Contrast enhanced ultrasound in diagnosis of hepatocellular carcinoma (HCC): role of wash-out in portal and late phases

Aim: According to International Guidelines, typical features of hepatocellular carcinoma (HCC) at imaging techniques are considered hyper-enhancement in arterial phase, iso/hypo-enhancement in portal-venous phase and hypo-enhancement in late phase. Aim of this study was to investigate the role of portal-late phase at contrast-enhanced ultrasound CEUS of small nodules (10–30mm) in cirrhosis.

Material and methods: 81 cirrhotic patients (66M; mean age 65.3y; 72 viral, 7 alcoholic, 2 cryptogenetic) with 99 newly developed small nodules (10–30mm) were included. CEUS was performed using dedicated technology (CnTI-Esaote) after injection of 2.4ml of SonoVue (Bracco), allowing detection of arterial and portal-late vascular patterns until disappearance of contrast. Diagnosis of HCC was established according to American Association Study of Liver Diseases (AASLD) guidelines, considering as imaging techniques computed tomography and magnetic resonance, not including CEUS. When imaging was not diagnostic and biopsy not feasible, 3month interval follow-up was considered.

Results: HCC was diagnosed in 69 of 99 nodules (69.7%): 31 confirmed by histology, 38 according to AASLD criteria. Of 69 HCC, CEUS was positive in arterial phase in 46 (66.6%): 17 (37%) were hyperechoic in arterial and hypoechoic in portal-late phases (hyper/hypo), 29 (63%) hyperechoic in arterial and isoechoic in portal-late phased (hyper/iso). In 23 HCC CEUS-pattern was isoechoic in all phases (iso/iso), therefore not diagnostic of HCC. 30 nodules resulted nonmalignant. In 21/30, CEUS-pattern was iso/iso: 16 were biopsied showing 11 cases of cirrhosis or macroregenerative nodules without dysplasia and 4 nodules with dysplasia. The remaining 5 did not progress to HCC over follow-up (mean 30 months). Pattern of remaining 9 nonmalignant lesions were hyper/iso: biopsies showed 1 cirrhosis, 3 cirrhosis with dysplasia and 5 high-grade dysplastic nodules (HGDN). Positive predictive value (PPV) for HCC of the hyper/hypo pattern was 100%, whereas that of hyper/iso was 76%. However, considering HGDN as HCC, PPV of hyper/iso increased to 89.5%. Negative predictive value of iso/iso pattern was 47.3%.

Conclusions: Wash out to hypoechogenicity at CEUS after arterial hyperechogenicity should be regarded as diagnostic for HCC. A pattern of late isoechogenicity after arterial hyperechogenicity despite not as specific as hyper/hypo still shows high diagnostic capability for HCC and identifies nodules which are already HCC or with strong tendency to progression to malignancy.