Twenty-seven hemangioblastomas of the central nervous system were treated at the Mayo Clinic with radiation therapy from January 1963 to August 1983. Six patients had von-Hippel Lindau syndrome, and four presented with polycythemia. The median age among the 15 males and 12 females was 48 years (range 20-68). Two clinical groups were apparent: those that received postoperative radiation therapy for clinically suspect, or microscopically positive margins (6 patients) and those who underwent therapy for gross residual disease (20 patients). One patient did not fall into either group because his initially unresectable tumor was treated with planned pre-operative radiotherapy to 40 Gy and was subsequently successfully cured by surgery. Because the combined modality approach did not allow assessment of local control with radiation alone, he was excluded from the gross residual cohort in terms of time-dose relationship analysis. The cohort with gross residual disease was particularly unfavorable as 12 of these patients had developed 17 local recurrences prior to radiation. Three had multiple lesions, and four had the von-Hippel Lindau syndrome. In-field disease control appeared to be improved when patients were treated more aggressively. Patients treated to a dose of 50 Gy manifested local control in 4/7 (57%) vs 4/12 (33%) in patients treated to less than 50 Gy. In-field local control was also better if patients received a TDF greater than 75 (local control in 66%) vs a TDF of 65-75 (local control in 22%). Actuarial analysis of in-field disease control showed more aggressive treatment improved control whether analyzed by dose level (greater than or equal to 50 Gy vs less than 50 Gy, or TDF greater than 75 vs less than 75). Four of the six patients who received radiation therapy for microscopically positive or clinically suspect margins achieved local control. Both patients manifesting in-field relapse were successfully surgically salvaged. Overall survival for the entire group of 27 patients was 85%, 58%, 58%, and 46% at 5, 10, 15, and 20 years, respectively. Recurrence-free survival was 76%, 52%, and 42% at 5, 10, and 15 years, respectively. Half of all in-field recurrences had occurred by 2 years, but the remaining half recurred from 5.6 to 14.4 years. Patients who developed in-field failure usually died from disease with a median survival of only 1.5 years, but surgical salvage was accomplished in 4/12. Hydro-myelia developed in two patients and required operation. Surveillance for systemic tumors also was important and revealed seven benign and four malignant tumors.(ABSTRACT TRUNCATED AT 400 WORDS)