Determining risks of having or developing medical conditions based on radiomics features extracted from automatically segmented 3d images of a skeletal muscle of a subject's trunk

The invention is notably directed to a computer-implemented method for determining risks of having or developing medical conditions, and/or complications thereof, in a subject. the method comprises: accessing a 3d image of a skeletal muscle of the trunk of the subject, wherein the 3d image has been previously obtained using a medical imaging technique; segmenting (s43 - s44) the 3d image using an automatic image segmentation process; extracting (s45) muscle radiomics features from the segmented 3d image; and executing (s50), based on predetermined sets of weights assigned to respective endpoints, a module to determine a score for each of the endpoints by weighting the extracted muscle radiomics features in accordance with a respective one of the predetermined sets of weights. the score is a risk of having or developing a medical condition or a complication thereof for the respective endpoint, while said medical condition is primarily linked to a system or an organ that is anatomically distinct from said skeletal muscle. the invention is further directed to related devices computerized systems and computer programs.