Fractal and multifractal analysis of atherosclerotic plaque in ultrasound images of the carotid artery

Abstract

Abstract Stroke is the cause of death following ischemic heart disease in 87% of cases. The chance of stroke rises with the severity of carotid stenosis and the thickening of the carotid artery due to the deposition of plaque. This study analyses the non-linear parameters of ultrasound images of the plaque in the carotid artery and classifies the images based on the textural features. The non-linear analysis is implemented via fractal and multifractal methods. The fractal dimensions and the lacunarity differ significantly for symptomatic and asymptomatic plaques. The occurrence of multifractal spectra and the scaling exponent function, i.e., interleaving sets of singularity strength, proves the multi-scaling property. The multifractal characteristics quantify the heterogeneity in the textural features, and this could be used for improving the classification of symptomatic and asymptomatic plaque. The results show the significance of fractal parameters of plaque in deciding the severity of plaque and hence, aiding the diagnostic process.