Gynecologic Oncology xxx (2014) xxx–xxx

YGYNO-975436; No. of pages: 5; 4C:

Contents lists available at ScienceDirect

Gynecologic Oncology

j ourna l homepage: www.e lsev ie r .com/ locate /ygyno

Characteristics of clear cell ovarian cancer arising from endometriosis:A two center cohort study

Giovanna Scarfone a, Alice Bergamini b, Stefania Noli a, Antonella Villa a, Sonia Cipriani c, Gianluca Taccagni d,Paola Vigano' e, Massimo Candiani b, Fabio Parazzini c, Giorgia Mangili e,⁎a Department of Obstetrics, Gynecology and Neonatology, IRCCS Fondazione Cà Granda, Ospedale Maggiore Policlinico, Milan, Italyb Università Vita-Salute, San Raffale Scientific Institute, Milano, Italyc Mario Negri Pharmacological Research Institute, Milan, Italyd Department of Surgical Pathology, San Raffaele Scientific Institute, Milano, Italye Obstetrics and Gynecology Unit, San Raffaele Scientific Institute, Milano, Italy

H I G H L I G H T S

• Clinical and histological features of the largest series of clear cell carcinomas arising in endometriosis are analyzed.• No difference in stage, grade, survival and endometrial cancer incidence was found between tumors arising or not in endometriosis.• Younger age, unilateral ovarian involvement and absence of ascites were more frequently found in cases arising in endometriosis.

⁎ Corresponding author at: Obstetrics and GynecoloInstitute, Via Olgettina 60, 20132 Milano, Italy.

E-mail address: mangili.giorgia@hsr.it (G. Mangili).

http://dx.doi.org/10.1016/j.ygyno.2014.03.0170090-8258/© 2014 Elsevier Inc. All rights reserved.

Please cite this article as: Scarfone G, et al, CGynecol Oncol (2014), http://dx.doi.org/10.1

a b s t r a c t

a r t i c l e i n f o

Article history:

Received 19 January 2014Accepted 10 March 2014Available online xxxx

Keywords:EndometriosisClear cell ovarian cancerEndometriosis associated ovarian cancer

Objective. Endometrioid and clear cell ovarian tumors have been referred to as “endometriosis associatedovarian cancers”. However, very few studies have compared clinical and prognostic features of endometriosis-associated cancers or cancers not associatedwith endometriosis according to specific histotypes.We have inves-tigated clinical and histological features of the largest published series of clear cell ovarian cancers arising inendometriosis using a retrospective database.

Methods. Seventy three patients with a primary diagnosis of either pure clear cell ovarian cancer and mixedendometrioid-clear cell ovarian cancer have been divided into two groups according to the detection of cancerstrictly arising from ovarian endometriosis or not (n = 27 and n = 46, respectively). Clinical and pathological

data have been compared.

Results. Patients with clear cell carcinomas arising from endometriosis tend to be significantly younger(51.4 ± 10.0 and 58.4± 11.2 years, p= 0.02). FIGO stage, laterality, prevalence of pure versusmixed histology,and presence of synchronous endometrial carcinoma were not significantly different between the two groups.Unilateral ovarian involvement was more frequent in cases arising in endometriosis (85% vs 63%, p = 0.04).Ascites was not found in any of the endometriosis-associated cancer cases vs 19.5% in patients without endome-triosis. The presence of endometriosis did not affect 5-year overall survival rates.

Conclusions. Endometriosis per se does not appear to be associated with a lower stage tumor or to predictprognosis in ovarian clear cell cancers. Unilateral involvement and reduced presence of ascites may be linkedto the cystic nature of endometriosis which frequently presents as monolateral and in which associated tumorsare more likely to be longer confined to the ovary before spreading.

© 2014 Elsevier Inc. All rights reserved.

Introduction

Endometriosis is a risk factor for epithelial ovarian cancer [1,2].Based on the results of a large international collaborative study, self-

gy Unit, San Raffaele Scientific

haracteristics of clear cell ov016/j.ygyno.2014.03.017

reported endometriosis was associated with an overall risk increaseof nearly 50% [Odds Ratio, 1.46; 95% confidence intervals (CIs), 1.31–1.63] [3]. This issue is of particular clinical relevance given the preva-lence of endometriosis, estimated to be around 5%, with a peak between25 and 35 years of age [4].

