Epacadostat plus pembrolizumab versus placebo plus pembrolizumab for advanced urothelial carcinoma: results from the randomized phase III ECHO-303/KEYNOTE-698 study

Abstract

Abstract Background Indoleamine 2,3-dioxygenase 1 (IDO1) levels correlate with poor outcomes in urothelial carcinoma (UC). IDO1 and programmed death-ligand 1 (PD-L1) are often co-expressed. Epacadostat is a potent and highly selective inhibitor of IDO1. In a subgroup analysis of patients with advanced UC participating in a phase I/II study, epacadostat-pembrolizumab treatment produced an objective response rate (ORR) of 35%. Methods ECHO-303/KEYNOTE-698 was a double-blinded, randomized phase III study of adults with metastatic or unresectable locally advanced UC with recurrence or progression following first-line platinum-based chemotherapy. Participants were randomized to epacadostat 100 mg twice daily (BID) plus pembrolizumab or placebo plus pembrolizumab until completion of 35 pembrolizumab infusions, disease progression, or unacceptable toxicity. The primary endpoint was investigator-assessed ORR per Response Evaluation Criteria in Solid Tumors version 1.1. Results Target enrollment was 648 patients; enrollment was halted early based on efficacy results from the phase III ECHO-301/KEYNOTE-252 study in metastatic melanoma. Forty-two patients were randomized to each treatment arm. Median duration of follow-up was 62 days in each arm. The investigator-assessed ORR (unconfirmed) was 26.2% (95% CI 16.35–48.11) for epacadostat plus pembrolizumab and 11.9% (95% CI 4.67–29.50) for placebo plus pembrolizumab. Two complete responses were reported, both in the placebo-plus-pembrolizumab arm. Circulating kynurenine levels increased from C1D1 to C2D1 in the placebo-plus-pembrolizumab arm and numerically decreased in the epacadostat-plus-pembrolizumab arm. The safety profile of epacadostat plus pembrolizumab was similar to that of pembrolizumab monotherapy, although a numerically greater proportion of patients in the combination vs. control arm experienced treatment-related grade ≥ 3 adverse events (16.7% vs. 7.3%). One patient in each arm died due to cardiovascular events, which were not deemed drug-related. No new safety concerns were identified for either agent. Conclusions Epacadostat plus pembrolizumab demonstrated anti-tumor activity and was generally tolerable as second-line treatment of patients with unresectable locally advanced or recurrent/progressive metastatic UC. Epacadostat 100 mg BID, when administered with pembrolizumab, did not normalize circulating kynurenine in most patients. Further study of combined IDO1/PD-L1 inhibition in this patient population, particularly with epacadostat doses that result in durable normalization of circulating kynurenine, may be warranted. Trial registration ClinicalTrials.gov, NCT03374488. Registered 12/15/2017.