vascular radiology kreel - bmj

Postgraduate Medical Journal (January 1970) 46, 19-31.

Vascular radiology in liver disease

Louis KREELM.D., M.R.C.P., F.F.R.

Royal Free Hospital, London, W.C. I

1/·' ~ SummaryVascular radiology of the liver has increased in scopeand function in recent years due mainly to theapplication of new techniques.

It is now possible to examine not only the inferiorvena cava and the portal venous system, but also thehepatic veins and the coeliac axis and superiormesenteric artery. Hepatic vein occlusion, portal veinpatency and collateral veins, as well as space-occupying lesions, can now be diagnosed with a fairdegree of accuracy.

These techniques have also helped in the under-standing of the altered haemodynamics of portalhypertension and can be used for treatment by intra-arterial perfusion of chemotherapeutic substances.

IntroductionThe full application of vascular radiology to the

diagnosis of disease of the liver has come somewhatlater than in other organs such as the kidneys andbrain for two reasons: (1) the rapid development ofliver biopsy and histological techniques appeared tomake other anatomical methods of diagnosis super-fluous and when any doubt existed laparotomy wasundertaken; and (2) it required the introduction ofselective arterial catheterization to show thatsufficient detail could be obtained to render this auseful diagnostic tool. At present most liver unitsmake extensive use, not only of spleno-portographybut also of arteriography, hepatic venography andcavography. These methods of vascular radiologymay not only provide an anatomical and patholo-gical diagnosis but in many cases also demonstratethe complex altered haemodynamics which resultfrom liver disease.The particular examination or sequence of radio-

logical examinations to be undertaken will dependnot only on the clinical problem, but also on theavailability of ancillary investigations such as isotopeor ultrasonic scanning, the whims and fancies ofparticular clinicians and the desire of radiologicaldepartments to become involved in this work. Asfar as radiologists are concerned it is no longer aquestion of extraordinary skill or complex equip-

ment because in most large hospitals there are'arteriographers' available and the equipment haslong since been installed.The technical factors that have been responsible

for the development of these procedures include theSeldinger method of percutaneous catheter insertion,the introduction of pre-formed radio-opaque cathe-ters, non-toxic contrast medium, and the availabilityof image intensification and television monitoring.There is now little doubt that these techniques can bequickly mastered in a modern radiology department-the great mystique of selective arteriography hasgiven way to a matter-of-fact approach with aconsequent greatly increased demand in cases withliver disease.The apparent complexity of vascular radiology of

the liver stems from the possible anatomicalapproaches that can be used. Not only is it possibleto visualize the hepatic veins, the splenic and portalvenous system and the hepatic arteries but this can beachieved by different methods. Indications and con-tra-indications for these are fairly well defined andwill be considered in each case. However, the actualtechniques will be referred to only briefly as theseare now fairly standard, considered in detail in mosttexts and in any case, like most practical procedures,are learnt by experience.

Hepatic venographyThe three hepatic veins drain into the inferior

vena cava, close to its entry into the right atrium andare best approached from the right medial basilic veinat the elbow via the subclavian vein, superior venacava and right atrium. An end hole N.I.H. 7 or 8cardiac catheter is preferred. The usual purpose ofthis is to obtain a wedged hepatic vein pressure todifferentiate pre- from post-sinusoidal portal hyper-tension. By injecting contrast medium down thecatheter (e.g. 20 ml of 45%/ sodium diatrizoate) awedged hepatic venogram can be obtained (Fig. 1).This examination has been used in cirrhosis where theportal vein often fills well in a retrograde fashionand in space-occupying lesions. However, it is mostvaluable in cases of hepatic vein occlusion (Budd-

copyright. on F

ebruary 20, 2022 by guest. Protected by

http://pmj.bm

j.com/

Postgrad M

ed J: first published as 10.1136/pgmj.46.531.19 on 1 January 1970. D

ownloaded from

http://pmj.bmj.com/

Louis Kreel

FIG. 1. Wedged hepatic venogram in cirrhosis of theliver. The sinusoidal pattern (s) is coarse and there isretrograde flow into the portal vein (pv) to the hilumof the liver. Contrast medium is present around thecatheter (c) lying in an hepatic vein.

Chiari syndrome) or veno-occlusive disease of theliver where specific diagnostic patterns may beproduced, such as a short length stenosis of thehepatic vein with a leash of surrounding collateralvessels (Fig. 2a). The hepatic veins proximal to theblock may be patent and dilated, filling out to theproximal tributaries. In other cases the larger hepaticveins may also be occluded. The appearance is thenthat of the demonstration of collateral hepatic veinsforming a coarse network pattern without contrastmedium being visible around the wedged catheter(Fig. 2b). The network pattern associated with theBudd-Chiari syndrome may also be comprised ofvery fine lines (Fig. 2c) and again no contrast will beseen in the vein around the catheter. Furthermore nocontrast medium penetrates to the portal venous

system. In veno-occlusive disease of the liver asimilar network pattern may occur but the largehepatic veins remain patent and contrast mediumcan be seen around the catheter as it lies in an

hepatic vein.

Inferior vena-cavographyThe inferior vena cava is intimately related to the

posterior aspect of the liver with liver tissue actuallyencircling this structure. It is thus not surprising that

changes in the size and consistency of this organ arefrequently reflected in the appearances produced byvisualization of the inferior vena cava. In posteriorlylying space-occupying lesions the inferior vena cavais displaced, indicating the site of the lesion. Onoccasion this may be the only positive radiologicalexamination. This occurs in medium sized lesionswhich impinge on the posterior surface of the liver,displacing the inferior vena cava forwards (Fig. 3).These posteriorly situated lesions also tend to causeobstruction of the vena cava with the production of acollateral circulation by way of the paraspinal andazygos systems. Large posterior space-occupyinglesions produce most dramatic displacement of thevena cava which acts as an excellent marker indelineating their medial and posterior limits. In theBudd-Chiari syndrome, the generalized enlargementdue to the swollen liver produces marked side-to-sidecompression of the inferior vena cava with very littlechange in its antero-posterior diameter. On the otherhand, the small contracted cirrhotic liver producescircumferential narrowing of this vessel and itsappearances are thus rather different in this condi-tion. The compression caused by the Budd-Chiarisyndrome, however, may become so severe as to leadto complete occlusion of the vena cava with acollateral network of channels being visualized.The demonstration of the inferior vena cava is a

simple procedure requiring only the percutaneouscatheterization of a femoral vein. Forty millilitres of45% sodium diatrizoate is injected by hand and fivefilms at 2-sec intervals are exposed. This should bedone in both frontal and lateral planes.

