Résumé : The 2003 St. Gallen consensus panel divided the many available adjuvant chemotherapy (CT) regimens into those with "standard efficacy" (ACx4, CMFx6) and those with "superior efficacy" (FA(E)Cx6, CA(E)Fx6, A(E)-->CMF, TACx6, ACx4--> paclitaxel (P)x4 or docetaxel (D)x4) but also greater complexity, toxicity and cost. This paper will summarize the latest information on long-term side effects of the "superior" regimens and 5-year benefits reported in taxane trials, including those of a "new" sequential regimen, FECx3--> docetaxelx3. Rapidly expanding evidence of marked heterogeneity in the magnitude of CT benefits according to the tumour oestrogen receptor (ER) status, a claim made for many years by IBCSG investigators, will be reviewed; it will lead to the conclusion that a revolution needs to take place in the way oncologists think about the CT added value and design adjuvant clinical trials. The conclusions proposed to the 2005 St. Gallen consensus panel are that: adequately dosed anthracycline-based CT regimens remain an acceptable standard for many women; a lower threshold for using taxanes in sequence or combination with anthracyclines (A) is justified in the presence of an ER-negative or low-ER tumour status, other aggressive biologic features (such as HER-2 overexpression), fear about A-induced cardiotoxicity; no recommendation can yet be made as far as the optimal taxane-A regimen, the best taxane or the best taxane schedule.