Matching-adjusted indirect comparison (MAIC) of relative efficacy for brigatinib vs. ceritinib and alectinib in crizotinib-resistant anaplastic lymphoma kinase (ALK+) non-small cell lung cancer (NSCLC).

Abstract

e20675 Background: Brigatinib (BRG), ceritinib (CER), and alectinib (ALEC) are second-generation ALK inhibitors developed for crizotinib (CRZ)-resistant ALK+ NSCLC. No randomized trial has directly compared the efficacy of these treatments. The goal of this analysis was to use an MAIC to indirectly compare progression-free survival (PFS) and overall survival (OS) for BRG vs CER and BRG vs ALEC in CRZ-resistant ALK+ NSCLC patients (pts). Methods: This analysis used pt-level data from arm B (180 mg qd with a 7-day lead-in at 90 mg, n=110) of the ALTA trial for BRG and published summary data from ASCEND-1/ASCEND-2 and NP28673 trials for CER and ALEC, respectively. PFS and OS curves from published data were digitized to generate virtual pt-level data for estimation of event rates. Pts in ALTA were assigned weights so that their weighted mean baseline characteristics matched baseline characteristics in each of ASCEND-1, ASCEND-2, and NP28673. Relative treatment effects pre- and post-matching were estimated using Cox proportional hazards models. Results: Prior to matching, baseline imbalances were noted in ECOG PS 2, prior chemotherapy, Asian race, smoking status, and best prior response to CRZ among trials. Weighting reduced effective sample sizes (ESSs) for ALTA pts in the matched comparisons. Relative efficacy results were robust and did not appreciably change after adjusting for the prognostic factors (Table 1). Conclusions: Both before and after matching, BRG had significantly longer PFS compared to CER and ALEC, significantly longer OS than CER, and a trend toward longer OS than ALEC. [Table: see text]