SplenoportographyPercutaneous splenovenography is the most fre-

quently used vascular examination in liver diseaseand a well established procedure. The details of thetechnique are fully considered in standard texts. Thebasic precautions are to exclude a bleeding tendencyand in no way to interfere with the free respiratoryswing of the splenic needle. All manipulationsassociated with the needle must be done in an apnoeicphase.The main indication for this examination is the

pre-operative demonstration of the portal vein incases of portal hypertension. For the successful per-formance of a portacaval shunt operation not onlymust the portal vein be of adequate calibre but thevenous wall must be normal. The width of the portalvein in cases of portal hypertension is 2-3 cm withno or only slight tapering as it approaches the hilumof the liver. Any narrowing of the portal vein isusually associated with thickening of the wall due toperiphlebitis. In portal hypertension there is anextensive collateral venous circulation with demon-stration of the coronary vein (Fig. 4) and often also

20

copyright. on F


http://pmj.bm

j.com/

Postgrad M


ownloaded from

http://pmj.bmj.com/


FIG. 2. Patterns seen in Budd-Chiari syndrome. (a) Hepatic venogram showing stenosis of anhepatic vein (hv) with dilatation of vein proximally and surrounding collateral veins. (b) Coarsenetwork pattern of collateral veins. (c) Fine network pattern.

the umbilical and mesenteric veins. The collateralvenous circulation in the abdomen can be so exten-sive as to produce a spontaneous 'porto-systemic'shunt with visualization of the inferior vena cava andthis shunt can produce 'spontaneous' porto-systemicencephalopathy. The demonstration of the coronarycollateral vein has recently become of greatersignificance, as this vein rather than the portal veinhas been used for creating a venous shunt to theinferior vena cava for the relief of haematemesis.

Non-visualization of the portal vein may occurunder two quite different circumstances. It may bedue to block of the portal vein or merely to retro-grade blood flow associated with portal hypertension.In both circumstances an extensive collateral venouscirculation is present. Radiologically these conditionsmay be indistinguishable on the splenovenogram,however in occlusion of the portal and splenic veinthere are 'bridging collaterals' which by-pass the siteof obstruction to show the intrahepatic portal

21

copyright. on F


http://pmj.bm

j.com/

Postgrad M


ownloaded from

http://pmj.bmj.com/

22 Louis Kreel

J~~~~~ ~ ~ ~ ~ ~ ~ ~~~~~~~~~~~~~ ..,...,

.:: ............. .. .............::u.:::::.:.II

*ffi~~~~~~~~ ~~ ~ ~~~~~~~~...;.2....i...

S:~~~~~~~~~~~'a''":..0.^:···.. .:.. .:: .: -

.. u0~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.:.........0_ ......................~~~~~~~~~~.......

..f.

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.^..

·:.............-.:S.

·. ::'::::.:::::·:::·.' :. :::·i:·:::·:::~i:··'··:·· .: :i.::i.::i:iiiiiii'F5I·i:l;l:: ::··I·:...... ....

_|...;~: -*B _ _ B U;~i~'::: :·l#iinitli eFidFig . ':;:::'::":::: '':" :''·:'' :::·':··: ······ ··''·''··:"

..;'''';""""""" '" ·;·:··· ..iii~i. ·l:::: I:v;eiliieiiI~a %Iw~i% l~aiiiiilI'~~~:"::''*.." :..:.

FIG. 3. Inferior vena cavogram showing forwarddisplacement (a) and compression of the right lateralmargin (b) of the inferior vena cava indicating a pos-teriorly lying liver mass which was due to actinomy-cosis. This lesion was not visible on splenoportographyor arteriography.

..j .Iq:I :i

ii:i":Ii:if

Ti:iiiiiii.i%i.l.i.5E.,.':' ..:.:.i.i........i.iI..: :i

ii;i:i: ..I

I.

FIG. 4. Spleno-venogram demonstrating a patent splenic(s) and portal vein (pv). There is also filling of thecoronary vein (c) which passes upwards over the lowerdorsal spine (arrow).

*'ii':i Y _'~~~~~~~~~~~~~aii8%1·B

:: i' g . .x

:; ' :.'.'.,.i.b0~~:;·::.

:e ::'........~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~........-.. ... ;,ii'i' ...`:'.-...,.. -.. .. ....X ..:::

11.:: f i:'11_B 919 Sbi. o 9.i.5

:nr: 7 7.9...s i·:'".^.:9in-i . :.:.:8: ........t,,4|t:~~~~~~iiii:·al

ia:Sg|.:.s~~~~~~~~~~~~~~~~~~~~~~~~~~·?81si::..i ···.··::. | | W:||N S~~~~~~~~~~~~~~~~~~~~·:·:i:

FIG. 5. Test injection in the steep supine oblique positionwith the right shoulder raised. The lead skin markersare superimposed on the spine and it can be seen thatthe middle marker is near the level of the origin of thecoeliac axis (c) which has a marked eccentric narrowingof the superior type due to the median arcuate ligamentof the diaphragm.

copyright. on F


http://pmj.bm

j.com/

Postgrad M


ownloaded from

http://pmj.bmj.com/


venous system. When this is of long standing it givesrise to numerous adjacent tortuous veins; the'cavernous transformation' of the portal vein. Ifthere is any doubt as to whether non-visualization ofthe portal vein is due to a true obstruction or toreversal of the blood flow, other procedures must beemployed, such as wedged hepatic venography,mesenteric arterio-venography or umbilical porto-graphy.

Spleno-portography also produces other informa-tion although this is frequently of a less satisfactorynature because it is often either equivocal or bettershown by other procedures. The demonstration ofspace-occupying lesions is such an example. Thesemay be shown by isotope or ultrasonic scanning aswell and often more clearly. Furthermore the natureof the lesion, whether solid or cystic, benign ormalignant, cannot usually be distinguished onspleno-portography. It must be mentioned thathepatomas not infrequently cause intrahepaticthrombosis of the portal vein or filling defectswithin it. Thus a bare area in the liver, associatedwith non-filling of an intrahepatic branch of theportal vein or a filling defect within it, should indi-cate the strong possibility of a hepatoma. Largetumours or cysts in the left lobe of the liver causedownward displacement and may also producenarrowing of the splenic vein, thus indicating the siteof such lesions.The altered pattern of intrahepatic branching in

cirrhosis producing a deficient branching pattern or'tree in winter' appearance is seldom of primeimportance as the diagnosis is essentially made fromhistological examination. However, the exact size ofthe liver can clearly be seen on spleno-portographyand as this is difficult to estimate by other methods,can be of considerable importance.

