Introduction Hepatic fibrosis is an important determinant of outcome in non-alcoholic fatty liver disease (NAFLD). The severity of non-alcoholic steatohepatitis (NASH) has relevance to clinical follow up intensity and assessment of the effectiveness of treatment interventions. Magnetic resonance imaging (MRI)-acquired T1 measurement has been shown to correlate with hepatic fibrosis and is also increased by inflammation, indicating potential as a tool to stratify NAFLD.
Methods Patients undergoing liver biopsy (LB) were invited for MRI and blood sampling prior to LB. Iron accumulation has been shown to reduce T1 independent of fibrosis, so an iron corrected hepatic T1 (cT1) was calculated using LiverMultiscan (Perspectum Diagnostics, Oxford). Histology was graded by two expert liver histopathologists according to the NASH-CRN system.
Results 50 patients had complete data sets following exclusions (1 inadequate biopsy, 3 MRI data unavailable). 28 (56%) patients were male. Median age 54 years. 26 (52%) patients had type-2 diabetes. Mean (±SD) body mass index was 33.6 (±5.1) Kg/m2. Median (IQR) of ALT, fasting glucose and cholesterol were 54 (52) u/L, 6.2 (4.0) mmol/L and 4.8 (1.5) mmol/L, respectively. 38 (76%) LBs demonstrated NASH and the remainder simple steatosis (SS).
Mean(±SD) cT1 for NASH and SS were 1007(±95) milliseconds (ms) and 907(±120) ms, respectively (p = 0.004). The correlation between cT1 and NAS was highly significant (Rho = 0.51, p < 0.0001). To differentiate NASH from SS, cT1 had an AUROC (95%CI) of 0.71 (0.53–0.89). To diagnose significant (>F1) and advanced (>F2) fibrosis cT1 had an AUROC (95% CI) of 0.65 (0.48–0.83) and 0.62 (0.47–0.78), respectively. To identify patients with SS and no significant fibrosis cT1 had an AUROC (95%CI) of 0.75 (0.56–0.93).
Conclusion Multiparametric MRI using LiverMultiscan has the ability to non-invasively stage fibrosis in patients with NAFLD. The additional benefit of this novel technology is the ability to concurrently establish the severity of NASH. The correlation of cT1 and NAS suggests that multiparametric MRI has potential for monitoring the effectiveness of treatment interventions in NAFLD.
Disclosure of Interest P. Eddowes: None Declared, N. McDonald: None Declared, N. Davies: None Declared, S. Semple: None Declared, S. Hübscher: None Declared, T. Kendall: None Declared, C. Kelly Employee of: Perspectum Diagnostics, M. Mavar Employee of: Perspectum Diagnostics, A. Herlihy Employee of: Perspectum Diagnostics, P. Newsome: None Declared, S. Olliff: None Declared, J. Fallowfield: None Declared, G. Hirschfield: None Declared