CD44 mediates constitutive type I receptor signaling in cervical carcinoma cells.

OBJECTIVE: The CD44 transmembrane glycoprotein family has been implicated in the growth and metastasis of numerous human cancers. CD44 may function in some cells through interactions with type I receptor tyrosine kinases, including erbB2. Here, we tested whether CD44 interacts with erbB2 and another type I receptor, the epidermal growth factor receptor (EGFR), in human cervical carcinoma tissues and cell lines and whether these interactions influence erbB2 signaling. METHODS: CD44, EGFR, and erbB2 colocalization were examined in 36 pT1b-pT2b cervical cancer cases and in the CaSki and SiHa cervical carcinoma cell lines by immunohistochemistry and laser scanning confocal microscopy. The role of CD44-EGFR-erbB2 interactions in erbB2 signaling was examined by immunoprecipitation and using antisense CD44 oligonucleotides. RESULTS: CD44, erbB2, and EGFR coexpression and colocalization were observed in 42% (15/36) of cervical carcinoma cases and in both cervical carcinoma cell lines. Colocalization occurred to an equivalent extent in all tumor grades examined. CD44 coimmunoprecipitated with erbB2 and EGFR in cervical carcinoma cell lysates, indicating that these proteins interact with each other. Reduction of CD44 expression inhibited constitutive erbB2 activity. High CD44 expression was linked to EGFR activity using dominant negative EGFR, suggesting that type I receptors may autoregulate their activity in these cells. CONCLUSIONS: Our data indicate that CD44 can mediate type I receptor function in cervical carcinoma cells that overexpress both CD44 and either erbB2 or EGFR and suggest a novel mechanism by which these proteins may contribute to cervical carcinoma tumor growth and metastasis.

Familial ovarian cancer and early ovarian cancer: biologic, pathologic, and clinical features.

Women with ovarian cancer have poor overall survival rates, largely because the disease is so often diagnosed at an advanced, less curable stage. Because women with early ovarian cancer experience good survival rates, there is great interest in the study and detection of early disease. Familial ovarian cancer has been relevant to the study of early ovarian cancer in two different ways. First, women from ovarian cancer families often undergo prophylactic oophorectomy to prevent development of this disease. These ovaries have been studied for pathologic or molecular features that might represent early preinvasive disease. Second, screening tests to detect presymptomatic ovarian cancer have selectively targeted this population because of the increased positive predictive value of these tests in this population. A review of the clinical, pathologic, epidemiologic, and molecular biologic aspects of familial ovarian cancer provides a background to facilitate understanding these issues.

Expression of CD44S and CD44v5 is more common in stage III than in stage I serous ovarian carcinomas.

Experimental evidence suggests that attachment of ovarian carcinoma cells to the peritoneal mesothelium involves the interaction between CD44 on ovarian carcinoma cells and hyaluronic acid on mesothelial surfaces. The authors therefore evaluated local and disseminated ovarian serous carcinomas for the expression of standard CD44 and CD44 splice variants CD44v5, CD44v6, and CD44v7/8. The relative amount of hyaluronic acid (HA) in stroma surrounding tumor nests also was studied. Using immunohistochemistry and archival tissue, 14 serous ovarian carcinomas confined to the ovary (stage I) and 14 serous ovarian carcinomas with peritoneal implants and positive peritoneal fluid (stage III) were stained with antibodies to standard CD44, CD44v5, CD44v6, and CD44v7/8. All tissues also were analyzed for HA using a HA binding peptide. Immunostaining was classified as focal or diffuse and graded from 1 to 4 based on intensity. Immunoreactivity for standard CD44 was seen in 5 of 14 (36%) stage I tumors and 10 of 14 (71%) stage III tumors. Similarly, immunoreactivity with CD44v5 was seen in 2 of 14 (14%) stage I tumors and 9 of 14 stage III tumors (64%). Hyaluronic acid was present in the stroma surrounding all stage I and III tumors, but was more intense in the stroma adjacent to metastatic implants from stage III carcinomas. Tumor cells were uniformly negative for intracellular HA. These results suggest that CD44S and CD44v5 are differentially expressed in early (stage I) and advanced (stage III) ovarian serous carcinomas and support previous studies that suggest a role for CD44 and stromal HA in the dissemination of ovarian epithelial cancer.

Primary ovarian dysgerminoma in a patient with a germline BRCA1 mutation.

