Get Chapter1 1

8/13/2019 Get Chapter1 1

1/47

Harald Becher .Peter N Burns

Handbook of Contrast EchocardiographyLeft ventricular function and myocardial perfusion

Peter N BurnsProfessor of Medical Biophysics and RadiologyUniversity of TorontoImaging ResearchSunnybrook and Womens Health Science Centre2075 Bayview AvenueToronto, OntarioCanada M4N [email protected]

Harald Becher

Professor of CardiologyUniversity of BonnRheinische Friedrich-Wilhelms-UniversittMedizinische Universittsklink und Poliklink IIKardiologie/PneumologieSigmund-Freud-Strae 2553105 [email protected]

Copyright 2000 by Harald Becher and Peter N Burns. This book is protected by copyright. All rights are reserved, whether the wholepart of material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproductionon microfilm.

This electronic copy was downloaded under the conditions of the End User License Agreement which accompanies it. Use and storageof this document signifies agreement with the terms of the Agreement. The files may not be altered without prior written permission ofthe copyright owners. No text, figures, tables or images may be displayed or reproduced, except for personal use, in any form or by anymeans, electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the copyright owners.

The use of general descriptive names, registered names, t rademarks, etc. in this publication does not imply, even in the absence of a specificstatement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use.


2/47


3/47

1 Contrast Agents for Echocardiography:Principles and Instrumentation

3

Shall I refuse my dinner because I do not fully

understand the process of digestion?

Ol iver Heaviside, 18501925


4/47

Introduction

Contrast agents for ultrasound are unique inthat they interact with, and form part of, the

imaging process. Contrast imaging cannot beperformed effectively without a basic under-standing of this interaction and how it isexploited by the new imaging modes that havebecome available on modern ultrasoundsystems. In this chapter we consider how echo-cardiography might benefit from a contrastagent, describe currently available agents andexplain their mode of action. We discuss theimpact of ultrasound contrast on echocardio-graphic techniques and instrumentation andconclude with the most recent developmentsin this rapidly evolving field.

1.1 The need for contrast agents in

echocardiography

1.1.1 B-mode imaging

It is well known that blood appears blackon ultrasound imaging. This is not becauseblood produces no echo, it is simply that thesound scattered by red blood cells at the lowdiagnostic frequencies is very weak, about

1,00010,000 times weaker than that fromsolid tissue, so lies below the displayed dynamicrange of the image. Amongst the roles that theimage plays in the ultrasound examination ofthe heart is the identification of boundaries,especially of those between the blood and thewall of the cavity. Identification of the entiremargin of the endocardium in a view of the leftventricle, for example, is an important compo-

nent of any wall motion study. Although in

some patients this boundary is seen clearly, inmany the endocardial border is poorly definedbecause of the presence of spurious echoeswithin the cavity. These echoes, which are

frequently a result of reverberation of the ultra-sound beam between the transducer and thechest wall and aberration of the beam in itspath between the ribs, result in a reductionof useful contrast between the wall and thecavity the blurred haze that is familiar tomany ultrasonographers. By enhancing the echofrom blood in these patients by using a contrastagent, the blood in the cavity can be renderedvisible above these artifacts. Because it is morehomogeneous than the wall and because it isflowing, this echo does not suffer from thesame artifacts and a clear boundary is seen,revealing the border of the endocardium(Figure 1).

If the echo from blood is enhanced by a con-trast agent, the signals obtained from a duplexDoppler examination of the vessel will be

similarly enhanced (Figure 2). A further rolefor such an agent immediately suggests itself. Ifthe echo from large blood vessels can be

2 Handbook of Contrast Echocardiography

Fig. 1 Contrast enhanced harmonic image of the

left ventricle shows the endocardial border clearly.


