OBJECTIVE Analysis of the results of radiotherapy in a large group of cerebral gliomas with identification of prognostic factors and the outcome with respect to different decades of treatment. METHODS Two decades (1979-1999) of radiotherapy in supratentorial astrocytic and oligodendroglial tumors (n = 821) at the University Hospital Groningen were retrospectively evaluated. Prognostic factors for survival were analyzed. Two decades of radiotherapy treatment were compared with respect to radiotherapy dose and treatment-field design. RESULTS Glioblastoma multiforme, including gliosarcoma, was the most frequent supratentorial glioma (n = 442) with a poor survival, i.e., median survival time (MST) 7 months, especially in patients > 50 years of age and with poor performance. Patients with good performance were selected for radiotherapy with an optimum dose of 60 Gy local-field irradiation. However, in patients with poor prognosis, no radiotherapy was applied or a shorter treatment scheme was given. Anaplastic astrocytomas (n = 131) were treated in the same way as glioblastoma multiforme. Over time, a decrease in radiation dose (from 60 to 45 Gy) and from whole brain irradiation to local-field treatment was observed, following the literature. In low-grade gliomas, prognostic factors for survival were age, performance, and extent of resection. Gemistocytic astrocytoma (n = 15) had an inferior survival compared to astrocytoma (MST 46 vs. 54 months), but a superior survival compared to anaplastic astrocytoma (MST 10 months). The presence of an oligodendroglial component in a glioma implied a superior survival compared to the astrocytic gliomas. The inherent biology of the glioma is reflected by the study of recurrent tumors with progression to higher grades of malignancy in 32-40% and by the histology of recurrent oligodendroglial tumors. In comparing two decades of radiotherapy in gliomas, no differences in survival were observed despite the technological improvements. However, reduction in long-term side effects was not evaluated, especially in low-grade gliomas which were treated in the second decade of the study with local fields only and a reduced radiotherapy dose using computerized three-dimensional (3-D) planning. CONCLUSIONS Radiotherapy does not cure cerebral glioma. Prognostic factors for survival are histopathologic classification and grading, age, and patient performance. Technological improvements do not improve survival, but possibly reduce late effects.