Germline mutations in the BRCA1 tumor suppressor gene are associated with increased risk for the development of ovarian cancer. All such cancers thus far reported have been of the epithelial histologic type. We identified an ovarian dysgerminoma in a 16-year-old woman (proband) with a family history of ovarian cancer during a review of histopathologic characteristics of ovarian cancers from women enrolled in the Gilda Radner Familial Ovarian Cancer Registry. Mutation analysis of DNA from this patient's peripheral blood leukocytes revealed a germline BRCA1 mutation (3312insG). The mutation was also present in the mother with breast cancer, a maternal aunt and a distant cousin with ovarian cancer, and a maternal grandfather and an uncle with skin cancer. The development of the proband's dysgerminoma may be unrelated to her germline BRCA1 mutation. Alternatively, such dysgerminomas may be caused by BRCA1 mutations, but occur so infrequently compared with epithelial cancers that they are seldom identified. Analysis of a larger series of ovarian germ cell tumors may resolve this question.

Histopathology of familial ovarian tumors in women from families with and without germline BRCA1 mutations.

Breast cancers from patients with germline BRCA1 mutations show characteristic histopathologic features. However, similar studies of BRCA1-associated ovarian cancers have reported inconsistent findings. Interobserver differences in histopathologic classification are a significant source of variation, and most studies have obtained histopathologic information from pathology reports rather than from review of histopathology slides. We therefore reviewed the histopathology slides and pathology reports to determine histologic type, grade, and stage for cancers of the ovary or peritoneum in 217 women from 126 families enrolled in the Gilda Radner Familial Ovarian Cancer Registry. Peripheral blood DNA from at least 1 affected member of each family was analyzed for BRCA1 mutations, and tumors from BRCA1 mutation-positive families were compared with those from BRCA1-negative families. Of 66 patients from 36 BRCA1-positive families, 64 had ovarian carcinoma, 1 had an ovarian carcinoma in situ, and 1 had a dysgerminoma. Of 151 patients from 90 BRCA1-negative families, 135 had ovarian carcinoma, 10 had ovarian borderline tumors, 3 had ovarian sex cord/stromal tumors, and 3 had primary peritoneal carcinoma. There were fewer grade 1 (P <.001) and stage I (P =.10) cancers in patients from BRCA1-positive families than in patients from BRCA1-negative families. Neither mucinous nor borderline tumors were found in the BRCA1-positive families. Ovarian cancers arising in women from BRCA1-positive families are more likely to be high grade and nonmucinous than cancers arising in women from BRCA1-negative families. The absence of borderline tumors in patients from BRCA1-positive families adds to accumulating evidence that BRCA1 mutations do not play a role in the development of these tumors. HUM PATHOL 31:1420-1424.

Comparative study of ovarian cancer histopathology by registry pathologists and referral pathologists: a study by the Gilda Radner Familial Ovarian Cancer Registry.

OBJECTIVE: The aim of this study was to evaluate whether there is a significant difference in the pathology diagnoses of women in the Gilda Radner Familial Ovarian Cancer Registry between the two expert Registry pathologists and the referral pathologist. Inaccuracies in verification that ovarian cancer did occur in family members could lead to unnecessary prophylactic surgery or genetic testing. METHODS: A retrospective review was performed of (1) site of malignancy; (2) histopathology of malignancy; (3) grade of malignancy; and (4) the presence or absence of malignancy between the Registry and referral pathologists. RESULTS: There was 95.3% complete agreement between the Registry and the referral pathologist on site of origin with a major difference in only 1.0% of the cases. In comparison of histopathology, there was a 61.7% complete agreement, and only 1.0% were considered major differences. There was 68.8% complete agreement in grade of the malignancy, whereas 2.3% were considered major differences. CONCLUSION: When constructing a family pedigree, it is important to obtain pathology reports to confirm the index case diagnosis of the presence or absence of ovarian cancer. However, because of the small percentage of major differences in diagnosis between the two Registry pathologists and the multiple referral pathologists, we believe genetic counselors and treating physicians can rely, in most instances, on the original histopathology report of verification of ovarian cancer without review of the original histopathology slides when recommending surveillance, genetic testing, and/or prophylactic surgery.

Ovarian carcinoma in situ with germline BRCA1 mutation and loss of heterozygosity at BRCA1 and TP53.

