Abstract

Background: Contrast-enhanced ultrasound (CEUS) is a promising diagnostic technique for hepatocellular carcinoma (HCC) diagnosis. While EASL HCC guidelines endorse its use in cases of inconclusive CT or MRI based on non-focused studies, the diagnostic potential of CEUS for HCC is not currently recognized by AASLD guidelines. This study aimed to assess the clinical impact of CEUS specifically in cases where liver lesions were indeterminate on MRI.

Methods: A prospective international multicenter validation study for CEUS LI-RADS (Liver Imaging Reporting And Data System) was conducted between January 2018 and August 2021. A total of 594 patients at risk for HCC were enrolled. CEUS was performed using intravenous administration of Lumason (Bracco Diagnostic) within 4 weeks of MRI, and the liver nodules were classified based on CEUS and MRI LI-RADS criteria. Tissue histology and CT/MRI imaging follow-up were used as the reference standard.

Results: A total of 545 nodules reached a final diagnosis based on the reference standard. Among them, 75 nodules with indeterminate MRI characterization (LR-NC: non categorizable, and LR-3: intermediate probability of HCC) were selected for analysis. Reference standard used was follow up CT/MRI (54, 72%), biopsy (14, 18.7%), and explant histology (7, 9.3%).

Among the nodules categorized as LR-NC on MRI (n=21), one was CEUS LR-1, was determined to be benign, all 6 CEUS LR-5 were confirmed to be HCC and 2 LR-M were malignant resulting in a clinical impact in 42.9% (9/21).

In the case of LR-3 on MRI (n=54), there were a total of 14 HCCs, with 7 CEUS LR-4 (probable HCC) and 7 CEUS LR-5. CEUS LR-1 (n=1) and LR-2 (n=6), were all benign (100% NPV). PPV of CEUS LR-5 for HCC was 100%. There was one CEUS LR-TIV, and biopsy confirmed poorly differentiated HCC. However, in the case of one CEUS LR-M, the follow-up MRI revealed LR-2, suggesting a false positive CEUS result.

When LR-NC and LR-3 lesions on MRI (n=75) were combined, CEUS LR-1 (n=2) and LR-2 (n=6), nodules were all benign, CEUS LR-5 (n=13) were all HCC. Two CEUS LR-M from LR-NC and one CEUS LR-TIV from LR-2 were malignant, resulting in a clinical impact in 32% (24/75). CEUS had one false positive LR-M case, which was benign on follow up MRI.

Conclusion: CEUS LI-RADS demonstrated high clinical impact in liver nodules with indeterminate MRI characterization accurately, identifying both non-malignant lesions, HCC and other malignancy in 32% of patients. These findings highlight the significant clinical value of CEUS in the characterization and diagnosis of liver lesions, particularly in cases where MRI findings are inconclusive or indeterminate.