Related Experiment Video
Updated: Jun 23, 2025
08:27
Novel Triple-Loop Technique for Suturing TFCC Injuries without Transosseous Tunnel
Published on: May 23, 2025
48
A deep learning-based algorithm for automatic detection of perilunate dislocation in frontal wrist radiographs.
Negin Majzoubi1, Rémi Allègre1, Cédric Wemmert1
1ICube UMR 7357 Université de Strasbourg, CNRS, 300 Boulevard Sébastien Brant, F-67412 Illkirch-Graffenstaden, Strasbourg, France.
Hand Surgery & Rehabilitation
|June 23, 2024
Summary
This study introduces a Deep Learning algorithm for automatically detecting perilunate dislocation in wrist X-rays. The AI model shows high accuracy, improving diagnosis of this wrist injury.
Area of Science:
- Radiology
- Artificial Intelligence
- Orthopedics
Background:
- Perilunate dislocation is a significant wrist injury requiring accurate diagnosis.
- Current diagnostic methods rely on radiograph interpretation, which can be challenging.
Purpose of the Study:
- To develop and evaluate a Deep Learning algorithm for automated detection of perilunate dislocation.
- To assess the algorithm's performance on anteroposterior wrist radiographs.
Main Methods:
- Two YOLOv8 deep neural network models were trained on 374 wrist radiographs.
- One model detected the carpal region; a second segmented critical areas for dislocation detection.
- Ensemble averaging was used to assign a probability of normal or pathological findings.
Main Results:
- The algorithm achieved an overall F1-score of 0.880.
- It demonstrated high performance in normal (F1-score 0.928, precision 1.0) and pathological (F1-score 0.833, recall 1.0) subgroups.
- The study indicates improved diagnosis through automated radiograph analysis.
Conclusions:
- Deep Learning algorithms can effectively detect perilunate dislocation in anteroposterior wrist radiographs.
- Automated analysis shows potential to enhance diagnostic accuracy and efficiency for this wrist injury.
- Further validation in larger datasets is warranted.
