Research www. AJOG.

org

ONCOLOGY
Are endometrial polyps true cancer precursors?
Tamar Perri, MD; Kurosh Rahimi, MD; Agnihotram V. Ramanakumar, PhD; Karen Wou, MD;
Dragana Pilavdzic, MD; Eduardo L. Franco, PhD; Walter H. Gotlieb, MD, PhD; Alex Ferenczy, MD

OBJECTIVE: The purpose of this study was to assess whether endome-
trial polyps (EMPs) represent cancer precursors.
STUDY DESIGN: Age standardized incidence ratios (SIRs) of histologi-
cally verified endometrial cancers (EmCas) were estimated in women
with EMPs and in women with uterine leiomyomata, which is a condition
that is unrelated to endometrial carcinogenesis. SIRs were calculated
as the ratio of observed to expected EmCas based on age-specific inci-
dence rates for female Montreal residents during the same period.
RESULTS: Of 1467 women with EMPs, 125 (8.5%) had EmCa. Of 1138
patients with uterine leiomyomata, 133 (11.7%) had EmCa. The SIRs of
EmCa for women with EMPs (odds ratio, 8.0; 95% confidence interval,
6.6 –9.5) were significantly lower than that in women with leiomyomata
(odds ratio, 19.1; 95% confidence interval, 16.0 –22.6). Abnormal
uterine bleeding was the main reason for evaluating patients with EMP
with or without associated EmCa.
CONCLUSION: The findings of higher EmCa incidence are consistent
with enhanced detection opportunity rather than with the endometrial
cancer precursor potential of EMPs.
Key words: cancer precursor, endometrial cancer, endometrial polyp,
leiomyomata, uterine bleeding

Cite this article as: Perri T, Rahimi K, Ramanakumar AV, et al. Are endometrial polyps true cancer precursors? Am J Obstet Gynecol 2010;203:232.e1-6.

E ndometrial polyps (EMPs) are lo-

calized exophytic overgrowths of
endometrial mucosa. They have not
been described in premenarcheal girls.
Their incidence peaks in the fifth decade
of life and gradually declines after meno-
pause. The prevalence of EMPs in the
general population has been estimated to
range between 7.8% and 25%.1-3 They
are, by far, the most frequent endome-
trial disease in menopausal women.
From the Division of Gynecologic Oncology,
Department of Obstetrics and Gynecology
(Drs Perri, Wou, Gotlieb, and Ferenczy), and
the Department of Pathology (Drs Rahimi,
Pilavdzic, and Ferenczy), Jewish General
Hospital; the Division of Cancer
Epidemiology, Departments of Oncology
and Epidemiology (Drs Ramanakumar and
Franco), McGill University, Montreal,
Quebec, Canada.
Reprints: Alex Ferenczy, MD, Professor of
Pathology and Obstetrics & Gynecology,
Department of Pathology, Jewish General
Hospital, 3755 Cote Sainte-Catherine Rd.,
Montreal, Quebec, Canada H3T 1E2.
alex.ferenczy@mcgill.ca.
Authorship and contribution to the article is
limited to the 8 authors indicated. There was
no outside funding or technical assistance with
the production of this article.
0002-9378/$36.00
© 2010 Mosby, Inc. All rights reserved.
doi: 10.1016/j.ajog.2010.03.036
EMPs are often associated with bleeding;
however, they may also be diagnosed in-
cidentally in hysterectomy specimens or
by screening or diagnostic endovaginal
ultrasonography or hysteroscopy.
Traditional teaching has suggested
that EMPs are cancer precursors.4 The
risk of EMPs that are associated with en-
dometrial carcinoma (EmCa) increases
with age, with the highest rate in patients
⬎65 years in whom 32% of the polyps
are associated with invasive or noninva-
sive malignancy.5,6 The prevalence of
malignant change in EMPs varies from
0.1-13%.2,3,5-10 Despite these observa-
tions, there is no credible evidence that
EMP per se is a true cancer precursor or
simply represents focal mucosal out-
growth with similar biologic characteris-
tics and behavior as nonpolyp-contain-
ing endometrium. It is also unclear
whether the perceived high rate of Em-
Cas in patients with EMPs is merely an
incidental finding during the diagnostic
workup of EMPs, because these patients
seek medical attention because of bleed-
ing and other symptoms and thus, are
subjected to endometrial evaluations.
The purpose of our study was to evaluate
the prevalence of EmCa in a large num-
ber of patients with EMPs and to com-
pare these findings with the expected
EmCa occurrence in these patients using
age-specific incidence rates for the Mon-
treal population within the same period.
We also aimed to evaluate whether can-
cer characteristics are different in pa-
tients with polyps, compared with can-
cers that are diagnosed in patients
without polyps.
P ATIENTS AND M ETHODS
Internal review board approval was
granted for the study by the Research and
Ethics Committee of the Jewish General
Hospital, which is an institution affili-
ated with McGill University in Montreal,
Canada. All patients who were diagnosed
consecutively with EMPs between the
years of 2000 and 2007 were identified
with the computerized database of the
Department of Pathology in our institu-
tion (n ⫽ 1880). Clinical and demo-
graphic data were extracted from patho-
logic requisition forms that included age at
diagnosis, reason for referral, menopausal
status, use of hormone replacement ther-
apy, tamoxifen therapy, presence of other
cancers, and means of endometrial evalu-
ation for histologic evaluation (eg, biopsy,
curettage).
All histologic slides were reviewed in-
dependently by 2 experienced gyneco-
logic pathologists. Discrepant cases were
reviewed by a third pathologist who was
blinded to the diagnoses of the first 2 pa-
thologists. In situations of disagreement,
the 3 pathologists reached a consensus

