Int J Gynecol Cancer 2008, 18, 465–469
Pathology of stage I versus stage III ovarian
carcinoma with implications for pathogenesis
and screening
A.V. YEMELYANOVA*, J.A. COSINy, M.A. BIDUSz, C.R. BOICEy & J.D. SEIDMAN*
*Department of Pathology and Laboratory Medicine and yDepartment of Obstetrics and Gynecology,
Washington Hospital Center, Washington, DC; and zDepartment of Obstetrics and Gynecology,
Walter Reed Army Medical Center, Washington, DC
Abstract.
Yemelyanova AV, Cosin JA, Bidus MA, Boice CR, Seidman JD. Pathology of stage I versus stage
III ovarian carcinoma with implications for pathogenesis and screening. Int J Gynecol Cancer 2008;18:465–469.
The progression of ovarian carcinoma from stage I when it is confined to the ovaries and curable to disseminated abdominal disease, which is usually fatal, is poorly understood. An accurate understanding of
this process is fundamental to designing, testing, and implementing an effective screening program for
ovarian cancer. Pathologic features of the primary ovarian tumors in 41 FIGO stage I ovarian carcinomas
were compared with those in 40 stage III carcinomas. The primary ovarian tumors in stage I cases, when
compared with stage III, respectively, were significantly larger (15.4 versus 9.8 cm), were less frequently
bilateral (12% versus 75%), more frequently contained a noninvasive component (88% versus 30%), had
a higher proportion of a noninvasive component (42% versus 8%), and were more often nonserous (83%
versus 20%) (P , 0.001 for all five comparisons). There are significant pathologic differences between the
primary ovarian tumors in stage I and III ovarian carcinomas that are very difficult to explain by a simple
temporal progression. These findings along with the growing body of literature suggest that early- and
advanced-stage ovarian cancers are in many instances biologically different entities. This knowledge may
have significant implications for our understanding of the biology of early- and advanced-stage ovarian
cancer and therefore on the development of screening strategies for ovarian cancer.
KEYWORDS:
ovarian carcinoma, prognosis, screening, staging.
Ovarian cancer kills more than 14,000 women annually in the United States, and the vast majority of these
deaths are in patients who present with advancedstage disease (FIGO stages III and IV)(1). Stage I invasive ovarian carcinoma (tumor confined to the ovaries)
is relatively uncommon (less than 20% of all invasive
carcinomas) and is associated with an excellent prognosis (nearly 90% or better 5-year survival for comprehensively staged patients(2–10)). The features and
mechanisms of progression of stage I to advanced
Address correspondence and reprint requests to: Jeffrey D. Seidman,
MD, Department of Pathology, Washington Hospital Center, 110 Irving
Street, NW, Washington, DC 20010, USA. Email: jeffrey.d.seidman@
medstar.net
Dr Yemelyanova is presently at Department of Pathology, Johns
Hopkins Hospital, Baltimore, MD.
doi:10.1111/j.1525-1438.2007.01058.x
2007, Copyright the Authors
Journal compilation # 2007, IGCS and ESGO
#
stage are poorly understood, and in fact, it is not
entirely clear that early- and advanced-stage ovarian
carcinomas represent the same disease at different
points in time. Understanding of the sequence of
events in ovarian cancer progression is important
because current screening strategies rely primarily on
ovarian imaging to detect what is destined to be latestage ovarian cancer at an earlier stage. The current
study compares and contrasts the pathologic features
of the primary tumors in early and advanced stage to
better understand and further characterize the pathways of ovarian tumorigenesis.
Materials and methods
This study was approved by the Institutional Review
Board of Washington Hospital Center (WHC). Consecutive FIGO stage I invasive ovarian surface epithelial
466 A.V. Yemelyanova et al.
carcinomas accessioned at WHC from 1991 to 2005
were compared with consecutive stage III cases. Only
patients who had their primary surgery at WHC were
included. Primary peritoneal carcinomas, atypical proliferative (borderline) tumors with microinvasion,
carcinosarcomas (malignant mixed müllerian tumors),
and tumors that were not clearly primary in the ovary
were excluded. Criteria for histologic type, primary
peritoneal carcinoma versus ovarian carcinoma, microinvasion, endometriosis, and grading were as previously reported(1,11–15).