It is also well known that tumors associated with endometriosisare confined to specific subcategories of disease, endometrioid andclear cell ovarian cancers, and the risk of detecting these histotypes isestimated to be about 4-fold higher in women with the disease [5].

arian cancer arising from endometriosis: A two center cohort study,

Fig. 1. A typical ovarian clear cell carcinoma of papillary type, grown into the lumen ofan endometriotic cyst lined by columnar endometrioid type cells. (Hematoxylin andeosin × 125.)

2 G. Scarfone et al. / Gynecologic Oncology xxx (2014) xxx–xxx

Controversy remains regarding the possibility that endometriosis-associated cancersmight represent a distinct entity from the correspon-dent typical histotype. Some studies have indeed suggested thatpatients with endometriosis tend to be younger, to be diagnosed in ear-lier stages and with lower grade lesions and to have an overall bettersurvival rate whereas other studies fail to confirm these findings[6–13]. Importantly, very few of these studies have evaluated clinicaland prognostic differences between endometriosis-associated cancersor cancers not associated with endometriosis according to the specifichistotype [8,9,13,14]. This might have prevented the identification ofcharacteristics potentially linked to a specific subcategory of disease.

In this context, in a recent study [8], we have analyzed a cohort ofpatients diagnosed with endometrioid ovarian cancer and confirmedthat women with endometriosis-associated cancer were significantlyyounger at the diagnosis, had a lower disease stage and a less prevalenthigh grade tumor. A difference in survival rate could not be confirmedbetween patients with and without endometriosis. Importantly, a syn-chronous endometrial precancerous or cancerous condition has beendetected at a significantly higher rate in endometriosis-associatedendometrioid cancers, possibly leading to various and novel interpreta-tions for the mechanisms of this cancer development.

Along this line, in this study, we have evaluated the clinical andprognostic features of endometriosis-associated ovarian cancers ofclear cell histotype and compared them with those of clear cell tumorswithout endometriosis.

Results of the present study considering themost numerous series ofcancers of clear cell andmixed endometrioid-clear cell histotype arisingin endometriosis and of our previous study on ovarian endometrioidtumors [8] tend to support the idea that endometrioid and clear cellcancers associated with endometriosis should be no more considereda single entity with similar risk and prognostic factors and biologicmechanisms of development.

Materials and methods

This is a retrospective study of 73 cases of clear cell ovarian carcino-ma consecutively observed at two centers, the Department of Obstetricsand Gynecology of the Fondazione Ospedale Maggiore Policlinico andthe Obstetrics and Gynecology Unit of the Scientific Institute SanRaffaele in Milan, Italy between 1990 and 2012. Institutional ReviewBoard approval has been obtained for this retrospective study.

All patients with a primary diagnosis of either pure clear cell ovariancancer or mixed endometrioid-clear cell ovarian cancer have been in-cluded in the study. Patients whose diagnosis was made elsewherewere excluded. Patients older than 75 were not included in survivalanalysis. Some data from 33 patients have been already included in aprevious paper [7].

All patients underwent surgery, received chemotherapy and werefollowed up at our institutions. Surgical staging was performed accord-ing to FIGO guidelines for ovarian cancer, including total abdominalhysterectomy, bilateral salpingo-oophorectomy, omentectomy andremoval of all macroscopic diseases [15]. Radical upper abdominal tech-niques were applied in selected cases to achieve optimal cytoreduction.All pathological analyses were performed by the same two gynecologicpathologists, who, in both institutions, are completely dedicated togynecological cases.

Patients were divided into two groups according to the detectionof cancer arising from ovarian endometriosis or not, on the basis ofpathological reports. The definition of ovarian cancer arising from endo-metriosis was given according to Sampson's [16] and Scott's criteria[17], that included: 1) the coexistence of carcinoma and endometriosisin the same ovary; 2) presence of tissue similar to endometrial stromasurrounding characteristic epithelial glands; 3) exclusion of ametastatictumor to the ovary; and 4) presence of benign endometriosis histologi-cally contiguous to themalignant tissue. Patientswith clear cell carcino-ma associated with but not arising in endometriosis were excluded.