232.e1 American Journal of Obstetrics & Gynecology SEPTEMBER 2010

www.AJOG.org
diagnosis using a multiheaded micro-
scope. Cases without adjacent endome-
trium to EMPs and polyp mimics (such
as polypoid carcinoma, hyperplasia or
cyclic endometrium, basalis and lower
uterine segment endometrium) were ex-
cluded (n ⫽ 413). Cases with poor he-
matoxylin-eosin staining were recut and
restained (n ⫽ 31). The number of his-
tologic slides with or without cancer and
polyp size (greatest diameter measured
with calibrated microscopic fields) was
recorded. Polyps were histologically
classified: 1, benign: functional or atro-
phic; 2, hyperplastic: simple hyperplasia
without atypia, complex hyperplasia
without atypia or both, or endometrial
intraepithelial neoplasia (EIN)/atypical
hyperplasia; and 3, cancerous. Cancer
cases were classified in the following
manner: 1, EMP with primary cancer
(cancer confined exclusively to EMP; ad-
jacent endometrium devoid of cancer);
2, EMP with synchronous or simulta-
neous cancer (cancer in EMP and adja-
cent endometrium); and 3, noncancerous
EMP with cancer only in adjacent endo-
metrium. The adjacent endometrium was
classified as proliferative, secretory, men-
strual, atrophic, disordered persistent pro-
liferative, nonatypical and atypical hyper-
plasia, or invasive adenocarcinoma.
EmCas and, if appropriate, sarcomas
were classified histologically according
to accepted nomenclature.4 Cancer cases
were graded and surgically staged ac-
cording to Fédération Internationale de
Gynécologie et d’Obstétrique (FIGO;
1988). All EmCa cases without EMP that
had been diagnosed between the years of
2000 and 2007 were identified with the
Pathology Department’s computerized
data base. The histologic slides were re-
viewed by 1 of the pathologists (A.F.) to
ensure the absence of EMP in these spec-
imens. Age at diagnosis, tumor grade,
and surgical stage were recorded.
The observed rate of associated malig-
nancy in patients with EMPs was strati-
fied by 5-year age groups and compared
with the equivalent expected incidence
of EmCa, with the assumption that can-
cer risk in these patients was the same as
that of Montreal women of the same age
and period of cancer diagnosis. Annual
age-specific rates were obtained from the
Quebec population-based cancer regis-
try for the years of 2001-2005. Because
incidence rates of cancer fluctuate over
time, we used 2 sets of estimates: 1 esti-
mate that assumed that our patients ex-
perienced the lowest incidence rates of
EmCa reported in the cancer registry for
female residents of Montreal (“conser-
vative”) and 1 estimate that assumed the
highest rates recorded during the same
period (“liberal”).
As a secondary objective, we examined
the extent of overdetection of EmCa in
EMP cases that resulted from the afore-
mentioned analysis with an analysis that
was based on another uterine disease
that may also prompt an incidental dis-
covery of EmCa. Therefore, we repeated
the same exercise for all patients who
were found with uterine leiomyomata on
pathologic investigation at our institu-
tion between the years of 2000 and 2007.
As for EMP cases, these were retrieved
from the same computerized database of
the Department of Pathology and re-
viewed as described earlier.
