The double vascular supply to the liver means that special care must be taken when scanning patients with vascular liver problems. Both supply vessels as well as the hepatic veins and sites for porto-systemic shunts must be scrutinised.

Arterial abnormalities are rare, though variants such as a dual supply are very common and important in liver surgery, especially for partial live-related transplantation. Occlusion is only encountered after liver transplantation. The anatomy of the portal vein is very constant but anomalies of the hepatic veins are common, especially accessory veins draining the right liver (segments 7 and sometimes 6) as well as the caudate lobe (segment 1). They are important in liver surgery and need to be mapped out as part of surgical planning using CT or ultrasound.

Portal hypertension can be caused by narrowings at different levels, the commonest being intrahepatic and resulting from cirrhosis. Pre-hepatic portal hypertension results from stenosis of the main portal vein or its tributaries and is a complication of carcinoma of the head of the pancreas and of splenectomy. Post-hepatic hypertension is seen in the Budd Chiari syndrome. Each produces different patterns of abnormal Doppler changes which are characteristic, with reduced or reversed flow in the main portal vein and marked increases in hepatic arterial flow, which eventually becomes the sole supply to the liver. In all types, extra hepatic porto-systemic shunts may form and the typical sites (gall bladder bed, retroperitoneum at the head of the pancreas, spleno-renal and spleno-gastric) need to be examined systematically with a view to understanding the altered haemodynamics. Except in the pre-hepatic types, the fetal umbilical vein may recanalise and decompress the portal system; it is seen as a channel in the position of the ligamentum teres and can be traced away from the liver passing inferiorly in the direction of the umbilicus. Here, varices may empty into subcutaneous veins that either take flow superiorly to anastomose with the superior epigastric veins or, more commonly, pass inferiorly into the pelvis and empty into retroperitoneal veins. In this situation, portal vein flow is faster than normal and there may be reversed flow in the right portal branch. If the portal vein occludes, flow may be restored via recanalisation or new vessels that open around the original portal vein, so-called cavernous transformation. The Doppler findings are complex, with numerous tortuous vascular channels in the porta hepatis.

The intrahepatic arterio-venous shunts that are so important in cirrhosis are usually microscopic and cannot be imaged with ultrasound. However, the rapid transit of blood from the hepatic artery to the hepatic veins can be tracked by using a bolus of a microbubble contrast agent as a tracer. In the normal, there is a substantial delay (> 12 secs) but this is much shorter in cirrhosis. This non-invasive test separates mild from severe chronic hepatitis and from frank cirrhosis and can be used to monitor these patients. Metastases produce a similar shift of the liver's blood supply in favour of the artery.

Surgical stents to shunt the blood from the portal to the systemic circulation may be inserted, the usual version being the transjugular intrahepatic porto-systemic shunt, which is highly echogenic on ultrasound. This may make Doppler evaluation difficult and there is always some degree of flow disturbance at the portal end of the shunt. Usually Doppler can be used to establish that the shunt is patent, but assessing restenosis is difficult because the flow may accelerate (across a stenosis) or slow (if the stenosis becomes haemodynamically significant). This confusing situation has led to angiography with pressure measurements being used instead of ultrasound.

Pathology of the hepatic veins is relatively uncommon, but occlusion occurs in the Budd-Chiari syndrome, which may be caused by membranous flaps in the upper IVC or occur in hypercoagulable states. Occasionally a tumour, such as a metastasis compresses one of the veins, leading to occlusion. If the occlusion is complete and of rapid onset, acute liver failure ensues and is usually rapidly fatal. Partial or incomplete forms lead to portal hypertension and Doppler can be used to map the ensuing haemodynamic changes, with absent or reversed flow in some of the hepatic veins and often features of portal hypertension with ascites.

Some rare vascular abnormalities are of interest. Veno-occlusive disease is most commonly the result of priming chemotherapy before bone marrow transplants. Unless severe enough to slow or reverse portal vein flow, it cannot be detected with ultrasound. The liver is often involved in haemorrhagic telangectasia (Osler-Weber-Rendu syndrome). The shunts may be large enough to be detectable on B-mode imaging but are more obvious with colour Doppler on which they form chaotic, fast flowing vessels in all parts of the liver. Peliosis is a rare disorder in which vascular lakes form throughout the liver and is associated with a variety of infections and with renal transplantation. Usually they are small and, since the flow velocity is low, undetectable on Doppler, although the liver may appear heterogeneous.