Measurement of Liver Blood Flow: A Review

The study of hepatic haemodynamics is of importance in understanding both hepatic physiology and disease processes as well as assessing the effects of portosystemic shunting and liver transplantation. The liver has the most complicated circulation of any organ and many physiological and pathological processes can affect it1,2. This review surveys the methods available for assessing liver blood flow, examines the different parameters being measured and outlines problems of applicability and interpretation for each technique. The classification of these techniques is to some extent arbitrary and several so called “different” methods may share certain common principles. The methods reviewed have been classified into two groups (Table 1): those primarily reflecting flow through discrete vessels or to the whole organ and those used to assess local microcirculatory blood flow. All techniques have their advantages and disadvantages and in some situations a combination may provide the most information. In addition, because of the many factors affecting liver blood flow and sinusoidal perfusion, readings in a single subject may vary depending on positioning, recent food intake, anxiety, anaesthesia and drug therapy. This must be borne in mind if different studies are to be meaningfully compared.


reason a portal congestio
index, which is the cross sectional area divided by the mean velocity, has been suggested by Moriyasu et al. 22 as being more useful than absolute flow values.

The Duplex scanner can also be a useful non-invasive tool for assessing patients with portosystemic shunts.Nelson et al. 23 studied patients before and after portosystemic shunting with both duplex scanning and angiography.They concluded Duplex was accurate in determining the direction of flow if an adequate tracing was obtained.Preoperatively it allows the determination of portal vein patency and direction of flow.Postoperatively most porto-caval and mesocaval shunts can be visualised as well as some Warren type shunts.As with many ultrasound appli- cations success is largely dependent on operator skill and experience.Patency is directly demonstrated by flow in the correct irection and a confirmatory sign is the demonstration of corresponding phasic patterns of blood flow in the portal vein and inferior vena cava.Other confirmatory signs are dilatation of the inferior vena cava proximal to portocaval and mesocaval shunts and dilatation of the superior mesenteric vein above a mesocaval shunt7.

Duplex is less useful in the assessment of hepatic and splanchnic arterial flow24.The vessels are relatively short, tortuous and deeply situated, making them difficult to image and the hepatic arterial supply is frequently multiple.Post liver transplan- tation the anatomy may be even harder to demonstrate and scanning of the hepatic artery is currently too time consuming and inaccurate to make it clinically useful in detecting hepatic arterial thrombosis.Intra-arterial Doppler flow probes are now being developed and combined with angiography may ultimately provide the best way to quantitatively measure hepatic and sp anchnic arterial flow4.

In summary, Duplex scanning potentially provides a non-invasive way of assess- ing liver blood flow in many clinical situations including the assessment of portal vein patency, direction of flow, surgical porto-systemic shunting and liver trans- plantation.Its accuracy has been validated in vitro and in experimental animals 25 '26 but problems do exist in using this technique in clinical practice (see appendix A).Improvements both in hardware and software as well as the development of colour flow mapping are likely to be reflected in a greater use of Duplex ultrasound in liver blood flow studies in the future.It offers one of the best approaches to the non- invasive assessment of portal flow but is not yet capable of reliably assessing hepatic arterial flow.

X-Ray angiography Angiography, or radiographic imaging of blood vessels, has a well established role in the diagnosis of liver disease and portal hypertension and is widely used in clinical practice for obtaining high quality vessel images27.As well as anatomical information, however, information on blood flow is also potentially available.

The techniques available for measuring flow using angiography are based on one of two principles.The first approach uses the principle that when an indicator is injected at constant rate into a blood vessel, the degree of dilution is proportional to blood flow, and the concentration in blood after mixing will be lower with higher flow and vice versa28.The main problem with this technique is accurate densito- metric calibration.The second technique involves the measurement of the time taken for the passage of a bolus of contrast material between two sites but unfortunately precise timing of the passage of a dispersing bolus is often difficult to achieve29.In a new approach to this problem flow is determined by computer analysis of contrast concentration profiles as a function of time and distance along a vessel segment29'3.Another solution has been to assess relative flow by using two injections and measuring superior mesenteric, hepatic and splenic arterial flow relative to cardiac output31'32.

