Automated VR Assessment of Suturing Skills in Pediatric Robotic Surgery

Overview

This study presents a virtual reality (VR) simulator for automated assessment of suturing skills in pediatric robotic surgery, specifically neonatal esophageal atresia repair. Automated scoring based on a reformulated suturing checklist showed substantial agreement with expert video-based evaluation, demonstrating feasibility for objective and repeatable performance assessment.

Background

Esophageal atresia is a rare congenital anomaly requiring complex neonatal surgical repair, with limited opportunities for surgeons to gain hands-on robotic suturing experience due to case rarity and technical difficulty. Robot-assisted surgery offers enhanced dexterity and visualization but is not widely adopted in neonates. VR simulation provides a controlled environment for surgical training and enables automated performance evaluation, which can complement subjective expert assessments and support skill acquisition in pediatric robotic surgery.

Data Highlights

Key Findings

• The VR simulator accurately reproduced a neonatal robotic suturing scenario with real-time tool and tissue interaction.
• Automated assessment was based on 11 measurable checklist items reformulated from a validated suturing checklist.
• Automated scoring achieved 67.3% accuracy and high precision (0.933) compared to expert video-based evaluation.
• Lower recall (0.560) indicated conservative scoring with fewer false positives.
• Higher agreement was found for metrics with clear geometric definitions; lower agreement occurred for parameters difficult to judge visually, such as needle insertion angle.
• The approach enables objective, repeatable, and observer-independent evaluation, reducing subjectivity inherent in manual video assessment.

Clinical Implications

Automated VR-based assessment can provide pediatric surgeons with objective, consistent feedback on robotic suturing skills, potentially enhancing training efficiency and skill acquisition. This method may supplement traditional expert evaluations, especially for parameters challenging to assess visually, and support real-time performance monitoring during simulation. Ultimately, such tools could improve surgical proficiency in rare and complex neonatal procedures.

Conclusion

The study demonstrates the feasibility of automated VR-based assessment for pediatric robotic suturing, showing substantial agreement with expert evaluations and offering a promising tool for objective surgical skills training. Future work will expand assessment to additional suturing phases for comprehensive performance evaluation.

References

1. Moorthy et al. 2003 -- Development and validation of a checklist for assessment of technical skills in laparoscopic surgery
2. Deie et al. 2020 -- Neonatal chest model for robotic esophageal atresia repair
3. SmartArm robotic platform documentation