PURPOSE Evaluation of the treatment outcome after radiosurgery (RS) alone or in combination with whole-brain radiotherapy (WBRT) with special attention to prescribed dose and its influence on local control and survival. PATIENTS AND METHODS Between September 1984 and January 1997, 236 patients with 311 brain metastases treated with radiosurgery met the following inclusion criteria: one to three brain metastases per patient; no previous WBRT; and Kamofsky performance status (KPS) > or = 50%. One hundred fifty-eight patients treated only with RS received a median dose of 20 Gy prescribed to the 80% isodose line; 78 patients received RS with a median dose of 15 Gy/80% and an additional course of WBRT. RESULTS For the entire series, overall median survival was 5.5 months, with control of CNS disease achieved in 92% of the treated brain metastases; the results were not significantly different between patients treated by RS with or without WBRT. However, in patients without evidence of extracranial disease, median survival was increased for patients who received WBRT (15.4 vs 8.3 months; P=.08). Additionally, there was a suggestion that increased doses for patients treated with RS only resulted in improved outcome. Four lesions were suspicious for radiation necrosis by magnetic resonance imaging (MRI); in one of the four lesions, radiation necrosis was confirmed histologically. The incidence of transient low-grade toxicity was 18%; symptoms could be treated by the temporary administration of steroids. CONCLUSION RS is an effective, noninvasive means of controlling brain metastases when used alone or in combination with WBRT. There is a trend for superior local control and especially in patients without extracranial disease for superior survival when RS is used in conjunction with WBRT. Randomized trials would seem to be warranted, comparing the benefit of RS with or without additional WBRT.