Detection of Dental Apical Lesions Using CNNs on Periapical Radiograph.

Area of Science:

• Dentistry
• Medical Imaging
• Artificial Intelligence

Background:

• Apical lesions are common chronic infectious dental diseases requiring time-consuming manual diagnosis from X-ray images.
• Current diagnostic methods can be hindered by image quality variations due to shooting angles or radiation doses.
• Automating repetitive tasks in diagnosis allows dentists to focus on critical treatment and patient communication.

Purpose of the Study:

• To develop and validate an automated lesion area analysis model using convolutional neural networks (CNNs).
• To reduce the time and improve the efficiency of diagnosing apical lesions in dental radiography.
• To establish a standardized clinical database for training and testing the diagnostic model.

Main Methods:

• A database of dental X-ray images was created with Institutional Review Board (IRB) approval.
• Image preprocessing involved a Gaussian high-pass filter and iterative thresholding to segment individual tooth images.
• A CNN migration learning model was trained using 70% of the image database and tested on the remaining 30%.

Main Results:

• The proposed CNN model demonstrated a practical diagnosis accuracy of 92.5% for apical lesions.
• The model successfully automated the identification and analysis of lesion areas in dental X-ray images.
• The automated approach significantly reduced the time required for diagnosis compared to manual methods.

Conclusions:

• The developed CNN model offers an efficient and accurate solution for the automatic diagnosis of apical lesions.
• This AI-driven approach has the potential to streamline endodontic treatment workflows.
• Further integration of such models can enhance diagnostic precision and patient care in dentistry.