The measurement of liver blood flow by angiographic techniques has largely been limited to hepatic arterial studies because catheter access to the portal system is not a routine procedure.Indirect portography, where contrast is injected into either the superior mesenteric rtery or coeliac artery and imaged as it passes out into the portal system results in generally poor images unsuitable for flow analysis.Following direct insertion of a catheter into the portal system, Sovak et al. 33 used a computer to calculate the displacement of lipoidal droplets per frame.The average velocity ranged from 15.5 to 24.4 cm/sec in 6 normal patients and decreased during inspiration.Recently Iwanaga et al. 34 used this method 35 as a "gold standard" to test the validity of a Doppler duplex system for measuring portal blood flow in 10 patients with liver disease and found a significant correlation (r=0.970) between the maximum portal blood flow velocity by duplex ultrasound and the mean velocity calculated from cineangiographic methods.

Although it requires vascular catheterisation X-ray angiography is still the modality of choice for critical morphological vascular studies.That X-ray angiogra- phy has not been widely used for measuring blood flow is due in part, we believe, to the us of inappropriate algorithms for processing the image data29'36.The method does, however, have great potential especially when combined with lower dose Digital Subtraction Angiography and new low osmolarity non-ionic contrast agents37.Minipuncture needles and catheters 38 have led to increased safety of the technique and the equipment and expertise is potentially available in many centres.


Nuclear magnetic resonance

Nuclear Magnetic Resonance (NMR) imaging is a noninvasive imaging modality that is rapidly gaining clinical acceptance, although widespread introduction has been delayed by expense.Flow detection with NMR spectroscopy has been xplored for more than 30 years39-41.When NMR imaging was first performed in the late 1970s signal loss was noted within arteries and attributed to high flow rates42.In the early 1980s, several causes of increased signal intensity were described, generally associated with slow flow in veins and dural sinuses43 '44.Understanding these flow phenomena has provided the basis for the development of specialized NMR imaging sequences intended to quantitate blood flow measure- ments.Several methods have been proposed to quantify blood flow 45-47 but at this stage their relative merits in terms of spatial and velocity resolution and image acquisition times have not been completely evaluated, nor has liver blood flow measurement been considered specifically.We suspe t, however, that this will be an area of great development in the future.


Dye Dilution Techniques

Plasma disappearance methods Attempts have been made since the middle of this century to measure liver blood flow by dye infusion methods48.Certain organic .dyesare extracted by the hepato- cytes and if the rate of extraction is measured, live blood flow can be calculated using Fick's Principle49.The liver plasma flow (LPF) is defined as:
LPF= FI/[Ca(t)-Cv(t)]
where, 1-I is rate of removal of the dye from the circulation by the liver in mg/min, Ca(t) and Cv(t) are the dye concentrations in mg/ml of the blood entering and leaving the liver respectively, leading to LPF measured in ml/min.LPF can be converted into liver blood flow if the value for the haematocrit is known.

The first dye used was bromsulphalein 5 but indocyanine green is now used more commonly as it is more specifically extracted by the liver51.The first measurements made using this substance employed the constant infusion method but Caesar et al. 51 have shown that analysis of plasma disappearance curves after a bolus injection gives nearly identical results.Hepatic extraction is usually measured by hepatic vein sampling, however, a method requiring only peripheral vein sampling and utilising pharmacokinetic modelling has been described and validated in normal subjects52.
his method has, however, been criticised when applied to patients after liver transplantation 53 and it may be unreliable in patients with liver disease54.

