Diagnosis of portal vein thrombosis in patients with hepatocellular carcinoma: efficacy of color Doppler sonography compared with angiography.

Abstract

The aim of this study was to determine the efficacy of color Doppler sonography in the diagnosis of portal vein thrombosis in patients with primary hepatocellular carcinoma. The findings on angiography were used as the gold standard for diagnosis. We compared the findings on color Doppler sonograms and hepatic angiograms in 18 patients with hepatocellular carcinoma and portal vein thrombosis and in 22 patients with hepatocellular carcinoma without portal vein thrombosis. In most patients, tumor-related thrombus of the portal vein was confirmed by autopsy or surgery. The sonographic criteria for diagnosing portal vein thrombosis included nonvisualization of portal vein flow, pulsatile flow in the thrombus, arterioportal shunts, and cavernous transformation of the portal vein. Nonvisualization of portal vein flow was the predominant color Doppler finding in all patients with portal vein thrombosis. Other flow abnormalities overlapped the finding of nonvisualization of portal vein flow in 10 of 18 patients with portal vein thrombosis. The sensitivity of color Doppler sonography was 100% for nonvisualization of portal vein flow, 89% for detection of pulsatile flow in the thrombus, 60% for detection of arterioportal shunts, and 100% for detection of cavernous transformation. Among these findings, pulsatile flow in the thrombus was diagnostic for pathologically proved neoplastic thrombi in the main portal vein; sensitivity and specificity were 89% and 100%, respectively, and accuracy was 96%. No portal vein that appeared normal on color Doppler sonograms was thrombosed on arteriograms. We conclude that color Doppler sonography is a useful means of imaging the vascular extension of a tumor-based thrombus and an accurate means of screening for portal vein thrombosis.