AI-Enhanced Simulator Training Validates Laparoscopic Surgical Skills

Overview

A novel AI-based system using DeepLabCut was developed to objectively evaluate laparoscopic forceps manipulation during simulator training. The system demonstrated high tracking accuracy and stability across different exercises, supporting its feasibility for surgical skill assessment.

Background

Laparoscopic surgery offers significant benefits over open surgery, including reduced pain, infection risk, and faster recovery, making it the preferred approach for many procedures. However, variability in surgical skill among practitioners can impact patient outcomes, highlighting the need for effective training. Simulation training has proven effective in skill acquisition, but traditional methods rely heavily on subjective guidance. The COVID-19 pandemic further limited intraoperative training opportunities, prompting exploration of AI-powered systems to enhance surgical education.

Data Highlights

Key Findings

• The AI system, based on DeepLabCut, successfully tracked forceps keypoints with an average error of 9.2 pixels compared to manual annotations.
• Tracking stability was confirmed qualitatively at critical points such as the forceps tip, hinge, and stick during needle trail, ring transfer, and suturing exercises.
• The system maintained accuracy across different backgrounds and exercise types, demonstrating scalability.
• Real-time or near-real-time feedback is feasible using an iPad® for video capture and display, enhancing training usability.
• Distinctive movement patterns, such as hesitation or retries, were identifiable, potentially differentiating skill levels among surgeons.

Clinical Implications

This AI-powered evaluation tool can provide objective, quantitative feedback on laparoscopic skills, supplementing traditional subjective assessments. It offers a scalable and accessible method for surgical training, especially valuable when direct intraoperative teaching is limited. Incorporating such technology may improve skill acquisition and ultimately patient outcomes.

Conclusion

The study demonstrates the feasibility of using AI-based posture estimation to objectively quantify laparoscopic forceps movements during simulator training. This approach holds promise for enhancing surgical education through precise skill evaluation and feedback.

References

1. DeepLabCut: markerless pose estimation of user-defined body parts with deep learning -- Mathis et al. 2018