We used Wilcoxon’s rank sum test to
compare means for interval-scaled vari-
ables and ␹2 tests for categoric or ordinal
variables. The significance of trends for
ordinal categories between groups was
tested with the Cochran-Mantel-Haens-
zel method. To examine EmCa occur-
rence among women with EMPs and
leiomyomata, we calculated age stan-
dardized incidence ratios (SIRs) and
95% CIs. Expected numbers were calcu-
lated based on the cumulative rates that
were derived from the annual age-spe-
cific incidence rates of endometrial can-
cer for Montreal from 2001-2005 and
cumulative age-defined person-time de-
nominators for the 2 patient groups,
EMPs, and leiomyomata. SIRs indicate
how much more common EmCa was in
each of the patient subsets (EMPs and
leiomyomata) than in female Montreal
residents of the same age.
R ESULTS
Table 1 shows the clinical characteristics of
the 1467 patients with EMP with and with-
out cancer and the initial diagnostic meth-
ods that were used. The mean number of
histologic slides ranged from 3 (endome-
SEPTEMBER 2010 American Journal of Obstetrics & Gynecology
Oncology Research
trial biopsy specimens) to 12 (hysterec-
tomy specimens). EMPs ranged from 2-5
cm (mean 1.8 cm) in the largest diameter.
Consensus agreement by 2 pathologists
was achieved in 1366 cases (93.1%), and
3-way consensus reading was carried out
in the remaining 101 cases. Women with
cancerous EMPs, whether primary or si-
multaneous (synchronous), were older
(P ⬍ .0001) than those women without as-
sociated cancer; the majority of occur-
rences were in postmenopausal women af-
ter the age of 55 years. Abnormal bleeding
was the most common reason for referral
in both groups but was significantly more
common in cancer-associated EMPs (P ⫽
.003). We detected endometrial malig-
nancy in 125 cases (8.5%) in which EMPs
were found on pathologic examination.
Primary and synchronous cancerous
EMPs accounted for 0.89% and 4.57% of
all EMPs, respectively; 3.07% EMPs were
noncancerous polyps with cancer in the
adjacent endometrium.
Table 2 shows the histologic types of
EMPs without and with carcinoma.
There was a significant difference in dis-
tributions of histologic categories be-
tween the 2 groups (P ⬍ .0001). The ma-
jority of noncancerous EMPs were of
the functional/atrophic type (71.2%).
Among the hyperplastic polyps, simple
hyperplasia without atypia dominated.
Endometrial intraepithelial neoplasia/
atypical hyperplasia occurred in only 55
of the 1386 evaluable EMPs (4.0%). Only
2 of 44 evaluable patients with cancer
and polyps (4.5%) were of the functional
type, and both were synchronous. The
endometrium adjacent to the EMP with-
out associated cancer was mainly of the
nonhyperplastic, proliferative, secretory,
or atrophic type (91%). In the 125 cases
of both primary and simultaneous can-
cerous polyps, the adjacent noncancer-
ous endometrium was mainly atrophic.
Table 3 shows the FIGO surgical
stages, grades, and histologic types of en-
dometrial cancers with and without
EMP and leiomyomata uteri. During the
same time period, 195 endometrial can-
cers were diagnosed in our hospital with-
out associated EMPs or leiomyomata. Of
1138 patients who were found to have
232.e2
Research Oncology www.AJOG.org

