Biliary Atresia: Color Doppler US Findings in Neonates and Infants

Abstract

To describe color Doppler ultrasonographic (US) findings in livers of neonates with biliary atresia (BA) and to compare them with US findings in livers of neonates with non-BA and control subjects. Institutional review board approval was obtained; acquisition of informed consent was exempted. US and color Doppler US findings were retrospectively reviewed in 64 patients with neonatal cholestasis and 19 control subjects. BA and non-BA were confirmed in 29 and 35 patients, respectively. Three pediatric radiologists assessed US and color Doppler US images, independently documented their findings, and resolved discrepancies by consensus. Triangular cord (TC) sign, gallbladder length, and hepatic artery and portal vein diameters were evaluated on US images. The presence of hepatic subcapsular flow was evaluated on color Doppler US images. Diagnostic value of TC sign and hepatic subcapsular flow in the diagnosis of BA were evaluated. Significance of hepatic artery and portal vein diameters in each group was assessed. In the diagnosis of BA, sensitivity and specificity of the TC sign on US images were 62% and 100%, respectively. On color Doppler US images, hepatic subcapsular flow was detected in all patients with BA and in five patients with non-BA. At the first review, there was a discrepancy between radiologists in interpretation of hepatic subcapsular flow in patients with non-BA. However, consensus was reached at the second review. There was no hepatic subcapsular flow in control subjects. Sensitivity and specificity of hepatic subcapsular flow on color Doppler US images were 100% and 80%-86%, respectively, on the basis of individual interpretations of reviewers. Sensitivity and specificity of hepatic subcapsular flow on color Doppler US images were 100% and 86%, respectively, on the basis of consensus reading. Mean diameter of the hepatic artery in patients with BA (2.1 mm +/- 0.7 [standard deviation]) was significantly larger than that in patients with non-BA (1.5 mm +/- 0.4, P < .001) and control subjects (1.5 mm +/- 0.4, P = .001). The presence of hepatic subcapsular flow is useful for differentiating between BA and other causes of neonatal jaundice.