Related Experiment Video
Updated: May 27, 2025
A Method to Estimate Cadaveric Femur Cortical Strains During Fracture Testing Using Digital Image Correlation
Published on: September 14, 2017
Impact of deep learning on pediatric elbow fracture detection: a systematic review and meta-analysis
Le Nguyen Binh1,2,3,4, Nguyen Thanh Nhu1,5, Pham Thi Uyen Nhi6
1College of Medicine, Taipei Medical University, Taipei, 11031, Taiwan.
Insights
Deep learning (DL) models show high accuracy in detecting pediatric elbow fractures, with pooled sensitivity of 0.93 and specificity of 0.89. Using ResNet architectures and expert-guided preprocessing enhances diagnostic performance for these common childhood injuries.
Area of Science:
- Artificial Intelligence in Medicine
- Pediatric Orthopedics
- Medical Imaging Analysis
Background:
- Pediatric elbow fractures are frequent injuries in children.
- Artificial intelligence (AI), specifically deep learning (DL), offers potential for improved fracture diagnosis.
- Systematic evaluation of DL model performance in pediatric elbow fracture detection is needed.
Purpose of the Study:
- To systematically evaluate the diagnostic performance of deep learning (DL) models for pediatric elbow fractures.
- To determine the pooled sensitivity, specificity, and AUC of DL models in detecting these fractures.
- To identify factors influencing DL model performance, such as preprocessing and architecture.
Main Methods:
- Comprehensive literature search of PubMed, EMBASE, and IEEE Xplore up to October 20, 2023.
- Inclusion of studies using DL models for elbow fracture detection in patients aged 0-16 years.
- Meta-analysis of extracted performance metrics: sensitivity, specificity, and area under the curve (AUC).
Main Results:
- Six studies met inclusion criteria for meta-analysis from 22 identified studies.
- Pooled sensitivity for DL models was 0.93 (95% CI: 0.91-0.96); pooled specificity was 0.89 (95% CI: 0.85-0.92).
- Pooled AUC was 0.95 (95% CI: 0.93-0.97), with performance influenced by preprocessing and model architecture.
Conclusions:
- Deep learning models demonstrate high accuracy in diagnosing pediatric elbow fractures.
- Recommended optimal performance through ResNet backbone architectures and expert-supervised manual preprocessing.
- DL holds significant promise for enhancing the accuracy and efficiency of pediatric fracture diagnosis.
Objectives:
Pediatric elbow fractures are a common injury among children. Recent advancements in artificial intelligence (AI), particularly deep learning (DL), have shown promise in diagnosing these fractures. This study systematically evaluated the performance of DL models in detecting pediatric elbow fractures.
Materials And Methods:
A comprehensive search was conducted in PubMed (Medline), EMBASE, and IEEE Xplore for studies published up to October 20, 2023. Studies employing DL models for detecting elbow fractures in patients aged 0 to 16 years were included. Key performance metrics, including sensitivity, specificity, and area under the curve (AUC), were extracted. The study was registered in PROSPERO (ID: CRD42023470558).
Results:
The search identified 22 studies, of which six met the inclusion criteria for the meta-analysis. The pooled sensitivity of DL models for pediatric elbow fracture detection was 0.93 (95% CI: 0.91-0.96). Specificity values ranged from 0.84 to 0.92 across studies, with a pooled estimate of 0.89 (95% CI: 0.85-0.92). The AUC ranged from 0.91 to 0.99, with a pooled estimate of 0.95 (95% CI: 0.93-0.97). Further analysis highlighted the impact of preprocessing techniques and the choice of model backbone architecture on performance.
Conclusion:
DL models demonstrate exceptional accuracy in detecting pediatric elbow fractures. For optimal performance, we recommend leveraging backbone architectures like ResNet, combined with manual preprocessing supervised by radiology and orthopedic experts.
