Artificial Intelligence in Pediatric Dentistry: A Systematic Review and Meta-Analysis

Area of Science:

• Pediatric Dentistry
• Artificial Intelligence
• Diagnostic Accuracy

Background:

• Artificial intelligence (AI) is increasingly vital in pediatric dentistry for enhancing diagnostics.
• AI applications include caries detection, risk prediction, and developmental assessments.

Purpose of the Study:

• To systematically review and meta-analyze the diagnostic performance of AI models in pediatric dental applications.
• To assess the clinical translation potential of current AI evidence.

Main Methods:

• Systematic search of major databases (PubMed, Scopus, Web of Science, Embase) following PRISMA-DTA guidelines.
• Inclusion of studies on AI diagnostic/predictive models in pediatric populations (≤18 years).
• Calculation of pooled sensitivity, specificity, and AUC using random-effects models, examining heterogeneity sources.

Main Results:

• Thirty-two studies included for qualitative synthesis; fifteen for quantitative analysis.
• Radiographic caries detection: pooled sensitivity 0.91, specificity 0.97, AUC 0.98.
• AI prediction models: pooled sensitivity 0.86, specificity 0.82, AUC 0.89; deep learning superior to traditional ML.

Conclusions:

• AI demonstrates significant potential to enhance diagnostic accuracy in pediatric dentistry.
• Clinical implementation is limited by insufficient external validation and methodological inconsistencies.
• Future research should focus on multicenter datasets, harmonized workflows, and explainable AI (XAI) for clinical translation.