Abstract 5711: The impact of combination oral azacitidine (CC-486) + pembrolizumab (PEMBRO) on the immune infiltrate in metastatic melanoma (MM)

Introduction: Although immune checkpoint inhibitors have improved survival for many with MM, most pts do not respond and median PFS is only 6 months for single agent therapy. Cancers subvert the cellular epigenetic machinery to facilitate immune escape. Epigenetic mechanisms of resistance are potentially reversible by DNA hypomethylating agents (HMA). Based on preclinical models, we hypothesized that CC-486, an oral HMA, will enhance response to PEMBRO in PD-1 naïve pts and reverse resistance in anti-PD-1 refractory pts. The aim of this study is to determine the safety, efficacy, and characterize the pharmacodynamic impact of CC-486 + PEMBRO on immune infiltrates in pts with MM. Experimental: NCT02816021 is an ongoing phase II study of CC-486 + PEMBRO in MM pts who are PD-1 naïve (Arm A) or who have progressed on prior PD-1 directed therapy (Arm B). Pts receive 300mg PO of CC-486 on days 1-14 and 200mg IV of PEMBRO q3 weeks. Serum and tumor biopsies are obtained at baseline, prior to cycles 3 and 5. Immune monitoring studies were performed by the Immunotherapy Platform at MD Anderson. Immune cell phenotyping by CyTOF was performed using 36 metal-conjugated antibodies targeting myeloid and T cell surface markers. FCS files were exported and manually gated for lymphocytes using FlowJo (version 10.1) and subjected to multidimensional phenographic analysis. Results: Thirteen pts (Arm A, n=6; Arm B, n=7) have been enrolled. Two of the 3 pts remaining on study are PD-1 naïve and have received 13 and 8 cycles of CC-496 + PEMBRO with partial responses by RECIST1.1 at 6 months, respectively. One pt on Arm B remains on study with stable disease after 4 cycles. The combination was considered safe after a run-in phase (6 pts/arm treated without DLTs) and the study is open to full accrual. The most common grade 3/4 AEs were leukopenia , neutropenia, vomiting, and diarrhea (2 each) with 1 grade 5 AE unrelated to treatment (hepatic rupture/bleeding due to progressive disease). Serum and tumor biopsies from 6 pts (3 per Arm) were available for interim analysis, with additional samples in process. Of these 6 pts, 1 pt (PD-1 naïve) responded to therapy by RECIST1.1. We evaluated the blood and tumor samples by CyTOF. In the tumor samples, we observed an increase in frequency of T cells in 2 PD-1 naïve patients but did not observe similar changes in PD-1 refractory patients in this small cohort. Similar data was found with immunohistochemistry. These changes were not observed