Segmentation of the heart and major vascular structures in cardiovascular CT images

Abstract

Segmentation of organs in medical images can be successfully performed with shape-constrained deformable models. A surface mesh is attracted to detected image boundaries by an external energy, while an internal energy keeps the mesh similar to expected shapes. Complex organs like the heart with its four chambers can be automatically segmented using a suitable shape variablility model based on piecewise affine degrees of freedom.In this paper, we extend the approach to also segment highly variable vascular structures. We introduce a dedicated framework to adapt an extended mesh model to freely bending vessels. This is achieved by subdividing each vessel into (short) tube-shaped segments ("tubelets"). These are assigned to individual similarity transformations for local orientation and scaling. Proper adaptation is achieved by progressively adapting distal vessel parts to the image only after proximal neighbor tubelets have already converged. In addition, each newly activated tubelet inherits the local orientation and scale of the preceeding one. To arrive at a joint segmentation of chambers and vasculature, we extended a previous model comprising endocardial surfaces of the four chambers, the left ventricular epicardium, and a pulmonary artery trunk. Newly added are the aorta (ascending and descending plus arch), superior and inferior vena cava, coronary sinus, and four pulmonary veins. These vessels are organized as stacks of triangulated rings. This mesh configuration is most suitable to define tubelet segments.On 36 CT data sets reconstructed at several cardiac phases from 17 patients, segmentation accuracies of 0.61-0.80 mm are obtained for the cardiac chambers. For the visible parts of the newly added great vessels, surface accuracies of 0.47-1.17 mm are obtained (larger errors are asscociated with faintly contrasted venous structures).

Automatic image-based segmentation of the heart from CT scans

“…Actually, CT is often preferred by diagnosticians since it provides more accurate anatomical information about the visualized structures, thanks to its higher signal-to-noise ratio and better spatial resolution. Although computed tomography was at one time almost absent in cardiovascular examinations, recent technological advances in X-ray tubes, detectors, and reconstruction algorithms, along with the use of retrospectively gated spiral scanning, have opened the doors to new diagnostic opportunities [6], enabling the non-invasive derivation of the aforementioned functional parameters [7,8]. Therefore, computed tomography becomes an important imaging modality for diagnosing cardiovascular diseases [9].…”

Sharing and reusing cardiovascular anatomical models over the Web: a step towards the implementation of the virtual physiological human project

“…The Web page shown in figure 2 will be displayed as a result of this selection. Currently, the list contains the following models: ABI model of the cardiac ventricles (Nielsen et al 1991;LeGrice et al 1997), INRIA model of the cardiac ventricles (Peyrat et al 2007), which includes both coarse-and fine-grain versions (Lamecker et al 2009), INRIA Aorta, Philips Whole-Heart (Lorenz & von Berg 2006), Philips Whole-Heart with Extended Vessels (Peters et al 2008), Philips Hamburg Left Atrium with multiple anatomical variations (Lorenz & von Berg 2006), Sheffield Aorta (Barber et al 2007) and finally the UPF-CISTIB Biventricular Heart Atlas (Ordas et al 2007). The original authors have set the accessibility of the geometric data for each model.…”

Automatic Intra-operative Generation of Geometric Left Atrium/Pulmonary Vein Models from Rotational X-Ray Angiography

“…For the pre-procedural CTA and MRA images, the entire heart was segmented to extract all four chambers, myocardium, and major vascular structures [17]. The CT model (mean mesh, R-table for the GHT and optimal boundary detection functions) was trained on independent training data [17].…”

“…The CT model (mean mesh, R-table for the GHT and optimal boundary detection functions) was trained on independent training data [17]. The specific MR model for the MRA pulse sequence used was trained with a leave-5-out approach on the 15 pre-procedural MRA images (including image calibration, Sec.…”