Fax: (011) 0030 210 3381511
Mucinous but not clear cell histology is associated with inferior survival in patients with advanced stage ovarian carcinoma treated with platinum-paclitaxel chemotherapy†
Article first published online: 27 JAN 2010
Copyright © 2010 American Cancer Society
Volume 116, Issue 6, pages 1462–1468, 15 March 2010
How to Cite
Bamias, A., Psaltopoulou, T., Sotiropoulou, M., Haidopoulos, D., Lianos, E., Bournakis, E., Papadimitriou, C., Rodolakis, A., Vlahos, G. and Dimopoulos, M. A. (2010), Mucinous but not clear cell histology is associated with inferior survival in patients with advanced stage ovarian carcinoma treated with platinum-paclitaxel chemotherapy. Cancer, 116: 1462–1468. doi: 10.1002/cncr.24915
See editorial on pages 1400-2, this issue.
- Issue published online: 3 MAR 2010
- Article first published online: 27 JAN 2010
- Manuscript Accepted: 13 JUL 2009
- Manuscript Revised: 25 JUN 2009
- Manuscript Received: 19 MAY 2009
- ovarian cancer;
- clear cell;
Mucinous and clear cell histology have been associated with adverse prognosis in ovarian carcinomas. The authors compared the outcome of these subtypes with that of serous tumors in patients who were treated with combination paclitaxel/platinum at their center.
Four hundred twenty patients with histologically confirmed, serous (n = 367), mucinous (n = 24), or clear cell (n = 29) ovarian carcinomas, International Federation of Gynecology and Obstetrics stage III or IV disease, and who were treated with paclitaxel/platinum after cytoreductive surgery were included in this analysis.
The median overall survival for each histological subtype was 47.7 months (95% confidence interval [CI], 37.7-57.7 months) for serous, 15.4 months (95% CI, 4.2-26.6 months) for mucinous, and 36.6 months (95% CI, 22.7-50.5 months) for clear cell carcinomas. Cox regression analysis showed that mucinous histology was an independent predictor of poor prognosis compared with serous tumors (hazard ratio, 0.360; 95% CI, 0.215-0.603; P = .001). In contrast, such a difference between clear cell and serous carcinomas was not found (P = .337). Median survival of patients with mucinous tumors and residual disease >2 cm was poor, averaging 7.1 months (95% CI, 4.6-9.6 months).
Mucinous but not clear cell histology is associated with significantly worse prognosis in advanced ovarian cancer treated with combination platinum/paclitaxel. Different therapeutic strategies should be studied in this entity. Cancer 2010. © 2010 American Cancer Society.
Epithelial ovarian cancer (EOC) is the fourth most frequent cause of cancer death in women and the leading cause of gynecologic cancer mortality, with around half of the cases occurring later than 65 years of age.1 Women more frequently are diagnosed with advanced stages (International Federation of Gynecology and Obstetrics [FIGO] stages III-IV). The standard treatment includes cytoreductive surgery and chemotherapy comprising a platinum analogue plus paclitaxel.2-5 Although platinum-based systemic chemotherapy has been proven to be effective, cure at these stages is infrequent because of the development of resistance to chemotherapy.6 Nevertheless, prognosis is variable, largely depending on the time of development of platinum resistance. Therefore, advanced ovarian cancer represents a heterogenous group regarding the outcome after initial management. The identification of prognostic factors may aid selection of patients who might benefit from different approaches.
Several clinical and histopathological factors influencing survival in advanced ovarian cancer have been identified, such as quality of primary cytoreduction, age, and tumor grade.7, 8 Histological subtypes have also been associated with different prognoses. In particular, clear cell and mucinous carcinomas have been identified in previous studies as adverse prognostic features.7-9 Both histological subtypes are rare, comprising <10% of epithelial ovarian cancers.10 It has been observed that both types have a poor response to chemotherapy, contrasting with that of serous histology, resulting in statistically lower 5-year survival rates.9, 11 Nevertheless, previous analyses suffer from limitations. The most important are the small number of patients included (as expected because of the rarity of these types) and the heterogeneity of chemotherapy used, because many studies have not included patients treated with the current standard platinum/paclitaxel. In addition, the quality of surgery has been improved with time, and older series may not reflect the current practice in advanced ovarian cancer.
In the present study, we studied the prognostic significance of mucinous and clear cell histology in stage III and IV epithelial ovarian cancer in a cohort of patients homogenously treated with a combination of platinum and paclitaxel after initial cytoreduction.
