Clinical Applications of 4D Flow MRI in the Portal Venous System

Abstract

Evaluation of the hemodynamics in the portal venous system plays an essential role in many hepatic pathologies. Changes in portal flow and vessel morphology are often indicative of disease.Routinely used imaging modalities, such as CT, ultrasound, invasive angiography, and MRI, often focus on either hemodynamics or anatomical imaging. In contrast, 4D flow MRI facilitiates a more comprehensive understanding of pathophysiological mechanisms by simultaneously and noninvasively acquiring time-resolved flow and anatomical information in a 3D imaging volume.Though promising, 4D flow MRI in the portal venous system is especially challenging due to small vessel calibers, slow flow velocities, and breathing motion. In this review article, we will discuss how to account for these challenges when planning and conducting 4D flow MRI acquisitions in the upper abdomen. We will address patient preparation, sequence acquisition, postprocessing, quality control, and analysis of 4D flow data.In the second part of this article, we will review potential clinical applications of 4D flow MRI in the portal venous system. The most promising area for clinical utilization is the diagnosis and grading of liver cirrhosis and its complications. Relevant parameters acquired by 4D flow MRI include the detection of reduced or reversed flow in the portal venous system, characterization of portosystemic collaterals, and impaired response to a meal challenge. In patients with cirrhosis, 4D flow MRI has the potential to address the major unmet need of noninvasive detection of gastroesophageal varices at high risk for bleeding. This could replace many unnecessary, purely diagnostic, and invasive esophagogastroduodenoscopy procedures, thereby improving patient compliance with follow-up. Moreover, 4D flow MRI offers unique insights and added value for surgical planning and follow-up of multiple hepatic interventions, including transjugular intrahepatic portosystemic shunts, liver transplantation, and hepatic disease in children. Lastly, we will discuss the path to clinical implementation and remaining challenges.

Keywords: 4D flow magnetic resonance imaging; cirrhosis; hemodynamics; portal hypertension; portal vein.

Conflict of interest statement

Conflicts of Interest

None of the authors have any relevant conflicts. Unrelated to this work, Dr. Reeder has ownership interests in Calimetrix, Reveal Pharmaceuticals, Cellectar Biosciences, Elucent Medical, and HeartVista. The University of Wisconsin receives research support from GE Healthcare and Bracco Diagnostics.

Figures

Fig. 1

Anatomical schematics of the portal venous circulation. (a) illustrates multiple forms of portosystemic collaterals in portal hypertension. Shown are GEV fed by reversed flow in the LGV, PUC from the LPV, IGV from the splenic circulation, and SRS with flow from the SV into the LRV. The portal venous circulation is highly variable. Common variants include (b) the LGV confluent with the SV instead of the PV and (c) the IMV confluent with the SMV instead of the SV. GEV, gastroesophageal varices; IGV, isolated gastric varices; IMV, inferior mesenteric vein; LGV, left gastric vein; LPV, left portal vein; LRV, left renal vein; PUC, paraumbilical collaterals; PV, portal vein; RPV, right portal vein; SMV, superior mesenteric vein; SRS, spontaneous splenorenal shunt; SV, splenic vein.

Fig. 2

Visualization and quantification of abdominal hemodynamics using 4D flow MRI: (Left) Time-averaged velocity images (Vx, Vy, and Vz) and MAG images from a 4D flow MRI acquisition in a healthy volunteer shown for a single coronal slice. These source images are used to semi-automatically create a vessel segmentation of venous (blue), arterial (red), and portal (purple) vasculature (middle). Within the boundaries given by the segmentation, velocity information can be portrayed via color-coded velocity streamlines and quantified. White arrows depict the direction of blood flow in portal vasculature. Volumetric flow rates of the SMV and the SV add up (0.42 mL/min + 1.09 mL/min = 1.51 mL/min) to the flow volume of the PV (1.6 mL/min) resulting in an error of 5.6% and confirming good data quality. MAG, magnitude; PV, portal vein; SMV, superior mesenteric vein; SV, splenic vein.

Fig. 3

Risk assessment of GEV. Velocity color-coded streamlines of portal vasculature and semi-transparent segmentation masks of arteries (red) and veins (blue) in the upper abdomen. Direction of flow is demarcated by orange arrows. (a) Healthy 46-year-old woman with no varices and hepatopetal flow in the SMV, SV, and portal vein. A FFC above 0 confirms the absence of portosystemic shunts between the measurement planes in SV, SMV and PVdist. (b) A 64-year-old man with varices at an endoscopically assessed low risk of bleeding. Note the hepatopetal flow in the LGV. (c) A 54-year-old man with high-risk varices. Hepatofugal flow is observed in the LGV and in gastroesophageal varices. FFC is below 0, reflecting increased shunting. This patient also has PUC supplied by the LPV. FFC, fractional flow change; GEV, gastroesophageal varices; LGV, left gastric vein; LPV, left portal vein; PUC, paraumbilical collaterals; PVdist, distal portal vein; SMV, superior mesenteric vein; SV, splenic vein.

Fig. 4

Velocity streamlines in the portal venous system before and after a meal challenge in a cirrhotic patient with large GEV and a healthy subject. Note the absence of relevant flow changes in the PV and the HA of the cirrhotic patient after the meal challenge. This is reflected by the relatively constant PVF. In contrast, there is a relevant increase in flow and velocity in the portal vein after the meal and a reduction in these parameters in the hepatic artery. This results in a substantial increase in the PVF after the meal, which is characteristic for healthy subjects. GEV, gastroesophageal varices; HA, hepatic artery; PV, portal vein; PVF, portal vein fraction

Fig. 5

Portosystemic shunts: (a) Patient with a large LGV with streamlines visualizing hepatofugal flow draining blood into GEV from where it drains into the AZY. Flow in the azygos vein measured 0.24 L/min, an independent risk factor for indicator for GEV at high risk for bleeding. Note the anatomical variant of the LGV that is confluent with the SV. (b) Patient with paraumbilical collaterals originating from the LPV as well as GEV that are supplied by the LGV. Again, flow reversal can be seen in the LGV that is confluent with the PV. Direction of flow is demarcated by yellow arrows. AZY, azygos vein; GEV, gastroesophageal varices; LGV, left gastric vein; LPV, left portal vein; PV, portal vein; SV, splenic vein.

Fig. 6

Assessment of TIPS placement in a 54-year-old man with nonalcoholic steatohepatitis, portal hypertension, and resultant refractory ascites. (a and c) Colored 3D segmentation masks of the complex difference angiogram and (b and d) velocity streamlines with arrowheads depict anatomy and hemodynamics, respectively. (a and b) before and (c and d) 2 weeks after TIPS procedure. Note the increase in blood flow in the portal vasculature after the procedure and the high velocities within the shunt. Disordered flow explains the signal dropout at the proximal end of the shunt. Ascites resolved after TIPS placement. IVC, Inferior vena cava; PV, portal vein; SMV, superior mesenteric vein; SV, splenic vein; TIPS, transjugular intrahepepatic portosystemic shunt.