Related Experiment Video
Updated: Jun 9, 2025
05:57
A Teleoperated Robotic System-Assisted Percutaneous Transiliac-Transsacral Screw Fixation Technique
Published on: January 6, 2023
2.2K
Deep learning prediction of error and skill in robotic prostatectomy suturing
N Sirajudeen1, M Boal1,2,3, D Anastasiou1,4
1Wellcome/ESPRC Centre for Interventional Surgical Sciences (WEISS), University College London (UCL), London, UK.
Surgical Endoscopy
|October 21, 2024
Summary
This study validates artificial intelligence (AI) models for assessing surgical skill and errors in robotic suturing. The AI approach offers a promising, objective method to improve surgical training and patient outcomes.
Area of Science:
- Robotics
- Surgical Technology
- Artificial Intelligence
Background:
- Manual assessment of surgical skill and errors is subjective and time-consuming.
- Artificial intelligence (AI) models are being developed for automated surgical assessment.
- This study aims to validate AI models for skill rating and error annotation in suturing gestures.
Purpose of the Study:
- To validate AI-driven skill rating and error detection in robotic surgery.
- To establish a benchmark for evaluating AI models in surgical skill assessment.
- To inform the development of objective automated assessment tools.
Main Methods:
- Utilized the SAR-RARP50 dataset of Robotic-Assisted Radical Prostatectomy (RARP) suturing videos.
- Annotated videos at the gesture level using Objective Clinical Human Reliability Analysis (OCHRA) as ground truth.
- Trained and tested vision-based deep learning models to estimate skill and identify errors.
Main Results:
- Demonstrated strong inter-rater reliability (r=0.70-0.89) and correlation (r=0.92) between objective assessment tools.
- AI skill estimation showed moderate correlation with established metrics (Spearman's rho 0.36-0.37).
- Error prediction models achieved a mean absolute precision of 37.14% and AUC of 65.10%.
Conclusions:
- This study is the first to apply detailed error detection and deep learning to real robotic surgical videos.
- The benchmark evaluation provides a foundation for AI in automated technical skill assessment.
- This approach shows promise for advancing objective surgical skill evaluation and reducing errors.
