Isosurfaces as deformable models for magnetic resonance angiography

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Cardiovascular Research

Background:

• Vascular disease alters lumenal shape, detectable via magnetic resonance angiography (MRA).
• Quantifying vascular shape from MRA is challenging due to image artifacts.
• Existing deformable models for vascular reconstruction often require complex surface mesh initialization.

Purpose of the Study:

• To develop a novel deformable isosurface methodology for accurate vascular shape reconstruction from MRA.
• To overcome limitations of prior models by utilizing trivial isosurface initialization for MRA.
• To improve the quantification of vascular stenosis and surface smoothness in MRA-derived models.

Main Methods:

• A deformable isosurface model was developed to conform to object boundaries in MRA images with minimal shape assumptions.
• External forces attract the surface to image edges, while a novel moment-based smoothing aligns surface triangle normals without translational motion.
• The method was tested on a digital arterial phantom and MRA data from human renal and carotid arteries.

Main Results:

• The deformable isosurface method demonstrated improved surface smoothness and more accurate stenosis measurements compared to standard isosurfaces.
• For renal arteries, stenosis measurements using the deformable model were within 12% +/- 6% of conventional X-ray angiography.
• The model successfully reconstructed vascular surfaces from MRA data of diseased arteries.

Conclusions:

• The proposed deformable isosurface method offers a robust approach for vascular shape reconstruction from MRA.
• This technique enhances the accuracy of stenosis quantification and improves the quality of vascular surface models derived from MRA.
• The method presents a significant advancement for analyzing vascular morphology in the presence of image artifacts.