Hepatocellular adenoma: comparison between real-time contrast-enhanced ultrasound and dynamic computed tomography
It has been reported that HCAs are most common in women who have taken oral contraceptives for long periods of time (Baum et al. 1973). The male-to-female ratio is 1:9, which is lower than that in our study (1:1). Our data coincide with reports from Asia (Kong et al. 2015; Hung et al. 2001) that suggested that a lower use of oral contraceptives in women and routine screening for hepatocellular carcinoma in men may result in a different male-to-female incidence ratio.
Several imaging modalities have been employed in the diagnosis of hepatic adenoma. However, no specific characteristics of HCAs can be identified on unenhanced imaging. Most HCAs are hypo-echoic or hyper-echoic on US and hypo-attenuated to the background liver parenchyma on CT. The mean size of HCAs in our study is larger than in previous reports (Zhu et al. 2011; Ricci et al. 2008). Although larger lesions are at a higher risk of hemorrhage, the echogenicity in our study were homogeneous. Unlike the spoke-wheel flow pattern of FNH, short rod-like or periphery vessels on color Doppler cannot proved effective diagnostic information for HCA (Kong et al. 2015).
On enhanced imaging, most reports showed that 45–60 % of HCAs were hyper-vascular in the arterial phase and hypo-enhanced in the portal or late phase. In our study, almost 70 % of lesions exhibited this enhancement pattern on CEUS and CECT, which can be encountered in HCC and hyper-enhancing metastases. Moreover, HCA with inhomogeneous enhancement is very difficult to discriminate from HCC in a non-cirrhosis liver. In case of this, future developments, such as elastography or new perfusion imaging, might be useful for classification of benign and malignant lesions (Wiggermann et al. 2013; Jung et al. 2007). On the other hand, 30.0 % of lesions on CEUS and 37.5 % on CECT appear to have sustained enhancement in later phases. Although this enhancement pattern suggests a benign tumor, a difficult differential diagnosis with FNH can also be encountered (Kong et al. 2015; Roche et al. 2015; Di Carlo et al. 2010; Kim et al. 2008; Dietrich et al. 2005; Bartolozzi et al. 1997).
This diagnostic dilemma could be resolved by specifically characterizing the dynamic imaging features. According to previous studies, HCAs have some special enhancement features, such as centripetal filling and persistent peripheral rim enhancement (Kim et al. 2008; Mathieu et al. 1986). The specific centripetal filling sign, which was discovered on arteriography, was first reported on CEUS by Kim (Kim et al. 2008). This is the main differential feature with FNH, which is characterized by centrifugal filling and central scarring. The detection rate ranges from 20 to 94 % (Laumonier et al. 2012; Kim et al. 2008; Kong et al. 2015), which is consistent with the rate of 55.5 % in our study. Laumonier reported that the inflammatory subtype of HCA had a peripheral rim of sustained enhancement with a hypo-echoic central area (Laumonier et al. 2012). In our study, this persistent peripheral rim enhancement feature was detected in 61.1 % of all HCAs and 50.0 % of inflammatory subtype on CEUS. The detection rate of this specific sign is lower in our cases because of the nine cases with unidentified subtype. However, none of the specific enhancement features discussed above was reported on CT. In our opinion, CEUS with inherent real-time scanning merit is better for display of those features. Although steatosis, calcification, necrosis and intro-tumor vessels on CT were reported in previous studies, these characteristics are not very sensitive (Burns and Wilson 2007; Hussain et al. 2006).
Our study has some limitations. First, it was a retrospective study based on small number of 18 cases, and only 14 of those were evaluated with CT. Second, as we focused on comparisons with CT, the problem of differential diagnosis with FNH or HCC was neglected. Therefore, limited information had been provided for the diagnostic accuracy of CEUS for this issue. Moreover, some imaging features vary depending on the subtype of HCA and/or tumor size. In our study, we failed to describe the correlation between imaging patterns and subtypes because of limited cases.
In conclusion, our study demonstrated that the enhancement level in three vascular phases of HCA on CEUS was consistent with that on CECT. Compared with CECT, the real-time capability of CEUS is superior for characterizing dynamic centripetal filling, peripheral rim of persistent enhancement, and delayed central washout. In addition, because it is radiation-free, readily availability, and easy to use, CEUS has been suggested as the first line imaging tool to diagnose HCA. Further studies using larger sample sizes and comparisons with MRI are required.