2022-RA-1297-ESGO A subgroup analysis of response rate by patient characteristics in patients with endometrial cancer receiving monotherapy dostarlimab in the GARNET trial

Introduction/Background

Clinical characteristics of patients have demonstrated that there may be independent predictors of response to cancer drug therapies. In this analysis, we evaluated objective response rate (ORR) by subgroups of clinical characteristics in patients with advanced or recurrent endometrial cancer who were treated with the anti-PD-1 dostarlimab.

Methodology

GARNET is a multicentre, open-label, single-arm phase 1 study. Patients were assigned to cohort A1 (mismatch repair deficient [dMMR]/microsatellite instability-high [MSI-H EC]) or A2 (mismatch repair proficient [MMRp]/microsatellite stable [MSS] EC) based on immunohistochemistry assessment. Patients received 500 mg of dostarlimab IV every 3 weeks for 4 cycles, then 1000 mg every 6 weeks until disease progression, discontinuation, or withdrawal. Patient baseline demographics (age and BMI), histology, and prior lines of therapies were collected for enrolled patients. ORR by BICR per RECIST v1.1 for prior lines of therapy and histology were pre-specified exploratory subgroup analyses, whereas age and BMI were post hoc subgroup analyses.

Results

153 patients with dMMR/MSI-H and 161 patients with MMRp/MSS EC were enrolled and treated. The efficacy-evaluable population included 143 patients with dMMR/MSI-H EC and 156 patients with MMRp/MSS EC with measurable disease at baseline and the opportunity for at least 6 months of follow-up. ORR for each subgroup (age, BMI, prior lines of therapy, and histology) in each cohort were similar to that of the ORR for each overall cohort (see table 1). Overlapping 95% CIs are observed for all the subgroups assessed.

Conclusion

The treatment benefit of dostarlimab was consistent across clinical characteristic subgroups on a per-cohort basis (dMMR/MSI-H response rates were consistently ≥40%, whereas MMRp/MSS response rates were between 8% and 20%). No correlation could be made between response rate and individual clinical characteristics. Given the small sample size of the subgroups, caution should be used when interpreting the results.