RESUMO
OBJECTIVE: The purpose of this study was to evaluate the accuracy of a new sonographic marker for the diagnosis of cirrhosis using hepatic vein wall changes. METHODS: A prospective pilot study evaluating 88 patients, 38 with cirrhosis and 50 with no evidence of liver disease, was undertaken. Hard copy sonograms of the hepatic veins were obtained and reviewed in a blinded fashion by 2 radiologists. The hepatic vein morphology was assessed by 3 parameters: hepatic vein wall straightness, uniformity of hepatic vein wall echogenicity, and visualization of a complete 1-cm hepatic vein segment. The 3 parameters were compared to evaluate sensitivity and specificity for the diagnosis of cirrhosis. Interobserver and intraobserver errors for each parameter were also calculated with kappa statistics to assess reproducibility. RESULTS: There was a strong correlation between altered straightness and nonuniformity of hepatic vein wall echogenicity and cirrhosis. The straightness parameter had superior sensitivity of 97% (95% confidence interval [CI], 85%-100%) and specificity of 91% (95% CI, 78%-97%) for diagnosis of cirrhosis. Uniformity of hepatic vein wall echogenicity was the next most useful parameter, with sensitivity of 88% (95% CI, 73%-97%) and specificity of 86% (95% CI, 72%-95%). The continuous 1-cm segment of the hepatic vein had sensitivity of 68% (95% CI, 49%-83%) and specificity of 91% (95% CI, 78%-97%). Hepatic vein evaluation was found to show both good intraobserver and interobserver error. CONCLUSIONS: Hepatic vein morphology on sonography, in particular, changes in the straightness and uniformity of hepatic vein wall echogenicity, is a new sign of cirrhosis, which may increase the overall accuracy of sonographic diagnosis of cirrhosis and which appears to have a moderately high degree of reproducibility.