Patient data were abstracted from charts (M.A.B.)
and follow-up was also obtained from public records
(social security death index). Histologic sections of
ovaries were reviewed by two gynecological pathologists (A.V.Y. and J.D.S.). One pathologist (J.D.S.) was
blinded to the original diagnosis and FIGO stage,
except in cases for which blinding was not possible
(cases for which the stage was obvious based on extraovarian tissue present on the slides of the primary
tumor). Surgical–pathologic stage was determined according to the FIGO criteria(16,17). Accordingly, stage I
includes tumors confined to the ovaries with or without tumor cells in peritoneal washings or ascites.
Evidence suggests that upstaging based on dense
adhesions to extraovarian tissue regardless of whether
tumor is present in those adhesions may be warranted
and appears to be common practice(2,18). However,
FIGO guidelines are not sufficiently clear to determine
whether this practice is permitted. We require microscopic evidence of tumor involving extraovarian tissue
to designate a patient as higher than stage I.
Histologically noninvasive epithelial components,
that is, possible precursor lesions, were sought in the
primary ovarian tumors. (Of note, it is an assumption
that morphologically benign or noninvasive components are in fact benign. It is possible that they may in
fact be in situ or invasive carcinoma with either ‘‘maturation’’ or other explanation for this appearance.) In
addition to endometriosis, a cystadenoma or atypical
proliferative (borderline) tumor was diagnosed when
areas that would qualify for such a diagnosis on their
own were identified. When such an area was identified, the percentage of the entire tumor occupied by
the benign and/or atypical proliferative component
(total noninvasive component) was estimated. For this
assessment, the gross pathologic description was evaluated in conjunction with the histology and in some
cases modified to account for large smooth-walled
cysts for which the sections of the cysts and solid areas
were not proportionally representative of the tumor.
These percentages were averaged for group comparisons. The ability of surgical pathologists to accurately
#
estimate proportions of tumor occupied by various
components has been previously validated(19). Ovaries
with small foci of unequivocal metastatic tumor
involving their surfaces or superficial cortex contralateral to a clearly primary ovarian tumor were considered bilateral. In such cases, only the primary tumor
was evaluated for precursor lesions.
Comparison of groups was performed using Fisher
exact test and Student’s t-test, as appropriate. Survival
curves were constructed according to Kaplan and
Meier. Data were analyzed by A.V.Y. and J.D.S.
Results
Forty-one stage I and 40 stage III patients were studied, and significant differences between these groups
were identified. The primary ovarian tumors in stage I
were significantly larger compared with those in
advanced stage (mean diameters of 15.4 versus 9.8 cm,
respectively; P , 0.001), more frequently unilateral
(88% versus 25%; P , 0.001), more diverse histologically (83% nonserous versus 20% nonserous; P ,
0.001), more likely to contain a discrete noninvasive
component (88% versus 30%; P , 0.001), and more
frequently associated with endometriosis (68% versus
30%; P , 0.001). The proportion of the primary tumor
(mean percentage) occupied by the noninvasive component, that is, a putative precursor lesion, was higher
in stage I than in stage III (42% versus 8%, respectively; P , 0.001). The results are summarized in
Tables 1 and 2.
All stage I patients had total abdominal hysterectomy, bilateral salpingo-oophorectomy with or without staging. Data on postoperative chemotherapy
were available for 26 patients: 18 patients received
platinum-based chemotherapy and 7 patients received
no chemotherapy. One patient who also had colon
cancer received 5-fluorouracil and received no specific
chemotherapy for ovarian carcinoma. Among the five
fatal cases, four received platinum-based chemotherapy and information on one was not available. Among
the nonfatal cases, 14 of 22 (64%) received platinumbased chemotherapy. All stage III patients had total
abdominal hysterectomy, bilateral salpingo-oophorectomy with staging/debulking surgery and received
postoperative platinum-based chemotherapy.
Among stage I patients, there were five tumor
deaths. The mean and median follow-up periods were
71 and 62 months, respectively (range 13–127 months).
