Doppler sonography of the liver is an important noninvasive means of investigating the hepatic vascular system and determining vessel patency and flow direction.

Technique

The hepatic vasculature can be assessed by both pulsed Doppler and color flow Doppler sonography. Pulsed Doppler allows assessment of the presence, direction, and velocity of blood flow within a selected sample volume. For velocity measurements, pulsed Doppler signals should be obtained from an angle of insonation of less then 60° so that errors in velocity calculation can be minimized. At times this may prove difficult, and hepatic vessels may need to be visualized from several approaches during the examination. Pulsed Doppler measurements are time consuming, as all areas of interest must be sampled individually and successively. Color Doppler provides a more rapid assessment of the presence, direction, and pattern of blood flow and may facilitate the detection of vessels that can be missed with gray-scale or pulsed Doppler sonography.

Portal venous system

Anatomy -The splenic and superior mesenteric veins join at the level of the pancreatic neck to form the portal vein. This vein courses in the hepatoduodenal ligament and bifurcates at the porta hepatis into right and left branches, which supply the right and left hepatic lobes.

The portal vein and its major branches are easily visualized using a right oblique intercostal approach. The extrahepatic portal vein also can be easily visualized using a subcostal, anterior abdominal approach (figure 1). Because of a thick investment of collagenous tissue, portal veins have highly echogenic walls. The portal vein is seen in about 97% of normal patients; failure to visualize it can suggest the presence of pathology, such as thrombosis.1

Doppler flow signature -Portal veins exhibit continuous low velocity flow which undulates slightly in response to respiration. Portal venous flow varies considerably with posture, exercise, and dietary state.2,3 Flow is normally directed in a hepatopetal direction (figure 1).

Abnormalities of the portal venous system

Portal venous hypertension -Portal hypertension develops when hepatopetal flow within the portal venous system is impeded. Cirrhosis and congestive heart failure account for the majority of cases of portal hypertension in the western world; other causes include hepatic vein occlusion, portal vein occlusion, and schistosomiasis.

Although the absolute documentation of portal hypertension requires an invasive procedure such as arteriography, Doppler sonography plays a useful role as a noninvasive method of assessing the portal venous system.

Initial studies suggested that portal vein dilatation may be a predictor of portal hypertension, but this measurement has been demonstrated to have a low sensitivity in detecting portal hypertension,4,5 as portal vein diameter is highly variable and can be affected by factors such as phase of respiration, ingestion of food, exercise, and by the presence of portosystemic collaterals. Sensitivity is increased by evaluating the response of the splenic or superior mesenteric veins to respiratory maneuvers.4

The measurement of decreased portal vein velocity by duplex sonography also has been used as a predictor of portal hypertension, but again, normal portal vein velocity is extremely variable and can be affected by many factors unassociated with portal hypertension. A "congestive index" has been shown to have a positive correlation with directly measured portal venous pressure; it is calculated from the ratio between the cross-sectional area (cm2) and the blood flow velocity (cm/sec) of the portal vein.6 In normal individuals, this ratio should not exceed 0.7.

It is the identification of the reversal of portal venous flow or of collaterals that most successfully allows the sonographic diagnosis of portal hypertension, as well as the detection of secondary findings such as splenomegaly and ascites. Sonography is reported to visualize 65 to 90% of portosystemic collaterals. The most common collateral pathway is the coronary gastroesophageal route, which is present in 80 to 90% of patients.7,8 The coronary vein is best seen on a longitudinal scan just medial to where the splenic vein joins the superior mesenteric vein. A dilated coronary vein (> 0.7 cm) is associated with severe portal hypertension.9

A patent paraumbilical vein which demonstrates hepatofugal flow also is an important predictor of portal hypertension (figures 2A and 2B). Although slow hepatofugal venous flow (25 cm/sec) may be detected in the ligamentum teres fissure in healthy subjects, hepatofugal flow with a velocity greater than 5 cm/sec or flow that is visualized with color Doppler sonography continuing anterior to the liver's surface is highly specific for the presence of portal hypertension.10 Other important portosystemic collaterals which may be detected by color Doppler include splenorenal, gastrorenal, and retroperitoneal collaterals.

