Artificial Intelligence-Based Pathomics: Technological Features and Clinical Applications

‘Pathomics’ is a combination of digital pathology and artificial intelligence that transforms pathology research from microscopic observation of cells in traditional pathology to digital pathology image analysis. Currently, large volumes of pathological image data can be filtered for the required tissue information in histopathology. This process involves the extraction of various pathological features such as nuclear characteristics, microenvironmental attributes, and individual cell analyses. Subsequently, a computer software is employed to analyze these features and construct digital pathomics models aimed at detecting biomarkers and molecular outcomes, thereby assisting in diagnosis and guiding clinical therapeutic strategies. Therefrom, this paper reviews the technical methods employed in pathomics and their characteristics, introduces the workflow of pathomics, discusses the clinical applications of pathomics in tumor diagnosis, treatment and prognosis, and emphasizes on the importance of the fusion of multi-source genomics, such as pathomics, imaging genomics and genomics.