Please cite this article as: Scarfone G, et al, Characteristics of clear cell ovGynecol Oncol (2014), http://dx.doi.org/10.1016/j.ygyno.2014.03.017

Data including age at diagnosis, clinical presentation, disease status,last follow-up and all pathological information, such as histology, stage,laterality, and presence of concurrent endometrial carcinoma werecollected from surgical and pathology reports. Stages higher than IIAwere classified as advanced, while lower stages were considered early.

All the abovementioned variableswere described for each of the twogroups and statistically compared. The Pearson Chi-square test or FisherExact test, as required, was used to assess the significance of differencesin clinical and pathological characteristics between the two groups.

Survival comparisons were obtained using the long-rank test in anunadjusted Kaplan Meier model, survival rates being calculated fromthe date of histological diagnosis. Finally, to account for the effect of po-tential confounding factors simultaneously, we used the Cox regressionmodel (after checking the proportional hazards assumption) to obtainthe hazards ratio (HR) and their corresponding 95% CIs. In all analysis,a p value of b0.05 was considered statistically significant. Follow-upwas updated in May 2013, being median follow-up of 68 months (95%CI, 55–103).

Results

Seventy-three patients met the inclusion criteria. Of these, n = 27(37%) cases have been affected by tumors arising in endometriosis(Fig. 1),while n=46 patients (67%) had no concomitant endometriosis.Patients and tumor characteristics of the two groups are summarized inTable 1.

In both groups, themost commondiagnosiswas incidental, followedby the detection of an abdominal mass. Symptoms did not differ statis-tically between the two groups except for presence of ascites that wasonly detected in patients without endometriosis (19.5% versus none ofthe patients in the endometriosis group).

Patients with endometriosis-associated clear cell carcinomas or ofmixed histology tend to be significantly younger than cases withoutendometriosis with mean age ± SD being 51.4 ± 10.0 and 58.4 ± 11.2for tumors arising in endometriosis and not, respectively (p= 0.02).

Among endometriosis-free patients, n = 23 (50%) were diagnosedin early stages while 50% in advanced stages. In the other group, n =18 (66.7%) were in early stages, and n = 9 (33.3%) were in advanced.

FIGO stages were not significantly different between the two groups(p = 0.16).

The prevalence of patients with pure clear cell histology and mixedhistology was respectively 76.1% and 23.9% in the endometriosis-freegroup and 70.4% and 29.6% in the endometriosis group. No statisticallysignificant difference was detected (p = 0.59).

arian cancer arising from endometriosis: A two center cohort study,

Table 1Clinical and pathological characteristics of clear cell carcinomas arising or not fromendometriosis.

Clinical features Endometriosis p-Value

No (n = 46) Arising (n = 27)

Age 0.02Mean ± SD (range) 58.4 ± 11.2

(41–89)51.4 ± 10.0(30–71)

Stage 0.17I–IIA 23 (50.0%) 18 (66.7%)IIB–IV 23 (50.0%) 9 (33.3%)

Histology 0.59Clear cell 35 (76.1%) 19 (70.4%)Mixed 11 (23.9%) 8 (29.6%)

Ovarian involvement 0.04Monolateral 29 (63.0%) 23 (85.0%)Bilateral 17 (37.0%) 4 (15.0%)

Laterality 0.55Right 15 (52.0%) 10 (43.0%)Left 14 (48.0%) 13 (57.0%)

SymptomsAbnormal uterine bleeding 4 (4.0%) 5 (2.2%) 0.28a

Constipation 3 (6.5%) 0 0.29a

Abdominal/pelvic mass 5 (10.8%) 7 (30.4%) 0.11a

History of adnexal cyst 2 (4.3%) 3 (13.0%) 0.35a

Ascites 9 (19.5%) 0 0.02a

Incidental diagnosis 19 (41.3%) 7 (30.4%) 0.18Paraneoplastic syndrome 4 (4.0%) 1 (4.3%) 0.64a

Abdominal pain 15 (32.6%) 4 (17.3%) 0.09Synchronous endometrialcancer

2 (4.0%) 0 0.53a

Year of diagnosis 0.141990–2000 16 (34.8%) 5 (18.5%)2001–2012 30 (65.2%) 22 (81.5%)

a Fisher exact test.

