Association between immune-related adverse events (irAEs) and atezolizumab efficacy in advanced NSCLC: analyses from the Ph III study OAK

Background: OAK, the first randomized Ph III study of atezolizumab (atezo; anti-PD-L1), demonstrated a superior survival benefit with atezo vs. docetaxel (doc) in NSCLC patients (pts) who had failed prior platinum therapy (HR 0.73; 95% CI: 0.62, 0.87). This analysis evaluates the benefit of atezo in pts with and without irAEs from the primary efficacy population (n = 850) of OAK. Safety data are reported separately. Methods: Pts were randomized 1:1 to atezo (1200 mg) or doc (75 mg/m 2 ) IV q3w. irAEs were defined using MedDRA Preferred Terms that included both diagnosed immune conditions and signs and symptoms potentially representative of immune related events, regardless of investigator-assessed causality. For this analysis, efficacy was evaluated in pts with and without irAEs in the atezo and doc arms; efficacy endpoints were OS, PFS and ORR. To overcome the inherent survivor bias between irAE subgroups (pts surviving longer may be more likely to have irAEs), OS was evaluated using a time-dependent (TD) Cox model. Additional exploratory analyses included atezo efficacy in pts who did vs. did not receive systemic steroids for irAEs. Data cutoff: July 7, 2016. Results: The incidence of irAEs in the atezo arm was 31% (25.0% grade 1 – 2, 6.2% grade 3 – 4, no grade 5). Baseline characteristics including PD-L1 expression on tumor cells or tumor-infiltrating immune cells were generally similar between irAE subgroups. OS per TD Cox model was in favor of atezo arm pts with irAEs vs. those without irAEs (HR 0.79; 95% CI: 0.60, 1.05). Median time to onset of first irAE was 1.6 mo; post irAE mOS was 17.3 mo (95% CI: 11.7, 21.2). 24 atezo arm pts (6%) required corticosteroid treatment. Median OS in pts who did vs. did not receive corticosteroids was 16.0 mo (95% CI: 10.3, 23.5; n = 24) vs. 21.9 mo (95% CI: 16.6, NE; n = 106), respectively. Median PFS was 5.9 mo (95% CI: 2.6, 14.5) vs. 5.4 mo (95% CI: 4.2, 8.8) and ORR was 29% (95% CI: 13, 51) vs. 21% (95% CI: 13, 30) in pts who did vs. did not receive corticosteroids. Conclusions: In this analysis, irAEs did not negatively impact the survival benefit of atezo. Further investigation on the impact of corticosteroids on atezo efficacy in randomized trials is needed. NCT02008227.