Background: In the Phase 3, randomized, open-label CASPIAN study, first-line durvalumab (D) added to etoposide plus either cisplatin or carboplatin (EP) significantly improved OS vs EP alone (HR 0.73 [95% CI 0.59–0.91]; p = 0.0047) in pts with ES-SCLC at the planned interim analysis. Here we describe treatment patterns and outcomes for pts according to brain metastases. Methods: Treatment-naïve pts (WHO PS 0/1) with ES-SCLC received 4 cycles of D 1500 mg + EP q3w followed by maintenance D 1500 mg q4w until disease progression (PD) or up to 6 cycles of EP q3w and optional prophylactic cranial irradiation (PCI; investigator’s discretion). Pts with either asymptomatic or treated and stable brain metastases were eligible. Brain imaging was suggested for pts with suspected brain metastases, but was not mandated at screening or during treatment. The primary endpoint was OS. Analysis of OS and PFS in pt subgroups with and without brain metastases was prespecified. Other analyses in these subgroups were post hoc. Data cutoff: Mar 11, 2019. Results: At baseline, 28 (10.4%) of 268 pts in the D + EP arm and 27 (10.0%) of 269 pts in the EP arm had known brain metastases; of these, only 3 pts (~11% of those with baseline brain metastases) in each arm received radiotherapy (RT) to the brain prior to study entry. D + EP consistently improved OS vs EP in pts with or without known brain metastases at baseline (HR 0.69 [95% CI 0.35–1.31] and 0.74 [0.59–0.93], respectively); PFS was also consistently improved with D + EP regardless of the presence or not of baseline brain metastases (HR 0.73 [0.42–1.29] and 0.78 [0.64–0.95]). Among pts without known brain metastases at baseline, similar proportions developed new brain metastases at first PD in the D + EP (20/240; 8.3%) and EP arms (23/242; 9.5%), despite 19 (7.9%) pts in the EP arm having received PCI. Overall, 48 of 268 (17.9%) and 49 of 269 (18.2%) pts in the D + EP and EP arms received RT to the brain subsequent to study treatment; rates remained similar across the D + EP and EP arms regardless of baseline brain metastases (11 of 28 [39.3%] and 11 of 27 [40.7%] pts with known baseline brain metastases, compared to 37 of 240 [15.4%] and 38 of 242 [15.7%] pts without known baseline brain metastases). Conclusions: In CASPIAN, OS and PFS outcomes were improved with D + EP vs EP regardless of baseline brain metastases, consistent with the ITT analyses. Rates of new brain metastases at first PD were similar between arms, although PCI was permitted only in the control arm. Rates of subsequent RT to the brain were also similar in both arms. Clinical trial information: NCT03043872