How Low is Too Low? Optimizing VENC in 4D Flow MRI for the Portal Venous System

Motivation: 4D flow MRI is a promising method for non-invasive characterization of gastroesophageal varices and portal venous flow but choosing the optimal velocity encoding (venc) a priori is difficult. Goal(s): Determine the optimal venc in a cohort of cirrhotic patients and analyze the feasibility of unwrapping low venc data using a 4D Laplacian method to utilize its superior velocity-to-noise ratio. Approach: We acquired low, medium, and high venc data and then compared aliasing in regular and unwrapped reconstructions compared with high venc data as the reference. Results: The unwrapped medium venc data had the best agreement with the high venc data with improved VNR. Impact: Knowledge of optimal venc for portal venous acquisitions will reduce the need for velocity scout images for 4D flow acquisitions. Additionally, phase unwrapping may also facilitate analysis of higher velocity flow in the hepatic artery from the same datasets.