5/47

enhanced by an agent, what effect will it haveon the small volume of blood that is in themicrovessels of the myocardium? The muscleitself appears bright on an ultrasound image, sowe can expect the additional brightness due tothe agent in the myocardial vessels to be very

small, yet if it is detectable, it would open thepossibility of using ultrasound to map therelative perfusion volume of blood in themuscle itself. As we shall see, it is normally notpossible to detect this tiny echo with existingtechniques, but with the aid of new methodssuch as harmonic and pulse inversion imagingand Doppler, myocardial perfusion volumeimaging becomes possible with ultrasound.

1.1.2 Doppler

Doppler forms an essential part of all echo-cardiography examinations. It is used both to

detect large volumes of blood moving slowly,such as in the cavities, and small volumes ofblood moving fast, such as in stenotic valvular

jets. It is also used in many vascular beds tovisualise the flow at the parenchymal level ofthe circulation, where blood vessels lie belowthe resolvable limit of the image. The detectionof such unresolved flow using Dopplersystems can be demonstrated simply by using aduplex scanner to create a power Dopplerimage of an abdominal organ such as thekidney, in which vessels that are not seen onthe greyscale image become visible using theDoppler mode. These vessels are the arcuateand interlobular branches of the renal artery,whose diameter is known to be less than100 m and therefore below the resolutionlimit of the image. However, as we progressdistally in the arterial system, the blood flows

more slowly as the rate of bifurcation increases,producing lower Doppler shift frequencies.The quantity of blood in a given volume oftissue also decreases, weakening the back-scattered echo. Eventually, a point is reached atwhich the vessel cannot be visualised and theDoppler signals cannot be detected. The myo-cardial perfusion bed lies beyond this point.

3Principles and Instrumentation

Conditions for successful Dopplerdetection of blood flow

Velocity of flow must be sufficient to give

detectable Doppler shift

Strength of blood echo must be sufficient for

detection

Tissue motion must be sufficiently slow that its

Doppler shift may be distinguished from that of

blood flow

Fig. 2 Enhancement of Doppler signals by amicrobubble contrast agent. The spectral display

shows an increase in intensity with the arrival ofthe agent in an artery following intravenous injec-tion.

Contrast in echocardiography: why?

To enhance Doppler flow signals from the

cavities and great vessels

To delineate the endocardium by cavity

opacification

To image perfusion in the myocardium


6/47

1.1.3 Doppler examination of smallvessels

To understand how contrast agents may help

to image myocardial perfusion we need toexamine these performance limits for theDoppler detection of blood flow. Two condi-tions must be satisfied before a Doppler signalcan be detected: first, the velocity of bloodmust be sufficient to produce a Doppler shiftfrequency that is distinguishable from that pro-duced by the normal motion of tissue, andsecond, the strength of the echo must providea sufficient signal at the transducer to allowdetection above the acoustic and electricalnoise of the system. For fast moving blood,such as that in a stenotic jet, it is the secondcondition that determines if a Doppler inter-rogation in spectral or colour modes fails toobtain a signal. If the echo is too weak, theexamination fails. The role of the contrastagent is to enhance the blood echo, therebyincreasing the signal-to-noise ratio and hence

the success rate of a Doppler examination.Figure 3 shows how with the addition of a con-trast agent, a colour Doppler system can showthe apical coronary vessels from a transthoracicview.

For small blood vessels, on the other hand, thediagnostic objective in using an ultrasoundcontrast agent is to detect flow in the circula-

tion at a level that is lower than would other-wise be possible in Doppler or greyscale modes.In cardiac applications the target small vesselsare those supplying the myocardium. Theechoes from blood associated with such flow at the arteriolar level for example exist in themidst of echoes from the surrounding solidstructures of the organ parenchyma, echoeswhich are almost always stronger than even the

contrast-enhanced blood echo. Thus, in order

to be able to image flow in the myocardium, acontrast agent is required that either increasesthe blood echo to a level that is substantiallyhigher than that of the surrounding tissue, or amethod must be conceived for suppressing theecho from non-contrast-bearing structures.

