Background: Adenosine, derived from ATP released by dying cancer cells in response to chemotherapy, binds to and activates the A_2a and A_2b receptors on immune cells, resulting in an ineffective anti-tumor immune response. Adenosine receptor blockade may therefore enhance the efficacy of chemotherapy when co-administered with a checkpoint inhibitor. AB928, the first clinical-stage small molecule dual A_2aR/ A_2bR antagonist, is highly potent and was well tolerated in dose escalation studies as a single agent or in combination with chemo/immunotherapy. Methods: ARC-4 (NCT03846310) is a Phase 1/1b, open-label study in participants (pts) with metastatic, locally advanced, or recurrent non-squamous NSCLC with no alternative or curative therapy option. Pts whose tumor has a genetic alteration for which targeted therapy exists must be chemo/immunotherapy naïve. Pts must have ECOG PS 0-1 and at least one measurable lesion. Two escalating doses of AB928 (75 or 150 mg) administered orally daily were given with standard doses of carboplatin, pemetrexed and pembrolizumab. Results: As of 27Dec19, 7 pts have received AB928 in Ph 1: 75 mg (n = 3), 150 mg (n = 4). Number of prior therapies range from 0 to 6 (median = 3). Most treatment emergent AEs (TEAEs) were Grade 1 or 2, with no Grade 5 events. The most common AEs were anemia, nausea and AST elevation. One pt experienced an SAE (Gr4 thrombocytopenia) that was at least possibly related to AB928. One pt discontinued before assessment due to adverse event (muscle weakness). Of the 6 patients with post-baseline disease assessments, all demonstrated decrease size of target lesion. Three patients (43%) achieved a confirmed partial response: one treatment naïve patient, one patient with EGFRmut disease that had previously progressed on treatment with erlotinib and osimertinib, and one patient who had progressed on therapy with ipilimumab and nivolumab. One additional pt had a > 30% decrease in target lesions and findings consistent with new lesions but opted to continue study treatment. Conclusions: Addition of AB928 to carboplatin, pemetrexed and pembrolizumab did not significantly add to the established safety profile of the standard agents. Combination treatment was associated with disease control in all evaluable pts, including responses in those with PD after TKI and immunotherapy. Expansion is on-going in pts with EGFRmut NSCLC that has failed treatment with TKI. Additional updates on the safety, clinical activity, and correlative biomarker results will be presented. Clinical trial information: 03846310.