Outcomes for Dostarlimab and Real-World Treatments in Post-platinum Patients With Advanced/Recurrent Endometrial Cancer: The GARNET Trial Versus a US Electronic Health Record-Based External Control Arm.

Background: Patients with advanced or recurrent endometrial cancer (EC) have limited treatment options following platinum-based chemotherapy and poor prognosis. The single-arm, Phase I GARNET trial (NCT02715284) previously reported dostarlimab efficacy in mismatch repair-deficient/microsatellite instability-high advanced or recurrent EC. Objectives: The objective of this study was to compare overall survival (OS) and describe time to treatment discontinuation (TTD) for dostarlimab (GARNET Cohort A1 safety population) with an equivalent real-world external control arm receiving non-anti-programmed death (PD)-1/PD-ligand (L)1/2 treatments (constructed using data from a nationwide electronic health record-derived de-identified database and applied GARNET eligibility criteria). Methods: Propensity scores constructed from prognostic factors, identified by literature review and clinical experts, were used for inverse probability of treatment weighting (IPTW). Kaplan-Meier curves were constructed and OS/TTD was estimated (Cox regression model was used to estimate the OS-adjusted hazard ratio). Results: Dostarlimab was associated with a 52% lower risk of death vs real-world treatments (hazard ratio, 0.48; 95% confidence interval [CI], 0.35-0.66). IPTW-adjusted median OS for dostarlimab (N=143) was not estimable (95% CI, 19.4-not estimable) versus 13.1 months (95% CI, 8.3-15.9) for real-world treatments (N = 185). Median TTD was 11.7 months (95% CI, 6.0-38.7) for dostarlimab and 5.3 months (95% CI, 4.1-6.0) for the real-world cohort. Discussion: Consistent with previous analyses, patients treated with dostarlimab had significantly longer OS than patients in the US real-world cohort after adjusting for the lack of randomization using stabilized IPTW. Additionally, patients had a long TTD when treated with dostarlimab, suggesting a favorable tolerability profile. Conclusion: Patients with advanced or recurrent EC receiving dostarlimab in GARNET had significantly lower risk of death than those receiving real-world non-anti-PD-(L)1/2 treatments.