A multicenter series of 350 intraductal breast cancers (DCIS) is reported. Mammography was the most sensitive test but suspicion arose only at palpation in 13% of cases whereas in 10% of cases biopsy was recommended for a benign lesion and DCIS was an unexpected finding. Mammography, physical examination and cytology must be combined to achieve optimal sensitivity. Systematic biopsy of apparently benign masses would increase DCIS detection rates but the cost-effectiveness of such a policy is questionable. A trend of conservative surgery was evident over time (from 1968-79, 28%; 1985-1989, 50%) but breast irradiation followed only in one fourth of the cases. The local recurrence rate was significantly higher in cases of limited surgery (with or without irradiation) with respect to mastectomy (1.2 vs 0.2 x 100 patient-years at risk). Most recurrences (7 of 8) in the conserved breast were infiltrating, but no recurrence was seen in subclinical DCIS cases. Three patients died of breast cancer after local recurrence in the conserved breast (2 cases) or mastectomy scar (1 case). Eligibility for conservative surgery of DCIS needs to be carefully discussed to avoid under-treatment. Contralateral breast cancer was recorded in 44 cases and the incidence of further metachronous cancer to the other breast was ten times higher than expected in normal breasts. Four patients died of contralateral breast cancer, free of ipsilateral recurrence. A careful follow-up of the contralateral breast in DCIS cases looks as important as surveillance of the conserved breast.