Introduction: Numerous cephalometric studies have been conducted to find predictors of OSA severity – all with the common handicap of relying solely on static radiographic measures. Using sleep MRI, we combined both static and dynamic measures to characterize OSA severity more accurately. Materials andmethods: This is a case-control studywith 15 severe OSA patients (AHI > 40) who are age, gender, and BMI matched with 15 controls (AHI 4% in oxygen saturation. Static midsagittal MRI slice was used to measure a priori predictors of OSA severity, including palatal plane, upper airway length, hyoid position (angle from Pogonion to hyoid to the posterior airway), and posterior airway space. Axial and coronal views of the oropharynx were then rated by blinded, independent reviewers, focusing on lateral pharyngeal walls only. The degree of collapse was rated as none (0), partial (1), or complete (2). Univariate analyses withwere conductedwith Pearson and Spearman tests with the a priori predictors as described. Multivariate analysis was then performed to identify independent predictors Results: The mean age, BMI, and AHI of the control group were 42.6, 29.7, and 5.28 . The mean age, BMI, and AHI of the severe OSA group were 42.4, 30.4, and 70.29. There were no statistically significant differences between the two groups. When compared with age, BMI, and gender-matched controls, palatal plane (r2 = 0.46, p = 0.01), hyoid position based on pogonionhyoid-airway angle (r2 = −0.72, p < 0.001), and lateral pharyngeal wall collapse from dynamic sleepMRI (r2 = 0.65, p < 0.001) correlated significantly with severity of obstructive sleep apnea. Upper airway length and posterior airway space did not reach statistical significance. In a multivariate analysis, both hyoid position and lateral pharyngeal wall collapse independently correlate with severity of OSA. Conclusion: While impractical to use sleep MRI for OSA patients to identify sites of airway collapse, this study has important clinical implications. Few static and dynamic predictors are needed to identify severe OSA. Characterizing hyoid position and identifying lateral pharyngeal wall collapse, even from routine nasopharyngoscopy, can be sensitive and cost-effective. Acknowledgements: The senior author of this study would like to thank the team led by Dr. Wang PC from Cathay General Hospital, Prof. LoMT from Central University, and Dr. Shih TT fromNational Taiwan University.
Read full abstract