The 5- and 10-year disease-specific survival rates were
91% and 82%, respectively. Among stage III patients,
there were 25 deaths. The mean follow-up period was
2007 IGCS and ESGO, International Journal of Gynecological Cancer 18, 465–469
Pathology of stage I versus stage III ovarian carcinoma
Table 1. Pathologic features of stage I versus stage III ovarian
carcinomas
Number of cases
Mean patient age (year)
Histologic type
Serous
Nonserous
Mucinous
Endometrioid
Clear cell
Transitional
Undifferentiated
Mixed
Bilaterality
Mean/median size (cm)
Grade invasive component
Low
High
Stage I
Stage III
41
57
40
62.5
7 (17%)
34 (83%)
8 (20%)
12 (29%)
11 (27%)
1 (2%)
0
2 (5%)
5 (12%)
15.4/14
32 (80%)
8 (20%)
1 (2%)
3 (8%)
1 (2.5%)
1 (2.5%)
1 (2.5%)
1 (2.5%)
30 (75%)
9.8/9
6 (15%)
35 (85%)
1 (2%)
39 (98%)
P
,0.001
,0.001
NS
NS, not significant.
48 months and the median was 47 months (range 1–81
months). The 5-year survival rate was 37%, and the
median survival was 49 months.
Discussion
Understanding the sequence of events in ovarian cancer
progression is critical to designing and implementing
effective screening programs for early detection. If undiagnosed stage I ovarian carcinoma typically progressed to stage III as is commonly assumed, the
current findings would indicate that with tumor progression, the primary ovarian tumor usually gets
smaller, loses most or all its noninvasive component,
and changes from nonserous to serous cell type. This is
an unlikely sequence of events. First, tumors of nearly
all sites generally enlarge with progression even if
tumor necrosis occurs. Second, the striking high proportion of the noninvasive component in a stage I
tumor, although arguably partially obliterated by rapid
growth of the invasive component, is inexplicably
reduced in absolute as well as in relative quantity in
advanced stage. Third, although mixtures of epithelial
cell types do occur in primary ovarian carcinoma, this
is infrequent and occurs in about 5% of patients(20,21).
The primary tumor in most advanced-stage ovarian
carcinomas is a pure serous carcinoma and generally
does not have features suggestive of origin in another
cell type(20).
Accumulating evidence from pathologic and molecular biologic studies supports a dualistic model of
ovarian carcinogenesis with different pathways for lowand high-grade tumors. The ‘‘high-grade pathway’’
#
Table 2. Presence of noninvasive component (putative precursor) in stages I and III ovarian carcinomas
Precursor/histologic
subtype
NS
,0.001
467
Tumor with benign
(cystadenoma)
or atypical proliferative
(borderline) component
Serous
Nonserous
Mucinous
Endometrioid
Clear cell
Transitional
Mixed
Undifferentiated
Tumors associated with
endometriosis
Serous
Nonserous
Mucinous
(seromucinous)
Endometrioid
Clear cell
Transitional
Mixed
Undifferentiated
Number of cases with
identifiable precursor
(noninvasive component)a
Mean % of total noninvasive
componenta
Stage I
24/41 (59%)
Stage III
P
8/40 (20%) ,0.001
3/7 (43%)
5/32 (16%)
21/34 (62%) 3/8
8 (100%)
1/1 (100%)
9/12 (75%) 2/3 (67%)
3/11 (27%) 0/1
0/1
0/1
1/2(50%)
0/1
—
0/1
28/41 (68%) 12/40 (30%) ,0.001
3/7 (43%)
25/34 (74%)
4/8 (50%)
8/32 (25%)
4/8 (50%)
1 (100%)
10/12 (83%) 2/3 (75%)
9/11 (82%) 1 (100%)
0
0/1
2 (100%)
0/1
—
0/1
36 (88%)
12 (30%)
,0.001
42
,0.001
8
a
Noninvasive component includes endometriosis, cystadenoma/
adenofibroma, and/or atypical proliferative (borderline) tumor.
accounts for the majority of ovarian cancers, which are
nearly always high-grade serous carcinomas believed
to arise de novo without a clearly identifiable precursor
lesion(22–26). The ‘‘low-grade pathway’’ is characterized
by an orderly progression from benign to atypical
proliferative to carcinoma in situ and finally invasive
carcinoma. The prototypical low-grade tumor in a
recently proposed model(22,23) is invasive low-grade
(micropapillary) serous carcinoma, which develops
from atypical proliferative (borderline) ovarian serous
tumors. However, invasive low-grade serous carcinoma is relatively uncommon, and in FIGO stage I,
rare in our experience. In one recent study and our
study of low-stage invasive ovarian cancers(27), a total
of 136 stage I and II cases, there was not a single case
of invasive low-grade serous carcinoma. The stepwise
progression from benign to atypical proliferative to
carcinoma is much more commonly characterized by
the heterogeneity of cell types and levels of neoplastic
progression (ie, atypical proliferative, carcinoma
in situ) seen in stage I tumors.
