As surgeons undergo training, they typically require guidance from seasoned doctors who can help refine their technique. Now, a new artificial intelligence (AI) system may soon offer valuable insights into a surgeon's performance, potentially reducing the need for direct supervision.

The Surgical AI System (SAIS) developed by researchers at California Institute of Technology (Caltech, Pasadena, CA, USA) and Keck School of Medicine of University of Southern California (USC, Los Angeles, CA, USA) aims to deliver objective evaluations of a surgeon's performance, with the ultimate goal of enhancing their skills and, consequently, their patients' outcomes. By analyzing a video recording of a surgical procedure, SAIS can determine the type of surgery being performed and assess the surgeon's execution quality. The system was trained on a vast collection of video data annotated by medical professionals, evaluating surgeons' performances down to individual discrete motions, such as handling a needle or manipulating tissue.

Once trained, SAIS was employed to review and assess the performance of surgeons during a wide array of procedures captured on video from various hospitals. Researchers hope that SAIS will guide surgeons on which skills need further development. To enhance the tool's usefulness, the AI system was designed to justify its assessments, enabling it to inform surgeons of their skill level and offer detailed feedback, supported by specific video clips.

During initial tests, researchers discovered an unintended bias in SAIS, causing it to rate surgeons' skill levels inaccurately based on an analysis of their overall movements. To address this issue, the AI system was directed to concentrate solely on relevant aspects of the surgical video. While this approach reduced the bias, it did not completely eliminate it, and researchers continue to work on refining the system.

"In high stakes environments such as robotic surgery, it is not realistic for AI to replace human surgeons in the short term," said Anima Anandkumar, Bren Professor of Computing and Mathematical Sciences and senior author of the studies. "Instead, we asked how AI can safely improve surgical outcomes for the patients, and hence, our focus on making human surgeons better and more effective through AI."

"Human-derived surgical feedback is not presently objective nor scalable," added Andrew Hung, a urologist with Keck Medicine of USC and associate professor of urology at Keck School of Medicine of USC. "AI-derived feedback, such as what our system delivers, presents a major opportunity to provide surgeons actionable feedback." 

http://www.gzjiayumed.com/en/index.asp