Computational fluid dynamics simulations and phase contrast-MRI validation of airflow in human pharynx during obstructive sleep apnea

Obstructive sleep apnea (OSA) is a frequent sleep disorder caused by the repetitive partial or complete obstruction of the upper airway despite the ongoing breathing effort. Numerical simulation is emerging as an important tool for the understanding of airflow conditions causing OSA. To ensure the reliability of numerical simulations, validation of the results is required. This study presents the comparison of pharyngeal flow patterns at different flow rates simulated using computational fluid dynamics (CFD) and measured using phase-contrast magnetic resonance imaging (PC-MRI). A computational model based on the stationary incompressible Navier-Stokes equation solved with finite element method was used. An anatomically accurate human pharynx model was used for the simulations. It was obtained from the 3D segmentation of computed tomography (CT) image data of a patient with OSA. The boundary conditions consisted of flow rates of 250 ml/s or 500 ml/s at the inlet and zero pressure at the outlet. PC-MRI was used to measure the flow patterns in a 3D-printed pharynx model based on the segmented CT data in a 3T MR system using an RF-spoiled gradient-echo sequence. A mixture of 55.27% glycerol in water was used as a test fluid at 25°C. An MR-compatible flow pump was used to achieve a flow rate of 88 ml/s or 177 ml/s, respectively, to replicate the inlet boundary conditions of the simulations. Comparable flow patterns obtained from CFD simulations and PC-MRI show that the stationary incompressible Navier-Stokes based model is a good predictor of the measured flow patterns. Below the minimum crosssectional area of the oropharynx, the flow patterns show a flow separation and a region of recirculation associated with the largest negative wall static pressure. The validation proves the reliability of CFD and further supports the use of CFD as a potential tool for the understanding of OSA.