From February 1989 to December 1992, 31 patients who presented with an initial pathological diagnosis of glioblastoma multiforme underwent tumor debulking or biopsy, stereotactic radiosurgery, and standard radiation therapy as part of their primary treatment. Presenting characteristics in the 22 men and nine women included a median age of 57 years, Karnofsky Performance Scale score median of 80, and median tumor volume of 16.4 cm3. Stereotactic radiosurgery delivered a central dose of 15 to 35 Gy with the isocenter location, collimator size, and beam paths individualized by means of three-dimensional software developed at the University of Wisconsin. The peripheral isodose line varied from 40% to 90% with a median of 72.5% and a mode of 80%. The mean follow-up period was 12.84 months with a median of 9.5 months. Statistical analysis was performed using Kaplan-Meier analysis and log-rank comparison of risk factor groups. The parameters of age, initial Karnofsky Performance Scale score, and biopsy were significantly different in patient survival from debulking; but no difference was noted between single and multiple isocenters and patterns of steroid requirement. Radiographic recurrences were divided by location into the following categories: central (within central stereotactic radiosurgery dose), 0; peripheral (within 2 cm of central dose), 19; and distant (> 2 cm), 4. There is no evidence of recurrence in five surviving patients. Actuarial 12-month survival was 37%, with a median survival of 9.5 months. These values are similar to previous results for surgery and standard radiotherapy alone. The results suggest that the curative value of radiosurgery is significantly limited by peripheral recurrences.