Efficient initialisation of distance-regularised level set without re-initialisation scheme and quantitative evaluation of IMT in B mode ultrasound common carotid artery images

Abstract

Segmentation of the Common Carotid Artery (CCA) wall is imperative for the estimation of the Intima Media Thickness (IMT) in B-mode longitudinal ultrasound images. IMT is considered as the prominent indicator of atherosclerosis and is defined as the distance between Lumen-Intima (LI) and Media-Adventitia (MA) interfaces of CCA vessel. This paper proposes an efficient initialisation scheme for Distance-Regularised Level Set without re-initialisation Evolution (DRLSE) method by utilising Fast Fuzzy C-Mean clustering (FFCM) technique to detect Li and MA interfaces accurately. Distance between the detected interfaces is determined by Polyline Distance Metric (PDM). Also, a quantitative and qualitative comparison among proposed method, manual method and snake method is presented. Experimental results show that the proposed method is robust in terms of region of interest selection (ROI), noise (speckle as well as salt and pepper) and images acquired from different machines. Inter-method, inter-operator and intra-operator variability have been studied by computing average, standard deviation, Coefficient of Variation (CV%) and via Wilcoxon’s signed-rank test (at p < 0.05). Reported results demonstrate close agreement between manual and proposed methods. The proposed method is semiautomatic and bears an advantage that a tight ROI is not required which encourages its use in clinical practice.