6530 Background: In the phase 3 KEYNOTE-048 trial ( NCT02358031 ) in R/M HNSCC (N = 882), 1L P vs E showed superior OS in PD-L1 CPS ≥20 and CPS ≥1 populations, noninferior OS in the total population, no PFS benefit, and favorable safety; 1L P+C vs E showed superior OS in CPS ≥20, CPS ≥1, and total populations, no PFS benefit, and comparable safety. Results of P or P+C vs E in incurable recurrent only, metastatic only, and R/M subgroups are shown. The metastatic only and R/M subgroups were combined and classified as metastatic. Methods: Patients with incurable recurrent (local and/or regional node recurrent disease) or metastatic HNSCC were randomly assigned 1:1:1 to P, P+C, or E. OS and PFS were estimated using the Kaplan-Meier method. Hazard ratios and 95% CIs were based on a Cox regression model with Efron’s method of tie handling with treatment as a covariate stratified by ECOG performance status, HPV status, and PD-L1 status. Data cutoff: Feb 25, 2019. Results: In the incurable recurrent only subgroup (n = 252), median OS was 11.5 vs 12.1 mo (P vs E) and 13.0 vs 11.1 mo (P+C vs E). In the metastatic subgroup (n = 620), median OS was 11.4 vs 9.7 mo (P vs E) and 13.0 vs 10.1 mo (P+C vs E). Median follow-up, OS, and PFS are shown in Table. Treatment-related adverse events (TRAE) in the incurable recurrent only subgroup: 49.4% (P), 94.7% (P+C), and 97.8% (E); grade 3-5 TRAEs rates: 16.0%, 78.7%, and 73.3%, respectively. TRAE rates in the metastatic subgroup: 61.6% (P), 96.4% (P+C), and 96.4% (E); grade 3-5 TRAEs rates: 17.6%, 69.5%, and 67.0%, respectively. Conclusions: P and P+C vs E were efficacious in the metastatic subgroup; in the relatively smaller subgroup of patients with incurable recurrent only HNSCC, the effect was less pronounced. Consistent with the total study population and regardless of disease state, the safety profile was favorable with P vs E and comparable with P+C vs E. These data further support use of 1L P or P+C in patients with R/M HNSCC. Clinical trial information: NCT02358031 . [Table: see text]