Comparison of Annual Dynamic Abbreviated MRI and Biannual Ultrasound for HCC Surveillance in a High-Risk Group: Two-Center Cohort Study

Introduction: Biannual ultrasound (US) is recommended for HCC surveillance, although its sensitivity for early-stage HCC is low. This study aims to compare the performance of annual dynamic abbreviated MRI (D-AMRI) and biannual US for HCC detection. Methods: This prospective, two-center cohort study enrolled participants eligible for HCC surveillance according to Korean guidelines between November 2018 and December 2021. Participants underwent four rounds of surveillance at 6-month intervals: both D-AMRI and US at the first and third rounds and only US at the second and fourth rounds. The diagnostic yield (DY) and false referral rate (FRR) were compared between the two modalities. The scan time of D-AMRI was measured using a stopwatch. Results: Of the 257 enrolled participants, 213 participants completed the protocol and were included in the analysis. Primary liver cancers were found in 32 participants, including 31 HCCs (22 very early-stage, eight early-stage, and one advanced-stage) and one intrahepatic cholangiocarcinoma. The DY of annual D-AMRI was higher than biannual US for all (6.4% [26/405] vs. 2.2% [9/405]), early-stage (6.2% [25/405] vs. 2.0% [8/405]), and very early-stage HCCs (4.7% [19/405] vs. 1.2% [5/405]) (p < 0.001 for all). No significant difference in the FRR was observed for all (2.5% [10/405] for both), early-stage (2.7% [11/405] for both) (p > 0.999 for both), and very early-stage HCCs (4.2% [17/405] vs. 3.5% [14/405], p = 0.564). The median scan time of D-AMRI was 13 min. Conclusion: Annual D-AMRI showed a higher DY than biannual US for detecting HCCs, without increasing the FRR.