Characterisation of liver disease using quantitative MRI: report of the imaging arm of the ELEGANCE study

Abstract

Evolving disease aetiology and rising prevalence of chronic liver disease are driving an increase in hepatocellular carcinoma (HCC). Non-contrast quantitative MRI (qMRI) has shown promising capability for characterising liver health and risk of liver related outcomes, but has not been evaluated in HCC surveillance or specifically examined in Asian populations. We explored the characteristics of qMRI in participants with chronic liver disease undergoing HCC surveillance as part of the prospective cohort study, ELEGANCE. Seven hundred sixty-nine participants underwent qMRI using LiverMultiScan to derive information relating to liver disease activity (iron-corrected T1, cT1), liver iron (from T2*) and liver fat content (through proton density fat fraction), as well as structural features of advancing disease such as, liver and spleen volumes, and the caudate to right lobe volume ratio. Three hundred twenty-eight participants had hepatitis B (HBV); 161 metabolic dysfunction-associated steatotic liver disease (MASLD); and 273 had cirrhosis (any aetiology). In HBV, qMRI identified that more than 40% had concurrent SLD. In MASLD, qMRI suggested 31% had MASH. In all aetiologies, cirrhosis was associated with elevated cT1, reduced liver fat, spleen enlargement, and caudate hypertrophy. A model incorporating these features demonstrated strong discriminatory performance for cirrhosis, particularly within the MASLD cohort (AUC = 0.88). In an exploratory analysis the same score demonstrated statistical power to discriminate between those with and without a subsequent HCC diagnosis. This study demonstrates the capability of qMRI for comprehensive characterisation of different liver disease aetiologies and the ability to identify those at highest risk of clinical outcomes.