Endometriosis N. events/N. of patients at riskNo 2/46 3/36 2/30 1/25 5/23Yes 1/27 0/22 3/21 1/18 1/16

Fig. 2. Kaplan–Meier survival functions in clear cell carcinoma cases arising or not inendometriosis.

3G. Scarfone et al. / Gynecologic Oncology xxx (2014) xxx–xxx

Bilateral ovarian involvement was detected in 17 patients (37%) inthe endometriosis-free group, compared to 4 (15%) patients with clearcell tumor or of mixed histology arising in endometriosis (p = 0.04).

In monolateral ovarian involvement, laterality did not differ be-tween the two groups (p= 0.55). The presence of a synchronous endo-metrial carcinoma was found only in 2 patients with clear cell ovariancancer without concomitant endometriosis.

The presence of endometriosis did not affect 5-year survival rateseven after taking into account potential confounding factors (Fig. 2,Table 2). Mean 5-year overall survival was 60.1 months (95% CI41–75) in endometriosis-free group versus 73.0 months (95% CI49–87) in the endometriosis group (log rank = 0.43). Overall survivalhas been also analyzed stratifying for selected characteristics(Table 2); none of them was significantly associated with an increasedoverall survival.

Discussion

The association between ovarian cancer and endometriosis has beenlargely documented [1,18], even if clinico-pathological mechanisms arestill for many aspects unclear. In particular, several studies have focusedon the link between endometriosis and clear cell or endometrioid his-tologies, suggesting that about one third of these cancers arise from en-dometriosis [5]. The two histologies are often considered as a singleentity, as “endometriosis-associated ovarian tumors”. However, thereis increasing evidence suggesting that endometrioid and clear cell ovar-ian carcinomasmay behave differently with respect to their associationwith endometriosis and that might even arise from distinct types of en-dometriosis with different cells of origin [19]. In a very interestingpaper, Kajihara and coworkers hypothesized that clear cell ovariancancers associated with endometriosis may arise from preexistingendometriosis derived from regurgitated endometrial cells whereasovarian Mullerian metaplasia may initiate neoplastic modifications

Please cite this article as: Scarfone G, et al, Characteristics of clear cell ovGynecol Oncol (2014), http://dx.doi.org/10.1016/j.ygyno.2014.03.017

in endometriosis-associated endometriod cancers. Support to this hy-pothesis derived from the analysis of hepatocyte nuclear factor (HNF)-1β, a nuclear factor associated with endometriosis whose expressionhas been found in epithelial cells of ectopic, eutopic endometrium andclear cell ovarian carcinomas, but not in ovarian cortical inclusioncysts nor in endometrioid ovarian cancer.

The potential diverse origins of different histotypes of endometriosis-associated cancers imply that they should be considered different clini-cal entities. As amatter of fact, very few studies have specifically studiedclinical and prognostic factors of tumors associated with endometriosisaccording to the specific histotype [8,9,13,14]. Two studies, one fromour group [8], have focused on the endometrioid histotype, suggestingthat endometriosis-associated endometrioid ovarian cancer has adifferent clinical behavior with respect to the non-endometriosis-asso-ciated tumor of similar histology. Both the studies agree that the firsttends to arise at younger ages, with lower stages, with a less prevalenthigh grade andwith a significantly higher incidence of synchronous en-dometrial tumors [8,9].

Along this line, the present study aimed at evaluating the clinical andprognostic features of clear cell histotype arising or not in endometri-osis. For this analysis, the larger series of clear cell or of mixed histotypetumors arising in endometriosis ever published for this rare tumor hasbeen enrolled. No difference in tumor stage, grade, survival rate and in-cidence of synchronous endometrial cancers was found between clearcell cancers arising or not in endometriosis. As a consequence, the re-sults of this study also support the idea that clear cell or endometrioidtumors associated with endometriosis should be considered differententities as clinical features vary greatly between the two. The only char-acteristic in common is represented by the occurrence at younger agescompared to the correspondent histologic type not associated with en-dometriosis. However, in this context, a potential role of a diagnosticbias could also explain, at least partially, the increased diagnosis at ini-tial stages for both histotypes of tumors. Indeed, patients with concom-itant endometriosis could be diagnosed earlier considering that thisdisease is frequently associatedwith specific symptoms and endometri-osis women tend to be better followed up whereas ovarian cancers arefrequently asymptomatic until advanced stages are reached [5].

