Background: In a previous analysis of KEYNOTE-189 (NCT02578680), we showed that tissue TMB (tTMB) assessed by whole-exome sequencing was not significantly associated with efficacy in either arm and that pembro + chemo improved outcomes vs placebo + chemo in both the tTMB ≥175 and tTMB < 175 mut/exome subgroups. Here, we explored the association of bTMB with efficacy in KEYNOTE-189. Methods: 616 patients (pts) were randomized 2:1 to pembro + chemo or placebo + chemo. bTMB was assessed in cfDNA using the Guardant Health Omni assay. Association of bTMB (continuous square root transformed) with outcomes in each arm was assessed using Cox proportional hazards models (OS, PFS) and logistic regression (ORR) adjusted for ECOG PS; statistical significance was determined at the 0.05 level without multiplicity adjustment. The clinical utility of bTMB on outcomes was assessed using the cutoff that most closely approximated the 175 mut/exome tTMB cutoff as determined by AUROC analysis. Data cutoff was 21 Sep 2018. Results: 235 (38%) treated pts had evaluable tTMB and bTMB: 160 in the pembro + chemo arm and 75 in the placebo + chemo arm. bTMB as a continuous variable was not significantly associated with OS or ORR for pembro + chemo (one-sided P = .229 and .051) or placebo + chemo (two-sided P = .641 and .069); bTMB was significantly associated with PFS in the pembro + chemo arm (one-sided P = .015) but not the placebo + chemo arm (two-sided P = .058). bTMB and tTMB scores were moderately correlated (r = .61). The bTMB cutoff that most closely approximated tTMB 175 mut/exome was 15 mut/Mb (AUROC 0.81, 95% CI 0.75-0.86). 178 (76%) pts had concordant bTMB and tTMB results—101 low and 77 high by both—whereas 57 (24%) had discordant results—21 high bTMB but low tTMB, 36 low bTMB but high tTMB. Pembro + chemo improved OS, PFS, and ORR vs placebo + chemo for bTMB ≥15 and < 15 mut/exome (Table). Conclusions: Similar to previous findings based on tTMB, bTMB has limited clinical utility in the setting of pembro with pemetrexed and platinum given as first-line therapy for metastatic nonsquamous NSCLC. Clinical trial information: NCT02578680.