BACKGROUND: The two-hit hypothesis for the genesis of cancer predicts that cancer can develop when the wild-type allele of a tumor suppressor gene is lost in an individual with a germline mutation in that gene. Neither loss of heterozygosity (LOH) for BRCA1 nor mutations of the TP53 (also known as p53) gene have been documented prior to invasion in ovarian cancers arising in women with germline BRCA1 mutations. Such documentation is difficult because lesions are rarely identified in ovarian epithelium. We, therefore, looked for LOH at microsatellite polymorphisms linked to the BRCA1 and TP53 tumor suppressor loci in an incidental carcinoma in situ of the ovary removed prophylactically from a woman with a germline BRCA1 mutation. METHODS: By use of laser-capture microdissection, we obtained pure populations of atypical ovarian epithelial cells and normal stromal cells. DNA was extracted, amplified with primers flanking polymorphic microsatellites linked to BRCA1 (D17S855 and D17S579) and TP53 (TP53 and D17S786), and analyzed for LOH at these microsatellites. We also tested for p53 expression in the abnormal epithelium by immunohistochemistry. RESULTS: Both of the markers linked to TP53 showed LOH, as did an intragenic BRCA1-linked marker (D17S855). The other microsatellite marker for BRCA1 was uninformative. Immunohistochemical staining with an antibody to p53 showed strong immunoreactivity confined to the atypical epithelium. CONCLUSIONS: BRCA1, as well as TP53, can undergo LOH prior to stromal invasion in BRCA1-associated ovarian cancer. Strong immunoreactivity for p53 suggests the presence of mutated p53 in these cells as well. These findings suggest that loss of function of these two tumor suppressor genes occurs early in ovarian carcinogenesis in BRCA1 mutation carriers.

Malignant peripheral nerve sheath tumors of the head and neck: analysis of prognostic factors.

Malignant peripheral nerve sheath tumors of the head and neck region are uncommon and may be associated with neurofibromatosis or occur in a sporadic manner. This is a retrospective review of 17 patients with these tumors who were treated at a single institution. Analysis of clinical and pathologic factors that influenced outcome was performed. There were 9 men and 8 women. Seven patients had a history of neurofibromatosis. Radiotherapy was implicated as a possible etiologic factor in 4 patients. The neck was the most frequently involved site. Overall survival at 5 years was 52%. Survival was improved for women and for patients with low-grade tumors. Age, tumor site, and size had no impact on survival. Survival was worse for patients with neurofibromatosis than for those with the sporadic form of the disease (P = 0.02). Survival was calculated by the method of Kaplan and Meier. The significance of such results was based on results of the log rank test. Local recurrence correlated with tumor size and resection margin status. No local recurrences occurred in those patients who had negative margins of resection and received adjuvant radiotherapy. Tumor grade was predictive of the development of distant metastases. Negative margins of resection are essential for obtaining local control, and the addition of adjuvant radiotherapy may be beneficial in this group. Salvage surgery for local recurrence is possible in some patients.

Prognostic significance of p53 and p21(waf1/cip1) immunoreactivity in epithelial cancers of the ovary.

OBJECTIVES: Theprognostic value of p53 expression in epithelial ovarian cancer remains unresolved. We hypothesized that prognosis may relate more to expression of p21(waf1/cip1), the major downstream effector of p53, which can also be induced through p53-independent mechanisms. We therefore studied the relationship of p53 and p21(waf1/cip1) expression in epithelial ovarian cancers to clinicopathological variables and prognosis. METHODS: Fixed, embedded tumors from 85 patients with untreated, primary epithelial ovarian cancer were immunostained with antibodies to p53 and p21(waf1/cip1). Expression was correlated with clinicopathological features and prognosis. Survival curves were calculated by the Kaplan-Meier method and compared using the log-rank test for p53, p21(waf1/cip1), and all combinations of expression of the two markers. RESULTS: Sixty-two percent of tumors expressed p53, and 42% expressed p21(waf1/cip1). There was no correlation between p53 and p21(waf1/cip1) expression. Advanced stage, grade, age >/=50, and p53 expression were associated with worse disease-free survival. Patients whose tumors were p53(+)/waf1(-), however, had a particularly strong association with poorer disease-free survival when compared with other combinations of p53 and p21(waf1/cip1) expression (P = 0.003). Neither p53, nor p21(waf1/cip1), nor combinations of expression were independently related to survival when histology, age, stage, and differentiation were considered. CONCLUSIONS: p53 expression in the absence of p21(waf1/cip1) expression is a better marker of poor prognosis than either p53 or p21(waf1/cip1) expression status alone in univariate analysis. Absence of independent prognostic significance may be related to the paucity of early stage cases in the current study.

p53, c-erbB, and Ki-67 expression in ovaries removed prophylactically from women with a family history of ovarian cancer.

