Efficacy, safety and biomarker analysis of durvalumab in patients with mismatch-repair deficient or microsatellite instability-high solid tumours

BackgroundIn this study we aimed to evaluate the efficacy and safety of the PD-L1 inhibitor durvalumab across various mismatch repair deficient (dMMR) or microsatellite instability-high (MSI-H) tumours in the Drug Rediscovery Protocol (DRUP). This is a clinical study in which patients are treated with drugs outside their labeled indication, based on their tumour molecular profile.Patients and methodsPatients with dMMR/MSI-H solid tumours who had exhausted all standard of care options were eligible. Patients were treated with durvalumab. The primary endpoints were clinical benefit ((CB): objective response (OR) or stable disease >= 16 weeks) and safety. Patients were enrolled using a Simon like 2-stage model, with 8 patients in stage 1, up to 24 patients in stage 2 if at least 1/8 patients had CB in stage 1. At baseline, fresh frozen biopsies were obtained for biomarker analyses.ResultsTwenty-six patients with 10 different cancer types were included. Two patients (2/26, 8%) were considered as non-evaluable for the primary endpoint. CB was observed in 13 patients (13/26, 50%) with an OR in 7 patients (7/26, 27%). The remaining 11 patients (11/26, 42%) had progressive disease. Median progression-free survival and median overall survival were 5 months (95% CI, 2-not reached) and 14 months (95% CI, 5-not reached), respectively. No unexpected toxicity was observed. We found a significantly higher structural variant (SV) burden in patients without CB. Additionally, we observed a significant enrichment of JAK1 frameshift mutations and a significantly lower IFN-gamma expression in patients without CB.ConclusionDurvalumab was generally well-tolerated and provided durable responses in pre-treated patients with dMMR/MSI-H solid tumours. High SV burden, JAK1 frameshift mutations and low IFN-gamma expression were associated with a lack of CB; this provides a rationale for larger studies to validate these findings.