2007 IGCS and ESGO, International Journal of Gynecological Cancer 18, 465–469
468 A.V. Yemelyanova et al.
The heterogeneity of cell types in stage I ovarian
carcinoma has been reported previously and, although
well known by gynecological pathologists and generally acknowledged in the pathology literature(21–28), is
often largely ignored in clinical studies. The peculiar
inverse relationship between tumor size and stage has
also been reported(29,30). The presence of a substantial
noninvasive component for endometrioid and mucinous carcinomas compared with serous is also known,
as is the propensity of clear cell carcinomas to be associated with large areas of endometriosis(24–28). Thus,
some of our findings are not new. These observations,
however, have only rarely been synthesized into an
analysis of the clinically and biologically more relevant division of stage I versus advanced-stage ovarian
carcinoma, that is, curable versus incurable cases.
When this argument is laid out, it becomes clear
that the progression of invasive ovarian carcinoma
from an early stage when it is confined to the ovaries
to widespread metastatic disease is poorly understood. In fact, the numerous and currently inexplicable
differences between the primary tumors suggest that
stage I and advanced-stage ovarian carcinoma are fundamentally different diseases. Hogg and Friedlander(31),
in a recent literature review, raised most of these
issues, but these views continue to receive very limited
attention(32).
Ovarian cancer is essentially incurable in advanced
stage and efforts are underway to develop a screening
test for earlier diagnosis at a ‘‘more curable stage,’’ that
is, FIGO stage I. The justification for this approach is
modeled after the highly successful screening programs for breast and colon cancer, which assist in
identifying these cancers at an earlier, more curable
stage. It is clear that the prognosis for early-stage ovarian cancer is excellent, and as noted earlier, comprehensively staged FIGO stage I patients have a 5-year
survival rate of nearly 90% or better(2–10). Therefore,
identifying ovarian cancers at this early stage when it
is potentially curable would appear to justify the
development of a screening test. However, the underlying concept that early- and advanced-stage ovarian
carcinomas are the same disease at different points in
time is an assumption that has not been tested and
therefore can only be supported by circumstantial evidence. This concept is counterintuitive and therefore
to an extent difficult to recognize on a casual perusal
of the literature on ovarian cancer. Attention to the
precise pathologic features of early- and advancedstage ovarian cancer makes it clear that there are
many assumptions, and therefore, much to be learned
about the progression of ovarian cancer from the curable to the incurable stage.
#
The criteria for a successful screening test are well
established. According to the World Health Organization, fulfillment of these criteria ‘‘are fundamental to the
integrity of the screening process in any country(33).’’
One of the main criteria is as follows: ‘‘The condition
sought should be an important health problem whose
natural history, including development from latent to
declared disease, is adequately understood(33).’’ Ovarian carcinoma does not appear to satisfy this criterion.
The reasons behind the disappointing reports of
screening trials for ovarian cancer may be in part explained by an incomplete understanding of the goals
of screening and the fundamental differences in earlyand advanced-stage tumors. In particular, most
screening trials for ovarian cancer result in the identification of predominantly advanced-stage carcinomas
as well as noninvasive (atypical proliferative or ‘‘borderline’’) and nonepithelial tumors. There is no compelling justification for earlier diagnosis through
screening of these latter types of tumors as they contribute minimally to ovarian cancer mortality. In
addition, when reported, most trials include screennegative women who develop ovarian cancer in the
interval period between screenings(34). These findings,
taken together, strongly suggest that ovarian neoplasms destined to be advanced stage are characterized by early metastatic growth when the primary
tumor is small. This is in stark contrast to stage I
tumors, which grow to significantly larger size than
their advanced-stage counterparts without metastatic
growth.
Available data suggest that the goal of screening
should be to identify stage I high-grade serous carcinomas, a finding which has been exceedingly uncommon in all reported trials. Although some stage I
carcinomas have been identified, most of these have
been nonserous, and the grade is frequently not reported. The current data, if confirmed, suggest that
screening methods based on ovarian imaging are
unlikely to significantly influence mortality and that
more sensitive screening tests that rely on parameters
other than tumor size are needed. Although we hesitate to be unduly pessimistic about screening for this
important worldwide public health problem, the data
suggest that reconsideration of the strategies for these
trials is in order.
