Background
Five-year update of the phase 3, randomized, double-blind KEYNOTE-671 study of perioperative pembrolizumab in resectable stage II, IIIA, or IIIB (N2) NSCLC. 797 participants with previously untreated resectable disease were randomized 1:1. Dual primary endpoints were event-free survival (EFS) and overall survival (OS).
Results
Interventions and follow up: Arm A: neoadjuvant pembrolizumab 200 mg q3w ×4 + platinum-doublet chemotherapy → surgery → adjuvant pembrolizumab q3w up to 13 cycles
Arm B: neoadjuvant placebo + platinum-doublet → surgery → adjuvant placebo
Primary endpoint: EFS (RECIST v1.1, investigator) and OS
mFollow up: 60.4 months
Results: EFS: 5-year 49.9% (95% CI 44.6–55.0) vs 26.5% (21.7–31.5), HR 0.58 (95% CI 0.48–0.69)
OS: 5-year 64.6% (59.5–69.2) vs 53.6% (48.3–58.6), HR 0.74 (95% CI 0.59–0.92)
Arm B: neoadjuvant placebo + platinum-doublet → surgery → adjuvant placebo
Primary endpoint: EFS (RECIST v1.1, investigator) and OS
mFollow up: 60.4 months
Results: EFS: 5-year 49.9% (95% CI 44.6–55.0) vs 26.5% (21.7–31.5), HR 0.58 (95% CI 0.48–0.69)
OS: 5-year 64.6% (59.5–69.2) vs 53.6% (48.3–58.6), HR 0.74 (95% CI 0.59–0.92)
Adverse events
Grade ≥3 treatment-related (primary analysis): 44.9% (pembrolizumab) vs 37.3% (placebo)
Immune-mediated grade ≥3: 7% vs 2%
Treatment-related deaths: 1% vs 1%
HRQoL: no detriment with pembrolizumab vs placebo at longer follow-up
Immune-mediated grade ≥3: 7% vs 2%
Treatment-related deaths: 1% vs 1%
HRQoL: no detriment with pembrolizumab vs placebo at longer follow-up
Conclusions
At 5 years, perioperative pembrolizumab nearly doubled EFS and continued to improve OS versus neoadjuvant chemotherapy and surgery alone. The durable, clinically meaningful benefit supports perioperative pembrolizumab plus neoadjuvant chemotherapy as a standard of care for resectable early-stage NSCLC.
Key Limitations
The perioperative design does not isolate the contribution of the neoadjuvant versus adjuvant pembrolizumab components; not all randomized patients proceeded to surgery, and optimal patient selection (e.g., by PD-L1 or pathologic response) remains undefined; cross-trial comparison with neoadjuvant-only and other perioperative regimens is confounded by design differences.
Clinical Context
Perioperative pembrolizumab is FDA-approved (October 2023) and NCCN-preferred for resectable stage II–IIIB NSCLC. It competes with neoadjuvant nivolumab + chemotherapy (CheckMate 816), perioperative nivolumab (CheckMate 77T), and perioperative durvalumab (AEGEAN); the choice of perioperative versus neoadjuvant-only strategy remains an active clinical question.