Although Doppler is an effective method forseparating the echoes from blood and tissue, itrelies on the relatively high velocity of theflowing blood compared to that of, say, thecardiac wall. Although this distinction whichallows us to use a highpass (or wall) filter toseparate the Doppler signals due to bloodflowfrom those due to the wall itself is valid forflow in large vessels or across the cardiac valves,

it is not valid for the myocardium, where themuscle is moving much faster than the bloodwhich flows in its vasculature. Thus theDoppler shift frequency from the movingmuscle is comparable to or higher than that ofthe moving blood itself. Because the wall filtercannot be used without eliminating both theflow and the muscle echoes, the use of Dopplerin such circumstances is defeated by the over-

whelming signal from tissue movement: the


Fig. 3 Contrast enhanced Doppler examiniationof the apical coronary vessels.


7/47


flash artifact in colour or the thump artifactin spectral Doppler. Thus myocardial flow can-not be imaged with conventional Doppler,with or without intravenous contrast agents. Anew Doppler modality is necessary: as we shallsee, harmonic and pulse inversion imaging

provide this (Figure 4).

1.2 Contrast agents forultrasound

The principal requirements for an ultrasound

contrast agent are that it should be easily intro-ducible into the vascular system, be stable forthe duration of the diagnostic examination,have low toxicity and modify one or moreacoustic properties of tissues which determinethe ultrasound imaging process. Although it isconceivable that applications may be found forultrasound contrast agents which will justifytheir injection directly into arteries, the clinical

context for contrast ultrasonography requires

that they be capable of intravenous administra-tion. As we shall see, these constitute ademanding specification for a drug, one thatonly recently has been met. The technology

universally adopted is that of encapsulatedbubbles of gas which are smaller than redblood cells and therefore capable of circulatingfreely in the body. These are the so-calledblood pool agents. Agents have also beenconceived that are taken up by a chosen organsystem or site, as are many nuclear medicinematerials.

1.2.1 Contrast agent types

Contrast agents might act by their presence inthe vascular system, from where they are ulti-mately metabolised (blood pool agents) or bytheir selective uptake in tissue after a vascularphase. Of the properties of tissue that influence

the ultrasound image, the most important arebackscatter coefficient, attenuation and acous-tic propagation velocity (1). Most agents seekto enhance the echo by increasing the back-scatter of the tissue that bears them as much aspossible, while increasing the attenuation inthe tissue as little as possible, thus enhancing

the echo from blood.

An ideal ultrasound contrast agent

Non-toxic

Intravenously injectable, by bolus or infusion

Stable during cardiac and pulmonary passage

Remains within the blood pool or has a well-

specified tissue distribution

Duration of effect comparable to that of the

imaging examination

Fig. 4 Myocardial perfusion imaged using inter-mittent harmonic power Doppler and Levovist.


8/47

1.2.1.1 Blood pool agentsFree gas bubbles

Gramiak and Shah first used injected bubblesto enhance the blood pool echo in 1968 (2).They injected saline into the ascending aorta

during echocardiographic recording and notedstrong echoes within the normally echo freelumen of the aorta and the chambers of theheart. Subsequent work showed that thesereflections were the result of free bubbles of airwhich came out of solution either by agitationor by cavitation during the injection itself. Inthis early work, many other fluids were foundto produce a contrast effect when similarly

injected (3, 4). The intensity of the echoes pro-duced varied with the type of solution used:the more viscous the solution, the more micro-bubbles were trapped in a bolus for a sufficientlength of time to be appreciated on the image.Agitated solutions of compounds such as indo-cyanine green and renografin were also used.

Most of the subsequent research into the appli-cation of these bubbles as ultrasound contrastagents focused on the heart, including evalua-tion of valvular insufficiency (5, 6), intracardiac

shunts (7) and cavity dimensions (8). Thefundamental limitation of bubbles produced inthis way is that they are large, so that they areeffectively filtered by the lungs, and unstable,so that they go back into solution within asecond or so. Hence this procedure was invasiveand, except by direct injection, unsuitable forimaging of left-sided cardiac chambers, thecoronary circulation and the systemic arterialtree and its organs.

