Patients and Methods Women with stages III to IV disease were stratified by coordinating center, maximal diameter of residual tumor, and intent for interval cytoreduction and were then randomly assigned among five arms that incorporated gemcitabine, methoxypolyethylene glycosylated liposomal doxorubicin, or topotecan compared with carboplatin and paclitaxel. The primary end point was OS and was determined by pairwise comparison to the reference arm, with a 90% chance of detecting a true hazard ratio of 1.33 that limited type I error to 5% (two-tail) for the four comparisons.
Results Accrual exceeded 1,200 patients per year. An event-triggered interim analysis occurred after 272 events on the reference arm, and the study closed with 4,312 women enrolled. Arms were well balanced for demographic and prognostic factors, and 79% of patients completed eight cycles of therapy. There were no improvements in either PFS or OS associated with any experimental regimen. Survival analyses of groups defined by size of residual disease also failed to show experimental benefit in any subgroup.
Conclusion Compared with standard paclitaxel and carboplatin, addition of a third cytotoxic agent provided no benefit in PFS or OS after optimal or suboptimal cytoreduction. Dual-stage, multiarm, phase III trials can efficiently evaluate multiple experimental regimens against a single reference arm. The development of new interventions beyond surgery and conventional platinum-based chemotherapy is required to additionally improve outcomes for women with advanced EOC.