Preliminary results of a phase II study of retifanlimab (PD-1 inhibitor) plus or minus epacadostat (IDO1 inhibitor) in combination with bevacizumab and hypofractionated radiotherapy for recurrent glioblastoma: NCT03532295.

2058 Background: Immunotherapies targeting the programmed cell death-1 (PD-1) pathway in recurrent glioblastoma (rGBM) have failed. We hypothesize that combining therapies targeting multiple immunosuppressive pathways with cytotoxic and antiangiogenic therapies will improve survival. Here, we evaluate the safety and efficacy of an anti-PD-1 monoclonal antibody (retifanlimab), hypofractionated radiotherapy (HFRT), and bevacizumab, with or without an oral IDO1 inhibitor (epacadostat), in patients with rGBM in a nonrandomized, noncomparative sequential two-arm Phase II study. Methods: This is an open-label Phase II study of 2 sequential cohorts (Table). Cohort A first examines retifanlimab + bevacizumab + HFRT in patients with IDH1/2-WT rGBM. After a toxicity monitoring period, Cohort B, which adds epacadostat, starts enrolling. Key inclusion criteria includes dexamethasone ≤ 4 mg/day at registration and candidacy for reirradiation. Previous bevacizumab use for radiation necrosis is permitted. The primary endpoint is OS probability at 9 months (OS-9). Secondary endpoints include PFS, OS, and toxicity. Exploratory endpoints include immunological phenotyping of blood and tissue. An increase of OS-9 from 38% (bevacizumab alone) to 60% is considered clinically relevant. Results: From 6/2020 to 12/2021, we have completed accrual for cohort A and the interim analysis results are presented here: 25 patients with rGBM enrolled, with 23 evaluable. Median age is 64.3 years (42.1-81.8), 30.4% female, 30.4% MGMT promotor methylated, median KPS 90 (range 70-100), baseline dexamethasone 0 mg (range 0-4) with 52.2% of patients on dexamethasone during the first 3 cycles, median baseline ALC 1,000 cells/µl (range 300-3,700). Patients received a median of 6 cycles to date (range 2-20). Median follow-up is 11.97 months per the reverse Kaplan-Meier method. Interim analysis shows a median PFS of 9.9 months (95%CI: 5.5 to not reached (NR)) and median OS of 12.2 months (95%CI 7.3-NR). Notably, Cohort A met the primary endpoint with an OS-9 of 71.4% (95%CI: 46.7% -86.1%). No dose limiting toxicities have been yet observed. Two treatment-related grade 3 toxicities have occurred (myositis, hypertension). Cohort B enrollment is ongoing, and correlative studies pending. Conclusions: Interim analysis suggests retifanlimab combined with HFRT and bevacizumab in patients with rGBM is well-tolerated and had encouraging OS and PFS at the time of data cutoff. Cohort B, which adds epacadostat, is currently enrolling. Clinical trial information: NCT03532295. [Table: see text]