Encapsulated air bubbles

To overcome the natural instability of free gasbubbles, attempts were made to encapsulategas within a shell so as to create a more stableparticle. In 1980 Carroll et al (9) encapsulatednitrogen bubbles in gelatin and injected theminto the femoral artery of rabbits with VX2tumours in the thigh. Ultrasound enhance-

ment of the tumour rim was identified.However, the large size of the particles (80 m)precluded administration by an intravenousroute. The challenge to produce a stable encap-sulated microbubble of a comparable size tothat of a red blood cell and which could survivepassage through the heart and the pulmonarycapillary network was first met by Feinstein etal in 1984 (10), who produced microbubbles

by sonication of a solution of human serumalbumin and showed that it could be visualisedin the left heart after a peripheral venous injec-tion. This agent was subsequently developedcommercially as Albunex (MallinckrodtMedical Inc, St Louis, MO).

A burgeoning number of manufacturers havesince produced forms of stabilised micro-

bubbles that are currently being assessed for


Fig. 5 The principle of Levovist. Air adheres tothe surface of galactose microparticles which 'size'the resulting bubbles to have a median diameterof about 4 m. Upon dissolution, the bubbles arecoated with a thin, permeable shell comprising

palmitic acid.


9/47

use as intravenous contrast agents for ultra-sound. Several have passed through Phase 3clinical trials and gained regulatory approval inEurope, North America and, more recently,

Japan. Levovist

(Schering AG, Berlin,Germany), is a dry mixture comprising 99.9percent microcrystalline galactose micro-particles, and 0.1 percent palmitic acid. Upondissolution and agitation in sterile water, thegalactose disaggregates into microparticleswhich provide an irregular surface for theadherence of microbubbles 3 to 4 m in size(Figure 5). Stabilisation of the microbubblestakes place as they become coated with palmiticacid, which separates the gas liquid interfaceand slows their dissolution (11). These micro-bubbles are highly echogenic and are suf-ficiently stable for transit through the pulmon-ary circuit. The estimated median particle sizeis 1.8 m and the median bubble diameterapproximately 2 m with the 97th centileapproximately 6 m (12). The agent is chemi-cally related to its predecessor Echovist

(Schering AG, Berlin, Germany) a galactoseagent that forms larger bubbles and which hasbeen used extensively without suggestion oftoxicity. Both preclinical (13) and clinical (14)studies with Levovist demonstrate its capacityto traverse the pulmonary bed in sufficientconcentrations to enhance both colourDoppler and, in some instances, the B-modeimage itself.

Low solubility gas bubbles

The shells which stabilise the microbubblesare extremely thin and allow a gas such as air todiffuse out and go back into solution in theblood. How fast this happens depends on anumber of factors which have been seen tovary not only from agent to agent, but frompatient to patient. After venous introduction,

however, the effective duration of the two

agents described above is of the order of a fewminutes. Because they are introduced as abolus and the maximum effect of the agent isin the first pass, the useful imaging time is

usually considerably less than this. Newer(sometimes referred to as second generation)agents, designed both to increase backscatterenhancement further and to last longer in thebloodstream are currently under developmentand early clinical use. Instead of air, many ofthese take advantage of low solubility gasessuch as perfluorocarbons, the consequent lowerdiffusion rate increasing the longevity of theagent in the blood. However, a price may bepaid for this stability in the reduced acousticresponsiveness of the agent (see 1.3.4.2).Optison (Mallinckrodt Medical Inc, St Louis,MO) is a perfluoropropane filled albumin shellwith a size distribution similar to that ofAlbunex (Figure 6). The stability of the smallerbubbles in its population is the probable causeof the greater enhancement observed with thisagent. Echogen (Sonus Inc, Bothell WA) is an

emulsion of dodecafluoropentane dropletswhich undergo a phase change, boiling tocreate gas which is stabilised into bubbles by anaccompanying surfactant. In practice, this


Fig. 6 Optison microbubbles photographed in

vitro with red blood cells (Courtesy Mallinckrodt Inc).


