A Viable Approach to Medical Image Processing for CFD Simulations of the Upper Respiratory Tract

Abstract

AbstractWe discuss a fast and easy approach to create the upper airway geometry of human lungs from medical images to perform Computational Fluid Dynamics (CFD) simulations. We have employed a combination of open-source and commercial image processing and CFD applications. From Computed Tomographic (CT) images of lungs available in the public medical repository, we were able to recreate the 3D structure of the upper airways up to 6 generations, which after postprocessing was used to investigate the flow pattern during respiration. The uneven surface of the real geometry, the curvature of the flow paths and asymmetric bifurcations lead to definite and distinguishable variation in flow pattern unlike that in studies with idealized lung geometries, such as that with the Weibel model. The nature of flow at various positions along the airways was investigated employing simple boundary conditions of constant inlet velocity at the truncated trachea and constant outlet pressure at the truncated bronchi (fourth generation). For the preliminary simulations, we could observe the presence of counter-rotating vortices (Dean flow) as well as asymmetric flow rate across the left and right lung (a 50% more flow toward the right bronchus).KeywordsMedical imagingImage segmentationAirway treeHuman lungRespirationCFDRespiratory tract3D SlicerAutodesk® Meshmixer™Ansys Fluent®