0551 Dynamic Upper Airway MRI (DUAM) of OSA May Be Better Than DISE for Detecting Complete Concentric Collapse

Abstract Introduction Currently, drug-induced sleep endoscopy (DISE) is mandatory for identifying palatal complete concentric collapse (CCC), which is associated with poor outcomes of hypoglossal nerve stimulation (HNS) therapy in obstructive sleep apnea (OSA). As a major inherent flaw, DISE lacks resolution along the superior-inferior axis of upper airway. We recently developed dynamic upper airway MRI (DUAM) methods to image at high spatiotemporal resolutions. Here we aim to compare the performance of DUAM against DISE for detecting CCC. Methods We recruited patients who previously received DISE to assess their candidacy for HNS. DUAM was conducted during natural sleep in contiguous axial slices (2.0x2.0x6.0 mm, 0.18 s), and OSA events were identified from concurrent physiological measurements. Segmentation of airway lumen by semi-automatic software was visually confirmed for veracity. Similar to the conventions in DISE, we determined the pattern of airway collapse by the changes of lumen contour from the times when the lumen was the smallest or absent. Results Six patients were studied by December 2025; four of them showed CCC during DISE. The other two patients both showed anterior-posterior collapse in velopharynx and partial lateral collapse in oropharynx, and both received HNS implants that did not effectively treat OSA. The collapse pattern was consistent across observed (more than 16) OSA events in each patient. In velopharynx and oropharynx, the collapse pattern (CCC vs. anterior-posterior vs. lateral) detected by DUAM was in full accordance with DISE. However, in both patients without CCC during DISE, at least one slice between velopharynx and oropharynx clearly showed CCC, concentric around the inferior part of soft palate. Such CCC could only be detected because the oral cavity space anterior to the soft palate was imaged. The video recorded during DISE was re-reviewed by experienced otolaryngologists and CCC was not identified. Conclusion DUAM can detect the CCC restricted to the upper airway segment between velopharynx and oropharynx, because of its high resolution (6 mm) along the superior-inferior axis and its ability to visualize the oral cavity space anterior to the soft palate. Such localized CCC can be missed by DISE and might be associated with negative HNS outcomes. Support (if any) NIH R01-HL173483 (Lingala)