Improving Caries Diagnosis with Targeted Nanoparticle-Enhanced AI-Assisted Intraoral Imaging

Abstract: This project responds to the need identified by NIDCR to develop improved methods to detect, monitor, and predict progression of dental caries (tooth decay) to improve human health. Worldwide, caries is the most common chronic disease affecting almost everyone. Dental disease is major cause of economic and social loss and leads to complications including pain, tooth loss and even death. For over a century, caries diagnosis has been performed visually and using a dental explorer, but early caries lesions are missed, and the explorer can cause cavitation. Like the world at large, dentistry is entering the era of Artificial Intelligence (AI), with some of the first applications to help dentists analyze X-Rays. While radiographs identify more advanced decay, they lack resolution to identify early disease. Newer methods for caries diagnosis show little benefit over visual/tactile exam and incur greater cost to dentist and patient. None of these can distinguish between active and inactive carious lesions, a critical need in modern dentistry, and a requirement for disease monitoring. With SBIR funding (Grant 5R44DE027903-03), we have successfully developed a clinically validated mouth rinse that uses targeted fluorescent biopolymer nanoparticles to illuminate early active caries lesions, formulated into an FDA- approved product, LumiCare™ (LC) Caries Detection Rinse. Our goal is to create an intraoral camera product that can collect clinical white-light and fluorescent images of caries that are illuminated with LC Rinse, which, in conjunction with an AI algorithm software, will provide a tool to effectively monitor carious lesion severity and therefore track the efficacy of medical remineralization treatments. Dental professionals would apply LC Rinse, image with the intraoral camera, and software would determine presence of fluorescence and provide quantitative scoring and diagnostic information (e.g., lesion depth, ICDAS severity, activity status) that can be tracked at each appointment. Images will be available to the dental professional as a communication tool with patients and as a record that can be used for insurance. Automated and objective early caries detection and monitoring will enable teledentistry and public health use cases and empower dentists and patients to use non-invasive medical treatments to improve oral health. Our preliminary research and customer discovery activities have demonstrated the potential of this technology. This project aims to prepare a Minimally Viable Product (MVP) optimized intraoral camera system that can collect images which can be analyzed with computer vision software to accurately detect fluorescence from LC-rinse illuminated caries lesions. Successful completion of the project milestones will establish feasibility of the intraoral camera design and regulatory testing requirements, which are critical milestones towards further clinical validation and commercialization. Phase II will involve AI model selection and optimization, in vivo testing and preparing for FDA 510(k) submission.