A randomized phase 2 study of durvalumab monotherapy and in combination with tremelimumab in patients with metastatic pancreatic ductal adenocarcinoma (mPDAC): ALPS study.

Abstract

217 Background: Durvalumab (D) and tremelimumab (T), monoclonal antibodies against PD-L1 and CTLA-4 immune checkpoints, have shown efficacy as monotherapy/combination therapy in multiple cancer types. Herein, we report a randomized phase 2 study to evaluate efficacy and safety of D monotherapy with or without T (D+T) in previously treated mPDAC. Methods: Part A was a lead-in safety and signal-seeking study with plans to expand to Part B as a nonrandomized or randomized controlled study pending efficacy signal. Eligible pts had progressive disease (PD) following front-line 5-FU- or gemcitabine-based therapy. In Part A, pts were randomized to D (1.5 g IV Q4W) or D+T (D 1.5 g IV + T 75 mg IV Q4W × 4 doses → D 1.5 g IV Q4W) for up to 12 months (mo) or until confirmed PD or unacceptable toxicity. Primary endpoint was investigator-assessed objective response rate per RECIST 1.1. Results: In Part A, 65 pts were randomized to D (n = 33) or D+T (n = 32). Due to a pretreatment death, 64 pts received therapy. Eleven pts (34.4%) in D+T and 10 (31.3%) in D had treatment-related adverse events (trAEs); 7 (22%) in D+T and 2 (6%) in D had grade ≥3 trAEs. Common trAEs: fatigue (12.5%), diarrhea (12.5%), and hypothyroidism (9.4%) in D+T; fatigue (9.4%), diarrhea (6.3%), and pruritus (6.3%) in D. Grade ≥3 trAEs were diarrhea (9.4%), fatigue (6.3%) in D+T and ascites (3.1%), hepatitis (3.1%), and increased lipase (3.1%) in D. In D+T, 3 (9.4%) pts and 1 (3.1%) in D discontinued therapy due to trAEs. No trAEs resulted in death. In D+T, 1 (3.1%) pt had a durable confirmed partial response (PR) > 12 mo and disease control rate (DCR) was 9.4%. In D, 2 (6.1%) pts had unconfirmed PRs, and the DCR was 6.1%. Median PFS in both arms was 1.5 mo, and median OS was 3.1 mo in D+T and 3.6 mo in D. Biomarker analyses (tumor mutation burden, PD-L1 and microsatellite status) are currently being evaluated. Conclusions: Typical safety profiles were observed for D or D+T in mPDAC. Part B was not enrolled as the threshold for efficacy was not met in Part A. D and D+T had modest activity in second-line non-selected mPDAC, underpinning the need for multimodal immune-based combinations to overcome intrinsic resistance in this disease. Clinical trial information: NCT02558894.