Abstract
This paper proposes an evaluation index for the classification of severe Obstructive Sleep Apnea Syndrome (OSAS) by the use of the tongue morphology and the cross sectional area of the narrowest upper airway evaluated by image analysis on the upper airway MRI. In general, polysomnography (PSG) is a gold standard evaluation for severe SAS, but sometimes the symptoms do not occur during PSG screening because of the first night effect. The morphology of the upper airway, on the other hand, gives much information about the severity of OSAS. So, the upper airway MRI is often used for medical diagnosis, but few evaluation indices have been reported objectively. We focused on the tongue region and considered two directions from the center to the edge of the tongue higher correlated with the severity and determined the cross sectional area in the narrowest upper airway. These features are weighted and linearly combined to predict the severity. Finally, the severe patients are detected by judging whether the prediction value is greater than or equal to 30. As a result, the true positive rate is 0.909 and the false positive rate is 0.476 for detecting the severe patients.
