Related Experiment Video
Updated: Aug 5, 2026
06:45
Automated Joint Space Detection Improves Bone Segmentation Accuracy
Published on: November 28, 2025
Automatic Classification of Anterior Talofibular Ligament Based on 2D Convolutional Neural Network
Feng Li1,2, Xiao-Shan Wang3, Ting Li4
1Department of Radiology, Tongji Hospital,Tongji Medical College, Huazhong University of Science and Technology, Wuhan, 430030, China.
Current Medical Science
|August 3, 2026
Summary
Convolutional neural networks (CNNs) show feasibility in automatically classifying anterior talofibular ligament (ATFL) injuries on MR images. This AI approach offers efficient and accurate ATFL evaluation, comparable to human radiologists.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Orthopedics
Background:
- Anterior talofibular ligament (ATFL) injuries are common ankle sprains.
- Accurate diagnosis of ATFL tears on MR images is crucial for effective treatment.
- Manual analysis of MR images can be time-consuming and subjective.
Purpose of the Study:
- To assess the feasibility of 2D CNNs for automatic ATFL classification on MR images.
- To develop and validate AI models for ATFL segmentation and tear detection.
- To compare the performance of AI models with junior radiologists.
Main Methods:
- Training a YOLOv11 model for ATFL segmentation on 560+96 MR images.
- Validating the segmentation model using Dice Similarity Coefficient (DSC).
- Training a 2D ResNet model for ATFL tear classification (normal, partial, total) on 1,103 MR images.
- Testing both models on an independent dataset of 420 images.
Main Results:
- The YOLOv11 segmentation model achieved a median DSC of 0.95.
- The automated workflow demonstrated 92.6% overall accuracy on the test set.
- Sensitivity (95.0%) and specificity (93.6%) for abnormal ATFL detection were comparable to a junior radiologist.
- Automated classification was significantly faster (3 minutes vs. 14 minutes).
Conclusions:
- 2D CNNs are feasible for automatic ATFL segmentation and classification on MR images.
- AI-powered analysis offers a potentially efficient and accurate tool for ATFL evaluation.
- This technology could aid in the timely diagnosis and management of ATFL injuries.
