Initially, preoperative chemotherapy had three aims, to improve the overall survival by an early exposure to systemic therapy, to evaluate in vivo the efficacy of chemotherapy and to monitor post-operative treatment, and to increase the rate of breast-conserving treatments. Eight randomised trials have been published. The general consensus about the results of these trials is that neoadjuvant chemotherapy did not improve overall survival, in comparison with an adjuvant treatment. Clinical and especially pathological complete response rates were correlated with the outcome of the patients and could be considered as an independent prognostic factor. Primary chemotherapy is a powerful tool to study the predictive factors, especially biological factors, of chemosensitivity. However, the usefulness of this approach to adjust post-operative treatments remains to be demonstrated. Finally, primary chemotherapy slightly increased the rate of breast conserving treatments, but is accompanied by a small increase in the risk of local recurrence. To propose a conservative treatment to a woman initially not suitable for a lumpectomy needs clear information about this enhanced risk.