Results from the phase 3 KEYNOTE-975 trial, presented by Manish Shah at ESMO 2026 during the Proffered Paper Session 1: Gastrointestinal Tumours, Upper Digestive, showed that adding pembrolizumab to definitive chemoradiotherapy (dCRT) did not significantly improve event-free survival in patients with previously untreated, unresectable esophageal or gastroesophageal junction (GEJ) cancer with PD-L1 combined positive score (CPS) ≥10.
The findings, reported in the trial’s second interim analysis, highlight the ongoing challenges of integrating immunotherapy into definitive chemoradiotherapy for locally advanced esophageal cancer.
Background
Definitive chemoradiotherapy is an established treatment option for patients with locally advanced esophageal cancer who are not candidates for curative surgery. However, the benefit of adding immune checkpoint inhibitors to this approach remains uncertain.
KEYNOTE-975 (NCT04210115) was designed to determine whether pembrolizumab could improve clinical outcomes when combined with dCRT in patients with unresectable esophageal squamous cell carcinoma (ESCC), esophageal adenocarcinoma (EAC), or GEJ cancer.
Study Design and Patient Population
The international, randomized, double-blind, placebo-controlled phase 3 trial enrolled 703 patients, who were assigned 1:1 to receive:
- Pembrolizumab plus dCRT: 349 patients
- Placebo plus dCRT: 354 patients
Eligible patients had previously untreated, locally advanced esophageal or GEJ cancer, an ECOG performance status of 0–1, and were ineligible for curative surgery but suitable for definitive chemoradiotherapy.
Pembrolizumab was administered at 200 mg every three weeks for eight cycles, followed by 400 mg every six weeks for five additional cycles. The control group received a matching placebo.
Chemoradiotherapy consisted of investigator-selected fluorouracil plus cisplatin with 50 or 60 Gy of radiotherapy, or FOLFOX with 50 Gy. The dual primary endpoints were investigator-assessed event-free survival (EFS) and overall survival (OS), evaluated in patients with PD-L1 CPS ≥10, CPS ≥1, and the overall study population.
Efficacy Results
At the January 21, 2026, data cutoff, the median follow-up was 47.1 months.
Among patients with PD-L1 CPS ≥10, adding pembrolizumab did not significantly improve event-free survival compared with placebo:
- Median EFS: 15.2 versus 14.7 months
- Hazard ratio: 0.97 (95% CI, 0.77–1.23)
- P = 0.3994
Overall survival results in the same subgroup also showed no apparent benefit:
- Median OS: 32.2 versus 35.0 months
- Hazard ratio: 1.00 (95% CI, 0.77–1.30)
- Nominal P = 0.5003
Because the prespecified EFS endpoint in patients with PD-L1 CPS ≥10 was not met, subsequent formal statistical testing of EFS in the CPS ≥1 and overall populations, as well as OS, was not performed.
Following these findings, the study was stopped.
Safety Findings
Grade 3 or higher adverse events were reported in 100% of patients in both treatment groups. Fatal adverse events occurred in 44 patients (13%) receiving pembrolizumab plus dCRT and 37 patients (10%) receiving placebo plus dCRT.
Treatment discontinuation due to adverse events was more frequent in the pembrolizumab group, occurring in 34% versus 27% of patients. Investigators reported no new safety signals, although the findings demonstrate the substantial toxicity burden associated with these treatment regimens.
Clinical Implications
KEYNOTE-975 is an important negative phase 3 trial in the development of immunotherapy-based treatment strategies for unresectable esophageal and GEJ cancer.
Despite the established benefits of immune checkpoint inhibitors in other treatment settings, adding pembrolizumab to definitive chemoradiotherapy did not improve the primary EFS endpoint in patients with PD-L1 CPS ≥10.
The findings do not support the routine addition of pembrolizumab to dCRT in this population based on the KEYNOTE-975 regimen.
Further research is needed to identify patients who may benefit from immunotherapy combined with radiation, explore alternative treatment sequencing, and develop more effective strategies for improving outcomes in unresectable esophageal cancer.
Conclusion
The phase 3 KEYNOTE-975 trial failed to demonstrate a significant event-free survival benefit from adding pembrolizumab to definitive chemoradiotherapy in unresectable esophageal or GEJ cancer with PD-L1 CPS ≥10.
With no apparent overall survival improvement in the same subgroup and higher rates of treatment discontinuation, the findings underscore the need for more effective and better-tolerated treatment approaches in this setting.