In conclusion, our data show that there are significant differences between stage I and advanced-stage
ovarian carcinomas beyond those suggested by a simple temporal progression. Most fatal ovarian cancers
arise de novo and have a specific histopathology. Typical examples of early-stage disease have a distinctly
different histopathology, and several lines of evidence
2007 IGCS and ESGO, International Journal of Gynecological Cancer 18, 465–469
Pathology of stage I versus stage III ovarian carcinoma
suggest that these latter tumors do not represent the
precursors of most fatal (advanced stage) cases. The
implications of this finding for the development of
a screening test are profound.
Acknowledgments
The authors thank Drs Thomas Godwin, Russell Vang,
Brigitte Ronnett, and Robert Kurman for their helpful
comments.
References
1 Seidman JD, Russell P, Kurman RJ. Surface epithelial tumors of the
ovary. In: Kurman RJ, ed. Blaustein’s pathology of the female genital
tract, 5th edn. New York, NY: Springer-Verlag, 2002:791–904.
2 Dembo AJ, Davy M, Stenwig AE, Berle EJ, Bush RS, Kjorstad K.
Prognostic factors in patients with stage I epithelial ovarian cancer.
Obstet Gynecol 1990;75:263–73.
3 Ahmed FY, Wiltshaw E, A’Hern RP et al. Natural history and prognosis of untreated stage I epithelial ovarian carcinoma. J Clin Oncol
1996;14:2968–75.
4 Trimbos JB, Schueler JA, van der Burg M et al. Watch and wait after
careful surgical treatment and staging in well-differentiated early
ovarian cancer. Cancer 1991;67:597–602.
5 Elit L, Chambers A, Fyles A et al. Systematic review of adjuvant
care for women with stage I ovarian carcinoma. Cancer 2004;101:
1926–35.
6 Gadducci A, Sartori E, Maggino T et al. Analysis of failures in
patients with stage I ovarian cancer: an Italian multicenter study.
Int J Gynecol Cancer 1997;7:445–50.
7 Bolis G, Colombo N, Pecorelli S et al. Adjuvant treatment for early
epithelial ovarian cancer: results of two randomised clinical trials
comparing cisplatin to no further treatment or chromic phosphate
(32P). Ann Oncol 1995;6:887–93.
8 Piver MS, Malfetano J, Baker TR, Hempling RE. Five-year survival
for stage IC or stage I grade 3 epithelial ovarian cancer treated with
cisplatin-based chemotherapy. Gynecol Oncol 1992;46:357–60.
9 Sainz de la Cuesta R, Goff BA, Fuller AF Jr, Nikrui N, Eichorn JH,
Rice LW. Prognostic importance of intraoperative rupture of
malignant ovarian epithelial neoplasms. Obstet Gynecol 1994;84:1–7.
10 Schueler JA, Cornelisse CJ, Hermans J, Trimbos JB, van der Berg
ME, Fleuren GJ. Prognostic factors in well-differentiated early-stage
epithelial ovarian cancer. Cancer 1993;71:787–95.
11 Bloss JD, Liao S-Y, Buller RE et al. Extraovarian peritoneal serous
papillary carcinoma: a case-control retrospective comparison to
papillary adenocarcinoma of the ovary. Gynecol Oncol 1993;50:
347–51.
12 Seidman JD, Soslow RA, Vang R et al. Borderline ovarian tumors:
diverse contemporary viewpoints on terminology and diagnostic
criteria with illustrative images. Hum Pathol 2004;35:918–33.
13 Clement PB. Diseases of the peritoneum. In: Kurman RJ, ed. Blaustein’s pathology of the female genital tract, 5th edn. New York, NY:
Springer-Verlag, 2002:729–89.
14 Malpica A, Deavers MT, Lu K et al. Grading ovarian serous carcinoma using a two-tier system. Am J Surg Pathol 2004;28:496–504.
15 Seidman JD, Horkayn-Szakaly I, Cosin JA et al. Testing of two
binary grading systems for FIGO stage III serous carcinoma of the
ovary and peritoneum. Gynecol Oncol 2006;103:703–8.
