The incidence of oral diseases, such as dental caries and periodontal diseases, is rising globally, including in Norway. This trend has serious health and economic consequences, significantly reducing the quality of life for those affected. It is estimated that half of the world’s population suffers from tooth decay, making it the most common health condition. Severe periodontal disease, which can lead to tooth loss, is also increasing worldwide. Dental treatment is often expensive, accounting for about 5% of total health costs and 20% of individual health costs in high-income countries. Identifying and treating abnormal changes at an early stage, before symptoms become noticeable, is crucial for reducing overall health costs. In Norway, around 75-85% of adults over the age of 18 visit their dentists annually for oral health checks. To enhance these check-ups, the AI-Dentify project aims to research, develop, test, and implement an automated computer-aided diagnostic tool to detect abnormalities as early as possible. This tool will help reduce variation in the interpretation of dental health images between dentists and enable objective grading and documentation of findings. By analyzing X-rays taken during dental check-ups with artificial intelligence (AI) technology, we can revolutionize dental health services and detect harmful changes early in a robust and objective manner.
Our recent work, published in the BMC Oral Health journal, focuses on using deep learning for proximal caries detection on bitewing X-rays. We trained three different object detection models using a dataset of 13887 bitewing images from the HUNT4 health study. Our models showed significant improvements in precision compared to dental clinicians, demonstrating the potential of AI to assist in caries diagnosis. Additionally, we have developed bone loss segmentation from orthopantomogram (OPG) images. Bone loss is a critical indicator of periodontal disease, and accurate segmentation is essential for early diagnosis and treatment planning. Our AI models have been trained to identify and segment areas of bone loss with high accuracy, providing valuable support to dental professionals in their clinical decision-making.
The integration of AI in dental diagnostics not only enhances the accuracy and efficiency of detecting oral diseases but also ensures consistent and objective assessments. This technological advancement is poised to transform dental care, making it more effective and accessible for patients worldwide.
-
The main ambition of the AI-Dentify project is to develop, test and validate a novel Boneprox software as a service platform, which is scalable to the global market. The solution will make it easy for the dental industry worldwide to use this as a complete solution on an everyday basis for decision support in oral disease diagnostics. The prevalence of non-communicable diseases particularly oral diseases are increasing worldwide including Norway. In Norway, every year more than 75% of adults visit their dentists for annual oral health check-up. We need to develop and integrate new research based computational models with the ability to automatically analyze data to retrieve useful information to aid clinicians in accurate diagnostics and decision making. Hence, we think modern-day supervised machine learning technique(s) may offer the promise to potentially solve this important issue. In dental science, a panoramic dental X-ray image (Orthopantomogram, known as OPG) captures both maxillary and mandibular dental arches and the surrounding structures including hypoid bone. We believe that implementing artificial intelligence based algorithms can reduce the rate of diagnosing false negatives, especially effective for inexperienced dentists for decision making. Thus, the overall goal of this project is to develop deep learning classification model based on computer-assisted diagnosis system using dental X-ray images to detect oral abnormalities such as caries and periodontitis, and used as a decision making tool.