ResearchIGA NT14134

Integrative development of a multimodal score for risk of relapse estimation of breast cancer patients

Breast cancer is a complex and heterogeneous disease for which several classification systems and genomic risk signatures have been proposed. Nevertheless, its clinical outcome and therapy response remain highly variable, indicating the presence of additional factors that should be accounted for. To the best of our knowledge, there is no method which jointly exploits the wealth of information that can be mined from histological images and molecular features in breast cancer. This project aims at exploring this new avenue for building risk scores, by extracting image descriptors that are

  1. correlated with genomic/molecular scores; and
  2. complementing these scores.

The first set of descriptors is meant to help the pathologist in assessing the tumor biopsies and, thus, to provide an improved and earlier evaluation of the tumor characteristics. The second set of descriptors is meant to help in building a multimodal risk score able to better stratify the patient population.