Related Experiment Video
Updated: Jul 17, 2025
Fracture Apparatus Design and Protocol Optimization for Closed-stabilized Fractures in Rodents
Published on: August 14, 2018
Developing an artificial intelligence diagnostic tool for paediatric distal radius fractures, a proof of concept
S Aryasomayajula1, C B Hing2, M Siebachmeyer2
1Kingston University, UK.
Insights
Convolutional Neural Network (CNN) software can diagnose pediatric wrist fractures from radiographs with 85% accuracy. This AI tool shows promise for improving diagnostic accuracy and streamlining care for children with suspected fractures.
Area of Science:
- Artificial Intelligence in Medicine
- Medical Imaging Analysis
- Pediatric Radiology
Background:
- Pediatric bone fractures are common, with wrist fractures posing diagnostic challenges due to growth plate complexities.
- A significant percentage of pediatric fractures are initially missed on radiographic examination.
Purpose of the Study:
- To develop and evaluate a Convolutional Neural Network (CNN) for detecting wrist fractures in pediatric patients.
- To assess the diagnostic accuracy of the CNN model compared to radiologist reports.
Main Methods:
- A CNN model was trained on 5,000 distal radius radiographs from children under 19.
- Transfer learning using a VGG16 model and data augmentation were employed to enhance performance.
- Hyperparameter tuning and comparison with radiology reports determined diagnostic accuracy.
Main Results:
- The final dataset comprised 15,498 images after augmentation.
- The CNN model achieved a diagnostic test accuracy of 85% after 20 epochs of training.
- The model demonstrated feasibility in identifying fractures in pediatric wrist radiographs.
Conclusions:
- A CNN-based approach is a feasible tool for improving the accuracy of diagnosing pediatric wrist fractures.
- This technology has the potential to reduce unnecessary hospital visits and optimize healthcare resource allocation.
Introduction:
In the UK, 1 in 50 children sustain a fractured bone yearly, yet studies have shown that 34% of children sustaining an injury do not have a visible fracture on initial radiographs. Wrist fractures are particularly difficult to identify because the growth plate poses diagnostic challenges when interpreting radiographs.
Methods:
We developed Convolutional Neural Network (CNN) image recognition software to detect fractures in radiographs from children. A consecutive data set of 5,000 radiographs of the distal radius in children aged less than 19 years from 2014 to 2019 was used to train the CNN. In addition, transfer learning from a VGG16 CNN pretrained on non-radiological images was applied to improve generalisation of the network and the classification of radiographs. Hyperparameter tuning techniques were used to compare the model with the radiology reports that accompanied the original images to determine diagnostic test accuracy.
Results:
The training set consisted of 2,881 radiographs with a fracture and 1,571 without; 548 radiographs were outliers. With additional augmentation, the final data set consisted of 15,498 images. The data set was randomly split into three subsets: training (70%), validation (10%) and test (20%). After training for 20 epochs, the diagnostic test accuracy was 85%.
Discussion:
A CNN model is feasible in diagnosing paediatric wrist fractures. We demonstrated that this application could be utilised as a tool for improving diagnostic accuracy. Future work would involve developing automated treatment pathways for diagnosis, reducing unnecessary hospital visits and allowing staff redeployment to other areas.
More Related Videos
07:35Creating Rigidly Stabilized Fractures for Assessing Intramembranous Ossification, Distraction Osteogenesis, or Healing of Critical Sized Defects
Published on: April 11, 2012
08:20A Reliable and Reproducible Critical-Sized Segmental Femoral Defect Model in Rats Stabilized with a Custom External Fixator
Published on: March 24, 2019