Prophylactically removed ovaries from 64 women were compared with those from 30 women with no known family history of ovarian cancer for expression of the p53 tumor suppressor protein, the c-erbB-2 oncoprotein, and for the proliferation antigen Ki-67. All analyses were performed without knowledge of the family history. Ki-67 was expressed in rare nuclei in both the surface and cyst epithelial cells, whereas p53 was expressed in rare nuclei only in cyst epithelial cells. Neither the proportion of positive cases nor the intensity of staining differed between groups. c-erbB2 was not expressed in surface or cyst epithelium in any case from either group. Nuclei of granulosa and granulosa lutein cells expressed both Ki-67 and p53. In conclusion, increased expression of p53, c-erbB2, and Ki-67 was not found in the epithelium of prophylactically removed ovaries, suggesting that increased expression occurs later in the development of carcinoma or invasive tumor evolves too quickly to identify expression of these proteins in preinvasive epithelium.

HER-2/neu oncogene amplification in stage I and stage III ovarian papillary serous carcinoma.

Oncogene amplification has been implicated in the genesis and progression of many cancers. Overexpression of the HER-2/neu proto-oncogene occurs in 20-30% of ovarian epithelial cancers, in which it may be of prognostic significance. Oncogene overexpression is traditionally studied using immunohistochemistry. In this study we used fluorescent in situ hybridization (FISH) to determine HER-2/neu amplification in ovarian papillary serous carcinoma and compared the frequency of amplification in two stages of the disease. Archival tissues from 23 cases of papillary serous ovarian carcinoma (9 cases of stage I and 14 cases of stage III) were analyzed by FISH using a HER-2/neu probe and a chromosome 17 centromere control probe. Determination of the level of amplification was performed according to the standard protocols of the Cytogenetics Laboratory at Rhode Island Hospital. Of the 23 cases successfully analyzed, the frequency of amplification among stage I tumors was 22% (2/9) and the frequency of amplification among stage III tumors was 71% (10/14). These results are significant (P = 0.036). The frequency of stage I tumors among amplified cases was 17% (27/12) and the frequency of stage III tumors among amplified cases was 83% (10/12). This study not only confirms the presence of a subset of ovarian papillary serous carcinoma with HER-2/neu gene amplification, but it also indicates that HER-2/neu oncogene amplification is more likely to be associated with a more advanced stage. Thus, the present data are consistent with the hypothesis that HER-2/neu amplification, similar to HER-2/neu protein over expression, is a prognostic marker of poor outcome.

Paradoxical correlations of cyclin-dependent kinase inhibitors p21waf1/cip1 and p27kip1 in metastatic colorectal carcinoma.

The cyclin-dependent kinase inhibitors (CDIs) p27kip1 and p21waf1/cip1 are key cell cycle-negative regulatory enzymes. The objective of this study was to correlate expression of p27kip1 and p21waf1/cip1 with survival, chemotherapy responsiveness, and expression of the proliferation marker Ki-67 in patients with advanced colorectal cancer. Immunohistochemistry was performed with antibodies to p27kip1, p21waf1/cip1, and Ki-67 on samples from 66 patients with metastatic colorectal carcinoma. Interpretation was performed by visual inspection and automated image analysis. Patients who obtained a response to chemotherapy had greater p21waf1/cip1 tumor staining with a mean of 10.0 positive cells/high-powered field, compared with 4.5 positive cells/high-powered field for nonresponders (P = 0.03). A positive Spearman correlation was seen between Ki-67 and p27kip1 (r = 0.48; P = 0.0001), as well as between Ki-67 and p21waf1/cip1 (r = 0.48; P = 0.0001). A trend toward shorter survival was seen in patients with positive specimens (median survival of 10 months for patients with both p27kip1- and p21waf1/cip1-positive specimens, compared with 22 months for patients with neither p27kip1- nor p21waf1/cip1-positive specimens). In contrast to that previously reported in normal colonic mucosa or early-stage colorectal cancer, we observed positive correlations of Ki-67 with both p27kip1 and p21waf1/cip1, a trend toward greater CDI staining indicating worse prognosis, and greater p21waf1/cip1 staining in tumors that were chemosensitive. These findings suggest that in the metastatic setting, CDIs may show altered function, compared with their role in the normal cell cycle.

