Integrative Radiomics for Precision Medicine (RADIOME)

The RADIOME team develops an integrative radiomics approach for precision medicine, leveraging medical imaging (MRI, CT, PET, ultrasound) in combination with clinical, biological, genomic, and pathomic data, using advanced artificial intelligence methods.

Our objective is to design robust, interpretable, and clinically transferable models to improve patient management in oncology. Our research focuses in particular on predicting prognosis, treatment response, adverse events, and recurrence risk, as well as on optimizing therapeutic strategies. Special emphasis is placed on independent external validation of the models and on their deployment in clinical practice. The team also aims to decode the biological information captured by AI algorithms by integrating data from multiple modalities and developing image representations that enable the identification of new biomarkers that are both interpretable and robust. Finally, the advanced analysis of whole-body PET/CT images fits within a systems medicine framework, integrating information from healthy organs and inter-organ metabolic networks to better understand tumor mechanisms and their interactions with the organism.

The team benefits from international recognition, built on major contributions to imaging biomarker harmonization, the development of radiomics and multimodal models, and innovative methods for modeling and interpretability. This expertise is further demonstrated through active involvement in reference initiatives (IBSI, CLEAR, METRICS), and strengthened by training activities and the wide, free dissemination for academic research of the software platform LIFEx (www.lifexsoft.org).