pared to women with (2.4%) or without

TABLE 1 EMPs (5.1%). Papillary serous carci-
Patients’ clinical characteristics noma occurred in similar frequency in
Endometrial polyps all 3 groups of women. Of the 13 cancer-
Without associated With associated Significance ous polyps with normal surrounding en-
Characteristics cancer (n ⴝ 1342) cancer (n ⴝ 125) P value dometrium, 9 polyps were FIGO grade 1,
Age at diagnosis, ya 56 (22–94) 66 (38–90) ⬍ .0001 well-differentiated endometrioid adeno-
..............................................................................................................................................................................................................................................
carcinoma; 3 polyps were papillary se-
Age at diagnosis, n (%) ⬍ .0001
..................................................................................................................................................................................................................................... rous, and 1 polyp was clear-cell
⬍45 y 214 (16.0) 7 (5.6) adenocarcinomas.
.....................................................................................................................................................................................................................................
45-54 y 400 (29.8) 11 (8.8) The age at diagnosis of patients with
.....................................................................................................................................................................................................................................
55-64 y 397 (29.6) 38 (30.4) EMP and leiomyomata is presented in
.....................................................................................................................................................................................................................................
Table 4. On average, women with EmCa
⬎65 y 331 (24.7) 69 (55.2)
.............................................................................................................................................................................................................................................. were older, regardless of whether they
Menopausal status, n (%) ⬍ .0001 had EMPs or leiomyomata, compared to
.....................................................................................................................................................................................................................................
Premenopausal 259 (19.3) 10 (8.0) women with either of these conditions
.....................................................................................................................................................................................................................................
Postmenopausal 660 (49.2) 104 (83.2) without cancer (P ⫽ .0001 for both
.....................................................................................................................................................................................................................................
comparisons).
Perimenopausal 23 (1.7) 2 (1.6)
..................................................................................................................................................................................................................................... Age-specific and overall SIRs for the
b
Unknown 400 (29.8) 9 (7.2) occurrence of EmCa among women who
..............................................................................................................................................................................................................................................
Reason for referral, n (%) .0002 had EMPs or uterine leiomyomata dur-
.....................................................................................................................................................................................................................................
Abnormal bleeding 703 (52.4) 83 (66.4) ing 2000-2007 are presented in Table 5.
.....................................................................................................................................................................................................................................
For all age groups and overall, the ob-
Ultrasound findings 158 (11.8) 17 (13.6)
..................................................................................................................................................................................................................................... served occurrence of EmCa in patients
Coincident finding during 193 (14.4) 5 (4.0) with these 2 conditions was greater than
hysterectomy for other would be expected for Montreal women
benign cause
..................................................................................................................................................................................................................................... of the same age and during the same pe-
Coincident finding during 59 (4.4) 7 (5.6) riod. The SIRs for EMPs were 13.8 (95%
hysterectomy for other
benign cause for
confidence interval [CI], 11.4 –16.4) and
malignancy 8.0 (95% CI, 6.6 –9.5) for conservative
.....................................................................................................................................................................................................................................
and liberal estimates, respectively.
Abnormal Papanicolaou test 12 (0.9) 4 (3.2)
..................................................................................................................................................................................................................................... Equivalently, SIRs for EmCa in patients
Tamoxifen users 23 (1.7) 0 with leiomyomata were 35.4 (95% CI,
.....................................................................................................................................................................................................................................
Hormone replacement 26 (1.9) 1 (0.8) 29.6 – 42.0) for the conservative esti-
therapy monitoring mate, and 19.1 (95% CI, 16.0 –22.6) for
.....................................................................................................................................................................................................................................
Cervical polyp 37 (2.8) 2 (1.6) the liberal estimate. The 95% CIs for the
.....................................................................................................................................................................................................................................
Other 131 (9.8) 6 (4.8) respective estimates for these 2 condi-
..............................................................................................................................................................................................................................................
tions are nonoverlapping, which indi-
Means of initial diagnosis, n (%) ⬍ .0001
..................................................................................................................................................................................................................................... cates that EmCas were significantly more
Hysteroscopy 121 (9.0) 2 (1.6) common than expected in patients with
.....................................................................................................................................................................................................................................
Curettage 102 (7.6) 2 (1.6) leiomyomata than in patients with EMPs
.....................................................................................................................................................................................................................................
Endometrial biopsy 794 (59.2) 16 (12.8) during the same period.
.....................................................................................................................................................................................................................................
Hysterectomy 325 (24.2) 105 (84.0)
..............................................................................................................................................................................................................................................
a
Data are given as median (range); b Not considered for testing significance. C OMMENT
Perri. Are endometrial polyps true cancer precursors? Am J Obstet Gynecol 2010.
We have undertaken this retrospective,
single-center study on a large number of
leiomyomata in their hysterectomy spec- .0009). The endometrioid adenocarci- consecutive, histologically verified EMPs
imens, 133 (11.7%) also had an EmCa. noma variant dominated the other types, with the aim of assessing their risk of be-
This was a higher frequency than the regardless of whether the cancer was ing associated with carcinoma. Of the
prevalence of EmCa in EMPs (8.5%). alone or associated with EMP or leiomy- 1467 evaluable EMP cases from 2000-
EmCa stage among patients with EMPs omata. Most endometrioid adenocarci- 2007, there were 125 (8.5%) associated
tended to be earlier than that of patients nomas were FIGO grade 1 and 2. Clear- with cancer, including 13 cases (0.89%)
without polyps (P for trend ⫽ .002). This cell adenocarcinoma was significantly confined exclusively to EMPs, and 112
was also true for EmCa stage among pa- more frequent in the group with coexist- cases (4.5%) that included both the
tients with leiomyomata (P for trend ⫽ ing uterine leiomyomata (17.3%), com- polyp and adjacent endometrium. In