Expression of the hyaluronan receptor, CD44S, in epithelial ovarian cancer is an independent predictor of survival.

Most ovarian carcinomas present at advanced stage, principally as the result of dissemination to peritoneal sites. Standard CD44 (CD44S) is the principal receptor for hyaluronic acid, and in vitro and animal studies have suggested that the attachment of ovarian carcinoma cells to the peritoneal mesothelium involves the interaction between CD44S on ovarian carcinoma cells and hyaluronic acid on mesothelial surfaces. We, therefore, analyzed a series of ovarian carcinomas for the expression of CD44S by immunohistochemistry to see whether expression of this receptor by tumor cells correlated with clinicopathological factors and measures of patient outcome. Fifty-six fixed, paraffin-embedded primary epithelial ovarian tumors were immunostained with antibody to CD44S. Membrane staining was considered positive, and results were correlated with stage, grade, age, histology, and survival. Twenty-two (39%) tumors were positive for CD44S. There was no correlation between CD44 expression and histological type, grade, age, or stage. However, CD44 expression was significantly associated with survival in both univariate (P = 0.003) and multivariate (P = 0.006) analyses. These results support a role for CD44S expression in the spread of ovarian epithelial cancer and suggest that expression of this molecule is a significant independent predictor of survival in women with this disease.

Trisomy 8 in stage I and stage III ovarian cancer detected by fluorescence in situ hybridization.

Ovarian cancer is the leading cause of death from gynecologic maligancy among women in the United States. In 1997, there were nearly 27,000 ovarian cancer cases with over 14,000 deaths. Recent attempts at early detection of ovarian cancer have been aimed at the identification of biomarkers that would indicate an underlining malignant process or reflect the biological behavior of the tumor. Our previous studies revealed that chromosome 8 copy number abnormality, especially trisomy, is common in several cancers. Archival tissues from 24 cases of papillary serous ovarian carcinoma (10 stage I and 14 stage III) were analyzed by fluorescence in situ hybridization (FISH) with a chromosome 8-specific alpha-satellite probe (Oncor, Gaithersburg, MD). The analysis was done according to standard protocols of the Lifespan Academic Medical Center Cytogenetics Laboratory at Rhode Island Hospital. Twenty-one of 24 cases (87.5%) were found to be trisomic for chromosome 8, if a cutoff point of >/=15% cells with three signals is adopted. Overall, 80% of stage I and 93% of stage III tumors had trisomy 8. This study confirms the presence of a high frequency of trisomy 8 in both early and late stages of the disease and suggests that trisomy 8 may be an early event in the multistep process leading to ovarian cancer. It is of interest to note that a higher frequency of trisomy 8 was found in a higher stage of disease, consistent with our previous results on breast cancer. Thus, additional FISH studies of ovarian tumors for chromosome 8 copy number assessment may be warranted.

Altered surface and cyst epithelium of ovaries removed prophylactically from women with a family history of ovarian cancer.

Despite intensive investigation, the nature of epithelial ovarian cancer precursors remains controversial. Because women with a strong family history of ovarian cancer have a high probability of developing ovarian cancer themselves, ovaries removed prophylactically from such patients provide an opportunity to identify early neoplastic changes. Ovaries removed from 64 consecutive patients undergoing prophylactic oophorectomy and from 30 women with normal ovaries and no known family history of ovarian cancer were examined by light microscopy for a number of histopathologic features and by image cytometry for abnormalities of the cyst and surface epithelium. All analyses were performed without knowledge of the family history. Seven benign, but no tumors of low malignant potential or malignant epithelial tumors were found in the prophylactic oophorectomy group. There were more cortical inclusion cysts in the prophylactically removed than controls ovaries (P = .016), but no other architectural features differed between the two groups. No abnormalities were found in the surface or cyst epithelium in either group by light microscopy. In contrast, image analysis identified differences in the nuclei between the two groups, indicating that those from the surface epithelium of prophylactically removed ovaries were larger and contained more heterogeneously dense chromatin than those of controls, and that nuclei of the cyst epithelium had more irregular outlines. Ovarian epithelium from prophylactically removed ovaries exhibit abnormalities that are only identified by image analysis, and which might represent early preneoplastic changes. Such ovaries may be useful for identifying early molecular changes in ovarian cancer.

Mutation analysis of BRCA1, TP53, and KRAS2 in ovarian and related pelvic tumors.