The prevalence of ascites and bilateral disease is the only other clin-ical discrepancies found between clear cell carcinoma arising and not in

arian cancer arising from endometriosis: A two center cohort study,

Table 2Five-year survival in cases with clear cell carcinoma arising or not from endometriosis.

Variable Endometriosis Cox regressiona

No (13 events) 5-year survival,% (95% CI)

Arising (6 events) 5-year survival,% (95% CI)

HR (95% CI) p-Value

Age≤54 84.0 (49–96) 80.7 (51–93) 1.07 (0.2–6.5) 0.94N54 51.2 (28–70) 57.1 (17–84) 0.83 (0.2–3.0) 0.78

StageI–IIa 94.7 (68–99) 92.3 (57–99) 1.38 (0.1–22.1) 0.82IIb–IV 21.1 (5–44) 44.4 (14–72) 0.68 (0.2–1.9) 0.46

HistologyClear cells 55.6 (34–73) 81.3 (52–94) 0.39 (0.1–1.4) 0.14Mixed 75.8 (30–94) 52.5 (12–82) 2.83 (0.5–17.4) 0.26

Year of diagnosis1990–2000 71.4 (41–88) 75.0 (13–96) NE2001–2012 53.1 (28–73) 72.6 (46–88) 0.46 (0.2–1.4) 0.17Overall 60.1 (41–75) 73.0 (49–87) 0.68 (0.3–1.8) 0.43

NE: non estimable due to low number of events.a Adjusted for stage, modeled in each strata of interesting variables, comparing endometriosis arising versus no endometriosis.

4 G. Scarfone et al. / Gynecologic Oncology xxx (2014) xxx–xxx

endometriosis being lower in the forms associated with the benign dis-ease. An explanation for this may be linked to the different origin of thetumor which may derive from an endometriotic cyst or from anadenofibroma [20]. Endometriosis presents more commonly as amonolateral ovarian cyst. Moreover, a more favorable outcome was ob-served in cystic clear cell carcinomas. Tumors growing intracysticallyare more likely to be confined to the ovary for longer periods of timebefore spreading, thereby allowing them to be diagnosed at lowerstage. This might also explain the increased frequency of ascites in tu-mors not associatedwith endometriosis. Thus, unilaterality and reducedpresence of ascites are most likely a consequence of an endometrioticand, therefore, cystic origin. These features were previously not foundto be associated with endometriosis-associated endometrioid ovariancancer [8].

To the best of our knowledge, there are only two other studieswhichfocused on prognostic implications of endometriosis in the context ofclear cell carcinomas and this is not surprising given the rarity of thistumor. In the study by Orezzoli et al. [13], eighty-four patients withclear cell carcinoma were identified with a 49% rate of coexisting endo-metriosis. However, only 15 of these tumors were found to arise in en-dometriosis while the others were not in contiguity with the tumor. Inline with our results, patients with clear cell cancer arising in endome-triosis tended to be younger, to have a pelvic mass at the diagnosisand no ascites. Contrary to our observations, patients with a tumor aris-ing from endometriosis significantly presented at early stage and theirmedian overall survival was remarkably greater than that of patientswith clear cell carcinoma without endometriosis although the limitednumber of patients did not allow for a statistical comparison of thisfactor. Discrepancies between the two studies may be due to theirsmall sample size and a type II error cannot be excluded. Moreover,the study byOrezzoli et al. encompassed a period of about 30 years dur-ing which adjuvant treatments varied greatly among the patients withor without endometriosis. Sixty percent of the patients in whom thetumor arose from endometriosis received platinum-based regimensin combinationwith taxanes compared to 39% of endometriosis-free pa-tients only. In this latter group, 11% of patients did not receive platinum-based regimens. Conversely, the current study, even though covering a22 year long period, only consider patients treated after platinum intro-duction in chemotherapeutic regimens. Thus, all patients have beenmore homogenously treated with platinum-based regimens.

