2531 Background: Nivolumab (nivo) and ipilimumab (ipi) were found in the recent phase III CheckMate 227 trial to have an overall survival benefit over chemotherapy for advanced non-small cell lung cancer (NSCLC). However, patients (pts) with untreated brain metastasis (mets) were excluded from that trial. Because 30% of NSCLC pts develop brain mets, we tested nivo/ipi with concurrent stereotactic radiosurgery (SRS) for NSCLC pts with active brain mets. Methods: We report the safety data from the phase I portion of an ongoing phase I/II single-institution trial in which one treatment group was given SRS with nivo (3 mg/kg) every 2 weeks (wks) plus ipi (1 mg/kg) every 6 wks x 4 cycles, followed by maintenance nivo (480 mg) every 4 wks until disease progression, unacceptable toxicity, or withdrawal of consent. Brain SRS was delivered within 7 days of initiation of nivo/ipi. The primary endpoints were safety and 4-month (mo) intracranial progression-free survival (PFS). Dose-limiting toxicity (DLT) was defined as > 15% intracranial toxicity (>G3 hypophysitis / neurologic toxicity) or > 30% extracranial toxicity (>G3 non-dermatologic non-lab toxicity or >G4 dermatologic / lab toxicity), refractory to medical management, assessed at 8 wks after treatment initiation. Target accrual for phase I was 10 evaluable pts, with enrollment suspended after every 5 pts for DLT assessment. Results: Since June 15, 2018, 13 pts were enrolled and 10 were evaluable for DLT. The median follow-up time was 6.8 mo (range 1.2‒18). As of January 6, 2020, only 1 pt had DLT-defined toxicity and thus the predefined stopping criteria were not met. This pt had G3 seizure right after SRS that resolved within a week, and then had increased but asymptomatic CNS edema 4 wks later. Aside from DLTs, 3 pts (25%) developed treatment-related G3 (elevated liver function tests, fatigue, nausea, adrenal insufficiency, and myocarditis) or G4 events (pneumonitis/acute respiratory distress syndrome in 1 pt with confirmed influenza at 7 mos after treatment initiation). This pt subsequently died of hemophagocytic lymphohistiocytosis (considered possibly related to the study drugs). Median intracranial PFS time was 9.7 mo, and the 4-mo intracranial PFS rate was 75%. Extracranial objective response rate was 33% in the 12 evaluable pts with a median response duration of 9.1 mo. Conclusions: Concurrent SRS withnivo/ipi was safe for pts with active NSCLC brain mets. Preliminary analyses of efficacy were encouraging for durable intracranial and extracranial response. Clinical trial information: NCT02696993 .