10/47

agent requires external preparation of the mix-ture (eg, the popping of a syringe by forcedwithdrawing of the plunger against its closedoutlet) to create the bubble population beforeinjection. SonoVue (Bracco Inc, NJ) usessulphur hexafluorane in a phosphorlipid shell.Definity (also known as DMP115, DuPontInc, Boston MA) comprises a perfluoropropanemicrobubble coated with a particularly flexiblebilipid shell which also shows improved stabili-

ty and high enhancements at low doses (15)(Figure 7). Other agents are under aggressivedevelopment (see Table).

1.2.1.2 Selective uptake agentsA perfect blood pool agent displays the sameflow dynamics as blood itself, and is ultimatelymetabolised from the blood pool. Agents canbe made, however, that are capable of provid-

ing ultrasound contrast during their meta-

bolism as well as while in the blood pool.Colloidal suspensions of liquids such as per-fluorocarbons (16) and certain agents withdurable shells (17) are taken up by the reticulo-

endothelial system from where they ultimatelyare excreted. There they may provide contrastfrom within the liver parenchyma, demarkingthe distribution of Kupffer cells (18). In thefuture, such agents with a cell-specific pathwaymay be used as a means both to detect and todeliver therapeutic agents to a specific site inthe cardiovascular system.

1.2.2 Using a contrast agent

There is no question that the majority of dif-ficulties that occur when a contrast agent isfirst used in a clinical echo laboratory can beattributed to problems with the preparationand administration of the injected materialitself. Many physicians and nurses used toadministering pharmaceutical drugs, includingx-ray contrast, cannot understand why the pre-paration and injection itself are so critical for


Fig. 7 Definity microbubbles seen flowing in ablood vessel. Intravital microscopy shows a capil-lary following intravenous administration of fluores-cent-labeled agent. Two microbubbles traverse thevessel with adjacent red blood cells (arrows) seenfaintly under the concomitant low-level trans-illumination.Courtesy Jonathan Lindner, University of Virginia, VA,

USA

Characteristics

Could not traverse

cardiopulmonary beds

Successful trans-

pulmonary passage

Improved stability

Controlled acoustic

properties

Formulation

Free gas bubbles

Encapsulated air

bubbles

Encapsulated low

solubility gas

bubbles

Particulate

(eg polymer shell)

gas bubbles

Generation

0

1

2

3

The evolution of contrast agentsfor ultrasound


11/47


Some ultrasound contrast agents

Manufacturer Name Type Development Stage

Acusphere AI-700 Polymer/perflourocarbon Early Development

Alliance/Schering Imavist Surfactant/ Clinical Development

perfluorohexane-air

Bracco SonoVue Phospholipid/ Late Clinical Development

Sulphur hexafluoride

Byk-Gulden BY963 Lipid/air Not commercially developed

Cavcon Filmix Lipid/air Pre-clinical Development

DuPont Definity Liposome/perfluoropropane Late Clinical Development

Molecular Optison Cross-linked human serum Approved in EU, US,

Biosystems/ albumin/perfluoropropane Canada

Mallinckrodt

Molecular Albunex Sonicated albumin/air Approved in EU, US

Biosystems/

Mallinckrodt

Nycomed Sonazoid Lipid/perfluorocarbon Late Clinical Development

Point Biomedical Bisphere Polymer bilayer/air Clinical Development

Porter PESDA Sonicated dextrose albumin/ Not commercially

perfluorocarbon developed

Quadrant Quantison Spray-dried albumin/air Pre-clinical Development

Schering Echovist Galactose matrix/air Approved in EU, Canada

Schering Levovist Lipid/air Approved in 69 countries,

including EU, Canada, Japan (not US)

Schering Sonavist Polymer/air Clinical Development

Sonus Echogen Surfactant/ Late Clinical Development

Pharmaceutical dodecafluoropentane


12/47


an ultrasound contrast study. Yet although theyare classified as drugs, all current agents arereally physical suspensions of bubbles in aninactive medium such as saline. The bubbles

are stabilised by shells that are sometimes onlytens of nanometres in thickness: they are physi-cally delicate and particularly susceptible todestruction by pressure or shear stress. The gasalso diffuses out of the bubbles with time andsometimes this process is faster outside thebody than once injected. In addition, thebubbles have a tendency to float and henceseparate from the solution that holds themover a period of time. As long as it is realisedthe bubble suspensions cannot be handled likean ordinary drug and require special care, it isrelatively easy to prepare the echo laboratoryfor their handling.

