PubMed1992
Computed tomography has widespread clinical application in the evaluation of the portal venous system, even though quantitative methods are impractical due to the inability to measure portal flow discrete from hepatic arterial flow, morbidity associated with the use of large volumes of iodinated contrast, and technical limitations. This represents a major disadvantage compared to Doppler ultrasound and magnetic resonance angiography. Qualitative applications include evaluation of portal vein patency, diagnosis of portal vein thrombosis, underlying inflammatory or neoplastic conditions, and evaluation of surgically created portosystemic shunts and collateral flow. Diagnostic criteria for portal venous thrombosis include nonopacification of the central portion of the portal vein, peripheral enhancement of the vein, and irregular periportal hepatic parenchymal enhancement. However, misdiagnosis is common, occurring in 16% of cases analyzed in one limited series, and periportal vein enhancement is now recognized as a nonspecific finding associated with underlying endothelial injury. Cavernous transformation of the portal vein and neoplastic invasion of the portal system are more reliably recognized. Computed tomography arterial portography demonstrates collateral pathways and arteriovenous shunts. Computed tomography has a sensitivity of 85% in detection of esophageal varices compared to endoscopy, but has the advantage of demonstrating splenorenal, gastrorenal, peripancreatic, pericholecystic, retroperitoneal and omental collateral vessels, and spontaneous large portosystemic shunts, with greater sensitivity than angiography. Computed tomography combined with Doppler ultrasound angiography remains popular, despite a lack of large-scale prospective efficacy studies demonstrating diagnostic superiority over other imaging techniques, largely because of its accessibility, and its detailed axial anatomic images providing an overview of multiple organ systems, and patency of major vessels.