Segmenting and Modeling the Human Airway with Voicing or Breathing Disorders

Abstract

Upper airway disorders encompass various conditions that hinder breathing and/or voicing. They may arise from factors such as past airway trauma, prolonged use of breathing tubes, tracheotomy history, vocal fold nerve damage, or specific autoimmune ailments. This work is focused on the segmentation and modeling of two prominent upper airway disorders: Obstructive Sleep Apnea (OSA) and Unilateral Vocal Fold Paralysis (UVFP). For UVFP, segmentation is carried out for three rabbit larynges in various vocal fold adduction conditions. For OSA, segmentation is carried out for a patient going through Ansa Cervicalis Stimulation (ACS), a novel surgical procedure to treat OSA. Our segmentation identifies the critical components that are needed to understand the upper airway disorders and their treatment. The OSA study also seeks to establish the theoretical and biomechanical foundation of experimental results. The finite-element model shows promising results in capturing the effects of the ACS on the movements of the tongue, epiglottis, pharynx wall, soft palate, and their interactions. Future work would aim to expand this study to include patient-specific features and more subjects to improve the robustness of the computational model and offer more insights into the ACS treatment.