Automated in vivo segmentation of carotid plaque MRI with Morphology‐Enhanced probability maps

Abstract

MRI is a promising noninvasive technique for characterizing atherosclerotic plaque composition in vivo, with an end-goal of assessing plaque vulnerability. Because of limitations arising from acquisition time, achievable resolution, contrast-to-noise ratio, patient motion, and the effects of blood flow, automatically identifying plaque composition remains a challenging task in vivo. In this article, a segmentation method using maximum a posteriori probability Bayesian theory is presented that divides axial, multi-contrast-weighted images into regions of necrotic core, calcification, loose matrix, and fibrous tissue. Key advantages of the method are that it utilizes morphologic information, such as local wall thickness, and coupled active contours to limit the impact from noise and artifacts associated with in vivo imaging. In experiments involving 142 sets of multi-contrast images from 26 subjects undergoing carotid endarterectomy, Numerous studies have shown that MRI exhibits high contrast for internal plaque features, but that combined information from multiple contrast weightings is critical for distinguishing all plaque components (1-6). Based on these studies, desirable combinations of contrast weightings and a set of image characteristics have emerged that can be used to segment plaque into its subcomponents. Manual segmentation using these characteristics has produced quantitative measurements of the relative volumes of necrotic cores, calcification, loose matrix, and fibrous tissue that correlate strongly with histologic assessments (5).Nevertheless, replacing subjective, manual segmentation with an automated segmentation alternative would have several benefits. Aside from saving time in image review, automated segmentation would reduce the considerable amount of training required to read these images and the corresponding inter-rater variability. Additionally, a viable, automated segmentation procedure would permit various combinations of contrast weightings and image characteristics to be objectively analyzed for accuracy in plaque characterization. Such studies have been conducted using automated segmentation of ex vivo endarterectomy specimens (6,7), but these results are difficult to translate to in vivo imaging, given the different constraints regarding acquisition time, resolution, contrast-to-noise ratio, and effects of blood flow. Recent efforts to demonstrate in vivo segmentation of some plaque components (8,9) are promising, but they have not yet been histologically validated.The goal of this work was to develop a flexible, multicontrast plaque segmentation technique that is suitable for objectively testing various approaches for measuring plaque composition in vivo and to validate the method with histology. Furthermore, the technique was developed to mimic the highly successful procedure used in manual review. Unlike other works (6,8,9) in which only the intensity values in each contrast weighted image were considered, and (7) in which spatial information was used just for minimizing pixel di...

Imaging of high-risk carotid artery plaques: current status and future directions

“…Validation of automated carotid plaque analysis utilizing predominantly 2D MR sequences has been completed. 13 Future work will extend the automated carotid plaque analysis to include all 3D MR sequences.…”

Rationale and Design of the Carotid Plaque in Human for All Evaluations With Aggressive Rosuvastatin Therapy (CHALLENGER Trial) Evaluation by Magnetic Resonance Imaging

“…19 Plaque components, such as the necrotic core, calcification, loose matrix, and fibrous tissue, will be analyzed by automated segmentation of the plaques as mean area for the artery and mean % of wall area on MRI. 20 …”

Carotid Plaque Stresses

“…Indeed, the characterization of atherosclerotic lesions using MRI approaches histological definitions (Cai et al, 2002). Recently, semi-automated tissue segmentation has been enabled (Liu et al, 2006). Furthermore, MRI has the ability to measure blood velocities through phase-contrast imaging (Firmin et al, 1990;McDonnell, III et al, 1994) and deformations using cine MR imaging (Draney et al, 2002;Metafratzi et al, 2002).…”

“…Recently, an approach utilizing computational morphology enhanced probability maps was described allowing for fully automated segmentation of carotid plaque morphology. (Liu et al, 2006) MR Contrast Weighting Plaque Component TOF T1W PDW T2W…”