Magnetic resonance elastography (MRE) in assessing hepatic fibrosis: performance in a cohort of patients with histological data.

Abstract

The purpose of this study was to compare fibrosis seen on liver biopsy to MR elastography (MRE) stiffness measurements in normal controls and patients with abnormal transaminases and chronic liver disease. The control group consisted of 22 healthy liver transplant donors who by definition had normal transaminases. The patient group (32 patients) was recruited from the Mayo Clinic Arizona hepatobiliary clinic over a 3-year span. All subjects underwent a liver biopsy as part of their evaluation and agreed to MRE within 35 days of biopsy. Non-parametric tests were used to compare the MRE-predicted liver fibrosis to the fibrosis noted on liver biopsy. Analysis included 54 subjects (32 patients with chronic liver disease and 22 healthy liver donor candidates). MRE median liver stiffness measurements increased per histologic liver fibrosis stage (one-way ANOVA p = 0.001), with significant correlation between increasing fibrosis stage and stiffness values. Median MRE for control group (2.13 kPa; mean = 2.3 ± 0.6 kPa) was significantly lower than for patient group (3.7 kPa; mean = 4.1 ± 2.1 kPa) (p = 0.003). Significantly, lower median stiffness was seen in zero-to-moderate (F0-2, n = 22) vs. severe fibrosis stages (F3-4, n = 10) 2.80 vs. 5.9 kPa, respectively (p < 0.05). Using a 3.7-kPa cut-off value, the predicted sensitivity and specificity for detecting F0-2 from F3-4 were 91% and 80%, respectively. Our analysis supports previous findings that MRE is a non-invasive and effective method for detection and assessment of liver fibrosis, particularly for discrimination between F0-2 stages and F3-4 stages. MRE may represent a valuable tool to finely discern hepatic fibrosis non-invasively.