Pirttiaho et al. 5 estimated liver blood flow by fast intravenous injection of indocyanine green (0.5 mg/kg body weight).They found that the liver blood flow in 5 normal patients was 1258 + 119 ml/min and that there was a close correlation (r-0.88) with dynamic 99mTc-sulphur colloid imaging but not with values obtained using the 133Xe clearance technique.

The advantage of dye clearance techniques is that they are relatively simple.Inaccuracies arise, however, when extrahepatic removal of the dye occurs 5 or when it is used in patients with liver disease54.These inaccuracies, combined with the development of other, more accurate methods, for measuring liver blood flow, have resulted in dye clearance methods being used less commonly.For many years, however, they were the best technique available and much pioneering work was done using them.


Radioisotopic methods

The co cept of using radioactive tracers to help in the assessment of liver blood flow and perfusion is attractive.Three basic groups of techniques have been described.

A. Diffusible Gas Tracers Kety 56 introduced the principle of "local tissue clearance" or "washout" of rapidly diffusing isotopes as a way of measuring blood flow.Initially small amount of radioactive 24Na was used but later inert and lipid-soluble gases such as 85Kr and 133Xe 57 were found to be more valuable with the cellular membrane not constituting a barrier to diffusion58.

Following injection of an arterial or portal venous bolus of gas dissolved in saline the elimination f these elements is in most situations only limited by the rate of capillary blood flow.Such an isotope will be eliminated in the form of a monoexponential function (giving a straight line when plotted on a logarithmic scale) if the tissue is uniformly perfused.Externally placed scintillation detectors are used to record the clearance curve.Fick's principle is then used in the analysis of the data and from a series of washout curves liver blood flow can be calculated.


B. Radio-labelled Colloids

In this technique colloid-bound radionuclides are administered intravenously and the rate constant of liver uptake is measured either by multiple blood sampling or external scintillation counting.The Fick principle is then applied to calculate blood flow, with the assumption that extraction efficiency is 100%.

Various colloids have been used with different radionuclides.The first work was with 32p labelled chromic phosphate which is a pure fl-particle emitter and cannot be detected by external counting59'6.It does, however, have a relatively high extraction efficiency at 95 %.Colloidal 198Au ha

been used for external c
unting 61'62 but has a lower extraction efficiency (80% or less).Now 99mTc-labelled sulphur colloid is most frequently used as it has a high extraction efficiency and can be counted externally.Dynamic images are acquired via a gamma camera and an on- line computer system and the assumption is made that the liver and spleen have an equal extr colloidal particles 63 and that this is close to 100%.Unfortunately, although these assumptions are probably valid in normal subjects, they may not be true in patients with liver disease64.Analysis of the time variation in liver activity is performed following bolus intravenous injection.The arterial and portal comp nents are separated by their times of arrival at the liver.Several different methods have been described to estimate fractional hepatic arterial flow using hepatic artery, portal vein, and reference organ time activity curves63-67.


C. Hepatosplenic Radionuclide Angiography

This technique is based on the use of 99mTc-pertechnetate which is not extracted by the liver.A bolus injection is given intravenously and dynamic images are acquired during its first pass phase68.The original technique 69 has been modified TM and is now reported to be more reproducible68.Sarper et al. 68 generated first-pass radioactivity versus time curves by following a rapid intravenous injection of 740 MBq of 99mTc-pertechnetate.For analysis, two tim points were identified: the arrival time of activity in the liver, to, and the time of maximum activity of the abdominal organs, tc.The former was estimated from the liver time-activity curve; the latter from the kidney time-activity curve and the accuracy of the method depends on there being normal perfusion of the kidneys which act as a reference organ.T e gradients of the liver curve from to + 7 seconds and from tc to tc + 7 seconds, calculated by linear least-squares regression analysis,, are Go and Gc respectively.A hepatic perfusion index (HPI) is then defined as:

HPI-G/(Go + Go).

This method was applied to 7 normal volunteers and 57 patients with biopsy proven cirrhosis.The HPI was 66 _+ 7% for the normals and ranged from 8-59 for patients with liver cirrhosis68.


