Improvement of Hemodynamic Analysis by 4D Flow MRI: Reynolds Resolution, Partial Volume Effect, Carotid Flow

Abstract

This thesis addresses thorough discussions on the improvement of hemodynamic analysis by four-dimensional (4D) flow magnetic resonance imaging (MRI). The 4D flow MRI is a versatile tool to obtain three-dimensional, three-component velocity information and flow geometry information simultaneously. Recent advances in 4D flow MRI have enabled to relate the cardiovascular disease on various vessels to hemodynamic parameters. Among various hemodynamic parameters, the WSS, a force exerted on a vessel wall in parallel, is thought to play an important role in clinical applications such as assessing the development of atherosclerosis. Nevertheless, the accuracy of WSS obtained with 4D flow MRI is rarely evaluated or reported in the literature, especially in the in vivo studies. This thesis proposes a novel and facile criterion, called Reynolds resolution, to assess the accuracy of WSS estimation in 4D flow MRI studies. To include diverse and extensive cases, we measured three different circular tubing flows with a diameter of 40, 8, and 2 mm by two MRI scanners, three radiofrequency coils, and five different spatial resolutions. In addition, the near-wall velocity gradient, required to be determined prior to the WSS, is calculated using two methods: assuming a linear velocity profile or quadratic velocity profile near the wall. The accuracy of WSS obtained using each method and tubing flow is evaluated against the theoretical WSS value. As a result, it is found that the value of Reynolds resolution is in logarithmic relation to the WSS error. This thesis also addresses a facile but reliable method to calculate the near-wall velocity gradient and consequently, improve the estimation accuracy of WSS. The method locates a vessel wall boundary within a voxel considering the partial volume effect (PVE) of 4D flow MRI data. Either a linear or quadratic velocity profile near the wall is assumed the same as the aforementioned paragraph. The three laminar circular tubing flows (utilized to determine the relationship between Reynolds resolution and the WSS accuracy) and a stenotic phantom flow are measured by 4D flow MRI with different spatial resolutions. It results that considering PVE improves the accuracy of WSS estimation, but only when the spatial resolution is high. For example, the WSS error is reduced from 24.6% to 17.5% with the consideration of PVE in case of the 2 mm tubing flow with a quadratic velocity profile. The improvement is maximized when 4D flow MRI has both high spatial resolution and high value of Reynolds resolution. It is also found that the assumption of a linear velocity profile near wall results in the best WSS estimation, but only with both high Reynolds resolution and high spatial resolution. Finally, the hemodynamics of severe carotid artery stenosis before and after CEA is investigated using 4D flow MRI. An in vitro experiment utilizing carotid phantoms, which underwent a procedure that emulated CEA with/without the patch repair, is performed with a high spatiotemporal resolution. Two hemodynamic parameters, the normalized time-averaged WSS (NTA; WSS; ) and the oscillatory shear index (OSI), are evaluated on the carotid surface, and elicit distinct changes in the carotids before and after CEA. The use of normalized parameter allows comparison of diverse cases with different conditions of hemodynamics and vessel geometry. An abnormal region of carotids is also introduced to account for continuous high-shear regions (high NTA; and low OSI) and stenosis-prone (chaotic low-shear) regions (low NTA; and high OSI). As a result, the stenosis-prone regions of the carotid with the patch are noticeably larger than the corresponding regions in no-patch carotid. A large recirculating flow zone found in the stenosis-prone region of the overexpanded internal carotid artery (ICA) of the postoperative carotid with patch partially blocks the flow path into ICA, and consequently the flow rate is not recovered after successful surgery, unlike an expectation. These results require further confirmation through long-term followup observations.