Information from randomized and nonrandomized studies does indicate that preoperative chemotherapy (primary medical therapy) results in a high tumor response rate and an increase in the use of breast conservation, despite great variation in the chemotherapeutic regimens used. The findings reported by Powles et al in this issue of the Journal of clinical oncology clearly confirm. Fisher's group studies (NSABP-18) and those of others in that regard. Despite the repeated demonstration that a putative increase in breast-conserving surgery results as a consequence, the proper role of preoperative therapy in the management of primary breast cancer remains unclear. This uncertainty relates to the fact that no new paradigm has been formulated to govern the use of such therapy and to replace the two independent paradigms that currently govern the treatment of breast cancer. One of these paradigms relates to the eradication of local manifestations of the disease by surgery and radiation therapy; the other serves as the exemplar for eliminating disseminated disease by systemic therapy. The absence of a clear understanding of the clinical role of preoperative therapy does not negate the importance of continuing to conduct biologic and clinical investigations using that therapeutic strategy as a model. However, it is our content that, at this time, preoperative therapy should be limited to patients in randomized studies.