Prospective Intraindividual Comparison of Magnetic Resonance Imaging With Gadoxetic Acid and Extracellular Contrast for Diagnosis of Hepatocellular Carcinomas Using the Liver Imaging Reporting and Data System

We intraindividually compared the efficacy of magnetic resonance imaging (MRI) with extracellular contrast agents (ECA-MRI) and MRI with hepatobiliary agents (HBA-MRI) for the diagnosis of hepatocellular carcinoma (HCC) using the Liver Imaging Reporting and Data System (LI-RADS). Between November 2016 and November 2017, we enrolled 91 patients with chronic liver disease who underwent both ECA-MRI and HBA-MRI within a 1-month interval for a first detected hepatic nodule on ultrasound. In total, 117 observations (95 HCCs, 19 benign lesions, and 3 other malignancies; median size, 18 mm) were identified with surgical resection. Two observers assessed two MRIs based on LI-RADS v2017, with consensus by a third observer. We then compared the diagnostic performance of LR-5 according to LI-RADS and modified LI-RADS. ECA-MRI had higher sensitivity (77.9% versus 66.3%) and accuracy (82.1% versus 72.6%) than HBA-MRI in the LR-5 category (P < 0.001). When applying either modified washout on the portal venous phase (PVP)/transitional phase (TP) of HBA-MRI or isointensity with a capsule during the PVP/delayed phase of ECA-MRI (illusional washout), 13 HCCs on HBA-MRI and 11 HCCs on ECA-MRI were correctly classified as HCC, while achieving 100% specificity. One cholangiocarcinoma was accurately classified only with HBA-MRI due to its targetoid appearance in the TP and hepatobiliary phase. Conclusion: ECA-MRI showed better sensitivity and accuracy than HBA-MRI for the diagnosis of HCC with LI-RADS. We achieved better diagnostic performance when applying a modified washout on PVP/TP HBA-MRI and an illusional washout on ECA-MRI than we did with conventional criteria.