Page 34 - Winter2021
P. 34

  Figure 6. The scattering mean free path (SMFP) is significantly lower in healthy rat lungs than in edematous and fibrotic rat lungs. Horizontal bar and X in the center, the median. Vertical lines (bottom to top): minimum, first quartile, third quartile, and maximum.
envelope statistics are also highly relevant for assessing and monitoring changes in lung structure because they are bound to be influenced by the distribution and size of the alveoli. The BSC is used to extract parameters of tissue microstructure by analyzing the power spectra of raw ultrasound data. Theoretical models of BSC account- ing for scatterer properties, such as scatterer diameter and scatterer concentration, are fitted to real ultrasound data, which allows one to estimate the parameters of the microstructure such as the size and density of the alveoli (Oelze and Mamou, 2016).
In a study on a rodent model of pulmonary fibrosis, it was demonstrated that ultrasound parameters related to the BSC and envelope statistics all showed significant dif- ferences between control lungs and edematous lungs and between control lungs and fibrotic lungs. These param- eters also correlated to the severity of pulmonary fibrosis as determined by histology (Lye et al., 2021).
Finally, ultrasound attenuation also reflects scattering and absorption and can therefore be used to inter- rogate lung tissue. Lungs exhibiting edema are less attenuating than fibrotic lungs due to the presence of fluid buildup. Higher frequencies are attenuated faster than lower frequencies. It is therefore pos- sible to measure the downward shift in frequency content in backscattered ultrasound signals in the form of a parameter called the backscatter frequency
shift (BFS) to quantify ultrasound attenuation in the lungs (Zenteno et al., 2016). In a study on rats with edematous or fibrotic lungs, the BFS proved able to discriminate between fibrosis and edematous lungs, suggesting that it could be used to increase the speci- ficity of diagnosis (Mohanty et al., 2020).
Can Specificity Be Achieved?
The question of diagnosis specificity is critical in lung ultrasound. Not only it is important to differentiate a healthy lung from a diseased lung, but it is also crucial to develop parameters that would be different for differ- ent interstitial diseases such as fibrosis and edema. This should be the pathway to true lung ultrasound diagnosis. Currently, in conventional ultrasound imaging of the lung, many interstitial diseases lead to the presence of B-lines, vertical artifacts that arise in a nonfully aerated lung.
However, by analyzing the raw ultrasound data, it becomes possible to characterize these artifacts based on their intensity, bandwidth, and native frequency (Demi et al., 2020; Mento et al., 2020). And what has emerged is that this kind of characterization has an impact in the dis- crimination of patients affected by fibrosis from patients affected by other lung pathologies. In other words, this type of approach adds specificity.
Figure 7, top left, shows Gaussian distributions for native frequency and bandwidth representing 13 fibrotic patients. Results from 26 patients are then jointly depicted in the two- dimensional (2-D) plot (Figure 7, top right), where the x-axis and the y-axis represent the mean native frequency and the mean bandwidth, respectively, of the B-lines observed in each patient. The patient IDs are written near the corresponding marker. In Figure 7, bottom left, histograms of the vertical artifacts total intensity for fibrotic and nonfibrotic patients are shown. The total intensity (Itot) is a recently introduced metric designed to study the vertical artifact dependency on several imaging parameters (Demi et al., 2020; Mento et al., 2020). Figure 7, bottom right, shows the 2-D plot of the patients’ distribution as a function of the mean native frequency and mean bandwidth of the vertical artifacts observed in each patient after the application of a 1.6 dB threshold to the mean total intensity. As seen, accounting for the total intensity has the effect of further improving the method’s specificity (Mento et al., 2020). Indeed, because of increasing absorption with fibrotic tissue, it is expected that a reduction in artifact intensity would be displayed in fibrotic patients.
34 Acoustics Today • Winter 2021

   32   33   34   35   36