A comprehensive computational model of sound transmission through the porcine lung

Abstract

A comprehensive computational simulation model of sound transmission through the porcine lung is introduced and experimentally evaluated. This "subject-specific" model utilizes parenchymal and major airway geometry derived from x-ray CT images. The lung parenchyma is modeled as a poroviscoelastic material using Biot theory. A finite element (FE) mesh of the lung that includes airway detail is created and used in comsol FE software to simulate the vibroacoustic response of the lung to sound input at the trachea. The FE simulation model is validated by comparing simulation results to experimental measurements using scanning laser Doppler vibrometry on the surface of an excised, preserved lung. The FE model can also be used to calculate and visualize vibroacoustic pressure and motion inside the lung and its airways caused by the acoustic input. The effect of diffuse lung fibrosis and of a local tumor on the lung acoustic response is simulated and visualized using the FE model. In the future, this type of visualization can be compared and matched with experimentally obtained elastographic images to better quantify regional lung material properties to noninvasively diagnose and stage disease and response to treatment.

FIG. 1.

Airway acoustic model showing one bifurcation.

FIG. 2.

(Color online) Experimental setup of SLDV measurement of lung surface motion.

FIG. 3.

(Color online) Lung CT images used for segmentation: (a) Airway network 1, (b) airway network 2.

FIG. 4.

(Color online) Geometry of the constructed airways: (a) Airway network 1, (b) airway network 2.

FIG. 5.

(Color online) Volume mesh of lung with airway tree: (a) With airway network 1, (b) with airway network 2.

FIG. 6.

(Color online) (a) Selected points on lung geometry for surface velocity amplitude comparison with experiment, (b) airway network 2 with terminal impedance specified.

FIG. 7.

(Color online) Lung normal surface velocity magnitude (dB m/s for 1 Pa input acoustic pressure) at 500 Hz (a) experiment; (b) simulation, lung with airway network 1; (c) simulation, lung with airway network 2.

FIG. 8.

(Color online) Lung normal surface velocity magnitude (dB m/s for 1 Pa input acoustic pressure) at 800 Hz (a) experiment; (b) simulation, lung with airway network 1; (c) simulation, lung with airway network 2.

FIG. 9.

(Color online) Real part of airway acoustic pressure (Pa): (a) 500 Hz, airway network 1; (b) 500 Hz, airway network 2; (c) 800 Hz, airway network 1; (d) 800 Hz, airway network 2.

FIG. 10.

Lung normal surface velocity magnitude (dB m/s for 1 Pa input acoustic pressure) at selected points [point locations are in Fig. 6(a)]. Key: —, experiment; ○○○, simulation.

FIG. 11.

(Color online) Stacked horizontal slices of the real part of the lung velocity (μm/s) in the anterior-posterior direction and airway acoustic pressure (Pa) (a) airway network 1 at 500 Hz; (b) airway network 2 at 500 Hz; (c) airway network 1 at 800 H; (d) airway network 2 at 800 Hz.

FIG. 12.

(Color online) Cross-section images of the real part of the lung velocity (μm/s) in the anterior-posterior direction (see arrow) (a) 500 Hz, (b) 800 Hz.

FIG. 13.

(Color online) Shear motion applied to the lung surface.

FIG. 14.

(Color online) Cross-section images of the real part of the lung velocity (mm/s) in the anterior-posterior direction (see arrow) at 500 Hz: (a) normal lung, (b) fibrotic lung.

FIG. 15.

(Color online) Cross-section images of the real part of the lung velocity (mm/s) in the anterior-posterior direction (see arrow) at 800 Hz: (a) normal lung, (b) fibrotic lung.

FIG. 16.

(Color online) Cross-section images the real part of the velocity (μm/s) in the anterior-posterior direction for the lung with tumor: (a) 500 Hz, (b) 800 Hz.