Predictors of neurologic and nonneurologic death in patients with brain metastasis initially treated with upfront stereotactic radiosurgery without whole-brain radiation therapy

Background. In this study we attempted to discern the factors predictive of neurologic death in patients with brain metastasis treated with upfront stereotactic radiosurgery (SRS) without whole brain radiation therapy (WBRT) while accounting for the competing risk of nonneurologic death. Methods. We performed a retrospective single-institution analysis of patients with brain metastasis treated with upfront SRS without WBRT. Competing risks analysis was performed to estimate the subdistribution hazard ratios (HRs) for neurologic and nonneurologic death for predictor variables of interest. Results. Of 738 patients treated with upfront SRS alone, neurologic death occurred in 226 (30.6%), while nonneurologic death occurred in 309 (41.9%). Multivariate competing risks analysis identified an increased hazard of neurologic death associated with diagnosis-specific graded prognostic assessment (DS-GPA) = 2 (P = .005), melanoma histology (P =.009), and increased number of brain metastases (P < .001), while there was a decreased hazard associated with higher SRS dose (P =.004). Targeted agents were associated with a decreased HR of neurologic death in the first 1.5 years (P = .04) but not afterwards. An increased hazard of nonneurologic death was seen with increasing age (P = .03), nonmelanoma histology (P<.001), presence of extracranial disease (P<. 001), and progressive systemic disease (P = .004). Conclusions. Melanoma, DS-GPA, number of brain metastases, and SRS dose are predictive of neurologic death, while age, nonmelanoma histology, and more advanced systemic disease are predictive of nonneurologic death. Targeted agents appear to delay neurologic death.