Caval Subtraction 2D Phase-Contrast MRI to Measure Total Liver and Hepatic Arterial Blood Flow Proof-of-Principle, Correlation With Portal Hypertension Severity and Validation in Patients With Chronic Liver Disease

Objectives: Caval subtraction phase-contrast magnetic resonance imaging (PCMRI) noninvasive measurements of total liver blood flow(TLBF) and hepatic arterial (HA) flow have been validated in animal models and translated into normal volunteers, but not patients. This study aims to demonstrate its use in patients with liver cirrhosis, evaluate measurement consistency, correlate measurements with portal hypertension severity, and invasively validate TLBF measurements. Materials and Methods: Local research ethics committee approval was obtained. Twelve patients (mean, 50.8 +/- 3.1 years; 10 men) with histologically confirmed cirrhosis were recruited prospectively, undergoing 2-dimensional PCMRI of the portal vein (PV) and the infrahepatic and suprahepatic inferior vena cava. Total liver blood flowand HA flow were estimated by subtracting infrahepatic from suprahepatic inferior vena cava flow and PV flow from estimated TLBF, respectively. Invasive hepatic venous pressure gradient (HVPG) and indocyanine green (ICG) clearance TLBF were measured within 7 days of PCMRI. Bland-Altman (BA) analysis of agreement, coefficients of variation, and Pearson correlation coefficients were calculated for comparisons with direct inflow PCMRI, HVPG, and ICG clearance. Results: The mean difference between caval subtraction TLBF and direct inflow PCMRI was 6.3 +/- 4.2 mL/min/100 g (BA 95% limits of agreement, +/- 28.7 mL/min/100 g). Significant positive correlations were observed between HVPG and caval subtraction HA fraction (r = 0.780, P = 0.014), but not for HA flow (r = 0.625, P = 0.053), PV flow (r = 0.244, P = 0.469), or caval subtraction TLBF (r = 0.473, P = 0.141). Caval subtraction and ICG TLBF agreement was modest (mean difference, -32.6 +/- 16.6 mL/min/100 g; BA 95% limits of agreement, similar to 79.7 mL/min/100 g), but coefficients of variation were not different (65.7% vs 48.1%, P = 0.28). Conclusions: In this proof-of-principle study, caval subtraction PCMRI-measurements are consistent with direct inflow PCMRI, correlate with portal hypertension severity, and demonstrate modest agreement with invasive TLBF measurements. Larger studies investigating the clinical role of TLBF and HA flow measurement in patients with liver disease are justified.