Cancer may be viewed as a genetic disease resulting from critical mutations that disrupt normal cell growth. To characterize the involvement of the BRCA1 and TP53 tumor suppressor genes and of the KRAS2 protooncogene in gynecologic cancer, mutation analysis of these genes was conducted in pelvic tumors of 85 patients that included 49 epithelial ovarian carcinoma cases. The 85 pelvic tumors contained 5 tumors with BRCA1 mutations, 33 with TP53 mutations, and 1 with a KRAS2 mutation. Each of the BRCA1 and KRAS2 mutations, and 25 of the TP53 mutations, were in ovarian carcinomas. Four of the BRCA1 mutations were germline and 1 was somatic. The 4 patients with germline BRCA1 mutations had an early age of disease onset (33-48 years) relative to the mean age of onset (58 years) of all 49 ovarian carcinoma patients, and 3 of these 4 patients had a family history of ovarian or breast cancer. None of the 4 tumors with germline BRCA1 mutations had a KRAS2 mutation or a TP53 mutation, despite a 51% frequency of TP53 mutations in the 49 ovarian carcinomas. Three of the 4 tumors with germline BRCA1 mutations retained a wild-type BRCA1 allele. The tumor with the somatic BRCA1 mutation contained a TP53 mutation and had no evidence for wild-type BRCA1 and TP53 alleles. These data suggest that both BRCA1 and TP53 were inactivated in 1 of 49 ovarian carcinomas. Moreover, mutational inactivation of both BRCA1 and TP53 did not occur in 4 tumors with a germline BRCA1 mutation. It has been proposed that tumorigenesis in cells with a heterozygous BRCA1 mutation requires inactivation of the wild-type BRCA1 and TP53 alleles, which results in genomic instability and acquisition of mutations in protooncogenes. Clearly, mutational inactivation of TP53 and the wild-type BRCA1 allele in ovarian tumors with a heterozygous, germline BRCA1 mutation is not an absolute requirement for tumor formation. It is possible that these alleles may be inactivated by nonmutational mechanisms or that other tumor formation pathways exist.

Metachronous benign solitary fibrous tumours of the pleura (localized "mesotheliomas"): a case report.

Solitary fibrous tumours are uncommon pleural tumours that are not related to mesotheliomas. They are typically benign and pedunculated and may grow to massive sizes. Surgical resection is usually curative. Tumour recurrence and metachronous development of multiple tumours are unusual. In this report a patient was treated for 3 benign solitary fibrous tumours of the pleura over 23 years. The authors hypothesize that this represents multifocal tumorigenesis as opposed to local tumour recurrence. The importance of complete surgical excision and lifelong radiographic follow-up are stressed.

Prognostic factors in extraovarian primary peritoneal carcinoma.

OBJECTIVE: To determine the pathological, clinical, and therapeutic factors which had prognostic significance in women with extraovarian primary peritoneal carcinoma (EOPPC). METHODS: A retrospective, clinicopathologic study was conducted of 75 women diagnosed with EOPPC. Diagnosis and assessment of prognostic pathological factors were based on the Gynecologic Oncology Group (GOG) criteria. Univariate and multivariate analyses were used to assess the following factors for their effect on overall survival: age, parity, presenting symptoms and signs, ascites, CA 125 level, history of oophorectomy, maximum ovarian dimension, histologic type, architectural and nuclear grades, number of mitosis and psammoma bodies, depth of ovarian invasion, estrogen and progesterone receptors (positive, negative), p53 overexpression (present, absent), performance status (GOG criteria), stage (FIGO criteria for ovarian cancer), debulking surgery (optimal versus suboptimal), first-line chemotherapy (platin-based without paclitaxel versus platin/paclitaxel), secondary cytoreduction, and second-line chemotherapy (paclitaxel-based versus no paclitaxel). RESULTS: The median overall survival of all patients was 23.5 months (95% CI 18.6, 39.8 months). The 5-year survival was 26.5% (SE 6.7%). p53 overexpression and estrogen and progesterone receptor positivity were demonstrated in 42.4, 50.0, and 6.3%, respectively. In univariate analysis, performance status, primary debulking surgery, stage, and age were significant on overall survival (P < 0.001, <0. 001, 0.004, and 0.012, respectively). In multivariate analysis, only performance status (P < 0.001) and primary debulking surgery (P = 0. 03) were independent prognostic factors. Conclusions. Overall survival in women with EOPPC is affected significantly by performance status and primary debulking surgery as independent variables. To improve survival, efforts should be made to achieve optimal tumor cytoreduction at primary surgery.