An Artificial Intelligence Coaching System to Improve Surgical Performance in Urologic Endoscopy - PROJECT SUMMARY-ABSTRACT Nephrolithiasis, or kidney stone disease, affects over 10% of the U.S. adult population, leading to significant healthcare burdens. As the primary treatment, urologic endoscopy via ureteroscopy (URS) varies considerably in outcomes and complication rates among surgeons, revealing a critical opportunity for surgical care quality measurement and improvement. This project proposes the development and validation of an Artificial Intelligence (AI) Surgical Coach system designed to enhance surgical performance and patient outcomes during URS procedures. Utilizing machine learning (ML) algorithms to analyze video data and neuroergonomic metrics from the operating room (OR), we aim to create a model predictive of surgical efficiency and clinical outcomes. Specifically, this project will: A) Model real-life intraoperative URS performance through surgical videos unobtrusively captured in the OR, expecting that ML models will be predictive of surgical efficiency and patient outcomes. B) Conduct high-fidelity simulations to establish the construct validity of our ML models across various levels of surgical expertise. C) Develop and evaluate the usability of an explainable AI (xAI) interface within the AI Surgical Coach, providing automated personalized feedback for surgeons' skills enhancement. D) Evaluate the effectiveness of the AI Surgical Coach through a randomized controlled trial (RCT), measuring the impact on surgeons’ URS skills improvement and sustainment. Our multidisciplinary team brings together expertise from surgery, medical education, human factors, and computer science to execute this innovative research. The anticipated outcomes include an AI-powered surgical coaching system that objectively assesses and provides actionable feedback, improving surgical performance and patient care in urologic endoscopy. The project will set the stage for a future multicenter R01 clinical trial and align with the National Institute of Biomedical Imaging and Bioengineering (NIBIB) - Division of Discovery Science and Technology (DDST) program's interest in AI-based virtual coaching for performance improvement in medical procedures.
