PURPOSE To define the natural history, and prognostic factors of patients with histologically unverified presumed gliomas diagnosed on CT or MR imaging, and treated with external beam radiotherapy. METHODS AND MATERIALS Retrospective review of 111 adults with histologically unverified presumed cerebral glioma treated with radiotherapy between 1974 and 1990. Using CT or MRI criteria alone 41 were presumed low grade, 63 high grade gliomas and 7 were unclassified. Survival results were compared to a cohort of 82 adults with histologically verified low grade gliomas treated over the same period with surgery and radiotherapy. RESULTS The 5 year survival probability of the whole cohort was 31%. Age, performance status, and the degree of contrast enhancement were independent prognostic factors for survival. Patients with presumed low grade glioma had a 5 year survival of 41% compared to 52% for patients with verified low grade glioma. After correction for prognostic factors no significant difference was found in the survival between patients with verified and unverified low grade tumors. One of 15 cases with subsequent histology, obtained at autopsy or salvage surgery, had nonglial pathology. CONCLUSION Patients diagnosed on the basis of clinical features and imaging as having presumed glioma have similar natural history and clinical behavior after treatment with radiotherapy to those with histologically confirmed gliomas. However, the results should not be taken as justification for avoiding biopsy. A proportion of patients may have nonglial pathology and new more effective treatment strategies for patients with glial tumors can only evolve on the basis of full diagnostic information including pathology.