PR KEYNOTE-048: Phase III study of first-line pembrolizumab (P) for recurrent/metastatic head and neck squamous cell carcinoma (R/M HNSCC)

Background: KEYNOTE-048 was an open-label, randomized phase 3 study of P or P + chemotherapy (C) vs EXTREME (E) as first-line systemic therapy for R/M HNSCC (NCT02358031). Methods: Patients (pts) with R/M HNSCC not curable by local therapy and with no prior systemic therapy (R/M setting) who provided a tumor sample for PD-L1 testing were randomized to P 200 mg Q3W, P+ C (cisplatin 100 mg/m2 or carboplatin AUC 5 Q3W+ 5-FU 1000 mg/m2/d for 4 d Q3W), or E (cetuximab 400 mg/m2 loading/ 250 mg/m2QW+ C) given until PD, unacceptable toxicity, 6 cycles (C), or 24 mo (P). Primary end points for P vs E and P + C vs E were PFS and OS in the PD-L1 combined positive score (CPS) ≥20 and ≥1 and total populations (pop). Cutoff date for this final PFS/interim OS analysis was Jun 13, 2018 (minimum follow-up, ∼17 mo). Results: 882 pts were randomized: 301 to P, 281 to P+ C, 300 to E. P was superior to E for OS in CPS ≥20 (N=255; median 14.9 vs 10.7 mo; HR 0.61 [95% CI 0.45-0.83]; P=0.0007) and ≥1 (N=512; median 12.3 vs 10.3 mo; HR 0.78 [95% CI 0.64-0.96]; P=0.0086); OS for P was non-inferior to E in the total pop (N=601). P did not prolong PFS in CPS ≥20 (P=0.5); per the analysis plan, no further PFS testing was done for P vs E. Confirmed ORR (P vs E) was 23% vs 36% for CPS ≥20, 19% vs 35% for CPS ≥1, and 17% vs 36% for the total pop; median DOR was 20.9 vs 4.2mo, 20.9 vs 4.5 mo, and 20.9 vs 4.5 mo. Gr 3-5 drug-related AE rates were 17% (P) vs 69% (E). P +C was non-inferior and superior to E for OS in the total pop (N=559; median 13.0 vs 10.7 mo; HR 0.77 [95% CI 0.63-0.93]; P=0.0034); OS for P+ C was not significantly superior to E in CPS ≥20 and ≥1 at this interim analysis. PFS was not prolonged with P +C (P=0.2). For P+ C vs E, confirmed ORR was 36% vs 36%, median DOR was 6.7 vs 4.3mo, and gr 3-5 drug-related AE rates were 71% vs 69%. Conclusions: For first-line R/M HNSCC, P significantly improved OS over E in the PD-L1 CPS ≥20 and ≥1 populations and was noninferior in the total population with favorable safety. P+ C significantly improved OS in the total population with safety comparable to E. P and P+ C responses were durable. These data support pembrolizumab and pembrolizumab + platinum + 5-FU as new first-line standards of care for R/ MHNSCC. The study continues to the final OS analysis.