Anatomical pathology has traditionally relied on the interpretation of histomorphological features under a light microscope by trained pathologists for diagnosis. Technological advancements have enabled the digitisation of tissue slides to produce high-resolution whole slide images, heralding the era of digital pathology (DP). Many laboratories around the world have incorporated DP into their routine workflows owing to the myriad applications it offers in facilitating tumour board discussions, remote reporting, teaching, and research. Most significantly, DP has engendered the field of computational pathology, a novel branch of histopathology incorporating artificial intelligence (AI) models. Computational pathology has been utilised in histomorphological quantification and diagnostic, predictive, and prognostic applications due to its potential to improve diagnostic accuracy, personalise treatment, and streamline workflows. Here, we highlight the work of Meier et al, Shen et al, and Lee et al, published in this journal in recent years, as they apply AI models to predict survival and treatment responses in gastric cancer, breast cancer, and diffuse large B-cell lymphoma, respectively. Collectively, these studies illustrate various approaches to incorporating AI into the DP pipeline and their potential clinical applications. Issues related to diagnostic accuracy, cost, patient confidentiality, and regulatory ethics still need to be addressed within the field. Despite this, the overall sentiment among pathologists is one of cautious optimism.
Tan-Garcia et al. (Mon,) studied this question.