International Breast Cancer Study Group, Bern, Switzerland (E-mail: mcastiglione@sakk.ch)
Ovarian cancer is the most common gynecological malignancy, as well as the leading cause of death from gynecological cancers in the western world. The survival of patients with advanced epithelial ovarian cancer has clearly improved with the advances in surgery and more effective chemotherapy. Because long-term survival and cure are possible in patients who present with advanced ovarian cancer, the therapeutic strategy generally consists of aggressive surgery combined with chemotherapy. The combination of carboplatin and paclitaxel has been accepted as the treatment of choice, based on the results of large prospective randomized trials. Most patients who present with advanced disease will achieve a clinical complete remission after initial therapy. Unfortunately, the majority of these patients will relapse, and most of the relapsed patients are incurable. Median time to progression (TTP) will vary between 18 and 24 months. Although not curative, treatment of patients with recurrent ovarian cancer can result in significant palliation and allow for good quality of life [1]. Several clinico-pathological parameters have been identified as prognostic factors in ovarian cancer, in particular tumor FIGO stage, volume of residual disease and grade of differentiation have proved to fulfill the criteria for a prognostic factor [2]. The identification of additional prognostic and predictive factors would be very helpful in order to better advise patients in the choice between standard therapy and participation in a clinical trial.
Molecular and biochemical parameters have been studied extensively and some have been proposed as possible prognostic or predictive factors. They should help to design better tailored treatments for patients who present with a new solid malignancy. The enzyme cyclooxygenase (COX), responsible for the conversion of arachidonic acid to prostaglandins, has drawn much attention. Cyclooxygenases are potent mediators of inflammation. While COX-1 is constitutively expressed in a wide range of tissues, COX-2 is cytokine inducible. The fact that COX-2 is inducible by pro-inflammatory cytokines and growth factors implies a role for COX-2 in both inflammation and the control of cell growth. The expression of COX-2 has recently been associated with carcinogenesis. In addition, the expression of COX-2 has also been shown to contribute to proliferation, invasion and metastasis in several solid tumors [3]. Expression of COX-2 has been associated with an invasive phenotype in a number of epithelial cancers, such as colorectal, breast, cervix, prostate, urothelial and gastric carcinomas. Furthermore, COX-2 expression has been identified as a predictive factor for poor response to radiation therapy in cervical carcinomas [4]. A recent study has implied the first clinical evidence that expression of COX-2 may also be an independent prognostic factor in human ovarian carcinoma [5]. In this issue of Annals of Oncology, Ferrandina et al. [6] present another study that supports the prognostic significance of COX-2 in ovarian cancer, and implies that the expression of COX-2 is a predictive factor for poor response in patients who present with advanced ovarian cancer and are treated with standard combination chemotherapy with paclitaxel and carboplatin.
Ferrandina et al. assessed the expression of COX-2 by immunohistochemistry in a prospective study of 87 patients with newly diagnosed advanced ovarian cancer. Only patients with measurable disease were included in the study to allow a precise response assessment. Patients were classified according to whether they were candidates for primary cytoreductive surgery or exploratory laparotomy only. All patients received three to six courses of cisplatin-based chemotherapy. Patients who were treated with primary cytoreductive surgery received six cycles of chemotherapy, whereas patients who underwent exploratory laparotomy mostly received three or four cycles before a secondary debulking surgery. COX-2 expression, defined as >10% of total tumor area and strong intensity [only an intensity score of 2 (grading scale 012) was considered positive] was found in 39 cases (44.8%). In both groups of patients, those undergoing cytoreductive surgery and those undergoing exploratory laparotomy only, COX-2 expression was found in a statistically significantly higher percentage of non-responding cases (P = 0.043 and P = 0.0018, respectively). In a multivariate analysis, only the expression of COX-2 and increased age remained as independent predictive factors of chemotherapy resistance. However, the prognostic value of COX-2 expression was only found in one of the two subgroups of patients. With a median follow-up of 25 months, COX-2-positive cases showed a statistically significantly shorter TTP and overall survival in the group who underwent exploratory laparotomy. The difference in TTP and overall survival according to COX-2 status was not significant in the group that underwent primary cytoreductive therapy.
Based on these findings, should we consider COX-2 expression as a new predictive factor, and possibly also prognostic factor, for patients with ovarian cancer?
Denkert et al. [5] have recently reported similar findings in a group of 117 patients with ovarian epithelial tumors. Patient selection was based on availability of tissue. Expression of COX-1 and COX-2 were analyzed by immunohistochemistry in 117 ovarian tumors and two normal ovarian tissues. Lesions were scored as positive for COX if there was either a diffuse staining or a focal expression in several clusters of cells. Expression of COX-1 was seen in 75% of the ovarian carcinomas, as well as in tumors of low malignant potential, and in 100% of the benign ovarian lesions and normal ovarian epithelium. COX-2 expression was only detected in the malignant lesions, in 42% of the 86 ovarian carcinomas, and in 37% of the 19 low malignant potential tumors, but not in the 12 cystadenomas or the two normal ovarian tissue samples. Moreover, in a univariate analysis, expression of COX-2 in invasive carcinomas was associated with a significantly reduced median survival time (52 months for patients with COX-2-negative carcinomas and 30 months for patients with COX-2-positive carcinomas; log rank test P = 0.04), this finding was even more pronounced in patients younger than 60 years of age (P = 0.004). In this study, a multivariate analysis confirmed expression of COX-2 as an independent prognostic factor for poor survival (relative risk 2.74, 95% confidence interval 1.385.47).
