| Abstract |
Purpose: This study compared the predictive performance of the relative enhancement index (REI) derived from gadoxetic acid (GA)-enhanced MRI with that of the functional liver imaging score (FLIS) in estimating liver function among patients with chronic liver disease (CLD) or liver cirrhosis (LC) by validating them with the albumin-bilirubin (ALBI) grade.
Materials and methods: We retrospectively examined 166 patients (79 women, 87 men; 57.4 years) who were diagnosed with LC or CLD and underwent GA-enhanced MRI between August 2020 and September 2023. The enhancement ratio (ER) is calculated using the formula: ER = [hepatobiliary phase liver signal (SI HBP20)-precontrast liver signal (SI pre)]/SI pre. The REI is calculated using the formula: REI = Liver Volume (LV) × ER. FLIS was assigned from the sum of three HBP image features, each scored between 0 and 2: liver parenchymal enhancement, biliary contrast excretion, and portal vein sign. Receiver operating characteristic (ROC) curve analysis was performed to determine the optimal cutoff values of ER, REI, and FLIS in differentiating between ALBI grades. The area under the curve (AUC), accuracy, sensitivity, and specificity were calculated for REI and FLIS to distinguish the ALBI grades. Spearman's rank correlation was used to evaluate the ER, REI, and FLIS correlations between the ALBI grades. To evaluate inter-reader reliability for LV, ER, REI, and FLIS, intraclass correlation coefficient (ICC) was used.
Results: ROC curve analysis showed that the optimal cutoff value of REI for predicting ALBI Grade 1 was 899-905 for readers 1 and 2 and 461-477 for ALBI Grade 3, respectively. REI performed best in predicting ALBI Grade 1, achieving an accuracy range of 94%-92.2%, sensitivity of 94.9%-94.1%, and specificity of 91.7%-87.5% for readers 1 and 2, respectively. All parameters showed high accuracy in distinguishing ALBI Grade 3 from other grades. However, REI outperformed the others, showing an accuracy range of 98.8%-97.6%, sensitivity of 94.4%-94.4%, and specificity of 99.3%-98% for readers 1 and 2, respectively. REI showed the best and very strong correlation with ALBI for both readers.
Conclusion: REI showed a very strong correlation with the ALBI grades for assessing liver function. It outperformed FLIS in predicting the ALBI grades, indicating its potential as a radiologic tool comparable to or better than FLIS in predicting liver function, especially given its dependence on liver volume.
|
