Background: Pembro is a potent anti–PD-1 monoclonal antibody. IPI, an anti–CTLA-4 antibody, has shown activity in advanced NSCLC. In melanoma, combined anti–PD-1 and anti–CTLA-4 treatment has shown robust efficacy and manageable toxicity. We report interim results from a phase 1 study evaluating pembro + IPI in patients (pts) with recurrent NSCLC. Methods: Pts with stage IIIB/IV NSCLC that recurred after ≤ 2 prior regimens received pembro + IPI every 3 wk for 4 cycles followed by maintenance pembro. Based on emerging data from the nivolumab + IPI advanced NSCLC study, doses were reduced from 10 mg/kg to 2 mg/kg for pembro and from 3 mg/kg to 1 mg/kg for IPI). Primary end point was safety and incidence of dose-limiting toxicities (DLTs) in the first 3 wk of dosing. Response was assessed every 6 wk per RECIST 1.1 by investigator review. Results: As of Dec 2014,17 pts were enrolled: 3 at pembro 10 mg/kg + IPI 3 mg/kg, 3 at pembro 10 mg/kg + IPI 1 mg/kg, and 11 at pembro 2 mg/kg + IPI 1 mg/kg. No DLTs or dose modifications were reported for the 15 pts treated at the time of analysis. 10 pts experienced drug-related AEs (DRAEs); none led to discontinuation or death. There were 2 gr 3 DRAEs, both rash. Gr 2 DRAEs were diarrhea and vomiting (n = 2 each) and chills, cough, decreased appetite, decreased weight, dehydration, depression, dysphonia, fatigue, myalgia, pruritus, and pyrexia (n = 1 each). Responses were seen in all dose groups among the 11 pts on treatment for ≥ 6 wk at the time of analysis, including 1 CR (9%) and 5 PRs (45%) (Table); all pts achieved disease control. 12 pts remain on treatment (range, 6 + to 26 + wk); 3 pts discontinued for PD. Conclusions: Preliminary data from KEYNOTE-021 cohort D demonstrate an acceptable toxicity profile and robust antitumor activity for pembro + IPI in pts with recurrent NSCLC.The use of lower pembro and IPI doses did not appear to negatively impact efficacy. Clinical trial information: NCT02039674

^aRECIST v1.1 confirmed + unconfirmed.