Portal vein thrombosis

As mentioned above, failure to visualize the main portal vein raises the suspicion of portal vein thrombosis. Portal vein thrombosis is associated with neoplasia, especially hepatocellular carcinoma, hypercoagulable states, and intraperitoneal inflammatory processes such as pancreatitis. In many cases, an underlying cause is not found.

Color Doppler sonography provides a rapid, noninvasive means of identifying portal vein thrombosis. It has been demonstrated to have high negative predictive value (0.98), so that a normal finding on color Doppler imaging virtually excludes the diagnosis.11 Although echogenic material may be seen within the portal vein lumen on gray scale sonography (figure 3), recently formed clot may be almost anechoic and may be overlooked unless color Doppler imaging is employed.12

In the color Doppler evaluation of suspected portal vein thrombosis, machine settings which maximize sensitivity to low flow and low wall-filter settings should be employed. Patients should be scanned during quiet respiration because the Valsalva maneuver, which occurs during breath holding, may abolish low flow entirely.13 Eating increases portal venous flow, and if portal flow is absent, postprandial scanning should be considered.13

Following acute portal vein thrombosis, the thrombus may resolve or the portal vein may undergo fibrosis. Chronic portal vein thrombosis may lead to "cavernous transformation" of the portal vein. In this condition, a single portal vein is not visualized, but multiple serpiginous vessels are seen in the region of the porta hepatitis (figure 4C) and can be identified as hepatopetal collaterals by using color (figures 4A and 4B) and pulsed Doppler (figure 4C) techniques.14

Congestive heart failure -Several studies15,16 have demonstrated that increased pulsatility of the portal venous Doppler signal may be seen in patients with congestive cardiac failure. A similar finding may be seen in severe tricuspid regurgitation.16

Transjugular intrahepatic portosystemic shunt (TIPS)-The transjugular placement of an intrahepatic shunt between the hepatic and portal veins is a relatively new nonsurgical treatment for complications of portal hypertension, including variceal bleeding and refractory ascites. Dysfunction of the shunt, either by occlusion or stenosis, occurs in up to 30% of patients within 6 months.17

Early detection of shunt dysfunction and prompt intervention are important to maintain patency and to prevent recurrent variceal bleeding. In patent, well-functioning stents, high velocity blood flow with peak velocities frequently greater than l00 cm/sec second can be seen.18 In stent stenosis, low velocity blood flow of 60 cm/sec or less may be detected within the entire stent, or there may be low velocity flow with focal velocity elevation at the

site of stenosis.18 Other changes in hepatic hemodynamics, such as interval change in direction of intrahepatic portal venous flow from hepatofugal to hepatopetal, interval change to hepatofugal flow in portosystemic collateral veins,19 or an interval decrease in stent blood flow velocity, also are suggestive of stent malfunction.

Hepatic artery

Anatomy -In the normal liver, 70% of the blood supply is derived from the portal vein; the hepatic artery provides the remaining 30%. The common hepatic artery typically is a branch of the celiac axis, but anatomical variations are common and are seen in up to 45% of patients. One common anomaly is the origin of the right hepatic artery from the superior mesenteric artery, as seen in approximately 11% of patients. The left hepatic artery also may have an anomalous origin, usually from the left gastric artery (10%).20 At the porta hepatis, the hepatic artery usually lies anteromedial to the portal vein (figure 5). If the artery is seen dorsal to the portal vein, an anomalous origin should be considered.13

Doppler flow signature -The hepatic artery and its tributaries demonstrate a low resistance arterial signal with a broad systolic peak and a large amount of forward flow during diastole.

