Patients with recurrent breast cancer can be divided into three categories: those with locoregional recurrence, those with distant nonvisceral recurrence, and those with visceral recurrence. Survival from time of first relapse is clearly dependent on these categories. Selection of therapeutic modality should be based on considerations of site recurrence, symptomatology, anticipated response to therapy, and expected toxicities related to the therapy. Chemotherapy is appropriate for patients who are either unlikely to respond to hormone therapy, quite symptomatic, clearly hormone refractory, or have rapidly progressive visceral disease. Previously untreated patients are likely to respond to chemotherapy, with no clear-cut marker or clinical category associated with increased or decreased likelihood of benefit. Studies are ongoing to identify markers for response to chemotherapy, with recent investigations focusing on HER-2/neu expression. Standard combination chemotherapeutic regimens, consisting of either cyclophosphamide/methotrexate/5-fluorouracil (CMF) or cyclophosphamide/doxorubicin/5-fluorouracil (CAF), are associated with response rates in untreated patients of 35% to 80% and in previously treated patients of 10% to 40%. Although CAF probably has a slightly higher response rate than CMF, the toxicity of CAF is substantially higher. Newer agents are effective in both previously untreated and treated patients with breast cancer. These include paclitaxel, docetaxel, vinorelbine (Navelbine; Burroughs Wellcome Co, Research Triangle Park, NC; Pierre Fabre Médicament, Paris, France), and amonafide. Furthermore, modulation of previously existing agents, designed to overcome resistance, has been tested. Only leucovorin/5-fluorouracil has apparent clinical benefit. A number of novel approaches are being designed or are currently being used in clinical trials. These include differentiating agents, anti-angiogenesis factors, antitumor antigen-based therapy, growth factor receptor/ligand therapy, and gene therapy.