OBJECTIVE To evaluate the outcomes of patients with giant cell tumor of bone (GCTB) treated with radiotherapy (RT) with or without surgical resection. METHODS We performed a retrospective review of the records from 25 consecutive patients with pathologically confirmed GCTB who had undergone RT between 1956 and 2000. RESULTS Patients ranged in age from 11 to 69 years (median 32); 16 were female and 9 were male. The anatomic distribution of lesions was as follows: cervical spine, 3; temporal bone, 1; thoracic or lumbar spine, 9; sacrum, 8; ilium, 1, and humerus, radius, and thumb metacarpal, 1 each. Tumors ranged in size from 2 to 20 cm (median 9.5) at their maximal dimension. Thirteen patients had been referred for RT for primary GCTB and 12 had been referred with locally recurrent disease after having undergone one or more other treatments. Fourteen patients had undergone RT for gross disease, and the remaining 11 had been treated with RT after gross total resection. In 10 of these 11 patients, the treatment margins were positive or uncertain. Radiation doses ranged from 25 to 65 Gy (median 46). At a median follow-up of 8.8 years (range 0.67-34), 7 patients had developed isolated local recurrence, 2 had developed isolated distant recurrence, and 3 had developed both. The actuarial 5-year overall and disease-free survival rate was 91% and 58%, respectively, and the actuarial 5-year local control and distant metastasis-free survival rate was 62% and 81%, respectively. Univariate analysis suggested that treatment for recurrent disease correlated with a lower disease-free survival rate (83% vs. 33%, p = 0.06), distant metastasis-free survival rate (100% vs. 64%, p = 0.08), and local control rate (83% vs. 42%, p = 0.08) at 5 years. Of the 12 cases of recurrence, 7 were ultimately successfully treated with additional salvage therapy. In 4 of these patients, salvage therapy included interferon-alpha 2b. CONCLUSIONS RT should be considered an adjuvant to surgery or as alternative therapy in cases of GCTB that are unresectable or in which excision would result in substantial functional deficits. When RT is used as primary therapy, the rate of local control seems to be satisfactory. In heavily pretreated patients, however, RT delivered as it was in this series can result in poor local control, and alternative therapies should be considered.