Advantages and limitations of radioisotope methods

The use of radio-l

elled diffusible gas t
acers has gained acceptance and popular- ity for measuring cerebral blood flow, where an inhalation technique can be used successfully.Unfortunately the technique is not

accurate for measuring li
er blood flow and the washout curves are frequently not monoexponentia171, perhaps due to recirculation of the tracer, the fact that liver tissue may not be homogenously perfused or because there is incomplete clearance of tracer during its first passage72.In addition, the need to catheterise either the hepatic portal or arterial system is a major disadva tage and one of the reasons why the technique has failed to gain widespred popularity in hepatic studies.Colloid bound tracers and radionuclide angiography can not provide absolute values for flow but they can provide valuable information about the relative contribution of the hepatic arterial and portal systems.Such an index may be of more interest than absolute flow values in certain disease states such as cirrhosis.

However, the background scatter of tracer and the affinity to fat which many tracers have, makes the measurements difficult to evaluate.Colloids also have a range of particle sizes and hence a range of values of extraction efficiency.


METHODS MEASURING TISSUE PERFUSION

Radioactive Microspheres If a bolus of tracer is well mixed in the afferent blood supplying an organ, then it will be distributed to different parts of the organ in exactly the same way as the blood which is transporting t.This is called the indicator fractionation principle.This principle has been used to quantify regional blood flow distribution using radio-labelled particles, diffusible indicators and autoradiography73'74.

Microspheres are chosen to be of a size (10-15/m) which will just lodge in the capillary circulation.The injection can be given some time before local distribution of the trapped spheres is measured, a procedure usually carried out post-mortem by taking biopsies of the tissue being studied and measuring radioactivity in a well counter.

Using this technique Greenway et al. 75 studied the regional distribution of portal and hepatic blood flow in the liver by injecting 14C amd 51Cr-microspheres into the portal vein and hepatic artery of 12 cats and 15 dogs.They found the liver homogeneously perfused from both systems in contrast to other work using different techniques76'77.The technique has also been used to study the vascularity of experimental liver tumours and in particular the relative role of portal and arterial blood supply to these tumours78-81.

The microsphere method is useful for providing values of blood flow in animal studies where sacrifice of the animal occurs.Impaction of microspheres must, however,

ffect local flow and with multiple injecti
ns before sacrifice this may become a significant source of artefact.To minimize this problem, injection quantities are made as small as possible, but this militates against accurate measurement of regional flow, especially in regions with low volume flow.Its major disadvantage is, however, that it cannot be used clinically.


Heat Exchange Methods

The concept of measuring tissue blood flow using heat clearance techniques was first suggested by Gibbs82.This method requires a heated thermocouple, main- tained at a certain temperature (2-4C) above that of the surrounding tissue, to be either placed onto the liver surface or inserted into the liver tissue.The tempera- ture of the needle is dependent on local blood flow increased perfusion tends to cool the needle, whereas reduced perfusion allows it to heat up.Measurement of the energy required to maintain the temperature increment constant therefore can be regarded as giving an indirect measurement of flow82-85.However, values will depend on the exact position of the probe and the

etabolic state of
the liver86.The method is, therefore, only a semiquantitative approach to flow and because of its invasiveness has not yet found widespread favour for liver studies.

Hydrogen Electrode Th

method was introduced by Auckland and Bower 87 and
further developed by Fieschi et a/. 88'89 and by Bozzao et al. 9. Molecular hydrogen is administered with the respiration gas until the tissue reaches saturation.The hydrogen supply is then turned off and its clearance rate is determined polarographically through platinum electrodes placed on, or into, the liver.A current is generated at the electrode surface by oxidation of molecular hydrogen to hydrogen ions.This current declines as hydrogen is removed and the steepness of the clearance curve correlates directly with the magnitude of the total liver blood flow and reflects perfusion within a radius of approximately 5 mm of the electrode.The calculation of tissue blood flow from hydrogen clearance curves is based on the theory developed by Kety, 91 and the method has been reviewed and simplified by Young92.

