Effective Caries Detection Through Artificial Intelligence in Radiographs and intraoral Images

Abstract

OBJECTIVES

The aim was to find a method that is more efficient and consistent in caries detection utilizing Al technology. A literature review was conducted using randomized clinical studies found on PubMed that focused on Al use in radiographs and intraoral images.

METHODS

The population of concern was dental patients and the intervention was the use of artificial intelligence. This study compared visual examination of caries with caries detection by artificial intelligence. The outcome explored was the effectiveness of artificial intelligence in detecting caries. The Pubmed database was used to search for keywords including Al, Machine Learning, Decay, Cavities, x-rays, and Diagnostic Imaging.

RESULTS

PubMed and keywords produced 34 studies. Of those studies, 5 were chosen for review, and 3 were used. Study 1 used a ResNet-18 deep learning model which was trained to provide radiographs. Study 2 used a convolutional neural network (CNN) model. Study 3 used deep learning based on CNN and consumer products, such as cell phones and cameras, to detect caries in low income areas where dental equipment and dental specialists are not readily available.

CONCLUSIONS

Artificial intelligence tools have detected caries with over 80-90% accuracy consistently in more than 3,000 radiographic images. Al used in conjunction with professional dental knowledge improves accuracy and efficiency of detecting caries in radiographs and clinical images. All three studies that we investigated showed Al technology to be an effective tool for detecting caries in radiographs. Further research proved that Al usage makes detecting caries more cost efficient and is able to be incorporated in low income populations where expensive technology is absent.

IMPLICATIONS

The findings from this research will improve the accuracy and speed of early caries detection, reduce diagnostic errors, and increase accessibility to dental care in underserved areas. This could reduce untreated caries in low-income communities.