Presentation : A compliant package with international standards for radiomics analysis of medical images






Precision oncology is an important field in cancer research as its main goal is to lead towards personalized treatments. The genomic heterogeneity of aggressive tumors usually results in heterogeneous tumor features at the anatomical scale that can be captured using quantitative medical image analysis. This analysis will help make accurate tumor phenotypic predictions. The extraction of the tissues characteristics from medical images is known as “radiomic analysis”. In this regard, our team, in collaboration with an international consortium called MEDomics, is developing a Python-coded software called MEDimage. This software offers flexible parameters to ensure a product adapted to the users needs to process medical images (e.g. filtering, cropping) and to extract radiomic features in an efficient way. However, validation of these image analyses remains a major challenge. Therefore, we integrated the procedures established by the International Biomarker Standardization Initiative (IBSI) into our MEDimage software. Our results are compliant with the IBSI. A version of MEDimage has already been released and will be tested by several collaborators to ensure the correct functioning of the package and the compatibility of the user needs with the possibilities offered by this technology. It is expected that this tool made for extracting rich and complex radiomic features will be used to conduct multicentric studies in precision oncology.