To evaluate and define treatment prognostic factors and selection criteria for patients with solitary brain metastasis treated with stereotactic radiosurgery. Between March 1990 and January 1995, 48 patients with solitary brain metastasis were treated with stereotactic radiosurgery at our tertiary care institution. Eleven patients were treated for progression or relapse of intracranial disease after whole-brain external beam radiotherapy or surgical resection. Patients were eligible for radiosurgery if they had a (1) Karnofsky Performance Status (KPS) of at least 70, (2) a solitary, radiographically distinct lesion < 4 cm in diameter, (3) lesion at least 1 cm from the optic chiasm, and (4) a reasonably well-controlled primary tumor. The factors subjected to univariate and multivariate analysis for local control and survival correlations were age, histology, control of systemic disease, surgical debulking, whole-brain irradiation, dose delivered, lesion volume, location (supratentorial vs. infratentorial), and new versus recurrent metastasis. Sex and initial KPS were also analyzed for survival. Local control was achieved in 39 (81%) patients. The only prognostic factor in the univariate analysis that significantly correlated with local control was new versus recurrent lesion (P = .009). Initial KPS of at least 80 and age < 65 years were significant with regard to survival in the univariate and multivariate analyses. Median survival after radiosurgery was 8 months. This study corroborates the usefulness of stereotactic radiosurgery for patients with a small solitary metastatic brain lesion. In our series, as well as in others, several characteristics (young age, optimal KPS, newly diagnosed metastasis, controlled or absent systemic disease) provide clues as to which patients will most benefit. Studies, including randomized trials, are needed to clarify patient and tumour characteristics that predict which patients will benefit from stereotactic radiosurgery.