OBJECTIVE To carry out a retrospective study of male breast cancer over a 22-year experience. METHODS Data from 121 male patients with breast cancer treated between the years 1972 and 1994 at the Surgical Clinic of Ankara Oncology Hospital were reviewed. Distribution of cases according to stage was: 2.5% stage I, 28.9% stage II, 55.4% stage III and 13.2% stage IV (AJCC staging method). The surgical treatment for 23 of the patients (19%) was Halsted's radical mastectomy or modified radical mastectomy. Seventy-three cases (60.3%) had total mastectomy without axillary node dissection and 25 (20.7%) had local tumour excision only. Seventy-two of 121 patients had adjuvant treatment. RESULTS In general the prognosis of men with breast cancer was worse than for women. In the analysis of patients in stages I, II and III-A (operable disease group), the 5-year survival rates were 73% in axillary node-negative patients and 77% in those with tumours sized under 5 cm (P<0.001). In these patients, univariate analysis demonstrated that axillary status (relative risk of death in positive status vs. negative=3.6), tumour size (relative risk in T3 vs. T1-2=2), surgical treatment type (relative risk in simple mastectomy vs. radical mastectomy=1.9) and adjuvant chemotherapy (relative risk if no chemotherapy=1.4) were statistically significant factors associated with survival. CONCLUSIONS Cox's regression model revealed that axillary status, tumour size and type of surgical treatment were the most important independent prognostic factors (P<0.001).