“…By tapping the computational power of artificial intelligence (AI) systems, a large amount of information that go beyond visual image assessment can be collected through the extraction and analysis of radiomics features from medical images, obtained in a routine setting with conventional imaging protocols (e.g., for tumor staging and follow-up). This approach has the potential to enhance options to carry out patient-specific diagnostic and prognostic evaluations, allowing e.g., to predict early treatment response and avoid undue over- or undertreatment depending on the biological aggressiveness of each patient-specific disease, and paving the way to a more individualized patient management [ 22 , 40 , 41 , 42 , 43 , 44 , 45 , 46 , 47 , 48 , 49 , 50 , 51 , 52 , 53 , 54 , 55 , 56 , 57 , 58 , 59 , 60 , 61 , 62 , 63 , 64 , 65 , 66 , 67 , 68 ]. Other advantages of radiomics over conventional biopsy include the possibility to capture the tumor tissue in its entirety (providing information regarding its clonal heterogeneity and the tumor microenvironment along with its stromal component), to guide biopsy towards specific tumor areas, and to be repeatable at virtually any time during the disease course, allowing longitudinal monitoring with improved opportunities of optimizing treatment strategies [ 39 , 43 , 44 , 45 ].…”