Artificial intelligence in urology training Enhancing annotation, feedback, and evaluation in robotic, laparoscopic,...

The integration of artificial intelligence (AI) into surgical training is rapidly evolving, driven by advancements in machine learning. This review aimed to map the current landscape of AI's educational applications in urology.

A systematic search of MEDLINE, PubMed, Embase, Cochrane, Scopus, and Engineering Village identified studies exploring AI applications in video-based surgical education and assessment. Search terms included AI, urologic procedures, and training/assessment components, and results were screened in Covidence^®. AI applications involving urologic procedures were included. For every study, two reviewers independently conducted screening. Data were synthesized thematically to evaluate AI's application in urology training.

Our search yielded 2774 studies, of which 59 relevant ones were identified. AI was most frequently applied with robotic-assisted radical prostatectomy (RARP), followed by robotic-assisted partial nephrectomy (RAPN). AI applications were broadly categorized into three domains: 1) annotation, where key anatomy and instruments from procedural videos are labelled; 2) feedback, such as recognizing surgical phases or monitoring surgical events; and 3) evaluation, where the surgical gestures are recognized or evaluated to stratify skill level and predict patient outcomes.

The emergence of AI use in urologic procedures underscores its transformative potential in procedural education and training. AI has wide applications in annotation, feedback, and assessment across different procedures. While prostatectomy dominates in the literature, the adaptability of AI frameworks exists across other urologic procedures. New, commercially available tools demonstrate promising results, making them potentially beneficial additions to urology training programs. Future efforts should focus on multicentric collaboration and longitudinal skill assessments.