1.2.2.1 PreparationIt should be appreciated that no two agents arealike in the way that they are prepared for injec-tion. Some vials contain bubbles that simply arereconstituted by the addition of saline, othersrely on the user to manufacture the bubblesduring the preparation process. These requiremechanical agitation or a more elaborate mixingprocedure that must be followed carefully.

Levovistis prepared by injection of sterile waterinto a vial containing a sugar/lipid powderfollowed by vigorous shaking of the vial byhand. The agent can be administered in one ofthree concentrations (200, 300 or 400 mg/ml),upon which the volume of water to be useddepends. It should be left to stand for about2 minutes after mixing. The more concentratedsuspensions are somewhat viscous, so thatbubble flotation is not a practical problem withthis agent. The bubbles are, however, sensitiveto the increase of pressure within the vial thatwould result from the injection of the water

into a closed space, so the vial is vented for theaddition of water and withdrawal of the agentby the use of a special cap provided with eachdose (Figure 8). The agent should be usedwithin about 30 minutes of preparation.

Optisonis kept in a refrigerator from which itshould be removed to bring it to roomtemperature before use. It is prepared by

simple shaking by hand and withdrawal from avented vial through an 18-gauge needle.Optison bubbles are buoyant and have atendency to rise quickly to the surface ofthe syringe unless it is gently but constantlyrotated before injection.

SonoVueis prepared by simple mixing withsaline, which is provided in the measured

syringe kit (Figure 9). The bubbles are also

Fig. 8 Preparation of Levovist. Sterile water ismeasured according to the concentration requiredand added to a vial of dry powder through a specialventing cap. The mixture is agitated by hand andleft to stand for two minutes before slow withdrawal

into the injecting syringe through the cap.


13/47


buoyant and the preparation of quite lowviscosity, so the bubbles also float rapidly.SonoVue can be infused effectively using aninfusion pump, but care should be taken to

ensure that the direction of the infused outputis vertical, so that bubbles do not becometrapped.

Definity, like Optison, is already mixed whenthe vial is manufactured. However, the bubblesare formed only after the vial has been agitatedfor 45 seconds in a mechanical shaker, which ismost conveniently located in the examinationroom (Figure 10 a, b). The agent is withdrawnfrom the vial by venting with a second 18-gauge needle and may be injected by syringe(Figure 10c) or infused by simple injection intoa drip bag (Figure 10 d). In this form Definityis unusually stable and the bubbles neutrallybuoyant, so that flotation is not a problem.

1.2.2.2 Preparing for injectionWith all agents, the use of smaller diameterneedles should be avoided because the bubblesare subjected to large pressure drop due to theBernoulli effect as the fluid exits the tip of thelumen. The faster the injection and the smallerthe diameter, the larger the pressure drop andthe greater the likelihood of damage to thebubbles. A 22-gauge or larger needle is best. It

should be borne in mind that the smaller the

Fig. 9 Preparation of SonoVue. The agent ispackaged as a dry powder and reconstituted usingsaline measured in a the syringe supplied.

a) b)

c) d)

Fig. 10 Preparation of Definity. The vial is agitatedby a mechanical device supplied specifically for thisagent (a, b). It is then withdrawn from the vial forbolus injection (c) or injected directly into a drip bag

for infusion (d).

