Inter- and intra-reader agreement for gadoxetic acid-enhanced MRI parameter readings in patients with chronic liver diseases

Objectives To examine inter- and intra-observer agreement for four simple hepatobiliary phase (HBP)–based scores on gadoxetic acid (GA)–enhanced MRI and their correlation with liver function in patients with mixed chronic liver disease (CLD). Methods This single-center, retrospective study included 287 patients (62% male, 38% female, mean age 53.5 ± 13.7 years) with mixed CLD (20.9% hepatitis C, 19.2% alcoholic liver disease, 8% hepatitis B) who underwent GA-enhanced MRI of the liver for clinical care between 2010 and 2015. Relative liver enhancement (RLE), contrast uptake index (CUI), hepatic uptake index (HUI), and liver-to-spleen contrast index (LSI) were calculated by two radiologists independently using unenhanced and GA-enhanced HPB (obtained 20 min after GA administration) images; 50 patients selected at random were reviewed twice by one reader to assess intra-observer reliability. Agreement was assessed by intraclass correlation coefficient (ICC). The albumin-bilirubin (ALBI) score, the model of end-stage liver disease (MELD), and the Child-Turcotte-Pugh (CTP) score were calculated as standards of reference for hepatic function. Results Intra-observer ICCs ranged from 0.814 (0.668–0.896) for CUI to 0.969 (0.945–0.983) for RLE. Inter-observer ICCs ranged from 0.777 (0.605–0.874) for HUI to 0.979 (0.963–0.988) for RLE. All HBP-based scores correlated significantly (all p