Atherosclerosis is characterized by the growth of fibrous plaques due to the retention of cholesterol and lipids within the artery wall, which can lead to vessel occlusion and cardiac events. One way to evaluate arterial disease is to quantify the amount of lipid present in these plaques, since a higher disease burden is characterized by a higher concentration of lipid. Although therapeutic stimulation of reverse cholesterol transport to reduce cholesterol deposits in plaque has not produced significant results, this may be due to current image analysis methods which use averaging techniques to calculate the total amount of lipid in the plaque without regard to spatial distribution, thereby discarding information that may have significance in marking response to therapy. Here we use Directional Fourier Spatial Frequency (DFSF) analysis to generate a characteristic spatial frequency spectrum for atherosclerotic plaques from C57 Black 6 mice both treated and untreated with a cholesterol scavenging nanoparticle. We then use the Cauchy product of these spectra to classify the images with a support vector machine (SVM). Our results indicate that treated plaque can be distinguished from untreated plaque using this method, where no difference is seen using the spatial averaging method. This work has the potential to increase the effectiveness of current in-vivo methods of plaque detection that also use averaging methods, such as laser speckle imaging and Raman spectroscopy.
|