Objective: Despite results from Gynecologic Oncology Group (GOG) 157 showing no statistically significant survival differences in patients treated with 3 versus 6 cycles of carboplatin and paclitaxel, further analysis of GOG 157 data suggested that certain early-stage epithelial ovarian cancers (EOCs) might benefit from extended chemotherapy. We sought to determine those stage I EOC cases at highest risk of failing 3 cycles of therapy.
Methods: All patients with surgical International Federation of Gynecology and Obstetrics stage I EOC operated on at the Mayo Clinic and The Ohio State University between January 1991 and December 2007 were identified through retrospective chart review. A cohort of patients who received 6 cycles of adjuvant carboplatin and paclitaxel chemotherapy was compared with a cohort of patients who received 3 cycles. Disease-free survival and disease-specific survival were primary outcomes analyzed.
Results: There were 107 patients who received either 3 or 6 cycles of adjuvant carboplatin and paclitaxel. Among all stage I EOCs, the number of cycles did not influence disease-free survival or disease-specific survival. The highest recurrence rate (7 [46.7%] of 15 cases) was among stage IC cases with fixed tumors and positive cytology and/or surface involvement. Among this cohort, 6 (66.7%) of the 9 patients who received 3 cycles recurred, whereas only 1 (16.7%) of the 6 patients who received 6 cycles recurred (hazard ratio, 5.97; 95% confidence interval [CI], 0.98-114.46; P = 0.05, Cox proportional hazards regression model) for an odds ratio of 3.94. The absolute risk reduction for 6 cycles in this highest risk cohort was 50%.
Conclusions: Patients with stage IC cancer and with fixed tumors and positive cytology and/or tumor surface involvement appear to have a higher risk of recurrence after 3 cycles (compared with 6) of platinum-based chemotherapy. The clinical behavior of this highest risk cohort implies a more aggressive tumor biology, and further understanding of such stage I EOCs is warranted.