Abnormalities of hepatic arterial flow

Portal hypertension -Increased hepatic arterial flow is seen in severe portal hypertension.21 In portal hypertension, as portal venous flow to the liver decreases, arterial flow increases,22 and the hepatic arterial system becomes more conspicuous, with vessels becoming larger in diameter and demonstrating higher velocity flow than normal hepatic arteries.21 "Corkscrew" arteries also have been demonstrated sonographically in patients with severe portal hypertension. These enlarged, torturous arteries appear similar to those frequently noted angiographically in patients with cirrhosis.22

Vascular liver tumors -Arterial enlargement also may be seen occasionally in vascular liver tumors such as focal nodular hyperplasia and hepatocellular carcinoma.23

Hereditary hemorrhagic telangiectasia (Osler-Rendu-Weber disease)-This autosomal dominant condition is characterized by telangiectases, arteriovenous fistulas, and aneurysms which involve the skin, mucosa, and the blood vessels of the lung, liver, and central nervous system.24 Sonography in patients with hereditary hemorrhagic telangiectasia should demonstrate enlarged torturous extra- and intrahepatic arteries with high velocity arterial flow.25,26 Large arteriovenous fistulas within the liver and "arterialization" of hepatic venous flow in the presence of these large AV fistulas also can be visible on sonography.

Hepatic veins

Anatomy -In most adults, three main hepatic veins (right, middle, and left) drain the liver (figure 6), with the middle and left veins merging just prior to their entry into the IVC.27 Variations in the branching of the hepatic veins and accessory hepatic veins are, however, relatively common. An inferior hepatic vein which drains the right hepatic lobe (segment VI), directly into the IVC is seen in 10% of patients and can sometimes be visualized on sonography.

Doppler flow signature -The hepatic veins demonstrate a complex triphasic Doppler waveform (figure 7) which is seen in all central systemic veins and is caused by fluctuation in right atrial pressure during the cardiac cycle. The hepatic veins have two periods of forward flow during the cardiac cycle, corresponding to the two phases of right atrial filling.13 A short-lived period of flow reversal is seen with right heart contraction.

Abnormalities of the hepatic veins

Hepatic venous obstruction, Budd-Chiari syndrome and hepatic venoocclusive disease -Budd-Chiari syndrome is a rare disorder resulting from occlusion of the hepatic venous system. The occlusion may be located at any level, from the hepatic venules to the inferior vena cava; occlusion at the venule level that is classically seen after bone marrow transplantation is sometimes known as hepatic venoocclusive disease.

Hepatic venous occlusion results in a classic triad of clinical symptoms: hepatomegaly, ascites, and abdominal pain. Budd-Chiari syndrome has been associated with polycythemia rubra vera, chronic leukemia, use of oral contraceptives, neoplasms, pregnancy, and congenital obstructing venous webs, but the exact cause cannot be determined in approximately two-thirds of cases.28

Although inferior vena cavography and hepatic venography remain the gold standard for investigating patients with Budd-Chiari syndrome, the use of color Doppler sonography as a screening tool has been advocated more recently, because it is a noninvasive, relatively inexpensive, and readily available technique.29,30 A positive correlation has been demonstrated with findings at venography.31

In acute Budd-Chiari syndrome, the liver is invariably enlarged and frequently inhomogeneous. There may be atrophy of the right lobe of the liver, hypertrophy of the lateral segment of the left lobe, and enlargement of the vena cava.32 If the hepatic veins are visualized by real-time sonography, thickened walls, areas of stenosis, irregularity, or proximal dilatation may be identified.33 Hepatomegaly, however, will often compress these vessels and render them invisible to real-time examination. In this situation, color Doppler can be used, as it allows visualization of compressed but patent hepatic veins so that the diagnosis of Budd-Chiari syndrome can be excluded.

The middle and left hepatic veins are best scanned in the transverse plane at the level of the xiphoid process because at this angle, the veins are oriented nearly parallel to the Doppler beam. For the same reason, a lateral intercostal approach is used to evaluate the right hepatic vein. Three hepatic veins must be found to exclude the diagnosis because involvement may be segmental. In Budd-Chiari syndrome, Doppler signals from the IVC and/or hepatic veins change from the normal phasic flow to absent, reversed, turbulent, or continuous flow due to interference of blood flow from these veins to the right atrium. Portal vein blood flow also may be affected and is characteristically either slowed or reversed.16

