The Development of an early detection tool for oral-dental radiation induced injury - Osteoradionecrosis (ORN) of the jaw is a severe complication of radiotherapy (RT) in head and neck cancer patients, leading to significant oral dysfunction and compromised quality of life. Early detection is critical for timely intervention, yet current clinical methods often fail to identify ORN before irreversible damage occurs. This F31 project leverages Al-powered radiographic analysis to enhance ORN detection and management by developing predictive models and automated tools to assist clinicians. In Specific Aim 1, I will establish a clinical benchmark by evaluating the consistency of clinician assessments of radiographic ORN. This benchmark will provide a gold standard for comparison with Al-based models. Specific Aim 2 will focus on developing an automated segmentation tool to segment ORN on radiographic images, using physician-generated contours as the ground truth. Specific Aim 3 involves building a predictive model based on radiographic density changes {delta-HU) to identify patients at higher risk of ORN early in the treatment process. The training goals of this award will provide me with essential sills in Al model development, medical imaging techniques, and clinical decision support tools, all within the specialized field of radiation oncology. Through mentorship at MD Anderson Cancer Center, an internationally recognized leader in cancer treatment and research, I will gain the expertise necessary to translate these innovations into clinical practice. This research aligns with the mission of NIDCR by advancing technology that reduces radiation induced oral morbidities, ultimately improving patient outcomes in dental and oral health related to radiotherapy.