232.e3 American Journal of Obstetrics & Gynecology SEPTEMBER 2010

www.AJOG.org Oncology Research

contrast, in 3.1% of patients, EmCa was

TABLE 2 confined solely to the endometrium that
Histology of endometrial polyps was adjacent to the EMP. During the
Status, n (%) same period, 195 cases of endometrial
Without With cancer were encountered without EMPs,
Histologic type cancer cancera Total, n (%) and 133 (11.7%) were diagnosed in 1138
Functional 295 (22.0) 2 (4.5) 297 (21.4) consecutive cases of uterine leiomyo-
..............................................................................................................................................................................................................................................
mata that were seen at our institution.
Atrophic 532 (39.6) 15 (34.1) 547 (39.5)
.............................................................................................................................................................................................................................................. These findings are consistent with the
Mixed 129 (9.6) 0 129 (9.3) hypothesis that EMPs per se do not con-
..............................................................................................................................................................................................................................................
Hyperplastic simple hyperplasia without atypia 306 (22.8) 18 (40.9) 324 (23.4) stitute a precursor lesion that carries a
..............................................................................................................................................................................................................................................
Hyperplastic complex hyperplasia without 6 (0.4) 0 6 (0.4) high risk of malignancy. Detection bias
atypia consequent to the diagnostic workup of
..............................................................................................................................................................................................................................................
Hyperplastic simple hyperplasia without atypia 27 (2.0) 1 (2.3) 28 (2.0)
EMPs explains the apparent increased
and complex hyperplasia without atypia risk of EmCa. When we considered ab-
..............................................................................................................................................................................................................................................
normal uterine bleeding (eg, symptom-
Endometrial intraepithelial neoplasia/atypical 47 (3.5) 8 (18.2) 55 (4.0)
hyperplasia atic patients), the high rate of endome-
..............................................................................................................................................................................................................................................
trial cancer in women with EMPs
Total 1342 (100) 44 (100) 1386 (100)
.............................................................................................................................................................................................................................................. seemed to be due to detection bias,
a
Only among evaluable cases. rather than endometrial cancer precur-
Perri. Are endometrial polyps true cancer precursors? Am J Obstet Gynecol 2010.
sor potential. Both EMPs and endome-
trial cancer produce abnormal uterine
bleeding that lead to histologic evalua-
tion of the endometrial cavity. This, in
TABLE 3 turn, leads to the discovery of both EMPs
Classification of endometrial cancers and associated cancer. Akin to our find-
ings, in studies in which distinction be-
Cancer tween primary and simultaneous can-
Alone With leiomyoma With polyps cerous EMPs was made, the rates of
(n ⴝ 195) (n ⴝ 133) (n ⴝ 125) cancer that was confined exclusively to
Classification n % n % n % EMPs was a rare event, on the order of
0.1-0.8%.2,3 Also, most studies failed to
FIGO stage
..................................................................................................................................................................................................................................... find an increased incidence of EmCa in
I 128 65.6 94 70.7 96 76.8 women with asymptomatic EMPs that
.....................................................................................................................................................................................................................................
II 20 10.3 17 12.8 13 10.4
.....................................................................................................................................................................................................................................
were discovered incidentally either by
III 30 8.7 5 12.8 10 4.0 screening ultrasonography or hysterec-
.....................................................................................................................................................................................................................................
tomy that was performed for causes
IV 17 15.4 17 3.8 5 8.0
.............................................................................................................................................................................................................................................. other than uterine bleeding.8,9,11,12 For
FIGO grade example, in a large multicenter study on
.....................................................................................................................................................................................................................................
1 85 43.6 50 37.6 69 55.2
.....................................................................................................................................................................................................................................
1152 consecutive asymptomatic women,
2 48 24.6 41 30.8 27 21.6 only 1 FIGO grade 1 endometrioid ade-
.....................................................................................................................................................................................................................................
nocarcinoma (0.1%) was confined ex-
3 25 12.8 8 6.0 8 6.4
.............................................................................................................................................................................................................................................. clusively to EMPs.12 This rate is similar
Histological type to that reported previously in asymp-
.....................................................................................................................................................................................................................................
Endometrioid 158 81.0 99 74.4 105 84.0
.....................................................................................................................................................................................................................................
tomatic postmenopausal women who
Serous 20 10.3 11 8.3 11 8.8 were not receiving hormone replace-
.....................................................................................................................................................................................................................................
ment therapy and who were participat-
Clear-cell 10 5.1 23 17.3 3 2.4
..................................................................................................................................................................................................................................... ing in endometrial screening studies (eg,
Mucinous 1 0.5 0 0 1 0.8 1.7 per 1000 women).13 The rate is also
.....................................................................................................................................................................................................................................
Carcinosarcoma 6 3.1 0 0 5 4.0
..............................................................................................................................................................................................................................................
much lower than the 0.8% rate that we
Total 195 100 133 100 125 100 and others found in symptomatic
..............................................................................................................................................................................................................................................
Some cell totals may not add up to the column total because of missing numbers. Staging information for non-surgical subjects
women.3
has been imputed based on the distribution of known status subjects. The staging information has been classified in the In agreement with detection bias, Fer-
following manner: early, stages I (Ia, Ib, Ic) and II (IIa, IIb); advanced, stages III (IIIa, IIIc) and IV (IVa, IVb). Only evaluable cases
are shown in the table.
razzi et al11 found that the prevalence of
Perri. Are endometrial polyps true cancer precursors? Am J Obstet Gynecol 2010. EmCa associated with EMPs was 10
times higher in women with abnormal

