LI-RADS for diagnosing hepatocellular carcinoma by contrast-enhanced US with SonoVue and Sonazoid—a single center prospective study

Abstract

Purpose

To compare the diagnostic performance of two algorithms for HCC diagnosis: SonoVue-CEUS based on CEUS LI-RADS version 2017 and a modified algorithm incorporating Kupffer-phase findings for Sonazoid-CEUS.

Methods

This single center prospective study enrolled high-risk patients for HCC. Each participant underwent SonoVue-CEUS and Sonazoid-CEUS. Each liver observation was assigned two LI-RADS categories according to each algorithm: SonoVue-CEUS LI-RADS and modified Sonazoid-CEUS LI-RADS. For the latter method, observations at least 10 mm with non-rim arterial phase hyperenhancement were upgraded LR-4 to LR-5 if there was no washout with a Kupffer defect and were reassigned LR-M to LR-5 if there was early washout with mild Kupffer defect. The reference standard was pathologic confirmation.

Results

Overall, 66 patients (mean age, 61.2 years ± 10.9; 54 male patients, 12 female patients) with 66 observations (mean size, 31 mm ± 16) were eventually enrolled. The results of Sonazoid-CEUS LI-RADS showed significant changes in sensitivity (82% vs. 65%, p < 0.001), accuracy (85% vs. 71%, p < 0.001) compared with the SonoVue-CEUS LI-RADS. There was no significant difference in specificity (93% vs. 87%, p = 0.26).

Conclusion

When incorporating Kupffer-phase findings, Sonazoid-CEUS LI-RADS had higher sensitivity without loss of specificity compared with SonoVue-CEUS LI-RADS.

Graphical abstract