PURPOSE: We performed this study to evaluate the usefulness and the limitation of magnetic resonance angiography in imaging portal vein and hepatic vein. MATERIALS AND METHODS: Magnetic resonance portography was performed in five normal subjects and seven patients with hepatomas, two patients with Budd-Chiari syndromes, one patient with liver cirrhosis and one patient with hepatic metastasis from stomach cancer. Magnetic resonance angiography was done with a 1.5-T Scanner. Breath-hold two-dimensional time-of flight images with spoiled gradient echo technique were acquired. Scan parameters were 34/8/40°(TR ms/TE ms/flip angle). The portal vein and the hepatic vein were selectively imaged by applying two presaturation bands. These images were then postprocessed by a maximum intensity projection algorithm. MRA findings were compared with ultrasonography in all cases of the hepatic disease, and conventional angiography (SMA portography) in four cases. RESULTS: In normal subjects, the splenic vein, intrahepatic and extrahepatic portions of the portal vein, and the hepatic veins were well visualized. In the patients with hepatic diseases, the varices (4 cases), the splenorenal shunts (2 cases), and the recanalized umbilical vein (1 case) were demonstrated. There were portal vein thrombosis in the cases of hepatomas and hepatic metastasis. In the cases of Budd-Chiari syndrome, the hepatic veins were not visualized and there were inferior vena cava obstructions with multiple collateral vessels. CONCLUSION: Compared with ultrasonography, MRP may be useful in evaluation of varices, splenorenal shunts, and other collaterals, though less accurate than conventional angiography. MRP can be useful as a noninvasive screening alternative in the evaluation of portal vein and hepatic veins.