Umbilical vein portographyThe portal vein can be shown by passing a

catheter along the umbilical vein remnant into theportal vein. The difficulty with this technique is thatnot infrequently the peritoneum is incised and,therefore, this procedure must be done in anoperating theatre. It is, therefore, also unsuitable inthe presence of ascites. However, the portal veinand its intrahepatic branches are densely shown bythis method and it is particularly suitable for thestudy of the intrahepatic portal venous pattern.Following this procedure the umbilical vein becomescompletely occluded and can thus only be used onceon any patient. Portal vein sepsis has also beenreported as a complication. It is particularly usefulfor space-occupying lesions and in showing theportal vein if this has not been demonstrated bysplenovenography.

Hepatic arteriographyAs with other organs, arteriography in liver disease

may be used for the demonstration of vascularlesions as such, lesions within the liver itself especi-ally space-occupying lesions, or for the demonstra-tion of the portal venous system. However, in hepaticarteriography it is absolutely essential to performselective arteriography and not just a free flushaortogram as with the latter method it is impossibleto obtain adequate visualization of either the liveror the portal venous system. A free flush aortogramis occasionally required for the demonstration of theorifices of the coeliac axis and mesenteric arteries,but then this must be done as skyline views in eitherthe steep oblique or lateral position.Furthermore in hepatic arteriography much larger

quantities of contrast medium are required than intoother organs. The catheter must, therefore, be ofadequate size for the quantity of contrast medium tobe injected in a short enough period of time. Thebarium impregnated polyvinyl catheter (O.P.P. 205or 240-Portex) has been found suitable and 40 mlof Conray '420' can be injected into the coeliac axisby hand with sufficient rapidity to obtain a gooddemonstration of the liver and portal vein.

Arteriographic techniqueThe details of the technique have been considered

in detail in previous publications and will, therefore,only be briefly mentioned. Basically the technique issimilar to selective catheterization ofother organs per-formed by the Seldinger method, using a pre-formedopaque catheter. Once the catheter is inserted intothe aorta the patient is turned into the supine obliqueposition with the right shoulder raised. Correctlyplaced opaque skin markers are extremely helpful inlocating the orifices of the coeliac axis and superiormesenteric artery. If, however, there is any difficultyor delay in entering these vessels, a test injectionmust be done in the steep oblique position to showthe position and calibre of these orifices and theirrelationship to the skin markers (Fig. 5). It thenbecomes a simple matter in most cases to direct thetip of the catheter into the appropriate artery.The two common difficulties encountered are

stenosis of the coeliac axis and anatomical variations.If a stenosis of the coeliac axis is encountered andthis occurs in approximately 12% of cases, one canoften still obtain an adequate examination byinjecting the superior mesenteric artery. However, itis advisable to inject a slightly less concentratedcontrast medium into this vessel (e.g. Hypaque 45%)as peripheral pooling of the contrast medium mayoccur in the small bowel with poor visualization ofthe portal vein. To some extent this can be overcomeby the use of 50 mg of tolazoline diluted in 5 ml in

23

copyright. on F


http://pmj.bm

j.com/

Postgrad M


ownloaded from

http://pmj.bmj.com/

Louis Kreel

normal saline, being given in 2 sec into the superiormesenteric artery about 30 sec prior to the contrastmedium. This is said to enhance visualization of theportal vein. In long standing coeliac axis stenosis,the gastroduodenal artery becomes greatly hyper-trophied as it acts as the collateral channel to fill thehepatic and splenic arteries.The other difficulty commonly encountered in

selective catheterization of this region is due to thecommon occurrence (20-25%) of anatomical varia-tions. The commonest variation is that the righthepatic artery arises from the superior mesentericbut the whole hepatic artery may arise in this way.The splenic and left gastric arteries then have aseparate origin from the aorta. Other variationsinclude a completely separate origin of the splenic orthe hepatic artery and a common origin for thecoeliac axis and superior mesenteric artery. Unlessthese variations are borne in mind, much time maybe wasted in searching for a non-existent orifice.Any doubts about the anatomical structure shouldbe immediately resolved by recourse to a test injec-tion and film.Without going into any great detail, a few other

technical factors of importance require mention.The film sequence used in any individual case isdependent on the nature of the clinical problem. Toshow lesions at the vascular orifices at the aorta thevery early films (4/sec for 2 sec) are all important andexposures are begun as the first 5 ml of contrast isinjected by means of a pressure pump (30 lb/in2 in a50-ml syringe). For intrahepatic lesions early filmsare also required, but for somewhat longer (2/sec for4 sec) and thus the first exposure is made as 20-25 mlof contrast medium has been injected by hand. Thelater phase is also important as the portal veinshould be demonstrated (one film every 2 sec fortwelve films). A steep supine oblique view with theleft shoulder raised must also be done for intra-hepatic lesions, but a shorter run is then adequate(2/sec fQr 4 sec and 1/sec for 6 sec). In arterio-portogr4phy the later phase is all important, but onefilm every other second is sufficient, however theexamination should cover a period of 30 sec.The injection of large quantities of contrast

medium into the coeliac or mesenteric artery causesa burning pain which, however, lasts only 15-20 sec.The procedure can thus be done under local anaes-thetic apart from the actual contrast medium injec-tion which requires an ultra-short-acting anaestheticsuch as Epontol.Uses and indications of hepatic arteriographyThe presence of a space-occupying lesion can be

determined by isotope scanning if it is larger than 3cm and similar sized lesions can be shown by ultra-sonic scanning. The advantages of arteriography are

that even smaller lesions can be shown provided thatthey are vascular and that in many cases arterio-graphy will also indicate whether space-occupyinglesions are benign or malignant. The other importantindication is the demonstration of the portal ormesenteric vein as a pre-operative procedure inportal hypertension. Arterio-portography is pre-ferred in cases where spleno-portography is contra-indicated or cannot be done as with an abnormallyprolonged prothrombin time, depressed plateletcount or after splenectomy. Another indication forthis procedure is when the portal vein has not beenshown by spleno-portography. Superior mesentericarterio-venography may then show either cavernoustransformation of the portal vein, indicating portalvein obstruction, or a sufficiently large mesentericvein to perform a shunt and even occasionally suc-ceed in showing a portal vein not shown by spleno-portography. Arterio-portography may also be usedto show the patency of a portacaval shunt (Fig. 6),particularly in cases which have also had splenec-tomy.

