Artificial Intelligence in Paediatric and Adolescent Fracture Detection: A Systematic Review and Meta-Analysis

Area of Science:

• Pediatric Radiology
• Medical Imaging
• Artificial Intelligence in Medicine

Background:

• Pediatric fractures are common but challenging to diagnose radiographically due to unique skeletal anatomy.
• Diagnostic errors in pediatric fracture detection are significant, impacting patient care.
• Developing advanced tools is crucial for improving diagnostic accuracy in pediatric radiology.

Purpose of the Study:

• To systematically review and meta-analyze the accuracy and efficiency of artificial intelligence (AI) in detecting pediatric fractures.
• To compare AI's diagnostic performance against human interpretation in children and adolescents.
• To assess AI's potential as an adjunct tool in pediatric fracture diagnosis.

Main Methods:

• Systematic review and meta-analysis following PRISMA guidelines.
• Searched PubMed, EMBASE, and Web of Science for studies published between 2019-2024.
• Included 11 studies evaluating AI for appendicular skeletal fractures in patients under 21 years.

Main Results:

• Standalone AI demonstrated statistically significantly higher sensitivity than human interpretation (mean difference: 0.04, p = 0.0005).
• AI showed non-inferior specificity compared to human interpretation.
• AI-assisted diagnosis significantly improved clinician sensitivity (mean difference: 0.07, p = 0.003).

Conclusions:

• AI exhibits high diagnostic performance for pediatric fractures.
• AI serves as a promising tool to enhance clinical efficiency and accuracy in diagnosing fractures in children.
• Further large-scale, multi-center prospective trials are needed to validate AI's real-world applicability.