Resumen
Cancer is a complex ensemble of morphological and molecular features whose role is still unclear. Moreover, their role may change in different areas of the same tumour. Artificial intelligence (AI) allows pathologists to go beyond human perception and bias and may help better understand how these features influence disease progression. Furthermore, by capturing variation intrinsic to the tumour, AI may improve the accuracy of current prognostic tools, such as Leibovich Score (LS), in predicting patient outcome and response to therapy. For these reasons, we studied clear cell renal cell carcinoma (ccRCC) tissue in which molecular features and their coexpression in the same cell were quantified and mapped using AI-based image analysis software. We demonstrated a novel approach for investigating ccRCC and revealed new potential biomarkers of prognosis which may also be able to direct patients towards the most appropriate personalised therapy.