Defining clinical and statistical improvement in consolidation single-arm trials in ovarian cancer.

Abstract

5032 Background: The design of single-arm consolidation or maintenance trials is challenging from an endpoint perspective. Studies are currently designed to show an improvement in PFS when considered against historical controls. PFS includes duration of second- line therapy (SLT), treatment-free interval (TFI), and time on protocol therapy (TPT). We illustrate how statistical power is affected by the duration of SLT. Methods: Efficacy data from four published single-arm consolidation therapies in second remission in ovarian cancer (Sabbatini et al, Gyn Onc 2010, 116[1]:66-71) are used for illustration. The studies were designed to show increase in median PFS from 9 to 13.5 months (mos). Three intervals: duration of SLT, TFI, and TPT were calculated. We assumed that PFS follows exponential distribution from start of SLT with a drop in hazard at start of protocol therapy. We calculated the sample sizes required to have 80% power, 33% or 40% hazard reduction, for a 1-sided comparison versus historical PFS of 9 mos using intent to treat paradigm. Results: Median duration of SLT was 4.5 mos (IQR 3.6-5.9), median TFI: 2.5 mos, TPT varied from 4 mos to 7.5 mos. Required sample sizes varied with duration of SLT (Table). Alternatively, if one accrues 35 patients with planned 33% hazard reduction and the duration of SLT delays the start of protocol consolidation therapy to 7.5 months, the power drops to less than 40%. To maintain 80% power in this scenario, either the protocol therapy would have to reduce the hazard by 57%, or sample size would have to increase to approximately 170 events. Conclusions: A longer duration of SLT is associated with lower statistical power unless the magnitude of benefit with the new investigational treatment is greater than expected, or the sample size is increased. Designs of consolidation trials should take into account the duration of SLT, by either excluding it from the definition of PFS or restricting it per protocol. Time of initiating consolidation 33% hazard reduction sample size 40% hazard reduction sample size 0 mos (at start SLT) 35 22 4.5 mos 90 55 6 mos 124 75 7.5 mos 170 103 No significant financial relationships to disclose.