Mitral annulus segmentation from three-dimensional ultrasound

Abstract

An accurate and reproducible segmentation of the mitral valve annulus from 3D ultrasound is useful to clinicians and researchers in applications such as pathology diagnosis and mitral valve modeling. Current segmentation methods, however, are based on 2D information, resulting in inaccuracies and a lack of spatial coherence. We present a segmentation algorithm which, given a single user-specified point near the center of the valve, uses max-flow and active contour methods to delineate the annulus geometry in 3D. Preliminary comparisons to manual segmentations and a sensitivity study show the algorithm is both accurate and robust.

Mitral Annulus Segmentation From 3D Ultrasound Using Graph Cuts

“…To avoid the variability inherent in user input, we developed an automated method which initializes contours by computing min-cuts on 2D graphs using the max-flow algorithm, with edge capacities derived from both P int and P ttd for added robustness. This method is preferred over our preliminary method which initialized generic contours in the projection images [33]. To overcome noise and anatomic variability in P int and P ttd , multiple spatial scales of the projection images are used to construct multiple resolution-specific contours.…”

Cardiac valve annulus manual segmentation using computer assisted visual feedback in three-dimensional image data

“…This agrees with a recent publication in semi-automatic segmentation of the annulus of the mitral valve, where the authors noted that manual segmentations using 2D slices "only allow access to a portion of the available information at each step. This requires the user to mentally interpolate the annulus surroundings, and often results in a lack of spatial coherence" [18]. Mental interpolation is made more difficult by the fact that the annulus is not planar, as discussed in section I.…”

“…A few methods for semi-automatic segmentation of valve annulae have been proposed in the last decade, e.g. [11], [13], [18], but manual segmentations have still been used for validation as a pseudo criterion (gold) standard. Therefore, annulus manual segmentation remains the dominating tool for the study of valve anatomy and physiology.…”

“…Annulae have usually been approximated in the literature by a closed curve, e.g. [18], or a toroid-like solid obtained from the concatenation of disks or spheres in each slice [8]. The curve must be allowed to be nonplanar, as valves such as the mitral have an annulus with a saddle shape [3], [4], the tricuspid annulus has a more planar saddle shape and is rounder in the projected plane [7], [15], and the aortic annulus has a three pronged crown-like circlet shape [9].…”

“…Regarding imaging modalities, echocardiography has been historically preferred in the literature for annulus segmentation, even in recent publications [16], [18], as it has good temporal resolution, is safe, cheap and non-invasive. However, if motion analysis is sacrificed, three-dimensional (3D) Magnetic Resonance Imaging (MRI) provides much better image quality than echocardiography, principally because of higher spatial resolution and lack of speckle texture.…”