Magnetic resonance elastography and biomarkers to assess fibrosis from recurrent hepatitis C in liver transplant recipients

Background. Imaging techniques evaluating liver stiffness (magnetic resonance elastography [MRE]) and biomarkers may be useful indicators of fibrosis stage in hepatitis C virus (HCV)+patients. Our aim was to compare the accuracy of MRE and biomarkers in staging fibrosis because of recurrent HCV in liver transplant (LT) recipients with hepatocellular carcinoma. Methods. Liver magnetic resonance imaging and MRE, FIBROSpectII, aspartate aminotransferase-to- platelet ratio index (aspartate aminotransferase [AST]: platelet index), AST:alanine aminotransferase ratio, and magnetic resonance imaging/MRE-guided biopsies targeting the stiffest regions (right and left lobes) were performed in HCV+LT recipients. Sensitivity, specificity, positive predictive value (PPV)/negative predictive value (NPV), and likelihood ratios were calculated for the best cutoff by receiver operating characteristic analysis. Results. Thirty-two recipients were included: 28 men, age 60 (±6.4) years, and time since LT 3.25 (±1.68) years. Both MRE (P=0.0001) and FIBROSpectII (P=0.009) were significantly different between no fibrosis and more than or equal to stage 1 groups, whereas aspartate aminotransferase-to-platelet ratio index and AST:alanine aminotransferase ratio were not different. Areas under the receiver operating characteristic curve were 0.87 for MRE and 0.84 for FIBROSpectII. MRE cutoff of 3.81 kPa had 87.5% sensitivity, 79.2% specificity, 58.3% PPV, and 95.0% NPV; FIBROSpectII cutoff of 42 had 87.5% sensitivity, 70.0% specificity, 53.8% PPV, and 93.3% NPV for detection of more than or equal to stage 1 fibrosis. Two patients had high MRE values because of unexpected acute rejection and portal vein thrombosis. Conclusions. MRE and FIBROSpectII are highly sensitive in detecting fibrosis due to recurrent HCV. Both are limited by the low specificity/PPV and confounding because of other graft complications. Values below the MRE and FIBROSpectII cutoffs, however, strongly suggest the absence of fibrosis and may avert the need for protocol biopsy staging.