Improving microsurgical suture training with automated phase recognition and skill assessment via deep learning

Area of Science:

• Medical Technology
• Artificial Intelligence in Medicine
• Surgical Education

Background:

• Microsurgical suturing requires extensive training and high precision.
• Objective assessment of surgical skill is crucial for effective training.

Purpose of the Study:

• To develop and evaluate a deep learning model for automated phase recognition and skill assessment in microsurgical suturing.
• To improve the efficiency and objectivity of surgical training programs.

Main Methods:

• Utilized modified Long-Range Recurrent Convolutional Networks (LRCNs) to process and segment microsurgical videos.
• Applied data augmentation and a skipping window strategy for frame selection.
• Trained and tested models on datasets of novice and expert surgeons.

Main Results:

• The deep learning models accurately classified surgical phases and distinguished between skill levels.
• Analysis of confidence and time spent in phases correlated with surgeon expertise.
• Models demonstrated good generalization capabilities across different datasets.

Conclusions:

• The proposed deep learning approach shows significant promise for objective skill assessment in microsurgery.
• This technology can enhance surgical training, leading to improved patient outcomes and personalized learning pathways.