Challenges to build an information system for a breast cancer risk score

Laurent Brisson (1, 2) , Emilien Gauthier (3) , Philippe Lenca (1, 2) , Franc ̧oise Clavel-Chapelon (4) , Stéphane Ragusa (4)

1 Lab-STICC_TB_CID_DECIDE
2 LUSSI - Département Logique des Usages, Sciences sociales et Sciences de l'Information
3 Lab-STICC - Laboratoire des sciences et techniques de l'information, de la communication et de la connaissance (UMR 3192)
4 IGR - Institut Gustave Roussy

Cancer has become the leading cause of death worldwide according to the World Health Organization. As a consequence, prevention and screening programs are set up by health authorities to decrease its incidence. Programs efficiency can be increased by targeting highest risk subset of the population. Efficient information systems capable of monitoring the population risk are thus needed. Constraints to build such a cancer risk score and their impacts on the information system are presented. As it will be shown, beyond risk score performance, a major constraint concerns the place of domain expert and the acceptability by end users. Readability then becomes an important criteria. It is shown that a simple k-nearest-neighbor algorithm can achieve good performance with the help of the domain expert. As an illustration, a risk score made of only four attributes is presented for the french population.

Domaines: Intelligence artificielle [cs.AI]

Laurent Brisson, Emilien Gauthier, Philippe Lenca, Franc ̧oise Clavel-Chapelon, Stéphane Ragusa. Challenges to build an information system for a breast cancer risk score. RCIS 2012: 6th International Conference on Research Challenges in Information Science, May 2012, Valencia, Spain. ⟨10.1109/RCIS.2012.6240437⟩. ⟨hal-03896759⟩