Two previous studies had failed to show expression of COX-2 in ovarian tumors [7, 8], but they had either not used immunohistochemistry, or had used different antibodies from the one used by both Denkert et al. and Ferrandina et al. More recently, two studies have demonstrated an expression of COX-2 in most ovarian carcinomas and borderline tumors, but unfortunately the authors did not correlate these findings with survival data [9, 10].
There are currently only two reports about the prognostic significance of COX-2 in ovarian carcinomas. While Denkert et al. showed a clear correlation between expression of COX-2 and a reduced median survival time, Ferrandina et al. only found this association in a subset of patients who underwent exploratory laparotomy only. There are several possible reasons for these different results. The Denkert report is a retrospective study, with the potential selection bias this may introduce. Even though Ferrandina et al. conducted a prospective analysis, eligibility criteria other than "a very homogenous series of stage IIICIV ovarian cancer patients with measurable disease" are not well defined and there may also have been a selection bias. In both studies, COX-2 expression was assessed by immunohistochemistry using the same antibody, but criteria for positivity were defined differently. This may explain why 75% of the ovarian carcinomas were positive for COX-2 in the Denkert report, while only 44% of the cases in the Ferrandina report expressed COX-2. In addition, from the report of Denkert et al. it is not evident whether treatment was taken into consideration for the multivariate analysis. Ferrandina et al. hypothesize that the lack of prognostic significance of COX-2 expression in the subgroup of patients who underwent cytoreduction may be due to the fact that a more invasive surgical procedure may modify the tumorhost interactions and tumor biology to a larger extent than exploratory surgery.
Ferrandina et al. are the first to report the predictive significance of expression of COX-2 in patients with advanced ovarian cancer treated with combination chemotherapy. While the results of this study show that expression of COX-2 may be a strong predictive factor for chemotherapy resistance in patients who present with advanced ovarian carcinomas and are initially treated with either cytoreductive therapy or exploratory laparotomy alone, this data should be interpreted with caution. Even though all patients were treated with the same combination chemotherapy, patients who underwent cytoreduction were likely to receive more cycles of chemotherapy than patients who underwent exploratory laparotomy only. In addition, assessment of response was different between the two groups; in both groups response was assessed by clinical examination, ultrasound and CA 125 assay; however, patients who underwent exploratory laparotomy at first and then secondary debulking also received a computed tomography scan and direct assessment of the response at the time of the secondary debulking. This renders the interpretation of the comparison of response to treatment in the two groups very difficult.
The adoption of COX-2 as a prognostic factor for poor survival may add information to the well accepted clinicopathological parameters and enhance research towards more tailored therapies. Whether any additional benefit will result from the inclusion of selective COX-2 inhibitors in the therapeutic plan of patients whose tumors express COX-2 is an interesting hypothesis, which needs to be explored.
Even though Ferrandina et al., and previously Denkert et al., have shown the prognostic significance of expression of COX-2 in patients with ovarian carcinomas, the interpretation of the results of this study needs to be very cautious. The number of patients is small and most of the analyses were performed on the subgroups. These results are indeed very stimulating and should be taken as a hypothesis-generating report. Large prospective studies are needed to validate the clinical utility of COX-2 expression as a prognosis factor for poor outcome and a predictor of chemotherapy resistance in patients with ovarian malignancies.
M. L. Nasi & M. Castiglione
International Breast Cancer Study Group, Bern, Switzerland (E-mail: mcastiglione@sakk.ch)
References
1. Ozols RF. Recurrent ovarian cancer: evidence-based treatment. J Clin Oncol 2002; 20: 11611163.
2. Hoskins WJ, Perez CA, Young RC. Principles and Practice of Gynecologic Oncology. Philadelphia, PA: Lippincott Williams and Wilkins 2000; 10051007.
3. Soslow RA, Dannenberg AJ, Rush D et al. COX-2 is expressed in human pulmonary, colonic, and mammary tumors. Cancer 2000; 89: 26372645.[ISI][Medline]
4. Ferrandina G, Lauriola L, Distefano MG et al. Increased cyclooxygenase-2 expression is associated with chemotherapy resistance and poor survival in cervical cancer patients. J Clin Oncol 2002; 20: 973981.
5. Denkert C, Kobel M, Pest S et al. Expression of cyclooxygenase-2 is an independent prognostic factor in human ovarian carcinoma. Am J Pathol; 160: 893903.
6. Ferrandina G, Lauriola L, Zanoni GF et al. Increased cyclooxygenase-2 (COX-2) expression is associated with chemotherapy resistance and outcome in ovarian cancer patients. Ann Oncol 2002; 13: 12051211.
7. Ristimäki A, Honkanen N, Jankala H et al. Expression of cyclooxygenase-2 in human gastric carcinoma. Cancer Res 1997; 57: 12761280.[Abstract]
8. Dore M, Cote LC, Mitchell A, Sirois J. Expression of prostaglandin G/H synthase type I, but not type 2, in human ovarian adenocarcinomas. J Histochem Cytochem 1998; 46: 7784.
9. Klimp AH, Hollema H, Kempinga C et al. Expression of cyclooxygenase-2 and inducible nitric oxide synthase in human ovarian tumors and tumor-associated macrophages. Cancer Res 2001; 61: 73057309.
10. Matsumoto Y, Ishiko O, Deguchi M et al. Cyclooxygenase-2 expression in normal ovaries and epithelial ovarian neoplasms. Int J Mol Med 2001; 8: 3136.[ISI][Medline]