To evaluate the feasibility of automated assessment of suturing skills in pediatric robotic surgery using a virtual reality (VR) simulator.
Key Findings:
Automated assessment achieved an accuracy of 67.3% compared to expert evaluations.
High precision (0.933) and moderate recall (0.560) were observed, indicating substantial agreement.
The automated method demonstrated conservative scoring behavior with a low false-positive rate.
Interpretation:
The VR-based automated assessment provides a more objective evaluation of suturing performance than traditional video-based methods, particularly for metrics difficult to judge from 2D video.
Limitations:
The study focused only on the needle positioning and driving phases, excluding other critical phases like knot tying.
The accuracy of automated assessment may vary for metrics that are challenging to evaluate visually.
Conclusion:
VR-based automated skills assessment can enhance the training of pediatric robotic surgeons by providing objective, repeatable evaluations and real-time feedback, with potential for future expansion to other surgical tasks.
