That AI Article Says What?? Publishing, Reviewing, & Evaluating AI Studies in Pathology

Abstract

AI studies in healthcare, and particularly those using digital pathology images, have skyrocketed over the past 3-5 years. Unfortunately, the same cannot be said regarding the dissemination of the widespread expert knowledge required for doing the proper evaluation of these studies, both from the peer reviewer aspect as well as the typical reader looking to learn more from AI publications. In this session, the Association for Pathology Informatics (API) will have key leaders in the digital pathology and AI space provide an overview of the clinical and technical questions one must ask when reading/reviewing these articles. Further, the editorial staff of the Journal of Pathology Informatics (JPI) will be present to comment specifically on what academic journals are looking for, and expecting, from AI and machine learning publications, focusing on common issues seen within the methods and results sections of these papers. Attendees will come away with not only a much better understanding of how to review these types of papers, but also how to submit their own publications in a more successful manner.