Contrast-Enhanced US Using Perfluorobutane for Diagnosing Small HCC in High-Risk Patients: Comparison with MRI LI-RADS Version 2018.

Abstract

Background Evidence supporting the incorporation of perfluorobutane (PFB) contrast-enhanced US (CEUS) into Liver Imaging Reporting and Data System (LI-RADS) in diagnosing small hepatocellular carcinoma (HCC) (≤20 mm) in patients at high risk for HCC remains limited. Purpose To compare diagnostic performance of PFB CEUS and MRI for small HCC in high-risk patients. Materials and Methods This multicenter retrospective study included high-risk patients who underwent PFB CEUS and concurrent MRI for evaluating liver nodules 20 mm or smaller from March 2020 to November 2023. PFB CEUS strategy A followed LR-5 criteria based on CEUS LI-RADS version 2017: observations that were at least 10 mm with nonrim arterial-phase hyperenhancement (APHE), with late and mild washout assessed up to 5 minutes postinjection; strategy B expanded on strategy A by also incorporating observations 10 mm or larger with nonrim APHE, no washout up to 5 minutes, and hypoenhancement on Kupffer-phase images. The diagnostic performance of these two CEUS strategies and MRI LI-RADS version 2018 was compared using a generalized estimating equations approach. Results A total of 365 patients (median age, 54 years; IQR, 47-61 years; 310 men) with 399 observations (median diameter, 16 mm; IQR, 12-19 mm; 252 HCCs, 41 non-HCC malignancies, 106 benign nodules) were included. Strategy B yielded higher sensitivity than strategy A (65.9% [166 of 252] vs 57.1% [144 of 252]; <i>P</i> < .001) without evidence of a decrease in specificity (91.8% [135 of 147] vs 93.9% [138 of 147]; <i>P</i> = .07). Compared with MRI, strategy B showed no evidence of a difference in sensitivity (65.9% [166 of 252] vs 72.6% [183 of 252]; <i>P</i> = .07) and specificity (91.8% [135 of 147] vs 90.5% [133 of 147]; <i>P</i> = .59). Strategies A and B showed higher specificity (97.4% [76 of 78] vs 89.9% [62 of 69]; 96.2% [75 of 78] vs 87.0% [60 of 69], respectively; both <i>P</i> = .04) but no evidence of a difference in sensitivity (54.8% [46 of 84] vs 58.3% [98 of 168] [<i>P</i> = .63]; 61.9% [52 of 84] vs 67.9% [114 of 168] [<i>P</i> = .39], respectively) in patients without cirrhosis versus in patients with cirrhosis. Conclusion In patients at high risk for HCC, PFB CEUS incorporating Kupffer-phase findings was effective for diagnosing small HCC, with diagnostic performance similar to MRI. © RSNA, 2026 <i>Supplemental material is available for this article.</i>