Quantification of common carotid artery and descending aorta vessel wall thickness from MR vessel wall imaging using a fully automated processing pipeline

Abstract

4 J. Magn. Reson. Imaging 2017;45:215-228.

Learning‐based automated segmentation of the carotid artery vessel wall in dual‐sequence MRI using subdivision surface fitting

“…Previous studies have attempted to develop automated methods for segmenting carotid artery wall in MR images. These algorithms range from 2D segmentations (Snake, discrete dynamic contour model, ellipse model, active shape model, B‐spline Snake) to 3D segmentations (cylinder model, intensity probability density function matching, graph‐cut). Compared with the method proposed by Ukwatta et al., we found that for the CCA, our method yielded a RMSED of 0.22 mm for the lumen contour (LC) and 0.40 mm for the outer wall contour (OWC), while their method yielded RMSED of 0.3 mm and 0.5 mm for the LC and OWC.…”

“…Three‐dimensional deformable models, supporting high interindividual anatomical variability of blood vessels, can be used to reproduce the real vascular geometry at single patient level for disease characterization. Algorithms based on cylindrical NURBS surface models have been proposed for segmenting the carotid lumen from Time‐of‐Flight (TOF) MRA and the carotid vessel wall from BB‐MRI . However, with this method, two separate tube models must be used to obtain the segmentations from Common Carotid Artery (CCA) to Internal Carotid Artery (ICA) and from CCA to External Carotid Artery (ECA), which may introduce inaccurate segmentations at the bifurcation region.…”

“…Algorithms based on cylindrical NURBS surface models have been proposed for segmenting the carotid lumen from Time-of-Flight (TOF) MRA 14 and the carotid vessel wall from BB-MRI. 15,16 However, with this method, two separate tube models must be used to obtain the segmentations from Common Carotid Artery (CCA) to Internal Carotid Artery (ICA) and from CCA to External Carotid Artery (ECA), which may introduce inaccurate segmentations at the bifurcation region. Recently, Kitslaar et al 17 introduced a hierarchical-tree model based on subdivision surface for the segmentation of branching vascular structures in computed tomography angiography (CTA) images.…”

Artificial Intelligence, Machine Learning, and Cardiovascular Disease

“…51 Similarly, calculation of left ventricular mass, papillary muscle identification, common carotid artery, and descending aorta measurements with fully automated AI programs have been performed with high accuracy and reproducibility. 52,53 AI-based algorithms therefore have already made substantial impact in diagnosing coronary artery disease, risk stratification, 32,33,44,45 cardiovascular imaging modalities such as echocardiography, 39,45 and cardiac MRI. 40,41,46,[51][52][53] However, it has to be noted that most of these studies are descriptive studies and are done at the state of the art centers.…”

“… 51 Similarly, calculation of left ventricular mass, papillary muscle identification, common carotid artery, and descending aorta measurements with fully automated AI programs have been performed with high accuracy and reproducibility. 52 , 53 …”