A comparison of MR elastography and 31P MR spectroscopy with histological staging of liver fibrosis.

ABSTRACT

OBJECTIVES: Conventional imaging techniques are insensitive to liver fibrosis. This study assesses the diagnostic accuracy of MR elastography (MRE) stiffness values and the ratio of phosphomonoesters (PME)/phosphodiesters (PDE) measured using (31)P spectroscopy against histological fibrosis staging. METHODS: The local research ethics committee approved this prospective, blinded study. A total of 77 consecutive patients (55 male, aged 49 ± 11.5 years) with a clinical suspicion of liver fibrosis underwent an MR examination with a liver biopsy later the same day. Patients underwent MRE and (31)P spectroscopy on a 1.5 T whole body system. The liver biopsies were staged using an Ishak score for chronic hepatitis or a modified NAS fibrosis score for fatty liver disease. RESULTS: MRE increased with and was positively associated with fibrosis stage (Spearman's rank = 0.622, P < 0.001). PME/PDE was not associated with fibrosis stage (Spearman's rank = -0.041, p = 0.741). Area under receiver operating curves for MRE stiffness values were high (range 0.75-0.97). The diagnostic utility of PME/PDE was no better than chance (range 0.44-0.58). CONCLUSIONS: MRE-estimated liver stiffness increases with fibrosis stage and is able to dichotomise fibrosis stage groupings. We did not find a relationship between (31)P MR spectroscopy and fibrosis stage. KEY POINTS Magnetic resonance elastography (MRE) and MR spectroscopy can both assess the liver. MRE is superior to ( 31 ) P MR spectroscopy in staging hepatic fibrosis. MRE is able to dichotomise liver fibrosis stage groupings. Gradient-echo MRE may be problematic in genetic haemochromatosis.