The conservation treatment of ductal carcinoma in situ (DCIS) is based on the surgical excision of the tumour together with irradiation of the remaining breast. Because short-term recurrence is almost certainly caused by residual tumour, an attempt should be made to verify the adequacy of the excision by assessing the specimen margin. The reliability of histologic margin assessment is influenced by the growth pattern of DCIS within the ductal tree and by the distance between tumour foci. Using an original stereoscopic technique, the present study of 60 mastectomy specimens shows that continuous and multifocal growth patterns are usual. A multifocal distribution (defined as gap of 4 cm or more between tumour foci) was found in only a single case. The growth pattern is related to DCIS type. Poorly-differentiated DCIS shows continuous growth, in contrast to the well-differentiated DCIS, which has a multicentric distribution. Irrespective of histologic type, however, only 8% of DCIS have a multifocal distribution with gaps greater than 10 mm. Therefore, with careful assessment, the likelihood of a false free margin seems theoretically low and should encourage the use of conserving treatment for eradicable DCIS.