[Molecular prognostic and predictive markers of breast cancer treatment]

Breast cancer heterogeneity has been deciphered during the last decade thanks to the use of high-throughput tools. A major clinical concern is the determination of the metastatic risk of the tumours (prognostic factor), but also the optimal choice of the treatment for a given patient and tumour (predictive factor). A significant advance has been obtained from the description of a novel molecular classification of breast cancers. This allowed a refinement in the determination of tumour groups displaying different prognoses. Presently, numerous gene signatures have been published and some of them are on the market, at least in the United States, but prospective validation studies are still ongoing. After a major enthusiasm, numerous questions have been raised, concerning especially the stability of these signatures. An improved knowledge of these limits is also an important factor for an optimal use of these data. It can be concluded that it should be possible to improve patients' care by integrating molecular and clinicopathologic data in a common approach on the one hand, and to further develop new biomarkers predictive of therapeutic efficacy on the other hand.