SEPTEMBER 2010 American Journal of Obstetrics & Gynecology 232.e4

Research Oncology www.AJOG.org

TABLE 4
Age at diagnosis
Endometrial Other Mean Median Range, Interquartile
cancera findings n age, y age, y y range
(a) No Polyps 1342 55.2 56 22–94 16 (48–64)
................................................................................................................................................................................................................................................................................................................................................................................
(b) No Leiomyomata 1005 46.6 46 22–87 11 (41–52)
................................................................................................................................................................................................................................................................................................................................................................................
(c) Yes Polyps 125 66.3 66 38–90 18 (59–76)
................................................................................................................................................................................................................................................................................................................................................................................
(d) Yes Leiomyomata 133 65.2 65 38–88 15 (60–75)
................................................................................................................................................................................................................................................................................................................................................................................
Yes None 195 65.7 67 28–91 18 (56–75)
................................................................................................................................................................................................................................................................................................................................................................................
a
Mean age comparisons for combinations (a) vs (c) and (b) vs (d) were both significant (P ⫽ .0001)
Perri. Are endometrial polyps true cancer precursors? Am J Obstet Gynecol 2010.

uterine bleeding than in their asymp- cally significant cancer risk factors, nant.14 In contrast, EIN, the true precur-
tomatic counterparts. The authors con- whereas polyp size was not. Finally, the sor of endometrioid adenocarcinoma is
cluded that the single most important pathogenesis of EMPs is inconsistent with a monoclonal glandular growth with a
risk determinant for finding cancerous its presumed increased malignant poten- cancer hazard ratio of 45- to 100-fold
polyps was bleeding. They found that tial. EMP develops as a result of monoclo- greater than primary cancerous EMPs.14
polyp size in symptomatic women also nal proliferation of genetically mutated Cytogenetically, ⬎50% have nonran-
carried a cancer risk, although of lesser stromal cells of the basalis layer of the en- dom chromosomal rearrangements in
degree than abnormal uterine bleeding. dometrium. The glands in EMPs are in- 6p21-22 region and in the 12q-13-15 and
Other investigators found age (post- duced secondarily like in those glands in 7q22 regions. Similar chromosomal al-
menopausal) and EMPs of ⬎1.5 cm to be the normal endometrium.14 Histologi- terations are found in other benign le-
statistically significant risk factors; cally, they are disorganized because of sions that are not known to be precursors
bleeding, hormone replacement ther- the focal and exophytic growth of (such as pulmonary hamartomas, lipo-
apy, and tamoxifen therapy failed to dis- stroma. The gland-lining cells are char- mas, and uterine leiomyomata).15
tinguish benign from malignant acterized by polyclonality in the same We addressed the possibility of detec-
EMPs.5,6 We found postmenopausal sta- manner as the adjacent normal endome- tion bias by calculating SIRs of EmCa in
tus, abnormal uterine bleeding, and age trium. They may be of the normal cyclic women with EMPs under the assump-
at diagnosis (⬎55 years) to be statisti- type, hyperplastic, and rarely malig- tion that they would have experienced

