Medially constrained deformable modeling for segmentation of branching medial structures: Application to aortic valve segmentation and morphometry

Abstract

Deformable modeling with medial axis representation is a useful means of segmenting and parametrically describing the shape of anatomical structures in medical images. Continuous medial representation (cm-rep) is a “skeleton-first” approach to deformable medial modeling that explicitly parameterizes an object’s medial axis and derives the object’s boundary algorithmically. Although cm-rep has effectively been used to segment and model a number of anatomical structures with non-branching medial topologies, the framework is challenging to apply to objects with branching medial geometries since branch curves in the medial axis are difficult to parameterize. In this work, we demonstrate the first clinical application of a new “boundary-first” deformable medial modeling paradigm, wherein an object’s boundary is explicitly described and constraints are imposed on boundary geometry to preserve the branching configuration of the medial axis during model deformation. This “boundary-first” framework is leveraged to segment and morphologically analyze the aortic valve apparatus in 3D echocardiographic images. Relative to manual tracing, segmentation with deformable medial modeling achieves a mean boundary error of 0.41 ± 0.10 mm (approximately one voxel) in 22 3DE images of normal aortic valves at systole. Deformable medial modeling is additionally demonstrated on pathological cases, including aortic stenosis, Marfan syndrome, and bicuspid aortic valve disease. This study demonstrates a promising approach for quantitative 3DE analysis of aortic valve morphology.

In-vivo heterogeneous functional and residual strains in human aortic valve leaflets

“…Deformable modeling : A model of the open AV, which in turn was based on 3DE data of patients not included in this study, was deformed to capture the geometry of the aortic leaflet surfaces in the multi-atlas segmentation result (Pouch et al, 2015c). The deformable model used here was a continuous medial axis representation (Yushkevich et al, 2006a), which enabled the extraction of a medial surface of each aortic cusp (i.e.…”

Fully Automatic 3-D-TEE Segmentation for the Planning of Transcatheter Aortic Valve Implantation

“…See http://www.ieee.org/publications standards/publications/rights/index.html for more information. among others) or the required level of user interaction (fully automatic [12], [15]- [17], [21], [22], semi-automatic [18]- [20], [23], [24] or interactive [14]). Among 3-D-TEE methodologies, Ionasec et al [12] proposed the first fully automatic algorithm for AV segmentation.…”

“…Although being fully automatic, no AV detection algorithm was required since the authors claim that the initialization can be derived directly from the ultrasound machine (recorded roll angle). Finally, Pouch et al [23] proposed a semiautomatic methodology to extract the AV root and leaflets using multi-atlas label fusion and template-based branching medial modeling. Their strategy requires five initial landmark points to register the target image with each atlas.…”

Image Segmentation and Modeling of the Pediatric Tricuspid Valve in Hypoplastic Left Heart Syndrome

“…To obtain a deformable medial template of the tricuspid valve, one of the 16 manual segmentations was selected at random and a triangulated mesh of the segmentation’s skeleton was created using the procedure described in [11]. To reduce potential bias associated with a single-subject template, the single-subject template was fitted to all the subject’s multi-atlas segmentations, and generalized Procrustes analysis was used to compute a new “average” template.…”

“…Next, in order to make consistent quantitative measurements, a deformable model of the tricuspid valve is warped to capture the geometry of the valve in the label map. The combination of label fusion and deformable modeling has a number of benefits that have been demonstrated in adult mitral and aortic valve applications [10, 11]: it exploits knowledge of tricuspid valve image appearance through the use of expert-labeled atlases; the deformable template has immutable topology, thereby preventing extraneous holes or artifacts in the output model; and the valve is represented volumetrically, as a structure with locally varying thickness. The goal of this work is to demonstrate the first semi-automated image segmentation and geometric modeling of the tricuspid valve in 3DE images from pediatric patients with HLHS.…”