MATERIALS AND METHODS
From our EOC database, which includes patients receiving first-line chemotherapy in our department since April 1994, we selected only patients with the following criteria: FIGO stages III-IV; serous, mucinous, or clear cell histology; and first-line chemotherapy with the combination of paclitaxel and a platinum compound. Patients receiving additional agents were also included. Only patients with histological confirmation of the diagnosis were included in this analysis. Patients with borderline or germ cell tumors were excluded from the analysis.
All patients were operated on, received chemotherapy, and were followed up at our institution (Alexandra Hospital, Athens, Greece). Patients who received their first course of chemotherapy up to December 2007 were included, to ensure at least 1 year of follow-up, as our database was updated in December 2008.
Staging was performed according to the FIGO staging system. Residual disease after surgical debulking was recorded according to the assessment by the surgeon. Tumor grading and histology were determined according to current FIGO guidelines. Performance status was evaluated according to the Eastern Cooperative Oncology Group (ECOG) standards immediately before the initiation of chemotherapy. Response to first-line therapy was defined by computed tomography scan and CA-125 at the end of chemotherapy; complete response (CR) was defined as no radiological evidence of disease and normal CA-125 levels (<37 IU/mL).
Analyses were performed using SPSS statistical software (SPSS for Windows, version 13; SPSS Inc., Chicago, Ill). For descriptive purposes, continuous variables were presented as means and standard deviations; for categorical purposes, percentages were used. Chi-square test was used for comparisons of baseline characteristics between the different histological subtypes.
Survival was calculated from the day of the initiation of treatment until the date of death or date of last follow-up. Overall survival rates were calculated using the Kaplan-Meier method, and the significance of differences was determined using log-rank test. The variables analyzed were age (<70 years vs ≥70 years), stage (III vs IV), grade (1-2 vs 3), residual disease (≤2 cm vs >2 cm), and performance status (0 vs 1-3). Survival of mucinous and clear cell carcinomas was compared with that of serous tumors. Cox proportional hazards model was used for univariate and multivariate analyses of survival. Variables with a P value <.1 in the univariate analysis were entered in the multivariate Cox regression model. Throughout analysis, the level of 5% was used to denote statistical significance.
Between 1995 and 2007, a total of 513 patients with stage III or IV ovarian cancer treated with paclitaxel/platinum were identified in our database. Histologically, tumors were grouped as 367 (71.5%) serous, 24 (4.7%) mucinous, 38 (7.4%) endometrioid, 29 (5.7%) clear cell, 29 (5.7%) poorly differentiated, and 26 (5%) unspecified. As specified, only patients with serous, mucinous, or clear cell histology were included in this analysis. The baseline characteristics of these patients according to the histological subtype as well as treatment administered are shown in Table 1.
|No. of patients||367||24||29|
|>70||76 (20)||5 (21)||6 (21)||.988||.998|
|≤70||291 (80)||19 (79)||23 (79)|
|IIIA||10 (3)||3 (13)||1 (3)||.047a||.339a|
|IIIB||25 (7)||1 (4)||0 (0)|
|IIIC||280 (76)||13 (54)||22 (76)|
|IV||52 (14)||7 (29)||6 (21)|
|1||22 (6)||5 (25)||NA||.003|
|2||140 (40)||11 (55)|
|3||191 (54)||4 (20)|
|Residual disease, cmb|
|≤2||87 (24)||11 (46)||2 (7)||.016||.036|
|>2||279 (76)||13 (54)||27 (93)|
|0||242 (66)||11 (46)||17 (59)||.042||.402|
|1||83 (23)||3 (12)||7 (24)|
|2||32 (9)||8 (33)||4 (14)|
|3||8 (2)||2 (9)||1 (3)|
|Carboplatin/paclitaxel||268 (73)||17 (71)||15 (52)||.895||.125|
|Cisplatin/paclitaxel||18 (5)||2 (8)||6 (20)|
|Cisplatin/paclitaxel/anthracycline||81 (22)||5 (21)||8 (28)|
|Prechemotherapy CA-125 level, IU/mLd|
|<37||55 (17)||7 (30)||4 (14)||.087||.776|
|≥37||273 (83)||16 (70)||19 (86)|
|Response to chemotherapye|
|CR||153 (59)||5 (28)||9 (41)||.013f||.556f|
|PR||71 (28)||3 (17)||8 (36)||.009g||.342g|
|SD||19 (7)||4 (22)||0 (0)|
|PD||15 (6)||6 (33)||5 (23)|
Median age of the present cohort was 61 years (range, 18-87 years). Carboplatin was administered at area under the curve 6,12 and creatinine clearance was calculated according to Cockroft and Gault13; cisplatin was administered at 75 mg/m2. Paclitaxel was administered at 175 mg/m2, whereas doxorubicin was administered at 50 mg/m2. All agents were administered every 3 weeks.
