Different Abbreviated MRI for Hepatocellular Carcinoma Diagnosis

This retrospective study enrolled untreated patients at risk for developing HCC who underwent gadoxetic acid-enhanced MRI from three independent cohorts from January 2020 to December 2022. Three AMRI protocols were extracted from the complete liver MRI protocol, including non-contrast (NC-AMRI), hepatobiliary phase (HBP-AMRI); and dynamic contrast-enhanced (DCE-AMRI). Two readers at each cohort independently interpreted each AMRI based on ancillary and main features from the Liver Imaging Reporting and Data System (LI-RADS) version 2018. For DCE-AMRI, two established algorithms were employed: modified LI-RADS (DCE-AMRIm) and LI-RADS v2018 (DCE-AMRIv2018). Non-HCCs were ascertained based on histologic and clinical follow-up data, while early-stage HCCs were confirmed post-operatively. Patient-level performance metrics were compared using bootstrap resampling and generalized estimating equations.

Inclusion criteria

• no previous history of HCC treatment
• no history of any other underlying malignancy besides HCC.

Exclusion criteria

• For HCC group: (a) without preoperative MRI within 3 months; (b) without postoperative pathological confirmation; (c) liver cirrhosis without histopathologic proof or radiologic evidence; (d) the size of HCC smaller than 1cm; (e) outside of Milan Criteria; and (f) suboptimal imaging quality.
• For non-HCC group: 1) lost follow-up, 2) insufficient follow-up period, or 3) suboptimal imaging quality.