OUP accepted manuscript

We sought to identify variates correlating with overall survival (OS) in patients treated with surgery (S) plus adjuvant stereotactic radiosurgery (SRS) versus definitive SRS for large (>4cc) brain metastases (BrM).We used univariate (UVA) and multivariate analyses (MVA) to identify survival correlates among eligible patients identified from a prospective registry and compared definitive SRS to S+ adjuvant SRS cohorts using propensity score-matched analysis (PSMA). Secondary outcomes were measured using the cumulative incidence (CI) method.We identified 364 patients; 127 and 237 were treated with S+SRS and definitive SRS, respectively. On UVA, SRS alone [HR1.73 (1.35,2.22) P<0.001), BrM quantity [HR 1.13 (1.06-1.22) (P<0.001)]; performance status (PS) [HR 2.78 (1.73-4.46) (P<0.001)]; extracranial disease (ECD) [HR 1.82 (1.37,2.40) (P<0.001)]; and receipt of systemic treatment after BrM therapy, [HR 0.58 (0.46-073) (P<0.001)] correlated with OS. On MVA, SRS alone [HR 1.81 (1.19,2.74) (P<0.0054)], SRS target volume [HR 1.03 (1.01,1.06) (P<0.0042)], and receipt of systemic treatment [HR 0.68 (0.50,0.93) (P<0.015)] correlated with OS. When PSMA was used to balance ECD, BrM quantity, PS, and SRS target volume, SRS alone remained correlated with worsened OS [HR 1.62 (1.20-2.19) (P=0.0015)]. CI of local failure requiring resection at 12 months was 3% versus 7% for S+SRS and SRS cohorts, respectively [(HR 2.04 (0.89-4.69) (P =0.091)]. CI of pachymeningeal failure at 12 months was 16% versus 0% for S+SRS and SRS.SRS target volume, receipt of systemic therapies, and treatment with S+SRS instead of definitive SRS correlated with improved survival in patients with large BrM.