TABLE 5
Endometrial cancers with polyps or fibroidsa
Endometrial cancers, n
Expectedb Standardized morbidity ratio (95% CI)
Disease Age, y Cases, n Observed Conservative Liberal Conservative Liberal
Endometrial polyps ⬍45 222 7 0.04 0.21 156.0 (62.5–321.5) 32.9 (13.2–67.7)
..............................................................................................................................................................................................................................................................................................................
45-54 417 11 0.70 1.63 15.7 (7.8–28.2) 6.7 (3.4–12.1)
..............................................................................................................................................................................................................................................................................................................
55-64 439 38 2.61 4.51 14.6 (10.3–20.0) 8.4 (6.0–11.6)
..............................................................................................................................................................................................................................................................................................................
ⱖ65 402 69 5.74 9.34 12.0 (9.4–15.2) 7.4 (5.7–9.3)
..............................................................................................................................................................................................................................................................................................................
Total 1480 125 9.09 15.69 13.8 (11.4–16.4) 8.0 (6.6–9.5)
................................................................................................................................................................................................................................................................................................................................................................................
Uterine leiomyomata ⬍45 406 3 0.07 0.37 40.8 (8.4–119.2) 8.1 (1.7–23.7)
..............................................................................................................................................................................................................................................................................................................
45-54 419 8 0.55 1.42 14.5 (6.3–28.6) 5.6 (2.4–11.1)
..............................................................................................................................................................................................................................................................................................................
55-64 167 55 0.99 1.72 55.3 (41.7–72.0) 32.0 (24.1–41.7)
..............................................................................................................................................................................................................................................................................................................
ⱖ65 146 67 2.14 3.47 31.4 (24.3–39.8) 19.3 (15.0–24.5)
..............................................................................................................................................................................................................................................................................................................
Total 1138 133 3.76 6.98 35.4 (29.6–42.0) 19.1 (16.0–22.6)
................................................................................................................................................................................................................................................................................................................................................................................
CI, confidence interval.
a
Represent the ratios of observed to expected numbers of endometrial cancers based on 2 sets of expected numbers of cases; b Expected numbers were calculated based on cumulative rates up to
the specified age that were derived from the annual, age-specific incidence rates of endometrial cancer for the city of Montreal during 2001-05; 2 sets of projected expected values were calculated:
1 set used the lowest (conservative) and the other set used the highest (liberal) annual age-specific rates for the 2001-2005 period to derive the cumulative rates up to the specified age.
Perri. Are endometrial polyps true cancer precursors? Am J Obstet Gynecol 2010.

