Artificial intelligence model for application in dental traumatology.

Area of Science:

• Dental Traumatology
• Artificial Intelligence in Dentistry
• Medical Imaging Analysis

Background:

• Healthcare systems are rapidly advancing with new diagnostic technologies, including artificial intelligence (AI).
• AI applications in dentistry are emerging, yet its use in dental traumatology remains limited.
• This study addresses the need for AI-driven tools in diagnosing dental fractures.

Purpose of the Study:

• To develop and evaluate a deep-learning, convolutional neural networks (CNN)-based model for detecting and classifying dental fractures.
• To assess the model's accuracy in differentiating various types of dental fractures using periapical radiographs.

Main Methods:

• A dataset of 72 periapical radiographs featuring 108 fractured teeth was curated and annotated by dental experts.
• Data augmentation techniques were employed to enhance the dataset's robustness.
• A CNN model was implemented using Python, with data split into 80% for training and 20% for testing, incorporating cross-validation.

Main Results:

• The AI model achieved high accuracy in distinguishing between different fracture types: uncomplicated crown fractures (96.0%), complicated crown fractures (96.3%), crown-root fractures (99.1%), and root fractures (97.2%).
• The overall accuracy for classifying all four types of dental fractures was 78.7%.
• The model demonstrated strong performance in differentiating specific fracture categories.

Conclusions:

• The proposed AI model exhibits excellent performance in classifying dental fractures, offering an innovative application in pediatric dentistry and dental trauma.
• The study recommends expanding the model to larger datasets and exploring its use with panoramic radiographs for broader clinical application.
• AI-powered diagnostic tools can significantly assist less experienced dentists in making accurate and timely treatment decisions for dental injuries.