Vascular lesions such as stenosis, aneurysms, angio-matous malformations and arterio-venous shunts areparticularly well demonstrated. These lesions areoften asymptomatic chance findings. In liver traumathe site of injury, active haemorrhage and a sub-or peri-capsular collection of blood may be shown.There has even been an example of the demonstra-tion of active haemorrhage into the biliary system.While the selective catheter is in position it is

possible to undertake blood-flow studies by usingradioactive xenon and surface counters. A furtherextension of the use of arteriography has been thecatheter placement for the subsequent administra-

FIG. 6. Venous phase of arterio-venography showing thesplenic (spl) and portal vein (p) draining into the inferiorvena cava (arrows indicate right lateral margin).

24

copyright. on F


http://pmj.bm

j.com/

Postgrad M


ownloaded from

http://pmj.bmj.com/


tion of intra-arterial infusion of cytotoxic agents.With care these catheters can be left in position forup to 3 weeks.Arterio-venography has also been used to show

the patency of a portacaval shunt and wheremalignant lesions have been demonstrated in otherorgans, such as the kidney, it is usually a simplematter to re-position the catheter to determinewhether deposits are present in the liver. It may bementioned that very definite changes may occur inthe coeliac axis and its branches as a consequence ofportal hypertension. Although at present these mayonly be considered as incidental observations, theydo, however, further our understanding of the alteredhaemodynamics in portal hypertension.

Space-occupying lesionsBenign conditionsOnce the presence of a mass lesion in the liver has

been determined its nature may be ascertained byvarious tests including hydatid and amoebic com-plimentation fixation, the presence of foetoglobulinsin hepatomas and by histology from needle biopsy.However, the exact extent of the lesion, its degree ofvascularity and its relationship to the hepatic arteryand portal vein can only be adequately shown byarteriography.

FIG. 7. Large hydatid cyst (hc) occupying the rightlobe of the liver with compression and displacement ofthe left lobe. The lower margin is indicated by arrowheads. The right hepatic artery (arrow) and its branchesare narrowed, displaced and stretched.

Hydatid disease of the liver may produce calcifica-tion particularly of a curvilinear type. On arterio-graphy it produces a 'bare' area over whichnarrowed, curved and stretched vessels can be seento be coursing (Fig. 7). The normal liver substance isdisplaced with compression of the arteries and veinswhich are equally affected. These cysts are usuallybest shown in the intermediate phase when botharteries and veins can be seen.

In the amoebic abscess the 'capsule' of the lesionis highly vascular and multiple small vessels arepresent, but these are regular in calibre and direc-tion. In the one case of actinomycosis of the liverexamined, the lesion was not visible arteriographic-ally or on spleno-venography but was shown byforward displacement of the inferior vena cava.

Congenital cystic disease of the liver producesmultiple small areas of vascular displacement whichmay be difficult to delineate. However, the associ-ated cysts in the kidney may be obvious on pyelo-graphy and on arteriography.The really difficult diagnostic problem is that of

the benign adenoma or fibroma of the liver. Theseare extremely slow growing tumours which mayreach a considerable size and are highly vascular onarteriography. These tumours may be the cause ofhypertension. The extreme vascularity with largefeeding arteries and irregular tortuous branchesresembles the solitary hepatoma. The definitivediagnosis will, of course, be made on histology.The other benign tumour which can also be shown isthe cavernous haemangioma of the liver. This lesion,however, produces characteristic angiographicappearances in that the slightly enlarged arteries arecrowded together and there are varix-like spacesshown on the late phase. These contrast-filled spacesare clearly defined and retain the contrast mediumfor a considerable time, even up to 30 sec after theinitial injection. Small angiomatous malformationsshow as well-defined localized areas of contrastpooling with a well-defined afferent arteriole.

Malignant lesionsThe hepatoma produces a variety of arteriographic

appearances and signs. At the one extreme there isthe highly vascular solitary mass with marked en-largement of the hepatic and branch arteries, in-creased number, size and tortuosity of the arteriolesand the production of small pools of contrastmedium in the mass with irregular areas of increaseddensity in the late phase. The adjacent vessels aredisplaced and compressed and these may also under-go considerable hypertrophy and tortuosity supply-ing the tumour as well. This is particularly so withthe right phrenic artery in posterior tumours in theupper part of the right lobe of the liver. The extreme

copyright. on F


http://pmj.bm

j.com/

Postgrad M


ownloaded from

http://pmj.bmj.com/

26 Louis Kreel

|~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~·6!'8.·.. ,.:.8o

m ;~~~~~~~~~~~~~~~~~~~~~~~~~~.::': ...:.:': .:.:.... ' .

