Diagnostic performance of LR-5 based on hypointensity on Gd-EOB-DTPA-enhanced MRI in the hepatobiliary phase for sHCC using LI-RADS v2018 criteria

Clin Radiol. 2024 Dec 19:81:106784. doi: 10.1016/j.crad.2024.106784. Online ahead of print.

Abstract

Aim: To investigate the value of the LR-5, which is based on hepatobiliary phase (HBP) hypointensity, for small hepatocellular carcinoma (sHCC) using LI-RADS v2018 criteria.

Materials and methods: From January 2015 to December 2021 in institution 1, and from January 2019 to February 2022 in institution 2, 239 patients at high risk for hepatocellular carcinoma (HCC) underwent contrast-enhanced MRI. Two radiologists independently evaluated the imaging features and classified them according to LI-RADS v2018 criteria, calculating the diagnostic performance of LR-5 based on consensus data. LI-RADS-m1: HBP hypointensity was used as an additional major feature along with the LI-RADS v2018. LI-RADS-m2: HBP hypointensity replaced nonperipheral "washout" in the portal venous phase. The definition of LR-DN was nodules pathologically diagnosed as high-grade dysplastic nodules (HGDN) were recategorized as LR-DN. The diagnostic performance of LR-5 was recalculated. The diagnostic performance of the LR-5 was compared using McNemar's test.

Results: Using LI-RADS v2018, LI-RADS-m1, and LI-RADS-m2 criteria for LR-5, the sensitivities were 82.67%, 86.22%, and 88.44%, the specificities were 82.00%, 66.00%, and 54.00%, and the accuracies were 82.55%, 82.55%, and 82.18%, respectively. After the addition of the LR-DN, the sensitivities of LR-5 in the above diagnostic model remained unchanged, with accuracies of 84.36%, 87.27%, and 88.36% and specificities of 92.00%, 92.00%, and 88.00%, respectively.

Conclusions: HBP hypointensity may improve the sensitivity of LR-5. We attempted to propose the LR-DN, HBP hypointensity may be used as a complement to washout as an additional major feature without significantly decreasing specificity.