Comparison of overall survival of lisocabtagene maraleucel (liso-cel) versus standard of care (SOC) adjusting for crossover in second-line (2L) R/R large B-cell lymphoma
Résumé
Introduction: TRANSFORM (NCT03575351) is a multinational, phase 3 study that compared efficacy and safety of liso-cel versus SOC, including overall survival (OS) as a key secondary endpoint, as a 2L treatment for patients (pt) with relapsed/refractory (R/R) large B-cell lymphoma who were eligible for transplant. The study design permitted pts treated with 2L SOC to receive liso-cel as third line (3L) treatment if protocol-defined conditions were met and requested by investigators. The intention-to-treat (ITT) analysis of OS for liso-cel versus SOC in TRANSFORM addressed a specific research question that did not account for treatment switching. Additionally, as real-world use of chimeric antigen receptor (CAR) T cell therapies in 3L may be less common compared with TRANSFORM, the ITT analysis may provide a conservative estimate of the treatment effect for liso-cel versus SOC on OS. Here, we conducted a complementary analysis of OS to adjust for crossover and estimate the relative effect of liso-cel versus SOC in the absence of 3L CAR T cell therapy.
Methods: An external control arm based on external data from the Collaborative Trial in Relapsed Aggressive Lymphoma (CORAL) phase 3 study (NCT00137995) was created using the inverse probability of treatment weighting (IPTW) method. CORAL provided relatively mature survival data and reflects OS in the absence of CAR T cell therapies, which were not available at the time CORAL was conducted. After aligning inclusion/exclusion criteria of the 2 studies, 258 pts from CORAL were retained. To control for confounding, the 2L SOC cohort from CORAL was weighted to match the baseline distributions of prognostic factors and treatment effect modifiers of the liso-cel population in TRANSFORM. Relative efficacy was estimated by fitting a weighted Cox regression model to survival data for the external SOC arm from CORAL and the liso-cel arm from TRANSFORM.
Results: After population adjustment via IPTW, the effective sample size (ESS) for SOC (CORAL) was 43.6% of the unadjusted sample size. The IPTW approach generally reduced imbalances in baseline characteristics between the 2 populations. The adjusted hazard ratio (HR; 95% CI) for liso-cel versus SOC was 0.50 (0.32–0.78), indicating that liso-cel was associated with a survival benefit in comparison with SOC in the absence of treatment switching (Table).
Conclusions: These analyses demonstrated that liso-cel is associated with prolonged OS in comparison with SOC under a scenario of no 3L CAR T cell therapy, and the consideration of external data is a suitable alternative to existing statistical methods to adjust for treatment switching.