Comparison of ultrasound, CT, and MR imaging in the evaluation of candidates for TIPS.

To compare ultrasound (US), CT, and MRI in the evaluation of hepatic vascular anatomy, portal and splenic venous flow, and collateral pathways (varices and spontaneous shunts) in candidates for transjugular intrahepatic portosystemic shunting (TIPS), 17 patients with history of refractory variceal bleeding or intractable ascites underwent duplex US, contrast-enhanced CT, and MRI before TIPS. The appearance of portal and hepatic anatomy was graded from 1 (not visible) to 4 (excellent visualization) independently by four radiologists. Presence and direction of portal and splenic venous flow, and presence and location of varices and spontaneous portosystemic shunts were also assessed. Results and effects of interobserver variation were assessed for significance using Friedman's ANOVA and Wilcoxon's signed-rank test. MRI yielded higher scores than CT or US for hepatic veins (P < .0001) and inferior vena cava (P < .0001). MRI and CT scored better than US for portal vein branches (P = .012) and splenic vein (P = .0038). All tests demonstrated the main portal vein well, with no statistically significant difference. US and MRI were more sensitive than CT for detecting portal vein flow and direction (US 76%, CT 0%, MRI 82%). MRI was most sensitive for splenic vein flow and direction (US 41%, CT 0%, MRI 76%). CT and MRI were more sensitive than US in detecting varices (US 5%, CT 50%, MRI 58%) and spontaneous shunts (US 13%, CT 75%, MRI 75%). Interobserver variation did not influence results significantly P = .3691). MRI provides the most useful information and may be the preferred single imaging test prior to TIPS.