Overall survival (OS) should be evaluated in all randomized cancer trials, even when not the primary end point, as it is clinically meaningful and measures both safety and efficacy. Crossover from the control to experimental arms in randomized trials may be incorporated to allow access to promising investigational treatments after progression on a control arm. The results of Study LIBRETTO-431, an ex-US multiregional, open-label, randomized, active-controlled trial of selpercatinib versus platinum-based and pemetrexed chemotherapy with or without pembrolizumab in patients with treatment-naïve advanced rearranged during transfection fusion–positive non–small cell lung cancer, highlight the challenges of interpreting OS in trials with high crossover rates and variable postprogression therapies. Trial results demonstrated a large improvement in progression-free survival with an acceptable safety profile, but were accompanied by an immature OS analysis with a hazard ratio of 1.26, favoring the chemoimmunotherapy arm. Regardless of the position of OS in the end point hierarchy, it is critical that OS analyses include prespecified plans for data collection and analytical methods to account for the potential impact of postprogression therapies on the interpretation of study results.
Duke et al. (Fri,) studied this question.