Clinical relevance and reproducibility of morphological MRI findings in liver cirrhosis: Results from ACCESS-ESLD.

OBJECTIVES

To validate the clinical relevance and reproducibility of morphological features on magnetic resonance imaging (MRI) in liver cirrhosis.

MATERIALS &

METHODS

This is a cross-sectional analysis of the baseline MRI examinations in a prospective multi-center study of patients with liver cirrhosis. Two radiologists independently assessed multiple morphological features and the overall degree of cirrhotic changes using a 5-point grading scale: 1 = none, 2 = subtle, 3 = moderate, 4 = evident, 5 = severe. Logistic regression was used to determine the association between imaging findings and clinical severity, using the model for end-stage liver disease (MELD) and Child-Pugh as reference standards. Inter- and intrareader agreement was assessed to evaluate reproducibility.

RESULTS

In our cohort of 150 patients (mean age, 64 ± 11 years, 43% women; 90% Child Pugh A) evident to severe cirrhotic changes on imaging were associated with MELD ≥ 10 and a Child-Pugh >5, odds ratios (OR) of 3.16 (95% confidence interval (CI): 1.54, 7.41, p = 0.008) and 4.15 (95% CI: 1.79, 9.63, p < 0.001) respectively. Reader agreement was substantial (κ = 0.59-0.64). Non-gastroesophageal portosystemic collaterals also demonstrated strong associations with clinical severity and consistent agreement between readers.

CONCLUSION

Morphological MRI features in liver cirrhosis are associated with established clinical markers of disease severity and demonstrate reasonable reproducibility, supporting their potential role in non-invasive disease assessment.