4D-Flow MRI for cardiovascular evaluation - application in routine clinical practice and integration in standard clinical workflow

Abstract

4D-flow MRI is a relatively novel imaging modality with growing applications in cardiovascular clinical workflow and research. The wealth of information obtained by 4D-flow MRI is unparalleled by conventional cardiac imaging modalities and has the potential to support the development of novel clinical parameters to improve diagnosis, assessment, and risk stratification of cardiovascular diseases. The studies reported in this thesis sought to examine the feasibility and validity of incorporating 4D-flow MRI and its derived clinical parameters into routine cardiovascular diagnostic workflow. In addition, the potential of 4D-flow MRI to improve the accuracy of CFD modelling of the LV was also examined.

Systematic examinations of existing evidence found several 4D-flow MRI derived parameters to have the potential of becoming clinical markers to improve the quantification of LV function and the risk stratification of aortopathy associated with bicuspid aortic valve. However, the existing evidence were all supported by small-scale studies, and substantial further investigations are required to prove the validity of these parameters in clinical use.

The methodology and pilot analysis of a large-scale non-inferiority study was reported with the aim to demonstrate clinical non-inferiority of 4D-flow MRI compared with conventional cardiac MRI, and the feasibility of incorporating 4D-flow MRI into routine clinical workflow by leveraging standardised imaging analysis protocols to reduce the dependence on MRI specialists. A large cohort investigation of bulk flow measurements in the large vessels comparing 4D-flow and 2D-PC supported 4D-flow as a potentially valid tool for clinical diagnosis of cardiac pathologies.

In CFD modelling of the LV, data derived from 4D-flow MRI was able to improve boundary condition modelling and overall model accuracy in a small-scale investigation. A clinically pragmatic and accurate model remains a challenging topic.