Related Experiment Video
Updated: Jun 26, 2025
07:56
A Full Skin Defect Model to Evaluate Vascularization of Biomaterials In Vivo
Published on: August 28, 2014
12.2K
Assessment of changes in vessel area during needle manipulation in microvascular anastomosis using a deep
Minghui Tang1,2,3, Taku Sugiyama4,5, Ren Takahari6
1Department of Diagnostic Imaging, Hokkaido University Faculty of Medicine and Graduate School of Medicine, Sapporo, Japan.
Neurosurgical Review
|May 9, 2024
Summary
A new deep learning algorithm objectively assesses surgical skill by analyzing vessel area changes during microvascular anastomosis, distinguishing expert from novice performance. This technology aids surgical education and patient safety.
Area of Science:
- Microsurgical techniques and surgical skill assessment.
- Application of artificial intelligence in medical training and evaluation.
Background:
- Successful microvascular anastomosis relies on precise needle handling to prevent vessel deformation.
- Objective methods for evaluating surgical skill, particularly "respect for tissue," are lacking.
- Current assessment often relies on subjective evaluations rather than quantifiable metrics.
Purpose of the Study:
- To develop a deep learning-based semantic segmentation algorithm for assessing vessel area changes during microvascular anastomosis.
- To objectively evaluate surgical skill related to tissue handling using video analysis.
- To differentiate between expert and novice surgeons based on quantitative video parameters.
Main Methods:
- A ResNet-50 deep learning model was trained using artificial vessel videos of microvascular anastomosis.
- The algorithm quantified vessel area changes, calculating coefficient of variation (CV-VA) and relative change per unit time (ΔVA).
- Tissue deformation errors (TDEs) were identified based on a ΔVA threshold and compared between expert and novice groups.
Main Results:
- The semantic segmentation model achieved high validation accuracy (99.1%) and Intersection over Union (0.93).
- Expert surgeons exhibited significantly lower CV-VA and ΔVA, indicating less vessel deformation.
- Experts made fewer TDEs and completed the task faster, with combined parameters showing complete discriminative power.
Conclusions:
- Deep learning-based semantic segmentation offers a novel, objective method for assessing microsurgical performance by analyzing vessel area dynamics.
- This approach provides quantitative metrics for "respect for tissue" during microvascular anastomosis.
- The developed algorithm can be integrated into computer-aided devices to enhance surgical education and improve patient safety.
