From January 1964 to January 1968, 737 patients with breast cancer were randomized at the Cancer Institute in Milan to undergo either Halsted mastectomy or Halsted mastectomy with internal mammary node dissection; 716 were considered evaluable. The series was part of an international cooperative study. The two groups of patients were comparable in age, menopausal status, quadrant distribution, and frequency of axillary metastases. No patients received postoperative radiotherapy or adjuvant treatments. At ten years, no difference was found in either in the length of disease-free or overall survival in the two series. The overall survival was 60.7% in patients treated with Halsted mastectomy and 57.0% in patients treated with Halsted mastectomy plus internal mammary dissection. In no subgroups was a statistically significant difference found. In the series treated by extended mastectomy, the incidence of internal mammary metastases was 20.5% (24.6% in cases with tumor in medial or central quadrants and 17.7% in cases with tumor in lateral quadrants). The follow-up of the patients treated by Halsted mastectomy showed that only 15 had a parasternal recurrence; the expected number was 75. In nine patients, parasternal recurrences were the first site of relapse of the disease. All of them were treated with radiotherapy, and in four a complete control was obtained.