovarian carcinoma. Gynecol Oncol. 1992;47:298–305.
11. Carcangiu ML, Chambers JT. Early pathologic stage clear cell
carcinoma and uterine papillary serous carcinoma of the endome-
trium: comparison of clinicopathologic features and survival. Int J
Gynecol Pathol. 1995;14:30–38.
12. Carcangiu ML, Tan LK, Chambers JT. Stage IA uterine serous
carcinoma: a study of 13 cases. Am J Surg Pathol. 1997;21:
1507–1514.
13. Chiesa-Vottero AG, Malpica A, Deavers MT, et al. Immunohisto-
chemical overexpression of p16 and p53 in uterine serous carcinoma
and ovarian high-grade serous carcinoma. Int J Gynecol Pathol.
2007;26:328–333.
14. Darvishian F, Hummer AJ, Thaler HT, et al. Serous endometrial
cancers that mimic endometrioid adenocarcinomas: a clinicopatho-
logic and immunohistochemical study of a group of problematic
cases. Am J Surg Pathol. 2004;28:1568–1578.
15. Egan JA, Ionescu MC, Eapen E, et al. Differential expression of
WT1 and p53 in serous and endometrioid carcinomas of the
endometrium. Int J Gynecol Pathol. 2004;23:119–122.
16. Inoue M, Okayama A, Fujita M, et al. Clinicopathological
characteristics of p53 overexpression in endometrial cancers. Int J
Cancer. 1994;58:14–19.
17. Iwasawa A, Nieminen P, Lehtinen M, et al. Human papillomavirus
DNA in uterine cervix squamous cell carcinoma and adenocarcino-
ma detected by polymerase chain reaction. Cancer. 1996;77:
2275–2279.
18. Jones MW, Kounelis S, Papadaki H, et al. Well-differentiated
villoglandular adenocarcinoma of the uterine cervix: oncogene/
tumor suppressor gene alterations and human papillomavirus
genotyping. Int J Gynecol Pathol. 2000;19:110–117.
19. Keating JT, Cviko A, Riethdorf S, et al. Ki-67, cyclin E, and
p16INK4 are complimentary surrogate biomarkers for human
papilloma virus-related cervical neoplasia. Am J Surg Pathol. 2001;
25:884–891.
r 2009 Lippincott Williams & Wilkins
P16 Expression in Uterine Adenocarcinomas
20. Kessis TD, Slebos RJ, Han SM, et al. p53 gene mutations and
MDM2 amplification are uncommon in primary carcinomas of the
uterine cervix. Am J Pathol. 1993;143:1398–1405.
21. Klaes R, Friedrich T, Spitkovsky D, et al. Overexpression of
p16(INK4A) as a specific marker for dysplastic and neoplastic
epithelial cells of the cervix uteri. Int J Cancer. 2001;92:276–284.
22. Kommoss S, du BA, Ridder R, et al. Independent prognostic
significance of cell cycle regulator proteins p16(INK4a) and pRb in
advanced-stage ovarian carcinoma including optimally debulked
patients: a translational research subprotocol of a randomised study
of the Arbeitsgemeinschaft Gynaekologische Onkologie Ovarian
Cancer Study Group. Br J Cancer. 2007;96:306–313.
23. Lax SF, Pizer ES, Ronnett BM, et al. Clear cell carcinoma of the
endometrium is characterized by a distinctive profile of p53, Ki-67,
estrogen, and progesterone receptor expression. Hum Pathol. 1998;
29:551–558.
24. Lax SF, Kendall B, Tashiro H, et al. The frequency of p53, K-ras
mutations, and microsatellite instability differs in uterine endome-
trioid and serous carcinoma: evidence of distinct molecular genetic
pathways. Cancer. 2000;88:814–824.
25. Lomo L, Nucci MR, Lee KR, et al. Histologic and immunohisto-
chemical decision-making in endometrial adenocarcinoma. Mod
Pathol. 2008;21:937–942.
26. McCluggage WG, Jenkins D. p16 immunoreactivity may assist in
the distinction between endometrial and endocervical adenocarci-
noma. Int J Gynecol Pathol. 2003;22:231–235.
27. Milde-Langosch K, Riethdorf S, Kraus-Poppinghaus A, et al.
Expression of cyclin-dependent kinase inhibitors p16 MTS1,
p21WAF1, and p27KIP1 in HPV-positive and HPV-negative
cervical adenocarcinomas. Virchows Arch. 2001;439:55–61.
28. Missaoui N, Hmissa S, Frappart L, et al. p16INK4A overexpression
and HPV infection in uterine cervix adenocarcinoma. Virchows
Arch. 2006;448:597–603.
29. Murphy N, Heffron CC, King B, et al. p16INK4A positivity in
benign, premalignant and malignant cervical glandular lesions: a
potential diagnostic problem. Virchows Arch. 2004;445:610–615.
