Detection of caries and determination of treatment needs using DentMA teledentistry: A deep learning approach

Abstract

Background: Teledentistry is considered capable of detecting dental caries remotely without direct contact with patients. Deep learning (DL) algorithms are trained with sufficient datasets to find patterns and models based on learning. By using a DL model, we propose a conceptual framework for the screening of dental caries using smartphones: the DentMA application, a new breakthrough in teledentistry technology. In this study, the DentMA teledentistry application was used for mobile screening for caries. Purpose: This study aimed to analyze the use of DentMA teledentistry to detect dental caries, enamel-dentin caries, and untreated caries, and to determine treatment needs in children. Methods: The participants of this study were 124 children aged 4–6 years. The study was conducted by having the participants’ mothers take intraoral clinical photos of the participants using the DentMA teledentistry application on their smartphones. For the photo to be taken, each participant was directed to sit upright, with the head looking straight ahead and the mouth open. Results: The results showed that DentMA teledentistry was capable of detecting dental, enamel-dentin, and untreated caries in children, and its ability to predict dental treatment needs was good (p < 0.005). Teledentistry screening using a mobile phone can detect not only caries but also a relationship between the complaints and the medical histories of patients with dental caries. Conclusion: The DentMA teledentistry application can detect dental caries in children according to the individuals’ complaints, including enamel-dentin caries and advanced caries, and can help determine treatment needs.