MILCOM – Radiomics for multiple myeloma

Diana MATEUS, Thomas CARLIER, Ludivine MORVAN, Françoise BODERÉ

But:  Développer des algorithmes d’apprentissage automatique permettant de lier de façon des biomarqueurs image (TEP/Scanner) au diagnostic et pronostic des patients de Myélome Multiple

Goal: Develop machine learning algorithms linking biomarkers from PET/CT images to predict the diagnose and prognosis of Multiple Myeloma patients

Multiple Myeloma is a type of cancer that proliferates within the bone marrow and the second most common type of blood cancer. Diagnose and treatment paths rely on the analysis of PET (Positron Emission Tomography) and CT (Computed Tomography) images.  In this project, we develop machine learning methods capable of extracting quantitative and reproducible measures from images with a predictive power on the survival probability or the time to the next event for multiple-myeloma patients