Soon after an acute thrombotic episode, intrahepatic collateral pathways develop, many of which are unique to Budd-Chiari syndrome and provide a definite diagnosis on color Doppler imaging. A "bicolored" hepatic vein is said to be pathognomic, indicating a proximal blockage.13

Sonography in Budd-Chiari syndrome has several limitations. Direct visualization of a hepatic vein web is difficult and venography may be needed in this situation, with possible treatment by balloon angioplasty.34 Involvement of the inferior vena cava also may be underestimated.31

Chronic liver disease -Liver parenchymal disease impairs the compliance of the hepatic veins, decreasing and flattening phasic oscillations. Flattening of phasic oscillations within the hepatic venous system is seen in 50 to 75% of patients with cirrhosis.35,36

Cardiac failure and tricuspid regurgitation-In right heart failure, increased systolic venous pressure causes the hepatic veins to dilate considerably. Normally, hepatic venous flow is directed towards the inferior vena cava, whereas in tricuspid regurgitation, a pronounced systolic reversal of hepatic venous blood flow is seen.

Conclusion

Doppler sonography provides an accurate, noninvasive method of assessing the hepatic vascular system, decreasing the need for hepatic angiography. Pulsed Doppler sonography allows the determination of flow direction and velocity within a selected sample volume, but is time-consuming and unable to provide a global visual display of Doppler information. Color Doppler sonography, on the other hand, provides a more rapid method of visually assessing the presence, direction, and pattern of blood flow within the liver. These two methods are complementary in the investigation of numerous hepatic vascular abnormalities, such as Budd-Chiari syndrome, portal hypertension, and vascular liver tumors. The use of Doppler sonography in the assessment of patients following surgery or after the TIPS procedure is also invaluable. AR

References

1. Merritt CBR: Ultrasonographic demonstration of portal vein thrombosis. Radiology 133:425-427, 1979.

2. Ohnishi K, Saito M, Nakayama T, et al: Portal venous hemodynamics in chronic liver disease: Effects of posture change and exercise. Radiology 155:757-761, 1985.

3. Bellamy EA, Bossi, MC, Cosgrove DO: Ultrasound demonstration of changes in the normal portal venous system following a meal. Br J Radiol 57: 147-149, 1984.

4. Bolondi L, Mazziotti A, Arienti V, et al: Ultrasonographic study of portal venous system in portal hypertension and after portosystemic shunt operations. Surgery 95:261-269, 1984.

5. Bolondi L, Mazziotti A, Arienti V, et al: Ultrasonography in the diagnosis of portal hypertension: Diminished response of portal vessels to respiration. Radiology 142:167-172, 1982.

6. Moriyasu F, Nishida O, Ban N, et al: "Congestion index" of the portal vein. AJR 146:735-739, 1986.

7. Burcharth F: Percutaneous transhepatic portography. I. Technique and application. AJR 132:177-182, 1979.

8. Nunez D Jr, Russell E, Yrizarry J, et al: Portosystemic communications studied by transhepatic portography. Radiology 127:75-79, 1978.

9. Lafortune M, Marleau D, Breton G, et al: Portal venous system measurements in portal hypertension. Radiology 151:27-30, 1984.

10. Wachsberg RH, Obolevich AT: Blood flow characteristics of vessels in the ligamentum teres fissure at color Doppler sonography: Findings in healthy volunteers and in patients with portal hypertension. AJR 164:1403-1405, 1995.

11. Tessler FN, Gehring BJ, Gomes AS, et al: Diagnosis of portal vein thrombosis: Value of color Doppler imaging. AJR 157:293-296, 1991.

12. Bolondi L, Gaiani S, Mazziotta A, et al: Morphological and hemodynamic changes in the portal venous system after distal splenorenal shunt: An ultrasound and pulsed Doppler study. Hepatology 8:652-657, 1988.

13. Grant EG, Schiller VL, Millener P: Color Doppler imaging of the hepatic vasculature. AJR 159:943-950, 1992.

14. Weltin G, Taylor KJW, Carter AR, et al: Duplex Doppler: Identification of cavernous transformation of the portal vein. AJR 144:999-1001, 1985.