The second study by Cuff et al. is poorly comparable with ours astheir definition of endometriosis-associated cancer was not strictlybased on Sampson's [16] and Scott's criteria [17]. In any case, endome-triosis per se did not appear to predict prognosis in their series of 53cases of clear cell cancers associated with the disease [14].

Please cite this article as: Scarfone G, et al, Characteristics of clear cell ovGynecol Oncol (2014), http://dx.doi.org/10.1016/j.ygyno.2014.03.017

In conclusion, based on the results from the present study andfrom studies focusing on ovarian endometrioid cancers [8,9], it is tempt-ing to speculate that clear cell and endometrioid carcinomas associatedwith endometriosis should be considered separate diseases withdifferent clinical behavior and prognostic factors. The more recentmolecular genetic analyses are not able to clarify whether the two sub-types of tumors derive from distinct types of endometriosis but theycertainly support important molecular differences between the two.A high frequency of ARID1A mutations has been detected mostlyin endometriosis-associated clear cell carcinomas (46–57%) and lessin endometriosis-associated endometrioid tumor (30%). As a matterof fact, loss of protein expression of the ARID1A tumor suppressorgene has been so far demonstrated in endometriosis adjacent toclear cell tumor samples [21]. Conversely, major genetic alterationsinvolved in endometriosis-associated endometrioid ovarian cancerare represented by PTEN, KRAS and β-catenin gene mutations. In ourprevious paper, we had indeed hypothesized a molecular parallelismbetween endometriosis-associated ovarian endometrioid cancer andtype I endometrial carcinoma that is typically characterized by muta-tions of the PTEN, K-RAS, PIK3CA and β-catenin genes [8].

Finally, as clear cell carcinomas exhibit very low estrogen receptorexpression while endometrioid cancer is predominantly positive for es-trogen receptor, the interactions between the iron-mediated oxidativestress due to the repeated hemorrhages in endometriosis [18,22,23]and the down-regulation of estrogen receptor expression have beensuggested to be involved in the development of clear cell carcinomas.

Future studies, particularly molecular investigations, should be un-dertaken to definitively clarify this aspect. This is even more importantconsidering that studies on endometriosis-associated ovarian cancer arealways facedwith a critical issue since the tumor, especially in advancedstages,might have obliterated the tissue of origin hiding any histologicalevidence of endometriosis. Thus, the better clarification of the geneticbackground of these two entities, which is a helpful clue to distinguishone from the other, remains a critical matter of investigation.

Conflict of interest statement

The authors declare no conflict of interest.

References

[1] Viganò P, Somigliana E, Parazzini F, Vercellini P. Bias versus causality: interpretingrecent evidence of association between endometriosis and ovarian cancer. FertilSteril 2006;88(3):588–93.

[2] Vercellini P, Viganò P, Somigliana E, Fedele L. Endometriosis: pathogenesis and treat-ment. Nat Rev Endocrinol Dec 24 2013. http://dx.doi.org/10.1038/nrendo.2013.255[Epub ahead of print].

arian cancer arising from endometriosis: A two center cohort study,

5G. Scarfone et al. / Gynecologic Oncology xxx (2014) xxx–xxx

[3] Pearce CL, Templeman C, Rossing MA, Lee A, Near AM, Webb PM, et al. Associationbetween endometriosis and risk of histological subtypes of ovarian cancer: a pooledanalysis of case–control studies. Lancet Oncol 2012;13(4):385–94.

[4] Viganò P, Parazzini F, Somigliana E, Vercellini P. Endometriosis: epidemiology andaetiological factors. Best Pract Res Clin Obstet Gynaecol 2004 Apr;18(2):177–200.

[5] Somigliana E, Viganò P, Parazzini F, Stoppelli E, Giambattista E, Vercellini P. Associa-tion between endometriosis and cancer: a comprehensive review and a critical anal-ysis of clinical and epidemiological evidence. Gynecol Oncol 2006;101:331–41.

[6] Boyraz G, Selcuck I, Yazicioglu A, Selcuk Tuncer Z. Ovarian carcinoma associatedwithendometriosis. Eur J Obstet Gynecol Reprod Biol 2013;170:211–3.

