KEYNOTE-C93/GOG-3064/ENGOT-en15: Phase 3, randomized, open-label study of first-line (1L) pembrolizumab (pembro) vs carboplatin (C) + paclitaxel (P) in mismatch repair-deficient (dMMR) advanced or recurrent endometrial carcinoma (a/rEC)

KEYNOTE-C93

Patient / Population Intervention / Exposition Comparaison Critère de jugement

In patients with dMMR advanced or recurrent endometrial carcinoma, first-line pembrolizumab improved progression-free survival compared with carboplatin plus paclitaxel.

N

299patients

Schéma

Phase 3, randomised 1:1, open-label trial in patients with inoperable stage III/IV advanced or recurrent endometrial carcinoma with centrally confirmed dMMR and no prior systemic therapy for advanced disease, comparing first-line pembrolizumab with carboplatin plus paclitaxel (NCT05173987).

Critère

Dual primary endpoints: progression-free survival by blinded independent central review and overall survival.

RésultatMedian PFS by BICR 38.2 months (95% CI 17.9-NR) with pembrolizumab vs 8.4 months (95% CI 6.8-10.4) with carboplatin plus paclitaxel, HR 0.57 (95% CI 0.42-0.79; one-sided P = 0.0002). 36-month PFS rate 52.8% vs 30.0%. Overall survival at the interim analysis: median not reached in either arm, HR 0.77 (95% CI 0.52-1.15); 36-month OS rate 70.7% vs 62.0%. ORR 64.1% vs 62.1%, complete response 33.8% vs 26.2%; median duration of response not reached vs 10.2 months. Between-group difference in GHS/QoL at week 12 of 7.02 (95% CI 2.54-11.50). Grade ≥3 treatment-related adverse events 18.1% vs 50.7%, most commonly anaemia 2.0% vs 18.1%.

Slomovitz BM, Backes F, Mackowiak-Matejczyk B, et al. Presented at ESMO Congress 2026, abstract LBA12, 26 October. Not yet published.

Discussion et critique

At a median follow-up of 36.5 months, chemotherapy-free pembrolizumab clearly prolonged PFS assessed by blinded central review, had markedly less grade ≥3 treatment-related toxicity and improved quality of life. Overall survival is still immature at this interim analysis, with a hazard ratio whose confidence interval crosses 1. Interpretation is also confounded because 85.2% of patients with progression on chemotherapy went on to receive PD-(L)1 inhibitors. Response rates were similar between arms, so the PFS gain reflects durability of response rather than more responses, and the open-label design weakens the quality-of-life comparison. The comparator was chemotherapy alone, not the chemotherapy plus immune checkpoint inhibitor regimen the abstract itself names as preferred, so the trial does not show how pembrolizumab alone compares with that standard.