ESMO Congress 2026

After nephrectomy, high-risk clear-cell renal-cell carcinoma went a long time without an adjuvant drug backed by randomised evidence. KEYNOTE-564 randomised 994 patients 1:1 to pembrolizumab (200 mg every 3 weeks, up to 17 cycles, about a year) or placebo. At the 2021 interim, median follow-up 24.1 months: 24-month disease-free survival 77.3% vs 68.1%, hazard ratio for recurrence or death 0.68 (0.53–0.87); 24-month overall survival 96.6% vs 93.5%.^[1] That is how adjuvant PD-1 got a positive phase 3.

Overall survival came later. The 2024 third interim, median follow-up 57.2 months: disease-free survival hazard ratio 0.72 (0.59–0.87), death hazard ratio 0.62 (0.44–0.87), 48-month overall survival 91.2% vs 86.0%.^[2] The 2026 five-year analysis, median follow-up 70 months: disease-free survival hazard ratio 0.71, 5-year disease-free survival 60.9% vs 52.2%; overall survival hazard ratio 0.66, 5-year overall survival 87.7% vs 82.3%.^[3] The disease-free survival hazard ratio sat between 0.68 and 0.72 across three looks. Other checkpoint inhibitors in the same window did not: CheckMate 914 part A, nivolumab plus ipilimumab versus placebo, 816 patients, disease-free survival hazard ratio 0.92 (0.71–1.19), p=0.53, grade 3–5 adverse events 38% vs 10%, four treatment-related deaths;^[4] IMmotion010, atezolizumab versus placebo, 778 patients, median disease-free survival 57.2 vs 49.5 months, hazard ratio 0.93 (0.75–1.15), p=0.50.^[5] Adjuvant kidney cancer is not ‘any immunotherapy’; the regimen that held is this pembrolizumab trial.

Clear-cell disease also has HIF-2α. LITESPARK-005 is advanced disease after immune-checkpoint and antiangiogenic therapy: belzutifan versus everolimus. Median progression-free survival 5.6 months in both arms, but 18-month progression-free survival 24.0% vs 8.3%, objective response 21.9% vs 3.5%; overall-survival hazard ratio 0.88, not significant.^[6] LITESPARK-022 added belzutifan to adjuvant pembrolizumab versus pembrolizumab plus placebo, 1,841 patients, median follow-up 28.4 months: disease-free survival hazard ratio 0.72 (0.59–0.87), 24-month disease-free survival 80.7% vs 73.7%; overall survival immature (hazard ratio 0.78); grade ≥3 adverse events 52.1% vs 30.2%.^[7] That question is pembrolizumab plus HIF-2α versus pembrolizumab, not versus surveillance.

This meeting’s KEYNOTE-564 tag is two stories. Gil Morgan, Natalia Gandur Quiroga and MeDucation are writing the KEYNOTE-564 final prespecified analysis: n=994, median follow-up 85.9 months; 7-year disease-free survival 57.2% vs 44.6% (hazard ratio 0.68), 7-year overall survival 82.3% vs 76.6% (hazard ratio 0.70); no new serious treatment-related safety signal. That is the same phase 3 stretched from 24 months to 7 years; the disease-free survival hazard ratio is still 0.68, the same point estimate as the 2021 interim. The other post is a reconstructed indirect comparison of KEYNOTE-564 and LITESPARK-022 (floor n=2,835): combination versus surveillance, disease-free survival hazard ratio 0.55 (0.45–0.69), 24-month disease-free survival 79.4% vs 67.8%, overall-survival hazard ratio 0.56. An ESMO Open reconstructed-IPD paper reports the same 0.55 and 0.56 and states that overall survival is immature and exchangeability is limited by race and incomplete PD-L1 reporting.^[8] 0.55 is not a randomised head-to-head of combination versus surveillance and does not rewrite adjuvant care; the 7-year analysis is the mature pembrolizumab-versus-placebo readout. The floor already frames the next question as selection: who must be treated, and who can be spared.