Analysis of articicial intelligence in the diagnosis of carious in interproximal radiographs: A literature review.

Abstract

This study aimed to explore the application of artificial intelligence in the early diagnosis of carious lesions in interproximal radiographs, analyzing its potential to optimize diagnostic accuracy and efficiency in dental practice. The methodology employed consisted of a comprehensive literature review, focused on the analysis of studies addressing the use of artificial intelligence, with an emphasis on Convolutional Neural Networks (CNNs), for the detection and classification of carious lesions in interproximal radiographs. The reviewed results demonstrated that AI, especially models such as U-Net, YOLO, and DenseNet121, offers an effective tool for caries detection, often surpassing the sensitivity and standardization of human diagnosis, particularly in the early stages of the disease. The identified benefits include faster and more accurate diagnoses, optimization of clinical time, and the possibility of less invasive interventions. However, challenges were observed related to the necessity and reliability of results (false positives and negatives), ethical and regulatory aspects, and professional resistance. Artificial intelligence represents a significant advance in caries diagnosis, with the potential to transform dentistry by acting as a complementary tool that assists and enhances the work of the dental surgeon. To fully realize this potential, continuous research and development efforts, improvements in data quality, and careful integration of technology with clinical expertise are essential, with a view to promoting oral health and improving patients' quality of life.