Recent studies have shown a survival benefit for patients with recurrent glioblastomas treated with stereotactic brachytherapy. On the basis of these encouraging results, we began a prospective study in 1987 to evaluate the use of brachytherapy in patients with newly diagnosed glioblastoma. Patients were considered eligible for this study if they met the following criteria: Karnofsky performance status 70% or greater; tumor size not greater than 5 cm in any dimension; a radiographically well delineated, supratentorial lesion not involving the ependymal surfaces; and pathologically confirmed glioblastoma. We treated 35 such patients between 1987 and 1990 with stereotactic brachytherapy as part of their initial therapy. The treatment protocol involved surgery, partial brain external-beam radiotherapy (59.4 Gy in 33 fractions), and stereotactic brachytherapy with temporary high-activity iodine 125 sources giving an additional 50 Gy to the tumor bed. Chemotherapy was not used in the initial management of these 35 patients. To compare our results with those obtained in a matched control group, we identified 40 patients with glioblastoma treated with surgery and external radiotherapy, with or without chemotherapy, between 1977 and 1986 at our institution. These patients had clinical and radiographic characteristics that would have made them eligible for the brachytherapy protocol. Survival rates at 1 and 2 years after diagnosis were 87% and 57%, respectively, for patients receiving brachytherapy versus 40% and 12.5%, respectively, for the controls (P less than .001). We conclude that stereotactic brachytherapy improves the survival of patients with glioblastoma when it can be incorporated into the initial treatment approach. Unfortunately, only about one in four patients with glioblastoma are suitable candidates for brachytherapy at the time of initial presentation.