Accurate determination of patient-specific boundary conditions in computational vascular hemodynamics using 3D cine phase-contrast MRI

In the patient-specific vascular CFD, determination of the inlet and outlet boundary conditions (BCs) is an important issue for a valid diagnosis. The 3D cine phase-contrast MRI (4D Flow) velocimetry is promising for this issue; yet, its measured velocities contain relatively large error and are not admissible as the BCs without any correction. This paper proposes a novel correction method for determining the BCs accurately using the 4D Flow velocimetry. First, we reveal that the error of the velocity measured by the 4D Flow at each measurement voxel is large but is distributed symmetrically. Secondly, our method pays attention to the incompressibility of the blood and the fact that the volume flow rate (VFR) in each vessel is constant on any cross sections. We reveal that the average of the cross-sectional VFRs integrated from many measurement voxel in each vessel is accurate despite the large error. Finally, we propose the novel correction method, which applies a smoothing to the measured velocities on each inlet or outlet boundary with a low-pass filter and then corrects them with the VFR. The results of the several phantom studies are presented to validate the accuracy of our method. A demonstrative analysis for an actual aneurysm is also presented to show the feasibility and effectiveness of our method.