Gouma et al. 77, using a hydrogen electrode applied to the surface of porcine liver, found that calculated liver blood flow measurements using this method gave much lower values than those obtained using he indocyanine green clearance method and in addition flow fell by over 90% if the hepatic artery was ligated.They concluded from these experiments that the surface of the liver is mainly supplied from the arterial system.Nishiwaki et a

93 used the hydrogen clearance met
od and transit-time ultrasonic blood flowmetry to investigate blood flow after liver trans- plantation in 40 mongrel dogs and found reductions in both hepatic arterial and portal venous flow after transplantation.The advantages of this method are that it can provide unlimited measurements of liver blood flow without significant alteration in physiological variables.There is no evidence that the administration of the hydrogen itse f significantly alters flow.Despite this, most investigators have not used the technique, mainly due to concern over the inflammability and explosiveness of pure hydrogen gas.In addition the method is not continuous, cannot handle rapit changes of flow and may reflect arterial rather than venous inflow77.It may also be inaccurate if the iver is not homogenously perfused.

Oxygen Electrode An oxygen electrode consists of a noble metal cathode maintained at a negative potential with respect to a reference electrode.It is placed on the surface of the organ to be studied and oxygen diffusing from the tissue to the cathode surface is reduced when the potential is applied, giving rise to a current94.A naked electrode is subject to "poisoning" by electrophoretic deposition of tissue protein on its surface but this can be prevented by covering the.electrode with a gas-permeable membrane95.When the oxygen consumption of the electrode is low, tissue oxygen is not disturbed and so a direct measurement of the partial pressure (pO2) is obtained96.However, if the oxygen consumption of the electrode is high the electrode will measure the rate of supply of oxygen to the tissue and this is dependent on local blood flow97 '98.Ji et al. 99 applied a microneedle electrode to rat liver and found that tissue pO2 values were different at periportal and perihepatic sites.Kram and Shoemaker 1 applied a "min

ture" oxygen electrode
in a single illustrative case to human cirrhotic liver to measure tissue pO2 and their instrument responded to both changes in local organ blood flow and arterial pO2.In these two studies electrode readings were not related to portal venous flow.We compared readings from a membrane covered (Clark type) flow dependent oxygen electrode applied to the surface of rabbit liver, with those from an electromagnetic flowmeter on the portal vein.We found that under operative conditions, changes in oxygen electrode readings correlated well with portal blood flow as measured by an electromagnetic flowmeter over a range of flow rates11.

This technique can give a continuous and instantaneous measurement of portal venous inflow when hepatic arterial inflow is undisturbed.However, it is invasive as the electrode must be applied directly onto the liver surface.In addition it only giv s a measure of flow in the tissue immediately below the electrode and no absolute value for flow can be calculated.Potentially, however, it can be used on human liver either at laparotomy or laparoscopy.


Laser Doppler

Laser Doppler is a relatively new technique (1972) 12 for measuring local blood flow.The device consists of a helium-neon laser and an optic fibre which transmits this light to the surface of the tissue to be studied.Light that is scattered by red blood cells undergoes a frequency shift and a portion of this spectrally-broadened light is transmitted back by a fibre light-guide to two photodetectors.This signal is analyzed and the relative portion of light which has undergone Doppler shift is proportional to velocity of blood flow.The microvascular bed consists of an intricate network of small blood vessels and hence the angle between the red cell velocity vectors and th beam propagation vectors of the scattered light can be regarded as random13.This technique gives a continuous measure of red cell motion in the outermost layer of the tissue under study with little or no influence on blood flow.The depth to which the beam penetrates varies with the tissue being studied 13'14'76 and flow is likely to be measured in a volume of approximately 0.6-1.3mm when the probe is applied to the liver surface.