30. Negri G, Egarter-Vigl E, Kasal A, et al. p16INK4a is a useful
marker for the diagnosis of adenocarcinoma of the cervix uteri and
its precursors: an immunohistochemical study with immunocyto-
chemical correlations. Am J Surg Pathol. 2003;27:187–193.
31. Nofech-Mozes S, Khalifa MA, Ismiil N, et al. Immunophenotyping
of serous carcinoma of the female genital tract. Mod Pathol. 2008;
21:1147–1155.
32. O’Neill CJ, McCluggage WG. p16 expression in the female genital
tract and its value in diagnosis. Adv Anat Pathol. 2006;13:8–15.
33. Paquette RL, Lee YY, Wilczynski SP, et al. Mutations of p53 and
human papillomavirus infection in cervical carcinoma. Cancer. 1993;
72:1272–1280.
34. Park TW, Zivanovic O, Theuerkauf I, et al. The diagnostic utility of
human papillomavirus-testing in combination with immunohisto-
chemistry in advanced gynaecologic pelvic tumours: a new
diagnostic approach. Int J Oncol. 2004;24:829–836.
35. Pirog EC, Kleter B, Olgac S, et al. Prevalence of human
papillomavirus DNA in different histological subtypes of cervical
adenocarcinoma. Am J Pathol. 2000;157:1055–1062.
36. Plunkett M, Brestovac B, Thompson J, et al. The value of HPV
DNA typing in the distinction between adenocarcinoma of
endocervical and endometrial origin. Pathology. 2003;35:397–401.
37. Reid-Nicholson M, Iyengar P, Hummer AJ, et al. Immunopheno-
typic diversity of endometrial adenocarcinomas: implications for
differential diagnosis. Mod Pathol. 2006;19:1091–1100.
38. Ronnett BM, Yemelyanova AV, Vang R, et al. Endocervical
adenocarcinomas with ovarian metastases: analysis of 29 cases with
emphasis on minimally invasive cervical tumors and the ability of
the metastases to simulate primary ovarian neoplasms. Am J Surg
Pathol. 2008;32:1835–1853.
39. Salani R, Kurman RJ, Giuntoli R, et al. Assessment of TP53
mutation using purified tissue samples of ovarian serous carcinomas
reveals a higher mutation rate than previously reported and does not
correlate with drug resistance. Int J Gynecol Cancer. 2008;18:
487–491.
www.ajsp.com | 1513
Yemelyanova et al
40. Sherman ME, Bur ME, Kurman RJ. p53 in endometrial cancer and
its putative precursors: evidence for diverse pathways of tumorigen-
esis. Hum Pathol. 1995;26:1268–1274.
41. Staebler A, Sherman ME, Zaino RJ, et al. Hormone receptor
immunohistochemistry and human papillomavirus in situ hybridiza-
tion are useful for distinguishing endocervical and endometrial
adenocarcinomas. Am J Surg Pathol. 2002;26:998–1006.
42. Tashiro H, Isacson C, Levine R, et al. p53 gene mutations are
common in uterine serous carcinoma and occur early in their
pathogenesis. Am J Pathol. 1997;150:177–185.
43. Tashiro H, Lax SF, Gaudin PB, et al. Microsatellite instability is
uncommon in uterine serous carcinoma. Am J Pathol. 1997;150:
75–79.
44. Tenti P, Romagnoli S, Silini E, et al. Human papillomavirus types
16 and 18 infection in infiltrating adenocarcinoma of the cervix:
PCR analysis of 138 cases and correlation with histologic type and
grade. Am J Clin Pathol. 1996;106:52–56.
45. Tommasino M, Accardi R, Caldeira S, et al. The role of TP53 in
cervical carcinogenesis. Hum Mutat. 2003;21:307–312.
1514 | www.ajsp.com
Am J Surg Pathol  Volume 33, Number 10, October 2009
46. Vang R, Shih I, Salani R, et al. Subdividing ovarian and peritoneal
serous carcinoma into moderately differentiated and poorly
differentiated does not have biologic validity based on molecular
genetic and in vitro drug resistance data. Am J Surg Pathol. 2008;
32:1667–1674.
47. Yemelyanova A, Ji H, Ronnett BM. Diffuse p16 expression
distinguishes uterine serous carcinoma (USC) from uterine endome-
trioid carcinoma (UEC) of endometrial origin but not HPV-related
endocervical adenocarcinoma (ECC). Mod Pathol. 2006;19:
202A–203A.
48. Yemelyanova A, Vang R, Seidman JD, et al. Endocervical
adenocarcinomas with prominent endometrial or endomyometrial
involvement simulating primary endometrial carcinomas: utility of
HPV DNA detection and immunohistochemical expression of p16
and hormone receptors to confirm the cervical origin of the corpus
tumor. Am J Surg Pathol. 2009;33:914–924.
49. Young RH, Clement PB. Endocervical adenocarcinoma and its
variants: their morphology and differential diagnosis. Histopatho-
logy. 2002;41:185–207.
r 2009 Lippincott Williams & Wilkins