Compared with serous tumors, mucinous carcinomas had a higher percentage of stage IV cases (29% vs 14%, P = .047), lower percentage of grade III cases (20% vs 54%, P = .003), higher percentage of residual disease ≤2 cm (46% vs 24%, P = .016), and fewer cases with performance status (PS) 0 (46% vs 66%, P = .042). In contrast, the only difference between clear cell and serous tumors was the lower percentage of residual disease ≤2 cm (7% vs 24%, P = .036).
After initial cytoreduction, 298 patients had measurable disease and could be assessed for response to chemotherapy. A striking difference in response to chemotherapy was observed between serous and mucinous carcinomas; mucinous tumors showed significantly lower CR and CR + partial response compared with serous tumors (59% vs 28%, P = .013; 77% vs 45%, P = .009, respectively). No significant difference was observed between serous and clear cell carcinomas.
The median follow-up for the entire cohort was 84 months. During follow-up, 254 patients died from the disease: 217 among serous tumors, 17 among mucinous, and 20 among clear cell tumors. The median overall survival (OS) for each histological subtype was 47.7 months (95% confidence interval [CI], 37.7-57.7 months) for serous, 15.4 months (95% CI, 4.2-26.6 months) for mucinous, and 36.6 months (95% CI, 22.7-50.5 months) for clear cell carcinomas. Log-rank test showed that mucinous carcinomas had significantly shorter median OS compared with serous tumors (P = .001) (Fig. 1). In contrast, such a difference between clear cell and serous carcinomas was not found (P = .337) (Fig. 2).
Because of the imbalance in baseline characteristics between the 3 histological subtypes, we performed multivariate analyses to study whether the differences in survival between mucinous or clear cell tumors and serous carcinomas were independent from the other important prognostic factors. These analyses confirmed that only mucinous carcinomas showed significantly inferior survival compared with serous tumors (P = .001) (Table 2), and no difference was observed for clear cell carcinomas (P = .560). Patients with serous tumors had a 64% reduction in the risk of death compared with those patients with mucinous tumors (hazard ratio [HR], 0.360; 95% CI, 0.215-0.603). The other factors that showed an independent favorable association with survival were ECOG PS 0 (HR, 0.591; 95% CI, 0.448-0.779) and residual disease ≤2 cm (HR, 0.413; 95% CI, 0.277-0.617).
|Variables||Median OS||Log-Rank P||Cox Regression P, HR (95% CI)|
|I, II, and III||53.6|
|Residual disease, cm|
|≤2||121.2||<.001||<.001, 0.413 (0.277-0.617)|
|0||60.4||<.001||<.001, 0.591 (0.448-0.779)|
|Serous||47.7||.001||.001, 0.360 (0.215-0.603)|
Stratification by baseline PS and residual disease after cytoreductive surgery was used to perform separate analyses of the importance of histology in different subgroups of patients with serous or mucinous histology. Although no interaction between these variables and the histological type was found, the median survival of patients with mucinous tumors and residual disease >2 cm was poor, averaging 7.1 months (95% CI, 4.6-9.6 months) (P for interaction = .224).