Laser Doppler has been studied in vitro by measurement of liquid flow through small-bore tubes and a coefficient of variation for readings of 6%15 confirmed its accuracy.In vivo it has been used extensively to measure skin blood ttow 16 but less has been written on its use and limitations for estimating liver blood flow76'107.

Arvidsson et al. 76, using Laser Doppler flowmetry in pigs, investigated liver blood flow and confirmed previous work 77 with hydrogen clearance methods that the liver surface is mainly supplied from the arterial system.Laser doppler has also been used to study blood flow to experimental liver metastases18.The advantages of the method are that it can provide an instantaneous and continuous measurement of microcirculatory flow in a way that does not alter flow.Its disadvantages are that the probe must be applied directly o to the liver, flow can not be measured in absolute units, the absolute volume of tissue measured is not known and only surface flow is assessed (needle probes have not been used in the liver as haematoma formation around the probe tip would make readings unreli- able).


CONCLUSION

Currently the measurement of hepatic blood flow and perfusion is fraught with difficulties.There are often large variations in both flow and perfusion measure- ments, not only between techniques but also between different groups using the same technique.Some of these differences may be due to the methods and conditions used and others are undoubtedly caused by complexities of the liver circulation which we do not yet understand.

Most importantly, we are still looking for a reliable method of non-invasively assessing liver blood flow in the clinical context.Duplex Do pler Ultrasound offers a good way of assessing portal vein flow but its many inaccuracies should be borne in mind.Clinical measurement of hepatic and splanchnic arterial flow is more difficult and duplex does not yet have the accuracy to perform this function reliably.Intraarterial Doppler flow probes, flow analysis of digital x-ray angiograms and NMR may have a role in the future.Radiolabelled colloids or hepatosplenic radionuclide angiography can provide valuable information about the relative contribution of the portal and arterial system but are still largely research tools and are less accurate in the presence of pathology.Currently, investigators are best advised to familiarise themselves with a range of techniques as it is apparent that no single method is able to fulfil all the requirements of either basic research or outine clinical practice.

APPENDIX A DOPPLER'S LIMITATIONS AND SOURCES OF ERRORS Although duplex instrumentation is now widely accepted as valid for most vascular applications, there are several disadvantages which have prevented total accep- tance of the method19.These have been well documented and summarised by Burns 16 and Merritt11.Briefly, they are as follows" (a) When used to measure laminar flow in a vessel and the beam width is less than the

ameter of the
essel, only the central portion of the vessel lumen will be insonated leading to an error in estimating velocity and hence flow.

(b) Errors can arise in estimating blood vessel diameter due to" (1) Imaging not being perpendicular to the longitudinal axis of the vessel.(2) Poor resolution of the imaging transducer.(3) Pulsatility of blood flow, causing variation in vessel diameter with time.(4) Observer variability1:.(5) Only a short length of vessel being available for imaging such as occurs with the portal vein13':5.

(c) Sampling problems" Nearly all current pulse mode Doppler machines are based on centre-line measurements of peak velocity and spectral 113 broadening Abnormalities may be missed due to failure to sample nearer the wall where flow disturbances are most likely to occur114'5.

(d) Errors due to beam angle" the magnitude of velocity is the function of the cosine of the intercept angle between beam direction and the blood vessel.The flow-volume rate calculation equation is a trigonometrical function of the angle between the beam direction and the blood flow.Errors vary considera- bly with t at angle, being minimal in the angle range 55-766.

