BACKGROUND: Brain metastases are common in patients with melanoma, and optimal management is not well defined. As melanoma has traditionally been thought of as "radioresistant," the role of whole brain radiation therapy (WBRT) in particular is unclear. We conducted this retrospective study to identify prognostic factors for patients treated with stereotactic radiosurgery (SRS) for melanoma brain metastases and to investigate the role of additional up-front treatment with whole brain radiation therapy (WBRT). METHODS: We reviewed records of 147 patients who received SRS as part of initial management of their melanoma brain metastases from January 2000 through June 2010. Overall survival (OS) and time to distant intracranial progression were calculated using the Kaplan-Meier method. Prognostic factors were evaluated using the Cox proportional hazards model. RESULTS: WBRT was employed with SRS in 27% of patients and as salvage in an additional 22%. Age at SRS > 60 years (hazard ratio [HR] 0.64, p = 0.05), multiple brain metastases (HR 1.90, p = 0.008), and omission of up-front WBRT (HR 2.24, p = 0.005) were associated with distant intracranial progression on multivariate analysis. Extensive extracranial metastases (HR 1.86, p = 0.0006), Karnofsky Performance Status (KPS) ≤ 80% (HR 1.58, p = 0.01), and multiple brain metastases (HR 1.40, p = 0.06) were associated with worse OS on univariate analysis. Extensive extracranial metastases (HR 1.78, p = 0.001) and KPS (HR 1.52, p = 0.02) remained significantly associated with OS on multivariate analysis. In patients with absent or stable extracranial disease, multiple brain metastases were associated with worse OS (multivariate HR 5.89, p = 0.004), and there was a trend toward an association with worse OS when up-front WBRT was omitted (multivariate HR 2.56, p = 0.08). CONCLUSIONS: Multiple brain metastases and omission of up-front WBRT (particularly in combination) are associated with distant intracranial progression. Improvement in intracranial disease control may be especially important in the subset of patients with absent or stable extracranial disease, where the competing risk of death from extracranial disease is low. These results are hypothesis generating and require confirmation from ongoing randomized trials.