232.e5 American Journal of Obstetrics & Gynecology SEPTEMBER 2010

www.AJOG.org
the prevailing rates of EmCa in Montreal
at the time of the diagnosis. Our finding
that patients with EMPs experienced a
risk of EmCa that was 8-14 times greater
than Montreal women of the same age
does not counter directly the “precursor
lesion” hypothesis. However, a second-
ary analysis with leiomyomata as a “con-
trol” uterine disease demonstrated that
the excess risk of EmCa in patients with
EMPs was, in fact, lower than that expe-
rienced by patients with leiomyomata
(19- to 35-fold higher). Furthermore,
the average age-adjusted excess risks of
EmCa for patients with leiomyomata
were significantly greater than those seen
in patients with EMPs, regardless of
whether conservative or liberal rates
were used. Barring a yet undiscovered
link between fibroid tumors and EmCa,
this finding lends strong support to the
notion of detection bias. Leiomyomata
and endometrial cancer histologically
are made of distinct cell types; the former
is mesenchymal, and the latter is epithe-
lial. The fact that EmCa is so much more
common among patients with leiomyo-
mata can be explained readily by the en-
hanced detection that is afforded by
screening and diagnostic examinations,
much like the situation in treating symp-
tomatic patients with EMPs.
There are limitations to our study.
First, it was retrospective and lacked a
complete review of medical records. Sec-
ond, our experience may not apply to
other practices because of the single-cen-
ter nature of the study. However, the
leading endometrial cancer risk factors
(such as age, symptoms, and hormone
therapy that included tamoxifen ther-
apy) were available from the pathologic
forms, and the findings confirmed the
experience of others. The strengths of
our study were the consensus histologic
ascertainment of a large number of con-
secutive cases, the quality of the cancer
registry data of the Province of Quebec
for the Montreal area, and the unique
opportunity to study detection bias di-
rectly by comparing the excess EmCa
risk in patients with EMPs with that in
leiomyomata, which is a condition that is
not a precursor step in endometrial
carcinogenesis.
In conclusion, our study results indi-
cate that EMP per se is not a cancer
precursor. The seemingly high rate of as-
sociation with EmCa is related to diag-
nostic bias in symptomatic women with
bleeding and not with an increased can-
cer potential. This was further supported
by the even stronger association between
the discovery of leiomyomata and
EmCA on pathologic evaluation. EMPs
and leiomyomata may represent condi-
tions that serve as risk indicator/marker
of preexisting EmCa in symptomatic
women. The pathobiologic findings of
EMPs (according to which the stroma
and not the glandular epithelium is
neoplastic) lends further support to the
lack of increased cancer potential of
EMPs. f 13. Koss LG, Schreiber K, Oberlander SG,
REFERENCES
1. Berlière M, Radikov G, Galant C, Piette PH,
Marbaix E, Donnez J. Identification of women at
high risk of developing endometrial cancer on
tamoxifen. Eur J Cancer 2000;36:S35-6.
2. Savelli L, De Iaco P, Santini D, et al. His-
topathologic features and risk factors for benig-
nity, hyperplasia and cancer in endometrial pol-
yps. Am J Obstet Gynecol 2003;188:927-31.
SEPTEMBER 2010 American Journal of Obstetrics & Gynecology
Oncology Research
3. Carlson JW, Mutter GL. Endometrial intraepi-
thelial neoplasia is associated with polyps and
frequently has metaplastic change. Histopa-
thology 2008;53:325-32.
4. Ronnett BM, Zaino RJ, Ellenson LH, Kurman
RJ. Endometrial carcinoma. In: Kurman RJ, ed.
Blaustein’s pathology of the female genital tract,
5th ed. New York: Springer; 2002:501-60.
5. Hileeto D, Fadare O, Martel M, Zheng W. Age
dependent association of endometrial polyps
with increased risk of cancer involvement.
World J Surg Oncol 2005;3:8.
6. Ben-Arie B, Goldshmit C, Laviv Y, et al. The
malignant potential of endometrial polyps. Eur J
Obstet Gynecol Reprod Biol 2004;115:206-10.
7. Goldstein SR. Controversy about the effects
and safety of SERM’s: the saga continues.
Menopause 2002;9:381-4.
8. Machtinger R, Korach J, Padoa A, et al.
Transvaginal ultrasound and diagnostic hyster-
oscopy as a predictor of endometrial polyps:
risk factors for premalignancy and malignancy.
Int J Gynecol Cancer 2005;15:325-8.
9. Shushan A, Revel A, Rojansky N. How often
are endometrial polyps malignant? Gynecol
Obstet Invest 2004;58:212-5.
10. Baiocchi G, Manci N, Pazzaglia M, et al. Ma-
lignancy in endometrial polyps: a 12-year experi-
ence. Am J Obstet Gynecol 2009;201:642.e1-4
11. Ferrazzi E, Zupi E, Leone FP, et al. How
often are endometrial polyps malignant in
asymptomatic postmenopausal women? A
multicenter study. Am J Obstet Gynecol
2009;200:235.e1-6.
12. Bakour SH, Khan KS, Gupta JK. The risk of
premalignant pathology in endometrial polyps.
Acta Obstet Gynecol Scand 2000;79:317-20.
Moussouris HF, Lesser M. Detection of endo-
metrial carcinoma and hyperplasia in asymp-
tomatic women. Obstet Gynecol 1984;64:1-11.
14. Mutter GL, Zaino RJ, Baak JPA, Bentley
RX, Robboy SJ. Benign endometrial hyperpla-
sia sequence and endometrial intraepithelial
neoplasia. Int J Gynecol Pathol 2007;26:
103-14.
15. Dal Cin P, Vanni R, Marras S, et al. Four
cytogenic subgroups can be identified in endo-
metrial polyps. Cancer Res 1995;55:1565-8.
232.e6