[7] Noli S, Cipriani S, Scarfone G, Villa A, Grossi E, Monti E, et al. Long term survival ofovarian endometriosis associated clear cell and endometrioid ovarian cancer. Int JGynecol Cancer 2013;23:244–8.

[8] Mangili G, Bergamini A, Taccagni G, Gentile C, Panina P, Viganò P, et al. Unravellingthe two entities of endometrioid ovarian cancer: a single center clinical experience.Gynecol Oncol 2012;126:403–7.

[9] Wang S, Qiu L, Lang JH, Shen K, Yang JX, Huang HF, et al. Clinical analysis of ovarianepithelial carcinoma with coexisting pelvic endometriosis. Am J Obstet Gynecol2013;208 [413 e1-5].

[10] Ogawa S, Kaku T, Amada S, Kobayashi H, Hirakawa T, Ariyoshi K, et al. Ovarian endo-metriosis associated with ovarian carcinoma: a clinicopathological and immunohis-tochemical study. Gynecol Oncol 2000;77:298–304.

[11] Kondi-Patifi A, Papakostantinou E, Iavazzo C, Griforiadis C, Salakos N, Gregoriou O.Clinicopathological analysis of ovarian carcinomas associated with endometriosis.Arch Gynecol Obstet 2012;285:479–83.

[12] Erzen M, Rakar S, Klancnick B, Syrjanen K. Endometriosis-associated ovarian carci-noma (EAOC): an entity distinct from other ovarian carcinomas as suggested by enested case–control study. Gynecol Oncol 2011;83:100–8.

Please cite this article as: Scarfone G, et al, Characteristics of clear cell ovGynecol Oncol (2014), http://dx.doi.org/10.1016/j.ygyno.2014.03.017

[13] Orezzoli JP, Russel AH, Oliva E, Del CarmenMG, Eichhorn J, Fuller AF. Prognostic im-plication of endometriosis in clear cell carcinoma of the ovary. Gynecol Oncol2008;110:336–44.

[14] Cuff J, Longacre TA. Endometriosis does not confer improved prognosis in ovariancarcinoma of uniform cell type. Am J Surg Pathol 2012;36:688–95.

[15] Pecorelli S. Revised FIGO staging for carcinoma of the vulva, cervix, and endometri-um. Int J Gynaecol Obstet 2009;105:103–4.

[16] Sampson J. Endometrial carcinoma of the ovary arising in endometrial tissue in thatorgan. Arch Surg 1925;10:1–72.

[17] Scott R. Malignant change in endometriosis. Obstet Gynecol 1953;2:293–9.[18] Vercellini P, Crosignani P, Somigliana E, Viganò P, Buggio L, Bolis G, et al. The ‘inces-

sant menstruation’ hypothesis: a mechanistic ovarian cancer model with implica-tions for prevention. Hum Reprod 2011;26:2262–73.

[19] Kajihara H, Yamada Y, Shigetomi H, Higashiura Y, Kobayashi H. The dichotomy in thehistogenesis of endometriosis-associated ovarian cancer: clear cell-type versusendometrioid-type adenocarcinoma. Int J Gynecol Pathol 2012;31:304–12.

[20] Veras E,Mao TL, Ayhan A,Ueda S, Lai H, HayranM, et al. Cystic and adenofibromatousclear cell carcinomas of the ovary: distinctive tumors that differ in their patho-genesis and behaviour: a clinicopathologic analysis of 122 cases. Am J Surg Pathol2009;33:844–53.

[21] Samartzis EP, Nosle A, Dedes KJ, Fink D, Imesch P. ARID1A mutations and PI3K/AKTpathway alterations in endometriosis and endometriosis-associated ovarian carci-nomas. Int J Mol Sci 2013;14:18824–49.

[22] Tanase Y, Yamada Y, Shigetomi H, Kajhara H, Oonogi A, Yoshizawa Y, et al. Modula-tion of estrogenic action in clear cell carcinoma of the ovary (Review). Exp Ther Med2012;3:18–24.

[23] Viganò P, Somigliana E, Panina P, Rabellotti E, Vercellini P, Candiani M. Principles ofphenomics in endometriosis. Hum Reprod Update 2012;18:248–59.

arian cancer arising from endometriosis: A two center cohort study,