1135P Comparison of intracranial (IC) response assessment criteria in patients (pts) with melanoma brain metastases (MBM) treated with combination nivolumab (NIVO) plus ipilimumab (IPI) in CheckMate 204

In CheckMate 204, NIVO + IPI showed high IC mRECIST objective response rates (ORRs) in pts with asymptomatic, unirradiated MBM and a lower but durable response in pts with symptomatic and/or steroid-requiring MBM. Response as a surrogate for progression-free survival (PFS)/overall survival (OS) has prompted the use of various response assessment criteria and cutoffs for target lesion size in MBM. In this exploratory analysis, ORRs and correlation of response to survival were examined by mRECIST, RANO-BM, RECIST, and volumetric response. Pts with metastatic melanoma and ≥ 1 unirradiated MBM (diameter, 0.5–3 cm) received NIVO 1 mg/kg + IPI 3 mg/kg Q3W × 4, followed by NIVO 3 mg/kg Q2W for ≤ 24 mo. IC ORRs were assessed using mRECIST (5 mm target lesion cutoff), RECIST (10 mm), RANO-BM (5 or 10 mm), and volumetric response (5 or 10 mm), per blinded review. IC PFS and OS using a 6-wk landmark were compared for responders vs nonresponders. IC ORR was numerically higher with mRECIST or volumetric assessment compared with RANO-BM or RECIST (Table). Responder vs nonresponder PFS and OS were significantly better across the different assessment criteria; mRECIST and volumetric response showed the strongest correlations (Table). mRECIST responders who were not RANO-BM 5 mm responders (n = 14) had similar OS to RANO-BM 5 mm responders. Among 41 pts with only target lesions < 10 mm, mRECIST ORR, and OS for the responders, was similar to the overall CheckMate 204 ITT population. This analysis supports mRECIST as a reliable assessment scale by showing strong differentiation of long-term benefit for responders vs nonresponders. Volumetric response also correlated with PFS/OS supporting future trial exploratory use. Responders with only MBM < 10 mm derived comparable PFS/OS benefit to ITT, supporting the inclusion of pts with smaller MBM in future trials.Table: 1135PAll pts (n = 119)mRECIST 5 mmRECIST10 mmRANO-BM5 mmRANO-BM10 mmVolumetricb5 mmVolumetricb10 mmORR, % (95% CI)45 (36–55)27 (19–36)34 (26–44)26 (18–35)40 (31–50)39 (31–49)PFS HR,a (95% CI)0.06 (0.02–0.16)0.18 (0.06–0.56)0.13 (0.06–0.29)0.25 (0.11–0.59)0.04 (0.01–0.11)0.07 (0.03–0.19)OS HR,a (95% CI)0.18 (0.07–0.45)0.22 (0.07–0.75)0.26 (0.10–0.71)0.34 (0.12–1.00)0.10 (0.03–0.35)0.11 (0.03–0.36)aResponder vs nonresponder; Cox proportional hazard model was used for hazard ratio (HR) calculation. bVolumetric response was defined as a 65% decrease in target lesions and progression as a 73% increase. Open table in a new tab