FIG. 8. Primary hepatoma (T) in a case of haemochro-matosis occupying most of the right lobe of the liverwith a central bare area, pathological radiating vesselsat the margin and compression of adjacent arteries.

~~~~~~~~~I E .....wB

8. g :; ° g

: : : ' ' : ::. . :d%~~

:. - .. '' .g

* ' g E~~~~~~j

FIG. 9. Avascular hepatoma in the lower part of theright lobe of the liver displacing normal liver tissueupwards (arrow). T, Tumour; c, coeliac axis; s, splenicartery; rh, right hepatic artery; Ih, left hepatic artery.

vascularity and contrast pooling may be presentonly in the more peripheral areas of the mass withrelative avascularity of its more central part. Thelatter appearance is common with the hepatomasupervening in haemochromatosis (Fig. 8).Another common appearance is that of a marked

increase in vascularity due to re-duplication ortriplication of the branch arteries and a markedincrease in the smaller peripheral arteries of the liver.

This appearance is essentially due to obstruction ofthe intra-hepatic portal vein by the tumour. Thispattern is easily recognizable when it affects a largearea of the liver but can be extremely difficult torecognize when confined to the more peripheralpart of the liver, especially at the top of the right lobe.

Occasionally the tumour may be shown as a fillingdefect in the portal vein, producing asymmetricalnarrowing. With vascular tumours an actual arterio-venous shunt is produced and the portal vein showsdensely in the arterial phase and can be seen to fillretrogradely. The filling defect or portal vein obstruc-tion may, of course, also be shown by spleno-portography.The single primary malignant liver tumour can

also be completely avascular (Fig. 9). This is especi-ally so with the cholangioma. The ultrasonic scan isvery helpful in this type of lesion, as it can distinguishsolid from cystic masses. Arteriographically this typeof lesion frequently produces the non-specificfeatures of vascular displacement and a bare area.In some cases irregular infiltration of arterial wallsor intra-hepatic arterial stenosis may be seen.The hepatoblastoma can produce multiple, rela-

tively discreet areas of hypervascularity with markedtumour staining in the late phase. This tumour maybe hormone secreting and has been associated withprecocious puberty in children.

Hepatic metastases may be extremely large,solitary and vascular and in many respects resemblethe solitary hepatoma (Fig. 10). The vascularity is,however, not usually as marked and the enlargementof the main hepatic artery is only slight. The smaller,or multiple, metastases frequently have a thick, densemargin with non-uptake of contrast in the centre ofthe lesion. The most vascular deposits arise from thecolon, kidney, retro-peritoneal sarcoma and fromthe malignant melanoma. Carcinoid tumours and theinsulinoma also produce vascular deposits, whichmay show as localized nodules of increased densityarising from a feeding arteriole.

Infrequently, the liver lesions may be demon-strated prior to delineation of the primary tumour.If the examination is extended to include films afterselective injection of the renal and superior mesen-teric arteries, the original lesions may then bedemonstrated (Fig. 1 la and b).Vascular lesionsDisease of the coeliac axis and mesenteric arteries

is common but, fortunately, asymptomatic in thevast majority of cases. Unsuspected stenosis of thecoeliac axis occurs in about 12% of cases and is acommon cause of 'failure' of selective catheterizationof this vessel. The usual cause is an atheromatousplaque near the orifice, but may be due to anextrinsic band, especially from the median arcuate

copyright. on F


http://pmj.bm

j.com/

Postgrad M


ownloaded from

http://pmj.bmj.com/

Vascular radiology in liver disease 27

~':~i!~:'~''~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~. .. ":'.· ~~~~~~~~~~4:.:~;....;:'

:'::',::,-:,:g-.:~::.?.. ..··:·::

FIG. 10. Large single vascular metastasis from a coloniccarcinoma producing marked displacement and spread-ing of hepatic arteries.

ligament of the diaphragm. The latter condition hasbeen associated with abdominal 'angina' said to be

caused by the 'steal' phenomenon. The blood supplyto the superior mesenteric is diverted up the gastro-duodenal to supply the hepatic and splenic vessels,producing relative anoxia of the intestine andcausing post-prandial pain.The appearance of this extrinsic lesion is charac-

teristic in that the narrowing is a slight distance fromthe orifice and the coeliac axis is displaced towardsthe aorta at this point, producing an eccentricnarrowing of the superior margin (Fig. 5).

Splenic artery aneurysms are the commonestincidental finding on coeliac axis arteriography.This lesion is particularly common in women overthe age of 50 years, but fortunately rarely leads tosymptoms, being well supported by peri-pancreaticand splenic hilar tissue. These aneurysms commonlycalcify but very rarely rupture. Rupture or leak ofthe aneurysm may be a cause of upper abdominalpain, blood loss and shock. There is a particularpredisposition for this to occur in the last trimester ofpregnancy. A leaking splenic artery aneurysm hasbeen diagnosed arteriographically. The particularinterest in liver disease is that splenic artery aneu-rysms are extremely common in association withlong-standing portal hypertension, irrespective ofthe cause (Fig. 12). Thus in the very large spleens ofportal hypertension, intrasplenic bifurcation aneu-rysms will be found in 40% of cases on selectivearteriography, but less commonly on the majorartery. Again, these aneurysms seldom rupture.

