Combining Artificial Intelligence and Indocyanine Green-Dyed Gauze for Dual-Navigation in Minimally Invasive Gastrointestinal Surgery: A Preliminary Study

This preliminary study evaluated the feasibility of a dual-navigation strategy that integrates artificial intelligence (AI)-assisted identification of safe dissection planes with indocyanine green (ICG)-dyed gauze-based localization of anatomical targets during minimally invasive gastrointestinal surgery. Fluorescence imaging is widely used for intraoperative guidance, and artificial intelligence-assisted prediction can contribute to enhancing surgical decision-making and reducing adverse events. The feasibility of this system was evaluated in 39 patients undergoing laparoscopic or robot-assisted gastrointestinal surgery. Data from the Eureka surgical AI system, used in conjunction with indocyanine green-dyed gauze, were retrospectively analyzed to identify loose connective tissue, indicating a safe dissection plane. Eureka successfully predicted loose connective tissue to identify safe dissection planes during tissue separation. In 23 cases, the gauze was more readily detected under near-infrared fluorescence than under white light, thereby serving as a precise intraoperative marker. These preliminary findings suggest that combining AI-assisted dissection-plane recognition with fluorescence-guided target localization is feasible and may support real-time orientation during minimally invasive gastrointestinal surgery.