S1178 Utility and Interpretation of the Quantitative MRI Metrics PDFF and cT1 as Biomarkers for Non-Alcoholic Steatohepatitis

Abstract

INTRODUCTION: Clinical guidelines for NASH require results from a liver biopsy for diagnosis and for monitoring disease progression. The well reported limitations with biopsy, such as associated risks and sampling error, coupled with patient preference, are driving investigation into non-invasive alternatives. MRI derived biomarkers proton density fat fraction (PDFF) and iron-corrected T1 mapping are gaining traction as emerging alternatives to biopsy for NASH. Our aim was to explore the correlations between iron corrected T1 mapping (cT1) and proton density fat fraction (PDFF) (from LiverMultiScanâ), with the histological components of NASH in a large cohort of cross-sectional data, in order to calibrate the measurement to histology, and to infer what might constitute a clinically meaningful change. METHODS: In a retrospective analysis of data combined from three previously published observational NASH studies, in which adult participants who underwent liver biopsy on suspicion of NAFLD or NASH and had an MRI scan measuring cT1 and PDFF (LiverMultiScan, Perspectum Ltd, UK), associations between imaging biomarkers and histology were tested using Spearman’s rank correlation coefficient (rs), and estimation of the causal relationships between the imaging variables and histology were performed using linear regression. RESULTS: N = 264 patients with mean age of 54 (SD:9.9), 39% females, and 69% with BMI ≥ 30kg·m−2 were included in the analysis. cT1 and PDFF both correlated with all features of the NAFLD activity score (NAS). cT1 was also positively correlated with Kleiner-Brunt fibrosis. Partial correlations, adjusting for steatosis, revealed cT1 correlated with inflammation and fibrosis, whereas PDFF did not, and both were still associated with the NAS, but correlation was weaker with PDFF than cT1. An estimated difference of 88ms in cT1, or 22% relative difference in PDFF was related to a 2-point difference in overall NAS. CONCLUSION: The correlations between cT1 and PDFF with the histopathological hallmarks of NASH demonstrate the potential utility of both cT1 and PDFF as non-invasive biomarkers to detect a pharmacodynamic change in NASH, with cT1 showing superiority for detecting changes in inflammation and fibrosis, rather than liver fat alone.Figure 1.: Box plots showing relationships between cT1 and PDFF with NAS (top row) and fibrosis (bottom row).Figure 2.: Example PDFF and cT1 parametric maps for patients with NAS = 1 (A) cT1 = 684ms, PDFF = 6.5%; NAS = 3 (B) cT1 = 833ms, PDFF = 16.9%; and NAS = 5 (C) cT1 = 916ms, PDFF = 18.5%.