Comparison of survival outcomes between dostarlimab and comparator treatments (tx) in patients (pts) with advanced/recurrent (A/R) endometrial cancer (EC) in England: Matching-adjusted indirect comparisons (MAICs).

e17534 Background: Limited tx options exist for pts with A/R EC who progress on/after platinum-based chemotherapy (PBCT). This study compared survival outcomes of pts with A/R mismatch repair deficient (dMMR)/microsatellite instability-high (MSI-H) EC receiving dostarlimab in Cohort A1 of the single-arm, Phase I GARNET trial, with an equivalent real-world (RW) cohort receiving current tx in England. The primary objective compared second-line (2L) overall survival (OS) between cohorts; a secondary objective was time to tx discontinuation (TTD; proxy for progression-free survival). Methods: An overall-2L-tx RW cohort was established using National Cancer Registration and Analysis Service data. Five tx-specific RW cohorts were identified (tx received at 2L in > 5% of pts) (Table). Clinical outcomes between GARNET and RW cohorts were compared using MAICs. Relevant prognostic variables were used to create 3 matching scenarios (SC): SC1: grade, histology, PBCT number; SC2: histology, PBCT number; SC3: race/ethnicity, stage at diagnosis, histology, prior surgery (MMR/MSI status not considered). OS was time from first dostarlimab dose (GARNET) or 2L RW tx (RW cohorts) to any-cause death; MAIC-adjusted hazard ratios (HR, 95% confidence interval [CI]) were estimated by weighted Cox proportional-hazards models. TTD after MAIC was summarized descriptively. Results: Effective sample size (ESS) assessed matching capacity; lower ESS in SC1 reduced its robustness (Table). Following matching, distributions of baseline characteristics were similar between cohorts. Across most matching scenarios, there was a lower risk of death in pts treated with dostarlimab compared to the overall RW cohort and for tx-specific RW cohorts (Table). Median TTD (months, 95% CI) was longer for pts treated with dostarlimab across all matching scenarios (SC1: 7.8 [4.1, not estimable (NE)]; SC2: 9.7 [5.5, NE]; SC3: 14.0 [6.4, 17.9]) than the overall RW cohort (3.4 [3.2, 3.4]). Conclusions: Results suggest that pts in GARNET with dMMR/MSI-H A/R EC receiving dostarlimab had better OS and longer TTD compared with pts receiving current 2L tx in England. Funding: GSK (217216; scott.r.goulden@gsk.com).[Table: see text]