Between July, 1984, and October, 1988, 263 patients (163 male, 100 female), aged from 4 to 83 years (mean 52 years), with malignant brain gliomas underwent surgical procedures: stereotactic biopsy in 160 and resection in 103 patients. There were 170 grade IV astrocytomas, 17 grade IV mixed oligoastrocytomas, 44 grade III astrocytomas, 22 grade III mixed oligoastrocytomas, and 10 malignant oligodendrogliomas. Overall median survival time was 30.1 weeks for grade IV gliomas, 87.7 weeks for grade III gliomas, and 171.3 weeks for malignant oligodendrogliomas. Multivariate analysis in 218 newly diagnosed cases revealed that the variables most strongly correlated with survival time were: tumor grade, patient age, seizures as a first symptom, a Karnofsky Performance Scale score of less than 70%, tumor resection, and a radiation therapy dose greater than 50 Gy. The proportions of patients receiving tumor resection versus biopsy in each of these prognosis factor groups were similar. Since most of the 22 patients with midline and brain-stem tumors were treated with biopsy alone, these were excluded. Considering 196 newly diagnosed patients with cortical and subcortical tumors, grade IV glioma patients undergoing resection of the contrast-enhancing mass (as evidenced on computerized tomography and magnetic resonance imaging) and postoperative external beam radiation therapy lived longer than those undergoing biopsy only and radiation therapy (median survival time 50.6 weeks and 33.0 weeks, respectively; Smirnov test, p = 0.0380). However, survival in patients with resected grade III gliomas was no better than in those with biopsied grade III lesions (p = 0.746). The authors conclude that, in selected grade IV gliomas, resection of the contrast-enhancing mass followed by radiation therapy is associated with longer survival times than radiation therapy after biopsy alone.