TY - JOUR
T1 - Acquisition of 3-D arterial geometries and integration with computational fluid dynamics
AU - Hammer, Steven J
AU - Jeays, Adam
AU - Allan, Paul L
AU - Hose, Rod
AU - Barber, David
AU - Easson, William J
AU - Hoskins, Peter R
PY - 2009/12
Y1 - 2009/12
N2 - A system for acquisition of 3-D arterial ultrasound geometries and integration with computational fluid dynamics (CFD) is described. The 3-D ultrasound is based on freehand B-mode imaging with positional information obtained using an optical tracking system. A processing chain was established, allowing acquisition of cardiac-gated 3-D data and segmentation of arterial geometries using a manual method and a semi-automated method, 3D meshing and CFD. The use of CFD allowed visualization of flow streamlines, 2-D velocity contours and 3-D wall shear stress. Three-dimensional positional accuracy was 0.17-1.8 mm, precision was 0.06-0.47 mm and volume accuracy was 4.4-15%. Patients with disease and volunteers were scanned, with data collection from one or more of the carotid bifurcation, femoral bifurcation and abdominal aorta. An initial comparison between a manual segmentation method and a semi-automated method suggested some advantages to the semi-automated method, including reduced operator time and the production of smooth surfaces suitable for CFD, but at the expense of over-smoothing in the diseased region. There were considerable difficulties with artefacts and poor image quality, resulting in 3-D geometry data that was unsuitable for CFD. These artefacts were exacerbated in disease, which may mean that future effort, in the integration of 3-D arterial geometry and CFD for clinical use, may best be served using alternative 3-D imaging modalities such as magnetic resonance imaging and computed tomography. (E-mail: [email protected])
AB - A system for acquisition of 3-D arterial ultrasound geometries and integration with computational fluid dynamics (CFD) is described. The 3-D ultrasound is based on freehand B-mode imaging with positional information obtained using an optical tracking system. A processing chain was established, allowing acquisition of cardiac-gated 3-D data and segmentation of arterial geometries using a manual method and a semi-automated method, 3D meshing and CFD. The use of CFD allowed visualization of flow streamlines, 2-D velocity contours and 3-D wall shear stress. Three-dimensional positional accuracy was 0.17-1.8 mm, precision was 0.06-0.47 mm and volume accuracy was 4.4-15%. Patients with disease and volunteers were scanned, with data collection from one or more of the carotid bifurcation, femoral bifurcation and abdominal aorta. An initial comparison between a manual segmentation method and a semi-automated method suggested some advantages to the semi-automated method, including reduced operator time and the production of smooth surfaces suitable for CFD, but at the expense of over-smoothing in the diseased region. There were considerable difficulties with artefacts and poor image quality, resulting in 3-D geometry data that was unsuitable for CFD. These artefacts were exacerbated in disease, which may mean that future effort, in the integration of 3-D arterial geometry and CFD for clinical use, may best be served using alternative 3-D imaging modalities such as magnetic resonance imaging and computed tomography. (E-mail: [email protected])
KW - 3-D
KW - artery
KW - flow-field
KW - optical tracking
KW - segmentation
KW - wall shear stress
KW - ultrasound imaging
U2 - 10.1016/j.ultrasmedbio.2009.06.1099
DO - 10.1016/j.ultrasmedbio.2009.06.1099
M3 - Article
C2 - 19828230
SN - 0301-5629
VL - 35
SP - 2069
EP - 2083
JO - Ultrasound in Medicine and Biology
JF - Ultrasound in Medicine and Biology
IS - 12
ER -