Prediction of Drug-Induced Sleep Endoscopy Outcome From Baseline Diagnostic Data Including Polysomnography

What this Trial Is About

Researchers are evaluating how to predict the site, pattern, and degree of upper airway collapse in patients with obstructive sleep apnea (OSA) using routine polysomnography data. The study focuses on developing and validating models that link airflow shape features from sleep studies with observations made during drug-induced sleep endoscopy (DISE), which is part of standard care. Understanding these collapse patterns is important because they affect the outcomes of different OSA therapies. The study will enroll 1000 patients with OSA who recently had polysomnography and will undergo DISE. In the first phase, a preliminary model predicting complete concentric collapse of the palate (CCCp) will be validated in 300 patients. Later, refined models using pooled data will be developed and tested in 700 patients. Throughout the study, patients will receive therapies as clinically indicated, and researchers will explore how the predicted and actual collapse sites relate to treatment effectiveness and adherence. Participants will undergo one DISE procedure lasting up to 40 minutes during drug-induced sleep. Researchers will collect and analyze polysomnography data alongside DISE results to develop prediction models for different collapse sites. The primary outcome is the presence or absence of CCCp during DISE, with secondary measures including collapse of the tongue base, lateral walls, and epiglottis. The study will monitor associations between collapse patterns and therapy outcomes. Participation involves data collection from existing clinical procedures and no experimental treatments are administered.

Pathophysiological Endotyping Using Baseline Polysomnography Data