Long-term comparative efficacy and safety of nivolumab plus ipilimumab relative to other first-line therapies for advanced non-small-cell lung cancer: A systematic literature review and network meta-analysis.

ABSTRACT

OBJECTIVES: To quantify the long-term comparative efficacy and safety of nivolumab in combination with ipilimumab (NIVO + IPI) relative to other immunotherapy (IO)-based regimens and chemotherapy in patients with first-line advanced non-small cell lung cancer (aNSCLC). METHODS: Phase 3 randomized controlled-trials (RCTs) with minimum 3-year follow-up evaluating IO-based regimens approved for first-line aNSCLC were identified via systematic literature review. Analytic populations were defined by levels of PD-L1 expression and histology. Due to presence of proportional hazards violations, time-varying hazard ratios (HRs) of overall survival (OS) and progression-free survival (PFS) were estimated via Bayesian fractional polynomial network meta-analysis. For safety endpoints, odds ratios (ORs) were estimated using indirect treatment comparisons (ITCs). RESULTS: CheckMate 227, KEYNOTE-189, KEYNOTE-407, KEYNOTE-024, KEYNOTE-042, and IMpower150 were included in the base case analysis. For OS and PFS, HRs of NIVO + IPI relative to other IO-based regimens trended downward over time across analytic populations. The 36-month OS HRs of NIVO + IPI versus comparators were 0.69 (95 % credible interval 0.47, 1.00) versus pembrolizumab + chemotherapy and 0.65 (0.45, 0.93) versus atezolizumab + bevacizumab + chemotherapy in the non-squamous and PD-L1 all-comers population; 0.73 (0.53, 1.02) versus pembrolizumab + chemotherapy in the squamous and PD-L1 all-comers population; and 1.05 (0.83, 1.32) versus pembrolizumab in the mixed histology and PD-L1 ≥ 50 % population. For PFS, 36-month HR point estimates ranged from 0.46 to 0.85 (only statistically significant versus pembrolizumab + chemotherapy in the squamous population; 0.46 [0.31, 0.69]). Adverse events (AEs) leading to discontinuation were not statistically significantly different between NIVO + IPI and pembrolizumab + chemotherapy, nor between NIVO + IPI and pembrolizumab monotherapy, although treatment-related grade ≥ 3 AEs were higher with NIVO + IPI than pembrolizumab monotherapy (OR = 2.21 [1.30, 3.75]). CONCLUSIONS: This study indicates trends towards long-term benefit with NIVO + IPI compared with other IO-based combinations, with manageable toxicities.