Ontology-based aPplications of mUlti-discipLinary knowlEdge on orodeNtal Conditions and outcomEs (OPULENCE) - PROJECT SUMMARY/ABSTRACT Orodental sequalae after radiation therapy (RT) for head and neck cancer (HNC) pose a significant problem for survivors, their oncology and dental providers, and the healthcare system. A recent prospective trial using advanced RT techniques and comprehensive pre-therapy dental evaluations found that within 2 years of finishing RT, 18% of patients suffered tooth loss, dose-dependent gingival recession with increased post-radiation caries (PRC) risk was frequently observed, and 6.4% developed bone exposure or osteoradionecrosis (ORN). These orodental toxicities result in diminishing quality of life, especially ORN which is associated with high symptom burden (i.e., dentition and activity limitations) and high financial costs. The combination of a lack of standards for reporting oncologic orodental sequalae, interoperability limitations between electronic medical and dental record systems, and poorly understood biological mechanisms of orodental disease, result in uncoordinated and information-deprived care. Clinically, there exists a knowledge gap on how to conceptualize and report the presence and/or severity of diseases such as ORN which has over 16 published staging systems, or PRC which is often measured using the Post-Radiation-Caries Index or International Caries Detected and Assessment System, both of which underestimate rates of PRC. Moreover, no consensus on assessing or reporting post-RT POD exists. The OPULENCE multi-institutional and multidisciplinary investigative team has started to bridge these unmet needs through published clinical standards for pre- and post-RT assessments and procedures, the development of clinical and image data-driven models for ORN, and contributions to the first publicly available operational ontology for oncology (i.e., for data harmonization and computable knowledge of general cancer care). Furthermore, we have published findings from a small prospective study which showed significant oral microbial shifts and reduced activation of oral neutrophils, changes that correlate with oral inflammation and POD, a known independent risk factor for ORN. Through the OPULENCE Proposal, we aim to open new horizons for early orodental disease diagnosis, toxicity mitigation, and data-driven management by leveraging artificial intelligence/machine learning methods for multi-domain ontology learning with corresponding common data elements highly expressive of the full (clinical, imaging, and biologic) spectrum of orodental disease (Aim 1), facilitating multidimensional, image-based FAIR data generation and application through informatics models of 2D-3D image transformation and NTCP models of PRC and POD (Aim 2), and prospectively identifying deep phenotypes and performing causal analysis on the interplay between biomarkers and post-RT orodental complications (Aim 3). All aims are highly synergistic with NIDCR future research initiative interests including investigations of personalized dental treatment practice prior to cancer therapy, developing risk assessment algorithms for early toxicity detection, monitoring and treating orodental complications, and enabling meaningful oncologic (i.e., medical) and dental care coordination.
