KEYNOTE-859 study of pembrolizumab plus chemotherapy for advanced HER2-negative gastric or gastroesophageal junction (G/GEJ) cancer: Outcomes in the protocol-specified PD-L1–selected populations.

Abstract

4014 Background: In the ITT population of the KEYNOTE-859 study of HER2-negative, advanced G/GEJ cancer (NCT03675737), pembrolizumab (pembro) + chemotherapy (chemo) significantly improved OS (HR 0.78, 95% CI 0.70-0.87; P < 0.0001), PFS (HR 0.76, 95% CI 0.67-0.85; P < 0.0001), and ORR (51.3% vs 42.0%; P = 0.00009) vs placebo + chemo at the protocol-specified interim analysis. The safety profile of pembro + chemo was as expected. We present efficacy outcomes of the protocol-specified PD-L1 combined positive score (CPS) ≥1 and CPS ≥10 populations. Methods: Eligible pts aged ≥18 y with HER2-negative, previously untreated locally advanced or metastatic G/GEJ adenocarcinoma, ECOG PS 0-1, and known PD-L1 CPS were randomized 1:1 to pembro 200 mg or placebo IV Q3W for ≤35 cycles, both given with investigator’s choice of 5-FU + cisplatin (FP) or capecitabine + oxaliplatin (CAPOX). Randomization was stratified by region (Europe/Israel/North America/Australia vs Asia vs rest of world), PD-L1 CPS (<1 vs ≥1), and chemo (FP vs CAPOX). Per protocol, the primary endpoint of OS and the secondary endpoints of PFS and ORR per RECIST v1.1 by blinded independent central review were tested in the PD-L1 CPS ≥1 and ≥10 populations. Data are from the interim analysis (median study follow-up, 31.0 mo). Results: At baseline, 618 (78.2%) of 790 pts randomized to pembro + chemo and 617 (78.2%) of 789 pts randomized to placebo + chemo had PD-L1 CPS ≥1; 279 (35.3%) and 272 (34.5%), respectively, had CPS ≥10. Baseline characteristics were generally consistent between treatment arms and populations. In the PD-L1 CPS ≥1 population, median OS was 13.0 mo (95% CI 11.6-14.2) for pembro + chemo vs 11.4 mo (95% CI 10.5-12.0) for placebo + chemo (HR 0.74, 95% CI 0.65-0.84; P < 0.0001), median PFS was 6.9 mo (95% CI 6.0-7.2) vs 5.6 mo (95% CI 5.4-5.7) (HR 0.72, 95% CI 0.63-0.82; P < 0.0001), ORR was 52.1% vs 42.6% ( P = 0.00041), and median DOR was 8.3 mo (range 1.2+ to 41.5+) vs 5.6 mo (1.3+ to 34.2+). In the PD-L1 CPS ≥10 population, median OS was 15.7 mo (95% CI 13.8-19.3) with pembro + chemo vs 11.8 mo (95% CI 10.3-12.7) with placebo + chemo (HR 0.65, 95% CI 0.53-0.79; P < 0.0001), median PFS was 8.1 mo (95% CI 6.8-8.5) vs 5.6 mo (95% CI 5.4-6.7) (HR 0.62, 95% CI 0.51-0.76; P < 0.0001), ORR was 60.6% vs 43.0% ( P = 0.00002), and median DOR was 10.9 mo (range 1.2+ to 41.5+) vs 5.8 mo (1.4+ to 31.2+). Among all treated pts in the pembro + chemo (n = 785) and placebo + chemo (n = 787) arms, immune-mediated AE incidence was 27.1% vs 9.3%. Conclusions: The addition of pembro to FP or CAPOX significantly improved OS, PFS, and ORR in the PD-L1 CPS ≥1 and ≥10 populations. Together with the efficacy and safety results from the ITT population, these data support pembro + chemo as a new first-line treatment option for pts with locally advanced or metastatic HER2-negative G/GEJ adenocarcinoma, regardless of PD-L1 expression. Clinical trial information: NCT03675737 .