Validated reconstructions of geometries of nasal cavities from CT scans

Developing validated objective measures of nasal airflow is paramount for improving the management of nasal obstruction. Nasal airflow can be studied objectively by simulating the flow in the computer. This requires a faithful reconstruction of the nasal geometry since the simulated airflow is sensitive to small variations of the complex shape of the nasal cavity. We here show that altering the geometry by less than 1 mm can change the airflow two-fold. We also show that a faithful reconstruction of the nasal geometry is possible with a threshold-based segmentation. Utilizing the known geometry of a CT phantom, we determine an optimal segmentation threshold of -450 HU. Changing this threshold by 100 HU alters the geometry by only about 0.1 mm. We use this verified segmentation to extract nasal geometries of three patients and simulate the respective airflows using the Lattice Boltzmann method. Using a simple model, we can predict how the reconstruction threshold affects the resistance to airflow. Since the segmentation and the simulation can be automated completely, this is an important step toward an objective analysis of nasal airflow based on CT scans.