Value of threshold growth for the diagnosis of hepatocellular carcinoma using LI-RADS.

Abstract

Background: The utility of threshold growth (TG) in hepatocellular carcinoma (HCC) imaging remains contentious across major guidelines. This study aimed to investigate the diagnostic implications of TG in HCC diagnosis using the criteria set by the Liver Imaging Reporting and Data System (LI-RADS).

Methods: In this single-center retrospective study, three radiologists independently evaluated pre-transplantation hepatobiliary agent-enhanced MR images and prior CT/MR images using LI-RADS v2018 in consecutive patients who underwent liver transplantation between January 2010 and November 2022. TG was defined as a ≥ 50% size increase in ≤ 6 months. Explanted livers served as reference standards. Frequencies of TG between HCCs and non-HCCs were compared using Fisher's exact test, and interobserver agreement was assessed using Fleiss κ statistics. The diagnostic performance of LI-RADS category 5 in the diagnosis of HCC was assessed with and without considering TG as a major feature. McNemar tests were used to compare results.

Results: The cohort included 158 patients (mean age, 59.1 ± 7.5 years; 130 males) with 280 observations (207 HCCs, 5 non-HCC malignancies, and 68 benign lesions). TG was identified in 44 (15.7%) observations. Interobserver agreement on TG was moderate (κ = 0.280). Incorporating TG as a major feature significantly enhanced the sensitivity of LI-RADS category 5 in diagnosing HCC (33.8% vs. 40.6%, p < 0.001) without compromising specificity (100.0% vs. 94.5%, p = 0.125).

Conclusions: Incorporating TG as a major criterion in LI-RADS category 5 enhanced the diagnostic sensitivity for HCC in liver transplant candidates with minimal impact on specificity. However, TG demonstrated a variable interobserver agreement.

Trial registration: Not applicable.