Shiva: Randomized Phase Ii Trial Comparing Molecularly Targeted Therapy Based On Tumor Molecular Profiling Versus Conventional Therapy In Patients With Refractory Cancer-Overall Survival (Os) Analysis.

11515 Background: The SHIVA trial is a multicentric randomized phase II trial comparing molecularly targeted therapy among 11 drugs based on tumor molecular profiling versus conventional therapy in patients with any type of cancer that is refractory to standard of care (NCT01771458). Only patients who had a druggable molecular alteration (DMA) identified on a mandatory tumor sample from a metastatic site using targeted sequencing, CGH and IHC were randomized. Cross-over was allowed at disease progression. The trial did not show any difference for its primary endpoint (PFS) [Le Tourneau et al., Lancet Oncol 2015]. We report here the OS of randomized and non-randomized patients. Methods: OS was estimated in the 4 following groups: 1) randomized patients, 2) patients for whom a DMA was identified but who were not subsequently randomized because they did not meet the randomization criteria (PS of 0 or 1, adequate organ function), 3) non-randomized patients because of the absence of DMA, and 4) non-randomized patients because no genomic analyses were performed. Since 70% of patients randomized into the standard arm eventually crossed over to the targeted therapy arm, all randomized patients were analyzed in group 1. The groups were compared in terms of patient characteristics using student and χ2 tests. OS was estimated using the Kaplan Meier method. Results: Among 741 patients included in SHIVA, 8 patients were included twice. Follow-up data were available for 680 out of the 733 patients. 197, 78, 222 and 183 patients belonged to groups 1 to 4, respectively. Median OS of the whole cohort was 7.9 months [95% CI: 7.0–9.1].As compared to non-randomized patients due to the absence of identified DMA, non-randomized patients with a DMA had a significantly worse prognosis: HR = 2.3 [95% CI: 1.7–3.0] (p > 0.0001) whereas randomized patients had a non-significant trend toward a better prognosis: HR = 0.85 [95% CI: 0.7–1.1] (p = 0.18). Conclusions: A statistically significant difference in OS was only observed in patients with a DMA who were not randomized. However, our analysis does not allow showing an intrinsic prognostic value of the DMA on OS.