Across the Europe and world, oral diseases present a major public health burden. Expenditure on the treatment of oral conditions often exceeds that of other diseases including cancer, heart disease, stroke and dementia, on average presenting about 10% of the annual health expenditure across the EU. Among the two most prominent oral disease and conditions are dental caries, which affects over 60% of school children and nearly all adults, while severe periodontal diseases affect from 15 to 20% of middleaged adults. Together, dental cavities and periodontal diseases form the two most prominent causes of tooth loss. An efficient management of the oral diseases could be achieved by novel state-of-the-art oral health screening systems that are suitable for clinical and home use. In this way, the oral health could be efficiently managed and monitored by each individual, lowering the burden on the healthcare systems and, at the same time, raising the awareness of people on importance of the everyday oral hygiene, the latter being particularly important in the case of children. However, achieving such ambitious objectives requires innovative and efficient solutions in several research fields. First, the information required to diagnose the most prominent oral disease must be acquired by noninvasive means. The most promising results for detection and quantification of diseased hard dental tissues were obtained by laser fluorescence (LF) and diffuse reflectance (DR) spectroscopy in the near infrared (NIR) spectral range. LF provides information on the mineral content, structure and presence of bacteria in the surface layers of hard dental tissues, while the less scattered NIR light penetrates deeper into the tissue and thereby complements the LF information. Furthermore, noninvasive functional and structural assessment of soft tissues can be efficiently conducted in the visible and the very near infrared spectral range. As effective screening for pathological and other changes of the oral cavity also requires detailed spatial information, hyperspectral imaging systems (HIS) that efficiently merge the spectral and spatial information seem as a perfect choice. However, the substantial differences among the three noninvasive optical techniques pose several challenges ranging from technical issues related to the spatial constraints imposed by the oral cavity, to the even more demanding analysis of the acquired multimodal hyperspectral images. Furthermore, imaging of the entire oral cavity may require acquisition of multiple partially overlapping hyperspectral images and the hyperspectral images of the oral cavity acquired at different time points will have to be registered into a common reference coordinate frame where the joint information can be efficiently analyzed and the temporal changes of oral cavity, resulting from progression of a disease or treatment, quantified. In the proposed project, we are going to focus on three main aspects of the oral cavity screening system: a) design of optical and illumination systems for acquisition of high fidelity multimodal hyperspectral images, b) development of novel robust image restoration and multimodal registration methods that will bring the hyperspectral images into a common reference coordinate frame where the acquired spectral, spatial and temporal information can be efficiently analyzed, and c) methods for reliable detection and quantification of incipient and secondary caries lesions and lesion activity from the restored and registered multimodal hyperspectral images covering the visible and NIR spectral range. With the proposed project, we expect to make a significant step towards screening systems for efficient management of oral health by reliable noninvasive diagnosis of oral diseases at the earliest stage. Such screening systems could significantly improve the quality of life among the aging European population and potentially result in huge costs savings associated to the treatment of oral diseases.
