Table 1.
Patient demographics and CT information.
Fig 1.
Common artifacts observed when segmenting the nasal cavity from CT scans.
(A) Coronal CT of patient #2 shows a patent ostiomeatal complex and a patent olfactory cleft (circles). (B) Contours of 3D reconstruction created with thresholding range -1024HU to -950HU show discontinuities in the airspace (circles). (C) In patient #1, the 3D reconstruction created with the same thresholding range has surface irregularities and noise (arrows). (D) In patient #3, the 3D model created using the thresholding range -1024HU to -200HU did not segment the walls of the ethmoid sinuses correctly.
Fig 2.
Effects of segmentation threshold on 3D reconstruction of the nasal cavity.
A wide range of segmentation thresholds (namely -800HU to -300HU) provides acceptable 3D reconstructions of the human nasal cavity. (A) Coronal CT of patient #1 after mucosal decongestion showing the middle turbinate, ethmoid sinuses, and part of the right maxillary sinus. (B,C,D) Contours of 3D models created using three different thresholds (-300 HU, -550 HU and, -800 HU). At the upper limit of acceptable thresholds (-300 HU, panel B), thin soft tissue walls are incorrectly identified as air (arrows). At the center of the range (-550 HU, panel C), a good 3D reconstruction is obtained with few or no artifacts. At the lower limit of acceptable thresholds (-800HU, panel D), narrow passages become partially or completely obstructed (circle) and irregularities appears at air-tissue boundary (asterisk).
Fig 3.
Outline of 3D reconstructions obtained with three different thresholds.
Coronal CT scan of patient #1 showing the outlines of the 3D reconstructions created with segmentation thresholds -300 HU (green), -550 HU (blue), and -800 HU (red). The close-up view (right-side panel) reveals a nearly uniform distance of 1 to 2 pixels between the models created with segmentation thresholds -300 HU and -800 HU.
Fig 4.
Airspace cross-sectional areas after mucosal decongestion with Oxmetazoline as a function of distance from nostrils.
(A) Definition of the relative distance from nostrils. (B,C,D) In all three patients, the airspace cross-sectional area increased systematically throughout the nasal cavity as the segmentation threshold was increased from -800HU to -300HU.
Table 2.
Unilateral volume of the human nasal cavity (nostrils to choana) in three patients with nasal airway obstruction (NAO) after mucosal decongestion with Oxmetazoline.
Table 3.
Unilateral surface area (cm2) of the human nasal cavity (nostrils to choana) in three patients with nasal airway obstruction (NAO) after mucosal decongestion with Oxmetazoline.
Fig 5.
Flow-pressure curve measured with rhinomanometry and calculated with CFD in 3D models reconstructed with segmentation thresholds of -300HU, -550HU, and -800HU.
Note the systematic increase in nasal airflow (reduction in nasal resistance) as the segmentation threshold increases from -800HU to -300HU.
Table 4.
Unilateral resistance (Pa.s/ml) of the human nasal cavity (nostrils to choana) measured at a unilateral flowrate of 125 ml/s in three patients with nasal airway obstruction (NAO) after mucosal decongestion with Oxmetazoline.
Fig 6.
Inspiratory streamlines (top) and air velocity colormap at coronal section D = 0.5 (bottom).
The main air stream flowed near the middle turbinate in patient #1 independently of the segmentation threshold (-300HU, -550HU, and -800HU). The right nostril was assumed to be blocked to reproduce rhinomanometry measurements of unilateral resistance in the left cavity (see text for details).
Fig 7.
The coronal section D = 0.5 was divided in three regions (inferior, middle, superior) for the analysis of intranasal airflow distribution.
The left and right cavities were analyzed independently. Each region corresponded to 1/3 of the nasal height. The inferior region corresponds to the nasal floor and the lower portion of the inferior turbinate. The middle region corresponds to the area surrounding the lower portion of the middle turbinate. The superior region corresponds to the olfactory cleft and upper portion of the middle meatus.
Fig 8.
Intranasal airflow distribution at coronal section D = 0.5.
In most nasal cavities, the middle region was the main airflow pathway. Intranasal airflow distribution was nearly independent of the segmentation threshold. Negative values correspond to regions of retrograde flow.