OBJECTIVE To define treatment selection criteria for patients with intracranial metastases treated with stereotactic radiosurgery. METHODS Between August 1989 and July 1993, 25 patients with intracranial metastases (28 lesions) were treated with stereotactic radiosurgery at the University of Florida. Thirteen patients were treated for progressive intracranial disease after external-beam radiotherapy, and 12 were treated with radiosurgery as an adjunct to initial treatment. Minimum eligibility criteria included histologic verification of primary disease, Karnofsky performance status 50% or greater, three or fewer intracranial metastases, radiographically distinct lesion(s) 4 cm or less in diameter, and reasonably well-controlled primary disease. Univariate and multivariate analyses tested the prognostic significance of Karnofsky performance status, lesion volume, number of lesions, treatment dose (both external beam and stereotactic), histology, site of primary disease, and time interval (less than or greater than 1 year) from primary diagnosis to development of intracranial metastasis or from treatment of intracranial disease to recurrence. RESULTS Local control was achieved in 84% of 28 lesions treated. The only significant prognostic indicator among the tested variables was the interval to development or recurrence of intracranial metastasis. CONCLUSIONS Although stereotactic radiosurgery improves local control rates and is likely to offer improved palliation for a select cohort of patients, the selection criteria for such patients remain poorly defined. Our data suggest that an interval of greater than 1 year from primary disease diagnosis to development of intracranial metastasis, or from treatment of intracranial metastasis to its recurrence, defines a patient cohort that is more likely to benefit from this treatment technique.