Abstract
OBJECTIVE: This study aims to present the initial clinical experience and evaluate the feasibility and safety of an artificial intelligence agent-based robotic-assisted bronchoscopy system for preoperative dye localization of pulmonary nodules before video-assisted thoracoscopic surgery. METHODS: This single-center, retrospective study analyzed 15 patients with peripheral pulmonary nodules who underwent preoperative dye localization using the artificial intelligence agent-based robotic-assisted bronchoscopy system (Polaris, LungHealth MedTech Company). Procedural efficacy was assessed by navigation success rate, localization time, and resection success rate. Safety was evaluated based on the incidence of complications. RESULTS: All 15 nodules (mean diameter: 1.03 ± 0.28 cm) were successfully navigated (100% navigation success). Dye localization was successful in 14 cases (93.3%), with 1 nodule resected anatomically. The median localization time was 10.0 minutes (interquartile range, 7.0-12.5). All nodules were completely resected with negative margins. No procedural complications, such as pneumothorax or hemorrhage, occurred. CONCLUSIONS: The artificial intelligence agent-based robotic-assisted bronchoscopy-assisted dye localization technique demonstrates high precision and safety for guiding video-assisted thoracoscopic surgery resection of peripheral pulmonary nodules. These promising results support its potential as a viable alternative to existing localization methods and warrant further validation in larger randomized trials to establish its clinical value.