The way we categorise and classify cancer types dictates not only the way we diagnose and treat patients but also many of our decisions on biomarker and drug development. In addition, cancer taxonomy proves the ground truth for future discoveries in the area of computational pathology and artificial intelligence. This editorial comment illustrates the relevance of cancer taxonomy in clinical and morphomolecular diagnosis, prognosis and therapeutic prediction; it shows its importance in identifying the epidemiology, aetiology and pathogenesis in oncology and explains its determinant role in computational tissue-based cancer diagnosis.