ESMO Congress 2026

Pembrolizumab's position in dMMR/MSI-H endometrial cancer was previously built on later lines: KEYNOTE-158 underpinned its approval in MSI-H endometrial cancer relapsing after at least one line of chemotherapy ^[7]; single-agent immunotherapy shows activity in the MSI-H recurrent setting, but the literature notes over 50% of patients do not respond ^[7][10]. KEYNOTE-C93/ENGOT-en15/GOG-3064 was designed precisely to move this option earlier, to first line and even chemotherapy-free ^[7]. US real-world retrospective data (n=228; first-line pembrolizumab monotherapy 80 vs carboplatin-paclitaxel 148) showed median rwTOT of 9.1 vs 3.5 months and rwTTNT of 16.6 months for pembrolizumab ^[8], suggesting feasibility but lacking randomized evidence.

What was actually reported at ESMO26 is LBA12 KEYNOTE-C93: a phase 3, randomized, open-label study, n=299, first-line pembrolizumab vs carboplatin + paclitaxel in dMMR advanced or recurrent endometrial carcinoma; primary endpoints are PFS by BICR and OS; median follow-up 36.5 months. Numbers from the live posts: median PFS 38.2 vs 8.4 months (HR 0.57); grade ≥3 TRAEs 18.1% vs 50.7%; median response duration NR vs 10.2 months; OS remains immature. Notably, there is no independent main-site news report for this topic so far; the above readings come only from live posts, and full abstract and subgroup details remain to be confirmed.

Relative to prior knowledge, C93 upgrades first-line immunotherapy monotherapy — previously supported only by later-line evidence ^[7] and real-world hints ^[8] — into randomized phase 3 controlled results: median PFS 38.2 vs 8.4 months, grade ≥3 TRAEs 18.1% vs 50.7%, and the NR median response duration suggests durability of benefit. This provides the most direct controlled evidence to date for chemotherapy-free first-line treatment in the dMMR population, echoing the study's design intent of moving immunotherapy earlier ^[7] and filling the causal-efficacy question real-world studies could not answer ^[8].

As for limitations, OS data remain immature and the survival benefit awaits longer follow-up. The literature also indicates that over 50% of MSI-H patients do not respond to single-agent immunotherapy ^[7], and MMR/MSI testing carries a discordance rate of about 3.8%, with subclonal or heterogeneous MMR loss showing a higher recurrence rate (38.9%) ^[9], which has practical implications for patient selection and biomarker definition. Subgroup analyses, biomarker details, and the impact of prior platinum adjuvant therapy are not disclosed in the available materials.