Diagnostic value of MRI-derived liver surface nodularity score for the non-invasive quantification of hepatic fibrosis in non-alcoholic fatty liver disease

Objectives To assess the accuracy of MRI-derived liver surface nodularity (LSN) score for staging of hepatic fibrosis in patients with non-alcoholic fatty liver disease (NAFLD). Methods Forty-seven patients with clinicopathological diagnosis of NAFLD who underwent 1.5-T liver MRI within 12 months of liver biopsy were included. Axial non-contrast T1-weighted 3D GRE was used for image analysis. LSN of the left lobe was measured using a custom semiautomated software. Histopathologic analysis (F0-F4) served as the reference standard for staging of fibrosis. Mann-Whitney test and Spearman's correlation coefficient were used to compare LSN scores between different stages of fibrosis and to assess the correlation. Diagnostic performance of LSN score for detection of significant (F2-F4) and advanced (F3-F4) fibrosis was assessed by receiver operating characteristics (ROC) curve.pvalue of less than 0.05 was considered statistically significant different. Results Twenty-one subjects had advanced fibrosis. The LSN scores among different stages of fibrosis were significantly different (p < 0.001). The correlation between LSN score and stage of fibrosis was also strong (rho = 0.71;p < 0.001). The areas under ROC curves for detection of significant and advanced fibrosis were 0.80 (95% CI 0.66-0.95) and 0.86 (95% CI 0.75-0.97), using a threshold of 2.23 and 2.44, respectively. This method showed 81% sensitivity and 88% specificity for detection of advanced fibrosis. Conclusion MR-based LSN score is a promising non-invasive objective tool for detection of advanced fibrosis in patients with NAFLD.