Fig. 5

AIRI roadmap (pipeline). The steps involved in the radiomics analysis workflow start with image acquisition and pre-processing and end with the application and evaluation of the outcome. After pre-processing, segmentation of the tumor regions (ROI) in MRI sequences is performed to then extract data from the segmented regions (feature extraction). Afterward, the most predictive features are selected (feature selection), and the ML model is trained with a specific algorithm (model training). AI has been incorporated to perform predictions of outcomes as the final step.