Pembrolizumab (pembro) plus platinum-based chemotherapy (chemo) for metastatic NSCLC: Tissue TMB (tTMB) and outcomes in KEYNOTE-021, 189, and 407

Abstract

Background: Pembro +chemo improves efficacy as 1L therapy for metastatic squamous and nonsquamous NSCLC, regardless of PD-L1 TPS. We explored the relationship between tTMB and efficacy in KEYNOTE-021 C and G (nonsquamous; NCT02039674), 189 (nonsquamous; NCT02578680), and 407 (squamous; NCT02775435). Method(s): tTMB was determined by whole-exome sequencing of tumor and matched normal DNA. For each study, the association of tTMB (continuous log10 transformed) with outcomes for pembro +chemo and for chemo was assessed in pts with evaluable tTMB using Cox proportional hazards models (OS, PFS) and logistic regression (ORR); statistical significance was determined at the 0.05 level (no multiplicity adjustment). Clinical utility of tTMB on outcomes was assessed using a prespecified cutpoint of 175 mut/exome. Primary study endpoints were OS and PFS in KEYNOTE-189 and 407 and ORR in 021. Result(s): tTMB was evaluable in 48% of treated pts in KEYNOTE-021 C and G, 48% in 189, and 56% in 407. Baseline characteristics and efficacy were similar in the tTMBevaluable and total populations. tTMB (continuous log10) was not significantly associated with efficacy of pembro +chemo or chemo (minimum P values across studies: 1- sided .072 for ORR, .052 for PFS, and .081 for OS for pembro + chemo; 2-sided .086, .055, and .475, respectively, for chemo). Pembro +chemo improved OS and PFS for tTMB >=175 and<175 mut/exome in KEYNOTE-189 and 407 (Table). In pembro + chemo-treated pts in KEYNOTE-021 G, ORR (95% CI) was 60.8% (38.5-80.3) in the 23 pts with tTMB<175 and 71.4% (47.8-88.7) in the 21 pts with tTMB >=175 mut/ exome. Conclusion(s): In this exploratory analysis, tTMB was not significantly associated with efficacy of pembro + platinum-based chemo or of chemo alone as 1L therapy for metastatic NSCLC, regardless of histology. Pembro +chemo showed survival benefit in the tTMB-high and low subgroups for both squamous and nonsquamous NSCLC.