Fixed and Adaptive Parallel Subgroup-Specific Design for Survival Outcomes: Power and Sample Size

Biomarker-guided clinical trial designs, which focus on testing the effectiveness of a biomarker-guided approach to treatment in improving patient health, have drawn considerable attention in the era of stratified medicine with many different designs being proposed in the literature. However, planning such trials to ensure they have sufficient power to test the relevant hypotheses can be challenging and the literature often lacks guidance in this regard. In this study, we focus on the parallel subgroup-specific design, which allows the evaluation of separate treatment effects in the biomarker-positive subgroup and biomarker-negative subgroup simultaneously. We also explore an adaptive version of the design, where an interim analysis is undertaken based on a fixed percentage of target events, with the option to stop each biomarker-defined subgroup early for futility or efficacy. We calculate the number of events and patients required to ensure sufficient power in each of the biomarker-defined subgroups under different scenarios when the primary outcome is time-to-event. For the adaptive version, stopping probabilities are also explored. Since multiple hypotheses are being tested simultaneously, and multiple interim analyses are undertaken, we also focus on controlling the overall type I error rate by way of multiplicity adjustment