#
469
16 International Federation of Gynecology and Obstetrics. Classification and staging of malignant tumors in the female pelvis. Acta
Obstet Gynecol Scand 1971;50:1–7.
17 Heintz APM, Odicino F, Maisonneuve P et al. Carcinoma of the
ovary. Int J Gynecol Obstet 2003;83(Suppl. 1):135–66.
18 Ozols RF, Rubin SC, Thomas GM, Robboy SJ. Epithelial ovarian
cancer. In: Hoskins WJ, Perez CA, Young RC, eds. Principles and
practice of gynecological oncology, 3rd edn. New York, NY: Lippincott,
Williams & Wilkins, 2000:981–1058.
19 Zaino RJ, Laskaris A, Whitney C, Sharkey FE. Morphometric analysis of endometrial adenocarcinoma: 2. A comparison of architectural
differentiation determined morphometrically with subjective grading. Int J Gynecol Pathol 1987;6:20–8.
20 Seidman JD, Horkayne-Szakaly I, Haiba M, Boice CR, Kurman RJ,
Ronnett BM. The histologic type and stage distribution of ovarian
carcinomas of surface epithelial origin. Int J Gynecol Pathol 2004;23:
41–4.
21 Heintz APM, Odicino F, Maisonneuve P et al. Carcinoma of the
ovary. In: Pecorelli S., ed. 26th Annual report on the results of treatment in gynecological cancer. Int J Gynecol Obstet 2006;95:S161–92.
22 Singer G, Kurman RJ, Chang H-W, Cho SKR, Shih I-M. Diverse
tumorigenic pathways in ovarian serous carcinoma. Am J Pathol
2002;160:1223–8.
23 Shih I-M, Kurman RJ. Ovarian tumorigenesis: a proposed model
based on morphological and molecular genetic analysis. Am J Pathol
2004;164:1511–8.
24 Bell DA, Scully RE. Ovary. In: Henson DE, Albores-Saavedra J, eds.
Pathology of incipient neoplasia, 3rd edn. Oxford, England, UK:
Oxford University Press, 2001:419–40.
25 Scully RE. Early de novo ovarian cancer and cancer developing in
benign ovarian lesions. Int J Gynecol Obstet Suppl 1995;49:S9–15.
26 Scully RE. Pathology of ovarian cancer precursors. J Cell Biochem
Suppl 1995;23:208–18.
27 Leitao MM, Boyd J, Hummer A et al. Clinicopathologic analysis of
early stage sporadic ovarian carcinoma. Am J Surg Pathol 2004;28:
147–59.
28 Chen S, Leitao MM, Tornos C, Soslow RA. Invasion patterns in
stage I endometrioid and mucinous ovarian carcinomas: a clinicopathologic analysis emphasizing favorable outcomes in carcinomas
without destructive stromal invasion and the occasional malignant
course of carcinomas with limited destructive stromal invasion. Mod
Pathol 2005;18:903–11.
29 Stratton JF, Pharoah P, Tidy JA, Paterson MEL. An analysis of ovarian tumor diameter and survival. Int J Gynecol Cancer 2000;10:
449–51.
30 Nakanishi T, Nawa A, Niwa Y, Nakamura S, Kuzuya K. Associations between tumor diameter and prognostic variables of epithelial
ovarian cancer. Eur J Gynaecol Oncol 2003;24:45–7.
31 Hogg R, Friedlander M. Biology of epithelial ovarian cancer: implications for screening women at high genetic risk. J Clin Oncol
2004;22:1315–27.
32 Rufford B, Jacobs IJ. Screening and diagnosis of ovarian cancer in
the general population. In: Gershenson DM, McGuire WP, Gore M,
Quinn MA, Thomas G, eds. Gynecologic cancer: controversies in management. New York, NY: Elsevier, 2004:355–68.
33 Holland WW, Stewart S, Masseria C. Policy brief: screening in
Europe. World Health Organization; European Observatory on
Health Systems and Policies. Geneva, Switzerland. Available
at: http://www.euro.who.int/Document/E88698.pdf. Accessed
December 30, 2006.
34 Rosenthal AN, Menon U, Jacobs IJ. Screening for ovarian cancer.
Clin Obstet Gynecol 2006;49:433–47.
Accepted for publication July 2, 2007
2007 IGCS and ESGO, International Journal of Gynecological Cancer 18, 465–469