An updated overall survival analysis with correction for protocol-planned crossover of the international, phase III, randomized, placebo-controlled trial of regorafenib in advanced gastrointestinal stromal tumors after failure of imatinib and sunitinib (GRID)

110 Background: The GRID study showed that regorafenib improves progression-free survival compared with placebo in patients with advanced GIST after failure of at least imatinib and sunitinib (HR 0.27; 1-sided pu003c0.0001; Demetri 2013). At the time of the primary analysis, no significant difference in the secondary endpoint of overall survival (OS) was observed (HR 0.77; p=0.199), but this result may have been confounded by the high rate of crossover to regorafenib (85%) of placebo patients at progression. We conducted exploratory analyses of updated OS data to assess the effect of correcting for this protocol-planned crossover. Methods: The data cut-off for this updated OS analysis was 31 January 2014 (2 years after the primary analysis). OS was corrected using two randomization-based methods: rank preserving structural failure time (RPSFT) and iterative parameter estimation (IPE); both methods are considered as best choice among all correction analytics. Hazard ratios and 95% CI were derived using the Cox...