Assessment of Apical Patency in Permanent First Molars Using Deep Learning on CBCT-Derived Pseudopanoramic Images: A...

Area of Science:

• Dentistry
• Artificial Intelligence
• Medical Imaging

Background:

• Root development and apical closure are crucial in endodontics, dental trauma, and age estimation.
• Digital health and AI offer new tools for dental diagnostics.

Purpose of the Study:

• To evaluate the effectiveness of deep learning Convolutional Neural Networks (CNNs) for automated assessment of apical region status in permanent first molars.
• To highlight a digital health application of AI in dentistry.

Main Methods:

• Retrospective analysis of 262 Cone Beam Computed Tomography (CBCT) scans.
• 147 anonymized dental images were cropped from pseudopanoramic radiographs.
• CNN model performance was evaluated using accuracy, precision, recall, F1-score, and ROC curves with AUC.

Main Results:

• CNNs achieved a precision, recall, and F1-score of 0.79 for open roots and 0.81 for closed roots.
• The macro average across all metrics was 0.80.
• Overall accuracy and Area Under the Curve (AUC) were also 0.80.

Conclusions:

• CNNs demonstrate significant potential for effectively assessing apical patency.
• AI-based image analysis from pseudopanoramic radiographs is a viable tool in dentistry.