Computer-Assisted Quantitative Analysis of Drug-Induced Sleep Endoscopy for Obstructive Sleep Apnea/Hypopnea Syndrome

Abstract

Objectives To use computer-assisted quantitative measurements of upper airway changes during drug-induced sleep endoscopy (DISE) and to correlate these parameters with disease severities and physiologic changes in patients with obstructive sleep apnea/hypopnea syndrome (OSA). Design A retrospective study. Setting Tertiary academic medical center. Patients and Methods A total of 170 patients who failed continuous positive airway pressure therapy and then underwent upper airway surgery were enrolled. All patients received polysomnography and DISE preoperatively. We used ImageJ 1.48v to obtain maximal and minimal measurements, including cross-sectional areas and anterior-posterior and lateral diameters at 4 anatomic levels (retropalatal, oropharyngeal, retroglossal, and retroepiglottic) under DISE, and then computed the percentage changes. We analyzed the clinical values of DISE changes by computer-assisted analysis in patients with OSA and any correlations between these changes and polysomnography parameters. Results The percentage changes of upper airway showed significant collapses at all 4 anatomic levels (all P < .0001). We also found that the changes at retropalatal levels were significantly greater and that retroglossal levels were significantly smaller, while the changes of anterior-posterior diameters at retroglossal levels showed a significant positive association with apnea-hypopnea index and desaturation index. However, there were no statistically significant correlations between upper airway changes and obesity. Conclusion Computer-assisted quantitative analysis could evaluate upper airway changes of OSA in an objective way and may help identify the sites of obstruction during DISE more accurately. Upper airway showed multilevel collapse with independent significant changes in patients with OSA, with the retropalatal and retroglossal levels playing important roles in particular.