An approach for patient-specific multi-domain vascular mesh generation featuring spatially varying wall thickness modeling. Issue 10 (16th July 2015)

• Record Type: Journal Article
• Authors: Raut, Samarth S.; Liu, Peng; Finol, Ender A.
• Is Part Of: Journal of biomechanics. Volume 48:Issue 10(2015)
• Page Start: 1972
• Page End: 1981
• Publication Date: 2015-07-16
• Subjects: Aneurysm -- Vascular mesh generation -- Cardiovascular biomechanics -- Multi-domain -- Finite element analysis; Animal mechanics -- Periodicals; Biomechanics -- Periodicals; Biomechanics -- Periodicals; Mécanique animale -- Périodiques; Biomécanique -- Périodiques; Electronic journals; 571.4305
• DOI: 10.1016/j.jbiomech.2015.04.006
• Languages: English
• ISSNs: 0021-9290

Abstract

Abstract: In this work, we present a computationally efficient image-derived volume mesh generation approach for vasculatures that implements spatially varying patient-specific wall thickness with a novel inward extrusion of the wall surface mesh. Multi-domain vascular meshes with arbitrary numbers, locations, and patterns of both iliac bifurcations and thrombi can be obtained without the need to specify features or landmark points as input. In addition, the mesh output is coordinate-frame independent and independent of the image grid resolution with high dimensional accuracy and mesh quality, devoid of errors typically found in off-the-shelf image-based model generation workflows. The absence of deformable template models or Cartesian grid-based methods enables the present approach to be sufficiently robust to handle aneurysmatic geometries with highly irregular shapes, arterial branches nearly parallel to the image plane, and variable wall thickness. The assessment of the methodology was based on i) estimation of the surface reconstruction accuracy, ii) validation of the output mesh using an aneurysm phantom, and iii) benchmarking the volume mesh quality against other frameworks. For the phantom image dataset (pixel size 0.105 mm; slice spacing 0.7 mm; and mean wall thickness 1.401±0.120 mm), the average wall thickness in the mesh was 1.459±0.123 mm. The absolute error in average wall thickness was 0.060±0.036 mm, or about 8.6% of the largest image grid spacing (0.7 mm) and 4.36% of the actual mean wall thickness. Mesh quality metrics and the ability to reproduce regional variations of wall thickness were found superior to similar alternative frameworks.