Enhancing Periodontal Treatment Through the Integration of Deep Learning-Based Detection with Bayesian Network Models

Area of Science:

• Dentistry
• Artificial Intelligence
• Medical Informatics

Background:

• Periodontal disease diagnosis relies on clinical examination and radiographic interpretation.
• Clinical decision support systems can improve diagnostic accuracy and treatment planning.
• Deep learning shows promise in analyzing medical images for disease detection.

Purpose of the Study:

• To develop and validate a Bayesian network (BN) clinical decision support model for periodontal care.
• To integrate deep learning for periodontal disease detection within the BN model.
• To assess the model's accuracy in treatment plan recommendations.

Main Methods:

• A Bayesian network (BN) model was constructed using clinical data.
• Deep learning, specifically Faster R-CNN, was employed for analyzing radiographic images to detect periodontal disease.
• The BN structure and probabilities were informed by both clinical data and AI-detected radiographic features.
• Receiver operating characteristic (ROC) curve analysis was used to evaluate model performance.

Main Results:

• The integrated model demonstrated high accuracy in identifying periodontal disease from radiographic images.
• The Bayesian network effectively incorporated deep learning outputs for comprehensive periodontal assessment.
• Receiver operating characteristic curve analysis confirmed the model's strong performance in guiding treatment plan recommendations.

Conclusions:

• Deep learning can be effectively integrated into Bayesian network models for enhanced periodontal disease detection.
• The developed clinical decision support system shows high accuracy and potential for improving comprehensive periodontal care.
• This AI-driven approach offers a promising tool for dentists in treatment planning and patient management.