LBA44Pembrolizumab with or without chemotherapy vs chemotherapy in patients with advanced G/GEJ cancer (GC) including outcomes according to Microsatellite Instability-High (MSI-H) status in KEYNOTE-062

Abstract Background KEYNOTE-062 (NCT02494583) was a randomized, study of 1L pembrolizumab (P) or pembro + chemo (P+C) vs chemo (C) in patients (pts) with PD-L1 combined positive score ≥1 (CPS ≥1), HER2-negative, advanced GC. Methods Eligible pts were randomized 1:1:1 to P 200 mg Q3W for up to 2 y, P+C (cisplatin 80 mg/m2 + 5-FU 800 mg/m2/d on d1-d5 Q3W [or capecitabine 1000 mg/m2 BID on d1-d14 Q3W per local guideline]) or placebo Q3W + C. Primary endpoints were OS in CPS ≥1 and CPS ≥10 for P+C vs C and P vs C and PFS (RECIST v1.1; central review) in CPS ≥1 for P+C vs C. ORR (RECIST v1.1; central review) in CPS ≥1 for P+C vs C was the secondary endpoint. The final analysis cutoff date was 26 Mar 2019. Results 763 pts (281 with CPS ≥10) were randomized to P+C (257), P (256), or C (250) (Table). Median follow-up was 11.3 mo. P was noninferior to C for OS in CPS ≥1 per prespecified margins. P vs C prolonged OS in CPS ≥10 (median 17.4 vs 10.8 mo; HR 0.69; 95% CI 0.49-0.97) but wasn’t tested per analysis plan. P+C vs C was not superior for OS in CPS ≥1 or CPS ≥10, with a favorable trend for P+C. In an exploratory analysis of pts with MSI-H tumors with CPS ≥1 (N = 50), median OS was not reached vs 8.5 mo for both P vs C (HR 0.29; 95% CI 0.11-0.81) and P+C vs C (HR 0.37; 95% CI 0.14-0.97). PFS was longer with P vs C (HR 0.72; 95% CI 0.31-1.68) and P+C vs C (HR 0.45; 95% CI 0.18-1.11). ORR was higher with P (57%) and P + C (65%) vs C (37%). Median DOR was 21.2 mo with P, not reached (P + C) vs 7.0 mo (C). Grade 3-5 drug-related AE rates were 17% (P), 73% (P+C), and 69% (C). Conclusions As 1L therapy for advanced GC, P was noninferior to C for OS in CPS ≥1 with clinically meaningful improvement for OS in CPS ≥10. P+C did not show superior OS and PFS in CPS ≥1 and OS in CPS ≥10. Clinical benefit was substantially enhanced in a small subset of pts with MSI-H tumors. The safety profile was more favorable for P vs C. Table:LBA44CPS ≥1P+CCPCaMedian, mo (95% CI)N = 257N = 250N = 256N = 250OSa12.5 (10.8-13.9)/ 11.1 (9.2-12.8)10.6 (7.7-13.8)/11.1 (9.2-12.8)HR (95% CI)/ b99.2% CI0.85 (0.70-1.03)0.91 (0.74-1.10)P = 0.0460.91b (0.69-1.18); NI margin = 1.2PFSa6.9 (5.7-7.3)/ 6.4 (5.7-7.0)2.0 (1.5-2.8)/6.4 (5.7-7.0)HR (95% CI)0.84 (0.70-1.02); P = 0.0391.66 (1.37-2.01)MSI-HN = 17/N=19N = 14/N=19OSaNot reached (3.6-NR)/8.5 (5.3-20.8)Not reached (10.7-NR)/8.5 (5.3-20.8)ORR, %64.7/36.857.1/36.8PFSaNot reached (3.6-NR)/6.6 (4.4-8.3)11.2 (1.5-NR)/6.6 (4.4-8.3)DOR, median, mo (range)NR