Between December 1981 and December 1988, 329 consecutive patients with stage I and II breast cancers who underwent wide excision (n = 261) or quadrantectomy (n = 68) with (n = 303) or without (n = 26) axillary dissection were referred to radiotherapy. Final margins of resection were microscopically free from tumor involvement in all cases. Radiotherapy consisted in 40-45 Gy over 4-4.5 weeks to the breast, with (n = 168) or without (n = 161) regional nodal irradiation of 45-50 Gy over 4.5-5 weeks. A mean booster dose of 15 Gy was delivered to the primary site by iridium-192 implant in 169 patients (group 1) or by electrons in 160 patients (group 2). Twenty-seven percent (n = 88) of patients received tamoxifen for > or = 2 years. Adjuvant chemotherapy was administered in 22% (n = 71) of patients. Groups 1 and 2 were not strictly comparable. Group 1 patients were significantly younger, had smaller tumors, were treated with cobalt at 5 x 2 Gy per week and axillary dissection was more frequently performed. Group 2 patients were more frequently bifocal and more frequently treated by quadrantectomy and tamoxifen, and irradiation used accelerator photons at 4 x 2.50 Gy per week. No difference in terms of follow-up and survival rates was observed between the two groups. For all patients the 5- and 10-year local breast relapse rates were 6.7% and 11%, respectively. No difference was observed regarding local control either by the electron or the iridium-192 implant boosts. Axillary dissection and age had an impact on the breast cosmetic outcome.(ABSTRACT TRUNCATED AT 250 WORDS)