Breast cancer is one of the most frequent cancers in women in Europe. Artificial intelligence can help physicians to make diagnostic and therapeutic decisions for breast cancer. However, most recent artificial intelligence methods (e.g. deep learning) are «black boxes» that cannot explain why the machine makes a given prediction. On the contrary, physicians need to understand the rationale of recommendations, in order to follow them. Here, we propose a visual case-based reasoning approach, allowing the visualization of both quantitative and qualitative similarities between cases. This approach was tested on 3 public datasets for diagnosis and one real datasets from the project for therapy, and presented to 11 physicians. This paper is a French summary of: Jean-Baptiste Lamy, Boomadevi Sekar, Gilles Guezennec, Jacques Bouaud, Brigitte Séroussi. Explainable artificial intelligence for breast cancer: A visual case-based reasoning approach. Artificial Intelligence in Medicine 2019(94):42-53s.