151:'· :1 ' I II s.is.ka. ·I I Iiilii··· .. ·:·i:· i(iii:lilaiiiiii·· '"'`' ..i;iiill.lib·i:ii i8iiiiii:'iii'iiiiilii.i liil:I·:I···::::da ilili8iili. ..:1'8. I'i:FiSiLCiiiiii'ii4iDii::··i:iillil:e:,e,·.:i:·i·:·: . ..i.ZiiiSiilii.iliiiii:::.·· ';·'·':'::':' i'i.ii;'''ij!'lla';;·.:·.:i;:·P·i·:.·"'lir ;l.i ii8f- ·:::-· ··.:::i·::P i.i.:J:ZZd:.·'·::·:·"; R:;::';: ·;:····:·::'::.::::::i:' ilH::l·:·····:·rtl:.CE.Er;I' .SSi 8.iB1-:i:iX.SPI-iiiiiil:iii:;l:;.il"I:·liil ii.iiliiiiii-i:···i· sii.i.i'.:li:l. :i:liiliiiifliliilpl3 lib%!ii8ii n:;aiiii::i: .I:':lv:lsiiiiiaiiYi!liil i?.Lilii.i.iifii::i:;iiaii:riiiiCiiEil "%iiiiii'·lill:I ]i:iii:ljl.i,,i,iiliY I·: iiiiPiFi6iii: :iii:;:i··'·' ilnli.i..-:i;:i:i!i:lii;i:i'li I· iPi.Bii;li'i:: iiiiiiiiil:9i:(·:: :I··iiHiiliilii:iii 'i iiliii·.·.ii,:::ii9.:e.-..'"':4'ii'i!'ii.iii&inis,, .: :il.:l::iiiliii:'.":";": ':"i'::'i.l'll:i:ii.iii' :I:iiiiil:i::ii:::i;iliiEillBi -1··· ,·:.i ·:.i·: ·I:l:i.·: :··il:inaCi.id.iFiitiiiiii.iliiCi·. .· ..;i.ii:ii"9IC8:l.l:hli;:ii.ii.i.i.l.iirii-*il:lil e::iili.ii.iii8::::i8::·:::::i·P ..:.ii·j.iOi:Ei.::.:··· ····:··:·:i·::':·s·:r:: ·.·i:... ·.i:·r·l;.i..·*::·is:.d:l·F* :il·;· ,iiiiliiilii%$9i.

:*1:8::6·1.. .:isi!liBhdE.iEi;i':'?''·83 .:·I:·.8fiii$.iiiiiiili..:::rii..i::.i.l."···: ···· :· ..... n:l·:l··l·B··. ";·:' --i rui: ·hi: ·ii;?·:lii p.·::i::·:i·l&:ii :·ii.iiiil ::·.·-: x·jel· '^·';··I·····.·:·:'·:j;l·······iiill;t.Eif.'..'j':6:":"': ··::Jreeas.:il:._3s..;i:;.*....".." '.I.J.;"..*`. ::.::i:.i.. :·· ··:::ilii.h;l.·'l::i""iff:gi:a:I:: :':: '''-':" iiiiiiiC!:·:·::·: ·':i'i.isiirii·:l:lil:ii.i.ii'.i';:...,ji: .i::i:::::::il·li···i'':'::iii.'·i $i.i:iiliii::·I· I'·iijglti:aii..i%i:;iHi;:il·;i:tii::I:, ::iil-ii:il ::iiaai·BifHilii.iiiiii:iiil.i8ii·:iisiii·.i.i'.ii:..· :.I.I ·;l'iiili::::ii.i:.li I:'''';ii.lij?ii i..:...:::.::· I::!iiaiiiiii·iiiliitii iLrli;a:;ssii·ii::iii:iitiiditii"i.ii::iZl8li:iii:l::::.i: ·i·::··i::.: dl .I:.iij:ji.ij$lj.jiiI;l;i!il;iiiii. ·liiiiiiiii:.:..::. :i:·i::.lillii:iiiliit :8:iRi: ·::;:::::· .·..::i·::s:i·:i2i::::i:lii:iiiiili i·;:..i:'l;i:i.:i·::·.·: ·:·.· ··:·g::-.:*::p::.::.:·: ·:·.i:: :··· '·ii;iiiilaniiilii: I:. ::l"::.:::iiiiill;i ii: li. i:K ii.i::iiiiiiiEiilliliiilii'iil',:.:il·ii·.l .I:.::j:ii:;i:.ii::ij.g: *lib .*l..eftt.ya8:lr ;: ;.:th: *rram:;·i-81 ;rjilsra-PB:"ii-., i"'. ."li""" ;--·-···:--;- i:!;-I; I.,:-:--------.4kB.:.:1·- 1;:

": ::6 'IFZf..ldll.i.ii;.,I::.:::-: ·· ·.....::.lij3f.S;i.j.pi:I::.a;-r..:i.:·:::····r:·::·:%:·..·ii.. :···::···· :·.I· SiUi91CilJislii:i::I·· ·. Ili:iii.:'I'i3:5i:18iliiiiI:;:::::: :i:iii.liii:i:.':·i·"·:· liili :i.JII' ···isi: ::: i'·iLii5i:··· ·iii ..II I.ni ilrii...··:i: ii' iiii:ii..i%l!PiLlti I .:P.:*: :·-di.:e·· ·::i:·::IPii.:.:.·::ir-

i..;e:·:.... .·":''''"'·..·:i.lii:p.ilajl r:·.. ·:Isi·s:e i:ii.ii:"' ·I· ::i.iiJ··..··.::?::;:::.i l.ltil.a..l..i.:i· ,i:lii.iiiil iiiliiilii.iiSii:iiil··i·lr

i?.·:riiiYiiiiiiiili$:;i:' ; : ·li,:il.iij":;ii;iiiiiii:·.·- i: ::::

di,r.unaps.: ···i:9·.·ii.:i.0..t·iI a.ii. gl.·18u.*h.pe.as.-*;:::.'.: i:!:i

i:ii.:i.,..-i. .t.rp .l'.gill.E.ZRSii .I,li.nskIl...a.T.ia &'I

C1- ".:.S.rra..· .4.'. ::" -.Y I .S.iiFIG. 11. (a) Multiple metastases in the liver in both the right and left lobes. The central areas are re-latively avascular with a contrast blush at the periphery showing as a thick rim of tissue. (b) Latephase of selective right renal arteriogram showing a hypernephroma with appearances similar to thesecondary deposits.

copyright. on F


http://pmj.bm

j.com/

Postgrad M


ownloaded from

http://pmj.bmj.com/

28 Louis Kreel

·i·: .~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~:'::::..!:::i!!'..:::~~~~~~~~~~~~~~~~~~~~~~~~~~:i'i~:.i;ii;·I:';i;i:··:lJ··a,~~~~~~~~~~~~~~~~~~i... . ..i~;~i~,,;.~~I·...*,?*i............~lL::::ii~E;:2:::::::..:::lil~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~:·:·I:·ii:·;:··~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~?iii~.

··;~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~:::i·"~~~~~~~~~~~~~~~ii.'.:~;i~~,ii. i~!:·il;·;::r' ?:.~i i:~

FIG. 12. There is a large aneurysm on the main splenic artery (s) and also small aneurysmson the branches at the hilum (h). Splenic artery aneurysms are commonly found in patientswith portal hypertension (as in this case) but are usually intrasplenic and at bifurcations.c, Coeliac axis.

tl11

_ 3.::,:~~ak:l et~~:e l~ lE-·:~:

SIISE~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ii:ji _ | 11| I 1|_ E:tIi_~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~·:_l~~~~~~~~~~"':': :'i:::'I:"'·I ::1::· ··'·· .gA_ isri|:.~~ ~ ~ iiaBI:illii·:,·i:i·;

i~iUli;t~~~~~~~~~~~~~~~~~~~~::: gidli:ii:liii:i:i :: .· ··-

E~~~~~~~~~~~~~~~~~~~~~~ggiiHailil·:.....:C:.E

&.X;...:.9~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~iii~j~iRM:,=,.°. ...... .R . . E J:-:~iiir::-i-

FIG. 13. Large aneurysm (a)ofthe hepatic artery which fills from the gastro-duodenal afterinjection into the superior mesenteric (sm) and then filling of the splenic artery with a smallaneurysm (arrow). There was marked atheromatous stenosis of the coeliac axis.

copyright. on F


http://pmj.bm

j.com/

Postgrad M


ownloaded from

http://pmj.bmj.com/


The main branch aneurysms may calcify, producingone or more curvilinear or egg-shell type calcifica-tions in the left hypochondrium.

Hepatic artery (Fig. 13) and portal vein aneurysmsalso occur. These arteries are less well supported bysurrounding adventitial tissue and have beenassociated with rupture. Aneurysms and stenosismay occur in association, particularly in the elderly.A rare but important cause of portal hypertension

is the arterio-venous fistula, most commonlyoccurring between splenic artery and vein. Thislesion is well demonstrated on arteriography andwhen present offers the surgeon an opportunity ofproducing a complete cure of the portal hypertension.

Arterio-venography in portal hypertensionAs was indicated previously, the pre-operative

