Quantitative MRI to Assess Portal Pressure in Patients With Liver Disease: Validation at 3 T

ABSTRACT Hepatic venous pressure gradient (HVPG) is the gold standard for assessing portal hypertension (PH). We previously demonstrated that a bivariate MRI model combining liver T 1 and splenic artery velocity at 1.5 T correlates with HVPG. This study aimed to evaluate this model at 3 T and investigate additional splanchnic MRI measures as potential non‐invasive markers of portal pressure. Data from 40 participants (15 metabolic dysfunction–associated steatotic liver disease [MASLD]/13 alcohol‐related liver disease [ArLD]/12 other, 55 ± 14 years) scanned at 1.5 T (retrospective) and 42 participants (21 MASLD/14 ArLD/7 other, 60 ± 12 years) scanned prospectively at 3 T who underwent clinical HVPG measurement were analysed. Liver and spleen iron‐corrected T 1 (c T 1 ) were acquired using fat‐suppressed spin‐echo echo‐planar T 1 mapping, and splanchnic haemodynamics were assessed using phase‐contrast MRI. Correlations between MRI parameters and HVPG were assessed, and previously derived univariate (liver c T 1 ) and bivariate (liver c T 1 + splenic artery velocity) models were evaluated using receiver operating characteristic (ROC) analysis for PH (HVPG ≥ 5 mmHg), clinically significant PH (CSPH; HVPG ≥ 10 mmHg) and increased risk of mortality (HVPG ≥ 15 mmHg). Liver c T 1 at 3 T showed a strong positive correlation with HVPG across the full range (0–23 mmHg; R = 0.73, p < 0.0001). After field adjustment, pooled 1.5 and 3 T liver c T 1 maintained a strong correlation with HVPG ( R = 0.76, p < 0.0001). At 3 T, the univariate liver c T 1 model outperformed the bivariate model, achieving AUROCs of 0.74, 0.86 and 0.91 for detecting PH, CSPH and HVPG ≥ 15 mmHg, respectively. In contrast, spleen c T 1 and superior mesenteric artery velocity correlated with HVPG only up to ~15 mmHg, after which values plateaued or declined. Liver c T 1 MRI at 3 T provides a surrogate of portal pressure across the full HVPG range and outperforms a previously proposed bivariate model. Spleen and splanchnic haemodynamic measures demonstrate a ceiling effect at higher portal pressures > 15 mmHg.