Table 1
1
Hepatic blood flow measurement techniques.
METHODS MEASURING BLOOD FLOWVelocity or Transit Time MethodsElectromagnetic FlowmeterDopier UltrasoundX-Ray AngiographyNuclear Magnetic ResonanceDye Dilution TechniquesPlasma Disappearance MethodRadioisotope techniquesMETHODS MEASURING TISSUE PERFUSIONRadiolabelled MicrospheresHeat ExchangeHydrogen ElectrodeOxygen ElectrodeLaser DopplerMETHODS MEASURING BLOOD FLOWVelocity or Transit Time Methods
AcknowledgementsWe thank Dr. S. Padayachee, Division of Radiological Sciences, Guy's Hospital School of Medicine, London, for valuable discussion on application of Doppler ultrasound to measure liver blood flow and also we thank Dr. C. Piasecki for valuable discussions on application of oxygen electrode to measure liver blood flow.D.J.H. is grateful for the support of the Leverhulme Trust.
Liver blood flow I.Intrinsic and nervous control of liver blood flow. P D I Richardson, P G Withrington, Gastroenterology, $. 11981

Liver blood flow: II. Effects of drugs and hormones on liver blood flow. P D I Richardson, P G Withr

gton, Gastr
enterology, $. 11981

Utilisation des forces electromotrices d'enrigistrement des variations de vitesse des liquids conducteurs: un nouvel hemodromograph sans palette dans le sang. P Fabre, C. R. Acad. Sci. 1941932

Measurement of flow and volume of blood. J G Webster, ch. 8Medical instrumentation: Application and Design. J G Webster, Boston: Houghton Mifflin, 1978

Methods of measuring blood flow. A M Harper, A R Lorimer, D L Th mas, Scientific Foundation of Anaesthesia. C Scurr, S Feldman, London: Heinemann, 1974

Miniature electromagnetic flow meter applicable to coronary arteries. E M Khouri, D E Gregg, J. Appl. Physiol. 182241963

Blood flow and blood velocity measurement in vivo by electromagnetic induction. D G Wyatt, Med. Biol. Eng. Comput. 221984

Significance and limitations of electromagnetic blood flowmetry, Experimental and Clinical results. H Meisner, K Messmer, Progr. Surg. 81970

Measurement of hepatic blood flow by bromsulphalein and by the electromagnetic flowmeter. T Drapanas, D N Kluge, W G Schenk, Surgery. 481960

The validity of chronic hepatic blood flow measurements obtained by the electromagnetic flow meter. J B Price, R C Britton, L M Peterson, J W Reilly, A R Voorhees, J. Surg. Res. 71965

The estimation of hepatic blood flow in man. S E Bradley, F J Inglefinger, G P Bradley, J J Curry, J. Clin.lnvest. 248901945

Study of the flow pattern in peripheral arteries by ultrasonics. Nihon Onkyo-gakka

S Satomura, 1959Shi15151

Duplex Doppler scanning in t
e pelvis and abdomen. K J W Taylor, P N Burns, Ultrasound Med. Biol. 111985

Noninvasive measurements of portal blood flow. K Okazaki, S Miyazki, S Onishi, Gastroenterology. 931987

Pulsed ultrasonic Doppler blood-flow sensing. D W Baker, IEEE Trans. Sonics Ultrasonic. 1731701970

The physical principles of doppler and spectral analysis. P N Burns, J. Clin. Ultrasound. 151987

Duplex Doppler examination of the liver and portal venous system. D B Koslin, L L Berland, J. Clin. Ultrasound. 151987

Pulsed Doppler flow as a criterion of portal venous velocity Comparison with cineangiographic measurements. K Ohnishe, M Saito, H Koen, T Nakayama, F Normura, K Okuda, Radiology. 1541985

Blood flow in the human portal vein. A cineradiographic method using particulate contrast medium. M Sovak, R L Soulen, F A Reichle, Radiology. 991971

Colour-coded carotid Doppler imaging: An angiographic comparison of 324 bifurcation. N Ackroyd, R Lane, L Dart, Aust. NZ. J. Surg. 541984

Effect of food intake and various extrinsic hormones on portal blood flow in patients with liver cirrhosis demonstrated by pulsed doppler with the octoson. K Okazaki, M Miyazaki, S Onishi,