15. Duerinckx A, Grant E, Perrella R: The pulsatile portal vein: Correlation of duplex Doppler with right artrial pressure. Radiology 176:655-658, 1990.

16. Hosoki T, Kuroda C, Tokunaga S, et al: Hepatic venous outflow obstruction: Evaluation of pulsed duplex sonography. Radiology 170:733-737, 1989.

17. Ferguson JM, Jalan R, Redhead DN: The role of duplex and colour Doppler ultrasound in the follow-up evaluation of transjugular intrahepatic portosystemic stent shunt (TIPS). Br J Radiol 68:587-589, 1995.

18. Foshager MC, Ferral H, Nazarian GK, et al: Duplex sonography after transjugular intrahepatic portosystemic shunts (TIPS): Normal hemodynamic findings and efficacy in predicting shunt patency. AJR 165:1-7, 1995.

19. Lafortune M, Patriquin H, Pomier G, et al: Hemodynamic changes in portal circulation after portosystemic shunts: Use of duplex sonography in 43 patients. AJR 149:701-706, 1987.

20. Michels NA: Newer anatomy of the liver: Variant blood supply and collateral circulation. JAMA 172:125-132, 1960.

21. Ralls PW: Color Doppler sonography of the hepatic artery and portal venous system. AJR 155:517-525, 1990.

22. Reuter SR, Berk RN, Orloff MJ: An angiographic study of the pre and post operative hemodynamics in patients with side-to-side portacaval shunts. Radiology 116:33-39, 1975.

23. Ralls PW, Johnson MD, Lee KP, et al: Color Doppler sonography in hepatocellular carcinoma. Am J Physiol Imaging 6:57-61, 1991.

24. Martini GA: The liver in hereditary hemorrhagic telangiectasia: An inborn error of vascular structure with multiple manifestations: A reappraisal. GUT 19:531-537, 1978.

25. Vilgrain V, Menu Y, Nahum H: Doppler sonography in Osler-Weber-Rendu disease (Letter). AJR 157:413-414, 1991.

26. Cloogman HM, Dicapo RD: Hereditary hemorrhagic telangiectasia: Sonographic findings in the liver. Radiology 150:521-522, 1984.

27. Cosgrove DO, Arger PH, Coleman BG: Ultrasonic anatomy of hepatic veins. J Clin Ultrasound 15:231-235, 1987.

28. Clain D, Freston J, Kreel L, et al: Clinical diagnosis of the Budd-Chiari syndrome. Am J Med 43:544-554, 1967.

29. Makuuchi M, Hasegawa H, Yamazaki S, et al: Primary Budd-Chiari syndrome: Ultrasonic demonstration. Radiology 152:775-779, 1984.

30. Grant EG, Perrella R, Tessler FN, et al: Budd-Chiari syndrome: The results of duplex and color Doppler imaging. AJR 152:377-381, 1989.

31. Menu Y, Alison D, Lorphelin JM, et al: Budd-Chiari syndrome: US evaluation. Radiology 157:761-764, 1985.

32. Torres WE, Whitmire LF, Gedgaudas-McClees K, et al: CT of hepatic morphologic changes in cirrhosis of the liver. J Comput Assist Tomogr 10:47-50, 1986.

33. Murphy FB, Steinberg HV, Shires GT III, et al: The Budd-Chiari syndrome: A review. AJR 147:9-15, 1986.

34. Meier WL, Waller RM III, Sones PJ: A Budd-Chiari web treated by percutaneous transluminal angioplasty. AJR 137:1257-1258, 1981.

35. Colli A, Cocciolo M, Riva C, et al: Abnormalities of Doppler waveform of the hepatic veins

in patients with chronic liver disease: Correlation with histologic findings. AJR, 162:833-837,

1994.

36. Bolondi L, Li Bassi S, Gaiani S, et al: Liver cirrhosis: Changes of Doppler waveform of hepatic veins. Radiology 178:513-516, 1991.

Dr. Belcher is with the Imaging Department at Medway NHS Trust in Gillingham, Kent, United Kingdom. Dr. Beidle is with the Department of Radiology at Louisiana State University Medical Center in New Orleans.