Dental age prediction from panoramic radiographs using machine learning techniques.

Area of Science:

• Pediatric Dentistry
• Artificial Intelligence in Healthcare
• Radiographic Analysis

Background:

• Dental age (DA) estimation is crucial in pediatric dentistry for growth assessment and treatment planning.
• Conventional DA methods are subjective and prone to variability.
• Automated approaches are needed to improve objectivity and efficiency.

Purpose of the Study:

• To develop and evaluate a deep learning (DL) model for automated dental age estimation.
• To assess the accuracy and interpretability of the DL model in classifying dental age groups.
• To explore the clinical utility of an AI tool for pediatric dental practice.

Main Methods:

• A dataset of 550 pediatric panoramic radiographs (ages 3-14) was used.
• The YOLOv11n-cls model was trained on 11 dental age groups.
• Data augmentation and AdamW optimizer were employed; performance evaluated using Top-1/Top-5 accuracy and Grad-CAM for interpretability.

Main Results:

• The DL model achieved 92.6% Top-1 and 99.5% Top-5 accuracy on the validation set.
• High performance was maintained on an independent test set, with most errors between adjacent age groups.
• Grad-CAM visualizations highlighted clinically relevant features, confirming model interpretability.

Conclusions:

• Deep learning, specifically the YOLOv11 model, demonstrates high performance for pediatric dental age prediction.
• The AI tool provides fast, accurate, and interpretable dental age classification.
• This AI tool is a promising adjunct for clinical integration in pediatric dentistry.