Quantitative hepatobiliary phase MRI–derived indices for assessment of hepatic functional reserve in chronic liver disease

Purpose To systematically evaluate the relationship between quantitative hepatobiliary phase (HBP) MRI–derived indices and established biochemical markers of liver function, and to assess their associations with ascites status and reproducibility in chronic liver disease (CLD). Methods In this retrospective single-center study, patients with clinically or histologically confirmed CLD who underwent gadoxetic acid–enhanced MRI were included. Quantitative HBP indices, including liver-to-spleen ratio (LSR), relative liver enhancement (RLE), liver-to-muscle ratio (LMR), and contrast uptake index (CUI), were calculated using standardized region-of-interest measurements. Associations with biochemical parameters (albumin, bilirubin, platelet count, sodium) and clinical scores (MELD-Na, ALBI, FIB-4) were analyzed. Multivariable regression models were constructed to evaluate associations between laboratory parameters and MRI-derived indices. Associations with ascites status were evaluated using receiver operating characteristic analysis; these analyses were interpreted as reflecting associations with ascites status within the same imaging/clinical framework. Interobserver and intraobserver reproducibility were assessed using intraclass correlation coefficients. Results HBP-derived indices demonstrated modest but significant associations with several biochemical markers of liver function. LSR showed the strongest association with MELD-Na, whereas CUI was most closely associated with ALBI grade. FIB-4 did not show significant correlations with MRI-derived indices. Patients with ascites had significantly lower values across all indices. ROC analysis demonstrated moderate discriminatory ability, with LSR yielding the highest AUC among the evaluated indices. Multivariable analyses indicated that different indices showed variable associations with biochemical parameters within the full model framework. Reproducibility analyses showed excellent interobserver and intraobserver agreement for most indices, with good interobserver agreement for LMR. Conclusion Quantitative HBP MRI–derived indices are reproducible imaging-derived indices that show consistent associations with markers of hepatic functional reserve. These metrics may provide complementary imaging-based information on hepatic functional status beyond conventional biochemical scores in CLD and should be interpreted as imaging-based associations rather than independent diagnostic tools.