Dental Caries Detection and Classification in CBCT Images Using Deep Learning.

Area of Science:

• Artificial Intelligence in Dentistry
• Radiology and Imaging
• Oral Health Diagnostics

Background:

• Dental caries diagnosis from cone beam computed tomography (CBCT) images is crucial for treatment planning.
• Current diagnostic methods can be subjective and time-consuming.
• The application of deep learning for dental caries detection in CBCT is an emerging field.

Purpose of the Study:

• To evaluate the accuracy of deep learning algorithms in diagnosing tooth caries.
• To assess the ability of deep learning to classify the extension and location of dental caries in CBCT images.
• To establish the novelty of using deep learning for dental caries detection in CBCT.

Main Methods:

• A dataset of 382 carious and 403 noncarious molar teeth from CBCT images was utilized.
• The dataset was split into development (training/validation) and test sets.
• A multiple-input convolutional neural network (CNN) was employed to analyze axial, sagittal, and coronal CBCT views for caries detection and classification.

Main Results:

• The CNN achieved high diagnostic accuracy for caries detection: 95.3% for carious and 94.8% for noncarious teeth.
• Sensitivity, specificity, and F1 scores for caries detection were above 92% for carious and 94% for noncarious teeth.
• The network demonstrated high performance in classifying caries extension and location.

Conclusions:

• Deep learning models accurately identify and classify dental caries in CBCT images with high sensitivity and specificity.
• This technology can significantly assist dental practitioners in improving diagnosis and treatment planning.
• AI-powered dental caries detection can enhance diagnostic accessibility, particularly in areas with a shortage of dentists.