A Comprehensive X-ray Dataset for Pediatric Ulna and Radius Fractures Analysis

Area of Science:

• Medical Imaging
• Artificial Intelligence in Medicine
• Pediatric Orthopedics

Background:

• Pediatric forearm fractures are common childhood injuries.
• Limited availability of standardized datasets hinders AI development and clinical validation.
• Lack of accessible data impedes research in pediatric fracture classification.

Purpose of the Study:

• To introduce the Pediatric Ulna and Radius Fractures (PediURF) dataset, a novel, publicly available resource.
• To facilitate the development and benchmarking of deep learning models for pediatric fracture classification.
• To provide a foundation for AI-driven clinical training and validation in pediatric orthopedics.

Main Methods:

• Compilation of over 10,000 de-identified pediatric forearm fracture images.
• Expert radiologist annotation and categorization into proximal, midshaft, and distal fracture types.
• Development and validation of URFNet, a dual-view classification model integrating anteroposterior and lateral radiographic views.

Main Results:

• The PediURF dataset offers a comprehensive resource for AI research.
• The proposed URFNet model achieved superior performance in classifying pediatric forearm fractures compared to other models.
• The dataset supports the development of robust deep learning algorithms for fracture diagnosis.

Conclusions:

• The PediURF dataset is a valuable, first-of-its-kind resource for advancing AI in pediatric fracture analysis.
• The URFNet model demonstrates the potential of dual-view classification for improved diagnostic accuracy.
• This work lays the groundwork for future deep learning applications in pediatric orthopedic imaging.