demonstration of the portal vein can be achieved bythis method and the indications for using this pro-cedure rather than spleno-portography have alsobeen mentioned. The portal and mesenteric veins arenever so densely shown as by spleno-portographyand the collateral circulation is often difficult to see.However, these features can be enhanced by using asubtraction technique which increases visualizationin the areas overlying the spine. As with spleno-portography, a normal portal vein, a narrowedportal vein or cavernous transformation of the portalvein can be demonstrated.

In assessing the portal venous system as demon-strated on arterio-portography, it must be remem-bered that selective injection of the coeliac axisnormally demonstrates the gastric and coronaryveins, while demonstration of the superior mesen-teric veins is usual by superior mesenteric arterio-graphy. Unlike spleno-portography, the demonstra-tion of these veins is thus not an indication of portalhypertension or a collateral circulation. For theradiographic diagnosis of a collateral circulation,veins must be shown which are not draining the areasupplied by the injected artery, such as the inferiormesenteric vein in both coeliac and superior mesen-teric arteriography or the mesenteric vein afterselective injection of the splenic artery, indicatingan abnormal direction of blood flow. Collateralvessels themselves are enlarged and often tortuous.The arterial pattern is also visualized on arterio-

venography and this shows definite changes associ-ated with portal hypertension. There is a strikingenlargement of the coeliac axis and splenic arteryand the splenic artery often becomes extremelytortuous, but the hepatic artery does not enlargeunless some other lesion such as a hepatoma ispresent. As has been mentioned earlier, splenicartery aneurysms are also common under these cir-cumstances and are probably the result of theprolonged increase in blood flow through the spleen.

In some of these cases there is a drop in the bloodpressure in the coeliac axis even though no stenosisis present. This is probably due to the marked bloodflow in this artery. By means of radioactive xenonstudies done while a selective catheter was in thecoeliac axis, it was possible to show that the increasein total blood flow through the spleen was of theorder of three to five times greater than normal.The hepatic arterial circulation appears to behave

rather differently. In some cases of portal hyper-tension not only is there no concomitant enlarge-ment of the hepatic artery as occurs in the coeliacand splenic, but there is a marked diminution in thecontrast medium flow into the liver and down thegastro-duodenal artery, with poor visualization of

FIG. 14. (a) Coeliac axis (c) arteriography in portalhypertension prior to a portacaval shunt showing con-trast medium up to the peripheral splenic (s) and phrenicvessels but only slight visualization of the hepatic artery(arrow). (b) Following a portacaval shunt there is promptfilling to the periphery of the liver as well as the spleenwhich has diminished in size. h, Hilum.

29

copyright. on F


http://pmj.bm

j.com/

Postgrad M


ownloaded from

http://pmj.bmj.com/

30 Louis Kreel

the intrahepatic vessels (Fig. 14). In some of thesecases it will be found that selective contrast injectioninto the superior mesenteric produces better fillingof the hepatic artery than direct injection into thecoeliac axis. There may even be filling of the splenicartery without a coeliac axis stenosis, indicating areversed flow or possible 'steal' effect through thesplenic circulation.These changes in the hepatic arterial circulation

can be shown to be functional by comparing theappearances before and after portacaval shunt.Following a successful portacaval shunt there is amarkedly increased visualization of the hepaticcirculation with a free and rapid flow of contrastmedium to the peripheral hepatic vessels. Wherexenon studies have been done after a successfulportacaval shunt operation, it was found that thereis no decrease in total blood flow through the spleeneven though the spleen had decreased in size,indicating a fall in the intrasplenic peripheral resis-tance. This is in keeping with an earlier observationthat the enlarged coeliac and splenic arteries ofportal hypertension do not decrease in size followinga successful shunt operation. It has also been notedthat the diminished contrast visualization of theliver in portal hypertension does not occur in extra-hepatic portal vein obstruction or in the presence ofjaundice.The most reasonable hypothesis at the moment to

account for these phenomena appears to be thatthere is a veno-arterial reflex between the portal veinand the hepatic artery. Increased portal pressureacting on a normal venous wall initiates the reflexwhich causes constriction of the peripheral hepaticarteries. This reflex does not operate if the portal veinis thrombosed or thickened by periphlebitis and isobviated in the presence ofjaundice. This reflex mayalso result in the diversion of blood flow to thesplenic artery (rather than the splenic circulationbeing a cause of the 'steal' effect). Of course, if theconstrictive hepatic arterial effects are irreversible,this would result in almost complete deprivation ofliver blood supply if a portacaval shunt operation iscarried out.

BibliographyTextbooksMAINGOT, R. (1969) Abdominal Operations, 5th edn. Apple-

ton-Century-Crofts, New York.NEBESAR, R.A., KORNBITH, P.L., POLLARD, J.J. & MICHELS,N.A. (1969) Celiac and Superior Mesenteric Arteries-A Correlation of Angiograms and Dissections. Churchill,London.

SCHOBINGER, R.A. & RUZICKA, F.F. (1964) Vascular Roent-genology. Macmillan, New York.

SHANKS, S.C. & KERLEY, P. (1970) A Textbook of X-rayDiagnosis, 5th edn. H. K. Lewis, London.

SHERLOCK, S. (1968) Diseases of the Liver and Biliary System,4th edn. Blackwell Scientific Publications, Oxford andEdinburgh.

