There has been growing interest in intraductal carcinomas of the breast (DCIS) over the last ten years mainly because of their increasing frequency and their difficult diagnosis. Their natural history is often surprising. For the time being, it is impossible to establish which proportion of DCIS might turn into an infiltrating carcinoma, and which factors are predictive for such a risk. These uncertainties are responsible for many controversies about their treatment. Based on a critical review of the latest publications, this paper deals with the possibilities of conservative treatment, challenging the remarkable results of total mastectomy (nearly 100% survival at 10 years). The risk of conservative treatment depends on the frequency of local recurrence, and on the potential vital risk of such recurrences, knowing that half of these recurrences will develop in an invasive, and no longer in situ, pattern. Randomized trials are being conducted on this question; they will not give an answer before the year 2000. In the mean time, conservative treatment seems to be reasonable for small low grade histologic lesions widely excised by surgery, and with rigorous possibilities of follow-up. The operation is followed by external irradiation. In case of recurrence, mastectomy has to be. It is not impossible that, performed under these conditions, a slight increase in mortality might follow such a strategy, thus heavily balancing the benefits of conserving the breast. Besides, surgical excision alone should only be performed as part of randomized trials, or for infra-centimetric lesions discovered by histology after resection of supposed benign lesions.