Level-set-based artery-vein separation in blood pool agent CE-MR angiograms

摘要: Blood pool agents (BPAs) for contrast-enhanced (CE) magnetic-resonance angiography (MRA) allow prolonged imaging times higher contrast and resolution. Imaging is performed during the steady state when agent distributed through complete vascular system. However, simultaneous venous arterial enhancement in this hampers interpretation. In order to improve visualization of arteries veins from steady-state BPA data, a semiautomated method artery-vein separation presented. method, central axis are used as initializations two surfaces that simultaneously evolve capture parts vasculature using level-set framework. Since can be close proximity each other, leakage evolving (venous) surface into (arterial) part inevitable. these situations, voxels labeled or based on arrival time respective surface. The evolution steered by external forces related feature images derived image data internal geometry level sets. paper, robustness accuracy three (based intensity, gradient, vessel-enhancement filtering) combinations them investigated tested seven patient datasets. To end, results with level-set-based segmentation compared reference-standard manually obtained segmentations. Best achieved applying combination intensity- gradient-based smoothness constraint curvature By datasets, it shown that, minimal user interaction, improved CE-MRA achieved.