Measurement and Analysis of Intracardiac Blood Flow and Vortex Ring Formation

Abstract: Increased understanding of the pumping mechanics of the heart are of great importance to develop diagnostics, treatment and prognostics for cardiovascular diseases. Blood flow in the heart is connected to its anatomy and function, and may therefore be a sensitive marker of cardiac health and disease. Therefore, the aim of this thesis is to develop and validate methods for quantification and visualization of intracardiac blood flow measurements using 4D phase contrast magnetic resonance velocity mapping (4D PC-MR). The thesis, which is based on five papers, aims to 1) validate measurement accuracy, 2) investigate vortex ring formation in the left ventricle (LV), and 3) evaluate and improve visualization of flow. For aim 1), 4D PC-MR stroke volume (SV) measurements in the aorta and main pulmonary artery were validated against 2D PC-MR at 1.5T and 3T, using two different 4D PC-MR sequences, one accelerated using SENSE and the other using k-t BLAST (Paper I). SENSE measurements showed good accuracy and measurements at 3T compared favorably to 1.5T. The k-t BLAST measurements showed a too high bias to be used for SV quantification. Furthermore, a phantom setup for validation of 4D PC-MR against independent measurements by particle imaging velocimetry (PIV) and planar laser-induced flourescence (PLIF) was developed and constructed (Paper II). The developed flow phantom showed excellent stability (R^2 =0.96, bias -0.06 +- unit 0.70 cm/s), making it suitable for validation of 4D PC-MR measurements. Validation of 4D PC-MR velocities against PIV show good agreement for mean velocities, but 4D PC-MR underestimates peak velocities by 8-25%. Vortex ring volume (VV) measurements with 4D PC-MR showed good agreement with PLIF. However, vortex ring mixing ratio (MXR) showed poor agreement. Due to possible differences between the phantom setup and in vivo vortex ring formation, further studies are needed to determine if MXR can be measured under in vivo flow conditions. For aim 2), a new method for quantification of vortex ring formation in the left ventricle using Lagrangian Coherent Structures was developed and implemented in software (Paper IV). Vortex ring volume was quantified in 15 healthy volunteers and 15 patients with heart failure. The vortex ring occupied 51 +- 7% of the LV blood volume in healthy volunteers, but only 26+-5% in the patients (p<0.001). This suggests that a larger volume of blood is static in the LV of the patients, with an associated increase in the risk of thrombus formation (Papers IV and V). The vortex ring mixing ratio (MXR), defined as the amount of blood pulled into the vortex ring due to its rotation divided by the total volume of the vortex ring, was also quantified in healthy volunteers and heart failure patients (Paper V). MXR was higher in the patients compared to the volunteers (33+-7% vs 19+-7%, p<0.001). For aim 3), a new method for visualization of 4D PC-MR blood flow measurements was developed and implemented in software. The new method, called Volume Tracking (Paper III), allows visualization of the motion of a blood volume through the heart. Volume Tracking gives incremental information about the blood flow compared to earlier used methods, e.g. by revealing a complex blood flow pattern in the right ventricle (RV) when compared to the LV. Additionally, the quality of particle tracing visualizations in 4D PC-MR accelerated using SENSE or k-t BLAST was evaluated (Paper I). No difference could be measured, showing that the higher acceleration, and therefore shorter scan duration, in k-t BLAST measurements can be used when the main goal is to visualize, and not quantify, blood flow.