Objective Performance Indicators During Robotic Right Colectomy Differ According to Surgeon Skill
Mishal Gillani,Manali Rupji,Terrah J Paul Olson,Patrick Sullivan,Virginia O Shaffer,Glen C Balch,Mallory C Shields,Yuan Liu,Seth A Rosen
DOI: https://doi.org/10.1016/j.jss.2024.07.103
Abstract:Introduction: Surgeon assessment tools are subjective and nonscalable. Objective performance indicators (OPIs), machine learning-enabled metrics recorded during robotic surgery, offer objective insights into surgeon movements and robotic arm kinematics. In this study, we identified OPIs that significantly differed across expert (EX), intermediate (IM), and novice (NV) surgeons during robotic right colectomy. Methods: Endoscopic videos were annotated to delineate 461 surgical steps across 25 robotic right colectomies. OPIs were compared among two EX, two IM, and eight NV surgeons during mesenteric dissection, vascular pedicle ligation, right colon and hepatic flexure mobilization, and preparation of the proximal and distal bowel for transection. Results: Compared to NV's, EX's exhibited greater velocity, acceleration and jerk for camera, dominant, nondominant, and third arms across all steps. Compared to NV's, IM's exhibited more arm swaps and master clutch use, higher camera-related metrics (movement, path length, moving time, velocity, acceleration, and jerk), greater dominant wrist pitch and nondominant wrist articulations (roll, pitch, and yaw), longer dominant and nondominant arm path length, and higher velocity, acceleration and jerk for dominant, nondominant, and third arms across all steps. Compared to NV's, EX/IM surgeons utilized more arm swaps, higher camera-related metrics (movement, path length, velocity, acceleration, and jerk), longer nondominant arm path length, and greater velocity, acceleration and jerk for dominant, nondominant, and third arms across all steps. Conclusions: We report OPIs that discriminate EX, IM, and NV surgeons during RRC. This study is the first to demonstrate feasibility of using OPIs as an objective, scalable way to classify surgeon skill during RRC steps.