Long-term results of the randomized, phase 3 KEYNOTE-412 trial of pembrolizumab (pembro) or placebo (pbo) plus concurrent chemoradiotherapy (CRT) for unresected, locally advanced head and neck squamous cell carcinoma (LA HNSCC).

Abstract

6013 Background: In the final efficacy analysis of the randomized, double-blind, phase 3 KEYNOTE-412 trial (NCT03040999), pembro + CRT did not significantly improve event-free survival (EFS) vs pbo + CRT (HR 0.83; 95% CI 0.68-1.03) in unresected LA HNSCC. We present results for KEYNOTE-412 with >2 yrs of additional follow-up. Methods: Adults with newly diagnosed high-risk unresected LA HNSCC (any T3-T4 [N0-N3] or any N2a-3 [T1-T4] larynx/hypopharynx/oral cavity/p16-negative oropharynx cancers and T4 or N3 p16-positive oropharynx cancer) were randomly assigned to receive CRT (70 Gy in 35 fractions + 3 cycles cisplatin 100 mg/m 2 Q3W) + 17 cycles of pembro 200 mg or pbo IV Q3W: first cycle 1 week prior to CRT, 2 cycles during CRT, then 14 cycles of maintenance. The primary end point was EFS assessed by blinded independent central review. The key secondary end point was overall survival (OS). Efficacy was analyzed in all randomly assigned pts (ITT population). Exploratory analyses included locoregional control (LRC), distant metastasis-free survival (DMFS), incidence of second malignancies in the ITT population, and efficacy in pts with PD-L1 CPS ≥1. Results: 402 pts were assigned to each arm; and 398 received ≥1 dose of study treatment in each arm. As of data cutoff date (August 21, 2024), median study follow-up was 74.4 mo (range, 63.7-88.1). EFS was longer with pembro vs pbo (HR 0.79; 95% CI 0.65-0.96). Overall, 186 (46.3%) and 217 (54.0%) EFS events occurred in the pembro and pbo arms, which represents an additional 15 events in the pembro arm and 25 in the pbo arm since the previous analysis. Full efficacy results for the ITT population are in the table. LRC HR was 0.80 (95% CI 0.57-1.14). Overall, 36 pts (9.0%) in the pembro arm and 45 (11.2%) in the pbo arm developed a secondary malignancy. In pts with PD-L1 CPS ≥1 (pembro, n = 339; pbo, n = 346), median EFS was 70.9 mo (95% CI 55.4-not reached [NR]) for the pembro arm and 48.3 mo (95% CI 26.8-66.8) for the pbo arm (HR 0.80; 95% CI 0.64-0.98); median OS was NR (NR; 95% CI NR-NR) for the pembro arm and NR (95% CI 70.0-NR) for the pbo arm (HR 0.84; 95% CI 0.66-1.06). The safety profile was consistent with previously reported adverse events at the time of the final analysis. Conclusions: At end of trial, with >2 yrs of additional follow-up, results showed a clinically meaningful EFS benefit with pembro + CRT versus pbo + CRT and no new safety signals in pts with LA HNSCC. Clinical trial information: NCT03040999 . Pembro + CRT(n = 402) Pbo + CRT(n = 402) EFS, median (95% CI), mo 71.8 (55.4-NR) 49.8 (26.8-66.2)  HR (95% CI) 0.79 (0.65-0.96)  5-yr EFS rate, % 54.7 47.2 OS, median (95% CI), mo NR (NR-NR) NR (74.3-NR)  HR (95% CI) 0.86 (0.70-1.07)  5-yr OS rate, % 64.4 59.8 DMFS, median (95% CI), mo NR (68.9-NR) 64.3 (49.8-76.0)  HR (95% CI) 0.80 (0.65-0.98)  5-yr DMFS rate, % 58.6 51.3