A controlled, prospective, randomized study evaluated the use of 1,3-bis(2-chloroethyl)-1-nitrosourea (BCNU) and/or radiotherapy in the treatment of patients who were operated on and had histological confirmation of anaplastic glioma. A total of 303 patients were randomized into this study, of whom 222 (73%) were within the Valid Study Group (VSG), having met the protocol criteria of neuropathology, corticosteroid control, and therapeutic approach. Patients were divided into four random groups, and received BCNU (80 mg/sq m/day on 3 successive days every 6 to 8 weeks), and/or radiotherapy (5000 to 6000 rads to the whole brain through bilateral opposing ports), or best conventional care but no chemotherapy or radiotherapy. Analysis was performed on all patients who received any amount of therapy (VSG) and on the Adequately Treated Group (ATG), who had received 5000 or more rads radiotherapy, two or more courses of chemotherapy, and had a minimum survival of 8 or more weeks (the interval that would have been required to have received either the radiotherapy or chemotherapy). Median survival of patients in the VSG was, best conventional care: 14 weeks (ATG: 17.0 weeks); BCNU: 18.5 weeks (ATG: 25.0 weeks); radiotherapy: 35 weeks (ATG: 37.5 weeks); and BCNU plus radiotherapy: 34.5 weeks (ATG: 40.5 weeks). All therapeutic modalities showed some statistical superiority compared to best conventional care. There was no significant difference between the four groups in relation to age distribution, sex, location of tumor, diagnosis, tumor characteristics, signs or symptoms, or the amount of corticosteroid used. An analysis of prognostic factors indicates that the initial performance status (Karnofsky rating), age, the use of only a surgical biopsy, parietal location, the presence of seizures, or the involvement of cranial nerves II, III, IV, and VI are all of significance. Toxicity included acceptable, reversible thrombocytopenia and leukopenia.