Predictors of overall survival (OS) in patients (pts) with melanoma brain metastasis (MBM) in the modern era.

9540 Background: The management and OS of pts with metastatic melanoma have improved due to new systemic therapies. However, relatively little is known about the use of these treatments (tx) and their association with OS in pts with MBMs. We reviewed a large cohort of MBM pts to assess how pt demographics, disease characteristics, and MBM tx impact OS in the current era. Methods: Under an institutional review board-approved protocol, retrospective data were curated and analyzed from pts diagnosed with, and received tx for, MBM from 2014 to 2018 at the MD Anderson Cancer Center (MDA). Pts diagnosed with uveal or mucosal melanoma or other cancers were excluded. Pt demographics; timing and features of initial melanoma dx; timing and features of initial MBM dx; prior, initial and subsequent tx; and OS were collected. OS was determined from MBM dx to last clinical follow-up (FU). Pts alive at last FU were censored. The Kaplan-Meier method and log-rank test were used to estimate OS and to assess univariate group differences, respectively. Multivariable (MV) associations of OS with variables of interest were investigated with Cox proportional hazards models. Initial treatment of MBM was assessed as a time-varying covariate. All statistical tests used a significance level of 5%. Results: A total of 401 MBM pts were identified. The median age at MBM dx was 61; 67% were male and 46% had a BRAF V600 mutation. At MBM diagnosis dx, most (70%) pts were asymptomatic; 70% had concurrent uncontrolled extracranial disease; 36% had elevated serum LDH. Prior tx included immunotherapy (IMT) for 39% and targeted therapy (TTX) for 17%. The median number of MBMs was 2; 31% had > 3 MBMs. Median largest MBM diameter was 1.0 cm, 9% had MBM > 3.0 cm, and 5% had concurrent leptomeningeal disease (LMD). Tx received after MBM dx included stereotactic radiosurgery (SRS; 53% as initial tx for MBM, 67% at any time after MBM dx), whole brain radiation therapy (WBRT; 16%, 35%), craniotomy (12%, 19%), IMT (37%, 74%), and/or TTX (22%, 40%). 31% received steroids during initial MBM tx. At a median FU of 13.4 (0.0 - 82.8) months (mos), the median OS was 15.1 mos, and 1- and 2-year OS rates were 56% and 40%. Notably, gender, time to MBM dx, and BRAF status were not associated with OS (univariate analysis). On MV analysis, clinical features associated with worse OS included increased age, increased primary tumor thickness, elevated LDH, > 3 MBMs, +LMD, +symptoms, and prior tx with IMT. Among tx used at any time after MBM dx, WBRT (HR 1.9, 95% CI 1.5-2.5) was associated with worse OS; SRS (HR 0.7, 95% CI 0.5-0.8) and IMT (HR 0.6, 95% CI 0.5-0.8) were associated with improved OS. Conclusions: In one of the largest cohorts of MBM pts described to date, OS has improved in MBM pts in the current era. Prognostic factors for OS include pt age, primary tumor and MBM features, prior tx, and tx for MBM. Additional analyses to assess the interaction of tx, disease features, and OS will be presented.