Evaluation of the Alveolar Crest and Cemento-Enamel Junction in Periodontitis Using Object Detection on Periapical...

Area of Science:

• Dental Radiology
• Medical Image Analysis
• Artificial Intelligence in Healthcare

Background:

• Periodontitis severity assessment relies on measuring alveolar crest (ALC) and cemento-enamel junction (CEJ) bone loss.
• Manual analysis of periapical radiographs (PAs) for bone loss is time-consuming and prone to errors.

Purpose of the Study:

• To develop an automated method for accurate periodontitis severity assessment from PAs.
• To reduce diagnostic time and improve the accuracy of bone loss evaluation in dental practice.

Main Methods:

• Image enhancement techniques applied to PAs.
• Object detection for single tooth extraction (max accuracy 97.01%).
• Instance segmentation for tooth bone and crown region extraction (accuracies 93.48% and 96.95%).
• Novel algorithm for automatic CEJ and ALC marking.

Main Results:

• Accurate extraction of single teeth and relevant anatomical regions.
• Automated marking of CEJ and ALC, facilitating bone loss assessment.
• Significant reduction in time required for dental diagnosis.

Conclusions:

• The developed automated method enhances the accuracy and efficiency of periodontitis diagnosis.
• This AI-driven approach improves healthcare quality by streamlining dental diagnostics.