SMITH, R. (1969) Progress in Clinical Surgery, Series IlI.Churchill, London.

Hepatic venographyCLAIN, D., FRESTON, J., KREEL, L. & SHERLOCK, S. (1967)

Clinical diagnosis of the Budd-Chiari syndrome. AmericanJournal of Medicine, 43, 544.

KREEL, L., FRESTON, J.W. & CLAIN, D. (1967) Vascular radio-logy in the Budd-Chiari syndrome. British Journal ofRadiology, 40, 755.

Inferior vena-cavographyCONTO, D., JR & Rocco, R.P. (1962) Inferior cavography in

chronic fibrosing liver disease. Hospital, 61, 107.FARINAS, P.L. (1947) Abdominal venography. American

Journal of Roentgenology, Radium Therapy, and NuclearMedicine, 58, 599.

FUCHS, W.A. (1961) The diagnostic value of cavography.Radiologica clinica, 30, 129.

Umbilical vein and mesenteric portographyBAYLY, J.H. (1964) Use of umbilical vein in the diagnosis of

upper gastro-intestinal bleeding. American Journal ofGastroenterology, 41, 235.

BAYLY, J.H. & CARBALHAES, O.G. (1964) Umbilical vein inadult: diagnosis, treatment and research. American Sur-geon, 30, 56.

CARBALHAES, O.G. (1959) Hepatoportografia por viaumbilical. Revista (le sanidad militar, 12, 42.

Spleno-portographyLEGER, L. (1967) Spleno-portography. Thomas, Springfield,

Illinois.ROUSSELOT, L.M., RUZICKA, F.F., JR & DOEHNER, G.A.

(1956) Portography in portal hypertension; its applicationin diagnosis and surgical planning. Surgical Clinics of NorthAmerica, 36, 361.

STEINER, R.E., SHERLOCK, S. & TURNER, M.D. (1957) Per-cutaneous splenic portal venography. Journal of the Facultyof Radiologists, 8, 158.

Arteriography in liver disease(a) Technique

EVANS, J.A. (1965) Techniques in the detection and diagnosisof malignant lesions of the liver, spleen and pancreas.Radiologic Clinics of North America, 3, 567.

KREEL, L. & WILLIAMS, R. (1964) Arteriovenography of theportal system. British Medical Journal, 2, 1500.

ODMAN, P. (1956) Percutaneous selective angiography of themain branches of the aorta. Acta radiologica, 45, 1.

ODMAN, P. (1958) Percutaneous selective angiography of thecoeliac artery. Acta radiologica, Supplement 159.

REDMAN, H.C. & REUTER, S.R. (1969) Angiographic demon-stration of portacaval and other decompression livershunts. Radiology, 92, 788.

SELDINGER, S.I. (1953) Catheter replacement of needle inpercutaneous arteriography: new technique. Acta radio-logica, 39, 368.

(b) Space-occupying lesionsABRAMS, R.M., BERANBAUM, E.R., SANTOS, J.S. & LIPSON, J.

(1969) Angiographic features of cavernous haemangiomaof the liver. Radiology, 92, 308.

BAUM, S., ROY, R., FINKELSTEIN, A.K. & BLAKEMORE, W.S.(1965) Clinical application of selective coeliac and superiormesenteric arteriography. Radiology, 84, 279.

BERDON, W.E. & BAKER, D.H. (1969) Giant haemangiomawith cardiac failure in the newborn infant: value of highdosage intravenous urography and umbilical angiography.Radiology, 92, 1523.

copyright. on F


http://pmj.bm

j.com/

Postgrad M


ownloaded from

http://pmj.bmj.com/

Vascular radiology in liver disease 31

EVANS, J.A. (1965) Techniques in the detection and diagnosisof malignant lesions of the liver, spleen and pancreas.Radiologic Clinics of North America, 3, 567.

FREDENS, M. (1969) Angiography in primary hepatic tumoursin children. Acta radiologica, 8, 193.

KREEL, L. (1969) Radiology in liver disease. British Journalof Clinical Practice, 23, 187.

KREEL, L., JONES, E.A. & TAVILL, A. (1968) A comparativestudy of arteriography and scintillation scanning in space-occupying lesions of the liver. British Journal of Radiology,41, 401.

MARGULIS, A.R., NICE, C.M. & RIGLER, L.G. (1956) Roent-gen findings in primary hepatoma in infants and children:analysis of 11 cases. Radiology, 66, 809.

MILANES, B., MCCOOK, J. & HERNANDEZ, A.L. (1953) Aorto-graphy and tumours of liver; preliminary report. Angi-ology, 4, 312.

SAMMONS, B.P., NEAL, M.P., ARMSTRONG. R.H.,JR & HAGER,H.G. (1967) Ten years experience with celicac and upperabdominal superior mesenteric arteriography. AmericanJournal of Roentgenology, Radium Therapy, and NuclearMedicine, 101, 345.

(c) Vascular and other lesionsBAUM, S., NUSBAUM, M., CLEARFIELD, H.R., KURODA, K. &TUMEN, H.J. (1967) Angiography in the diagnosis of gastro-intestinal bleeding. Archives of Internal Medicine, 119, 4.BOIJSEN, E., JUDKINS, M.P. & SIMAY, A. (1966) Angio-graphic diagnosis of hepatic rupture. Radiology, 86, 66.

BRON, K.M. & REDMAN, H.C. (1969) Splanchnic arterystenosis and occlusion (incidence, arteriographic andclinical manifestations). Radiology, 92, 323.

DEUTSCH, V. (1968) Compression of the coeliac trunk and theangiographic evaluation of its haemodynamic significance.Clinical Radiology, 19, 309.

HARJOLA, P.T. & LAHTIHARJU, A. (1968) Coeliac axis syn-diome. American Journal of Surgery, 115, 864.

(d) Portal hypertensionBOIJSEN, F., EKMAN, C.A. & OLIN, T. (1963) Coeliac and

supe ior mesenteric angiography in portal hypertension.Acta chirurgica Scandinavica, 126, 315.

GITLIN, N., GRAHAME, G., KREEL, L. & SHERLOCK, S. (1968)Arteriography, xenon blood flow and pressure studies inportal hypertension before and after portacaval shunt.Proceedings of the llird Meeting European Association forthe Study of the Liver, Modena, p. 22.

KREEL, L. (1967) The recognition and incidence of splenicartery aneurysms. M.D. Thesis, University of the Wit-watersrand.

KREEL, L., GITLIN, N. & SHERLOCK, S. (1969) Hepatic arteryangiography in portal hypertension. American Journal ofMedicine. (In press.)

RUZICKA, F.F. & Rossi, P. (1969) Arterial portography:patterns of venous flow. Radiology, 92, 777.

WILLIAMS, R., CONDON, R.E., WILLIAMS, H.S., BLENDIS, L.M.& KREEL, L. (1968) Splenic blood flow in cirrhosis andportal hypertension. Clinical Science, 34, 441.

copyright. on F


http://pmj.bm

j.com/

Postgrad M


ownloaded from

http://pmj.bmj.com/

vascular radiology kreel - bmj

Documents