A New Technique of Virtual Modeling of Individual Patient Airways in Obstructive Sleep Apnea

Abstract

Objectives:

Up to 14% of the US population is estimated to have obstructive sleep apnea (OSA) related to obesity. Other than continuous positive airway pressure, treatments have had variable results since the exact site(s) of obstruction and the optimal method of modifying those sites with devices or surgery can be difficult to establish. We introduce a technique for modeling the upper airway that shows the behavior of the airway and predicts the location of collapse in OSA.

Methods:

Accurate modeling of air to solid fluid-structure interaction, challenging to model in the past, was accomplished using ANSYS software applied to a 3-dimensional rendering of an OSA patient’s airway (with an AHI of 87.6/h) using thin-cut computed tomographic scan data. Viscoelastic properties of the pharyngeal walls, tongue and palate, not previously well characterized, were measured using 5 fresh porcine cadaver heads with a Bose ElectroForce test instrument with Dynamic Mechanical Analysis software. Various dynamic displacements were applied to tissue samples and associated forces measured simultaneously to obtain dynamic material properties.

Results:

Areas of lowest pressure during inspiration were identified as likely regions of collapse during apneic episodes, such as the base of the tongue with pressures as low as –4730 Pa (–48.23 cm H₂O), similar to values reported in human esophageal pressure measurements in upper airway resistance syndrome.

Conclusions:

This novel investigation virtually and accurately models the upper airway in OSA, and allows virtual modification of the airway to predict effects of treatment.