Comprehensive Insights into Artificial Intelligence for Dental Lesion Detection: A Systematic Review

Area of Science:

• Utilizes artificial intelligence (AI) and deep learning for enhanced diagnostic accuracy in dentistry.
• Focuses on automated detection of various dental lesions from multiple imaging modalities.

Background:

• Growing demand for AI in healthcare necessitates robust automated diagnostic systems.
• Deep learning models show significant promise for dental lesion detection in various imaging techniques.

Purpose of the Study:

• To systematically review the utilization of deep learning methods for detecting dental lesions.
• To analyze AI approaches, data augmentation, and challenges in AI-based dental lesion detection.

Main Methods:

• Conducted a systematic review following PRISMA guidelines, analyzing 29 primary studies (2019-2024).
• Searched multiple databases including IEEE, Web of Knowledge, Springer, ScienceDirect, PubMed, and Google Scholar.
• Addressed research questions on AI objects, state-of-the-art approaches, data augmentation, and challenges.

Main Results:

• Identified five lesion types: periapical, cyst, jawbone, dental caries, and apical.
• Common deep learning models include U-Net, AlexNet, and YOLOv8.
• Flipping, rotation, and reflection were key data augmentation techniques; challenges include data integration, quality, generalization, and overfitting.

Conclusions:

• Deep learning models offer potential for improved accuracy and clinical workflow support in dental diagnostics.
• Solutions for AI-based dental lesion detection must be generalizable across image types, data volumes, and quality variations.