Adjuvant chemotherapy is widely used, but its performance is not optimal. Two subgroups of patients do not get any benefit from adjuvant chemotherapy: the first one comprises patients who are already cured by locoregional treatment alone and the second one patients who do not profit from adjuvant chemotherapy because of resistance to the regimens employed. To improve the cost/benefit of this treatment strategy, we have two means: one is to improve the sensitivity of prognostic factors to be able to select a specific group with a good signature that does not need adjuvant treatment; the second is to identify predictive factors that may help us to select the optimal therapeutic strategy or the optimal regimen or drug for individual patients. New technologies of microarray revealed several genetic profiles. A large randomized trial (Microarray In Node-negative Disease may Avoid ChemoTherapy, MINDACT) will compare the information obtained with the genomic profiling and the classical clinico-pathologic index (St Gallen); the objective is to allow women not to be treated with adjuvant chemotherapy if their genomic signature is good. Another trial (EORTC 10994) is conducted in order to show that in cases of p53 mutated tumors, neoadjuvant chemotherapy with docetaxel is more efficient than an anthracycline-containing regimen. A supplementary study will evaluate gene profile predicting for p53 status. So, new genomic prognostic factors are still in development and seem very promising for optimizing the indications for adjuvant chemotherapy.