Traditionally, clear cell and mucinous histologies have been associated with a worse overall survival compared with serous ones.7 Nevertheless, information on the prognostic significance of histology in the context of changing therapeutic standards is limited. We retrospectively studied a large series of homogenously treated patients, all with FIGO stage III or IV. In the interpretation of our results, the following factors should be taken into consideration: retrospective analyses can introduce selection biases, and treatment at relapse was not homogenous. Despite these limitations, we believe that our series is representative of contemporary practice in ovarian cancer. All patients were operated on at our center, which has a high volume of ovarian cancer, and histopathological evaluation was performed by highly experienced pathologists. The median survival of 47 months in serous carcinomas is comparable to that of recent series in the taxane era.2, 3, 7 The composition of our cohort regarding stage, histology, and grade distribution are also the expected for otherwise unselected patients. Finally, our findings regarding prognosis are in line with other studies, showing that performance status and residual tumor volume are consistently independent predictors of prognosis.7-9
Previous data on the prognostic significance of histologic subtypes were only partially confirmed. Whereas mucinous histology was associated with significantly worse survival, clear cell carcinomas were not. The reason for this discrepancy is unclear. Most series showing an association of clear cell type with adverse prognosis have included a sizeable proportion of patients who did not receive paclitaxel.9, 10 The introduction of this agent has changed the prognosis in advanced ovarian cancer, and may have had an impact on this particular histological subtype. Indeed, in a small series of 31 patients, a significantly higher risk ratio and longer OS were observed for patients with clear cell carcinomas who received paclitaxel.14 This could be attributed to the different molecular mechanisms underlying the resistance of these 2 agents. Clear cell has a platinum-resistant phenotype, which could be partially attributed to p53 mutations described in this neoplasm.15 The addition of paclitaxel may reverse the relative resistance of clear cell tumors to platinum therapy, because p53 mutations were not shown to negatively affect the antitumor effect of this combination.14, 15 In 1 of these studies in particular, the addition of paclitaxel resulted in equally high response rates among patients with mutated p53 with serous or clear cell histology,15 consistent with our results. It should be underlined that the abovementioned data can only explain our results to a point, because p53 mutations occur in the minority of clear cell carcinomas. In contrast to these previous findings, in a recent retrospective study, including only patients who received platinum/paclitaxel, clear cell histology was associated with significantly shorter survival than serous carcinomas.7 The small sample size of clear cell tumors in our series may account for the lack of statistical significance of the difference in OS between serous and clear cell types. Nevertheless, the following points should also be taken into consideration. The median survival for serous and mucinous carcinomas was identical to ours, but clear cell tumors had a notably shorter median survival of 24 months, compared with 36 months in our cohort. Reported baseline characteristics were also similar, suggesting that other, yet unidentifiable, factors may play a role in the prognosis of this particular histological subtype.
Paclitaxel-platinum chemotherapy was associated with poor results in mucinous tumors compared with serous and clear cell tumors. The association of mucinous histology with higher occurrence of stage IV disease and PS > 0 could be responsible for this outcome. Nevertheless, this histological subtype was also associated with favorable characteristics, such as more differentiated tumors and higher percentage of residual disease ≤2 cm. Furthermore, a multivariate analysis showed that mucinous histology was an independent factor of poor prognosis. This is consistent with results from recent studies identifying this histological subtype as an adverse prognostic feature irrespective of other important factors, such as surgical outcome.16 Again, differences in molecular profile could, at least partially, explain this striking difference. P53 mutations are rare in mucinous histotypes, suggesting that the combination of a taxane with carboplatin or cisplatin may be inadequate, as outlined above. It has been suggested that other treatments used in mucinous gastrointestinal carcinomas might be more useful.17 Preclinical data have suggested a synergism of oxaliplatin and 5-fluorouracil in cisplatin-resistant experimental models.18 This synergism results from down-regulation of the excision-repair cross-complementation group 1 by 5-fluorouracil, resulting in enhanced sensitivity to oxaliplatin. Furthermore, irinotecan, another effective agent in gastrointestinal malignancies, has been used in small series for the treatment of these tumors with encouraging results.19 Nevertheless, these strategies have not been validated in prospective randomized trials yet.
Median survival for patients with mucinous carcinomas and residual disease >2 cm was extremely poor, averaging 7 months, with only a 2.5% probability of being higher than 9.6 months. Our results suggest that patients with ovarian tumors of mucinous histology are not ideal candidates for interval debulking surgery. Maximal surgical effort is currently the best option as the initial management of these patients. If this is not possible, the inclusion in clinical studies may represent the best option for these patients. In this context, it is encouraging that a higher percentage of residual disease ≤2 cm was observed among mucinous tumors compared with serous carcinomas.
In conclusion, histology remains an important determinant of prognosis in advanced ovarian cancer, but the significance of each histological subtype may be changing in the context of modern therapy for this disease. The treatment of mucinous tumors remains a challenge. Currently, optimal cytoreduction represents the most important factor for long-term remission and should be sought aggressively in these patients.
CONFLICT OF INTEREST DISCLOSURES
The authors made no disclosures.