Résumé : Breast cancer complexity has long been known and investigated. After a first classification of the disease based on histology features, starting from the 1980s breast cancers have been distinguished on the basis of oestrogen receptor expression and later according to HER2. By 2000 the "microarray revolution" had shown that the phenotypic differences between breast cancers were a reflection of their mRNA expression profiles, while the more recent "genomic revolution" is revealing the genomic bases of breast cancer heterogeneity. However, how this huge amount of data and knowledge translate into clinically relevant practice is currently not clear. In the present review we discuss how the different breast cancer classification methods might translate into improved clinical guidelines with regard to staging, therapy, and follow up of patients with breast cancer.