++
HIGHLIGHTS
Three major areas of surgical training in which artificial intelligence is being explored include intraoperative feedback, postoperative analysis, and simulation training.
The use of artificial intelligence for surgical training attempts to overcome limitations of manual assessment, namely that manual assessment can be very time consuming and expensive and that manual assessment is inherently subjective because it relies on raters observing and evaluating performance.
Automated assessment can be performed by replicating well-validated manual scoring systems and then augmenting those systems with quantitative metrics of performance captured by sensors.
++
Surgery is a complex task that involves both technical and decision-making skills. Acquiring surgical competence is critical for ensuring an optimal outcome. A systematic review of the incidence of adverse events during surgery has shown it to be over 14.4%, and 14% of these adverse events were either fatal or severe. Preventable adverse events were present in 5.2% of the cases, highlighting the role of surgical expertise in reducing the potentially avoidable harm to patients’ health.1
++
Gaining surgical competence is a lengthy, multifaceted process that takes many years to complete.2 Surgical residency programs in North America, of which the current model was originally inspired by Halsted and later Churchill as a structured apprenticeship model, involve immersing surgical trainees in a clinical environment under graduated supervision. Over time, the training methodology has evolved to include simulation-based training,3 establishment of skills laboratories in residency programs, and the development of a national skills curriculum jointly by the American College of Surgeons (ACS) and the Association for Program Directors in Surgery (APDS). The ACS/APDS curriculum consists of 3 phases: (1) basic/core skills and tasks, (2) advanced procedures, and (3) team-based training.4 Surgical education has benefited tremendously from recent advances in technology, such as artificial intelligence (AI), computer vision, and virtual/augmented reality. Some of the most common applications of this technology in surgical training include the following:
++
Intraoperative feedback: As one of the most powerful teaching tools, objective feedback during the operation can significantly help the residents learn the technical skills through real-time performance assessment.
Postoperative analysis: Recorded videos from the operation can be a valuable educational resource.
Simulation training: Skills can be developed by training on various types of simulators such as benchtop part-task trainers, full procedural simulators on a cadaver, or virtual reality–based surgical trainers.
++
To take full advantage of any of these methods, the performance of the trainee needs to be evaluated based on some proficiency criteria.5 For this purpose, several scoring systems have been developed for assessment. For instance, the Objective Structured Assessment of Technical Skills (OSATS)6 is a widely used evaluation tool and has been extensively studied and validated for the manual grading of surgical skills in the operating room for different procedures.7 The OSATS global rating scale is based on the cumulative performance score on 7 domains ...