Preparing a contrast agent: tips

Establish machine settings and scan patient

before mixing the agent

Use vents: never withdraw from or inject agent

against a closed space

Watch for flotation of bubbles: gently shake the

vial and syringe after preparation

Follow the manufacturers directions carefully


14/47

needle lumen, the slower the injection shouldbe to avoid inadvertent bubble destruction.Bubbles can also be destroyed by pushing thesyringe plunger against a closed line: if thishappens accidentally, the dose should be dis-carded. Because the total injection volume forsome studies can be less than 1ml, a flush isneeded to push the agent into the centralvenous stream. 510 ml saline administered

through a 3-way stopcock at the end of a shortline which allows free movement of the syringefor mixing is often best. It is good practice forthe saline to traverse the right angle bend of thestopcock, not the agent, which should beconnected along the direct path, again to avoidbubble destruction (Figure 11).

1.2.3 Administration methods

1.2.3.1 BolusFigure 12 shows dose-response measurementsmade using a dedicated Doppler probe posi-tioned on the brachial artery of a patient afteran injection of Optison. A first-pass peakenhancement of 30 dB is followed by a steady,exponential wash-out of about 3 minutesduration, which is typical of the kinetics of ablood pool microbubble agent. The hugeenhancement at the peak usually causes over-

loading of the Doppler receiver, creating spec-

tral or colour blooming. Sonographers com-monly evade this by waiting until the wash-outphase, in which a more suitable enhancementcan be obtained for a longer period. It is clear,however, that this is an inefficient way to usethe agent because the majority of bubbles donot contribute to the collection of diagnosticinformation. Decreasing the bolus volumedoes not necessarily help. I f the dose of

Optison is increased progressively from


Administering a contrast agent: tips

Establish the venous line before preparing the

agent

Use a 22-gauge or larger cannulation; cubital

vein is often best

Use a 3-way stopcock if preferred but never push

filled syringe against a closed valve

Watch for bubble flotation: keep the syringe

moving between injections

Inject slowly: follow the manufacturers

recommendations

Fig. 12 Example of typical time enhancementcurve measured in the brachial artery of a patientafter injection with Optison. Note that the effectiveduration of enhancement in this case is about 3minutes.

Fig. 11 A three way stopcock, line and flush readyfor injection of Optison.


15/47


Fig. 13 The enhancement curves of Optison with increasing dose. Note that increasing the dose in-creases both the peak enhancement and the wash-out time of the agent.

Fig. 14 Peak enhancement as a function ofadministered dose. This graph shows that the effectof the agent at its peak does not increase with

large doses. This may be due to 'bolus spreading'.

Fig. 15 Duration of enhancement as a function ofadministered dose. Note that the effect of largerdoses is to extend the duration of enhancement,

rather than increase the height of the peak effect.


16/47


10200 l/kg (Figure 13), we see the peakenhancement increase slightly and the wash-out time increase more dramatically. Note thatincreasing the dose by a factor of 10 does not

have the same effect on the peak enhancement(Figure 14). Instead, it is the wash-out time ofthe agent, reflected by the duration of enhan-cement that is increased (Figure 15). The areaunder these curves represents the product of

the enhancement and the time of enhancementand may be regarded as a crude measure of theintegrated effect of the agent. If this is plottedagainst the dose we see (Figure 16) that increas-

ing the volume of a bolus causes this value torise in a roughly linear manner. It is on thisobservation that the effort to infuse agents isbased.

1.2.3.2 Infusion

Although there is a definite role for bolus injec-tion where short duration, maximum effect isrequired from the agent (eg in LV opacification

for the wall motion studies of 2.4), infusing acontrast agent offers the possibility of enhance-ment for a length of time that more closelymatches that of the imaging examination.Agents may be infused by slow injection, by useof an injection pump, by dripping the dilutedagent through an IV line, or by use of an infu-sion pump of the kind used to titrate IV drugs.Figure 17 compares SonoVue administered as abolus (Figure 17 a) and as an infusion throughan IV pump (Figure 17 b). Figure 18 shows theresults of a slow manual injection of Levovist,with an enhancement of 15 dB effectivelylasting for more than 9 minutes. It can be seenthat the effect of infusion is to maintain anenhancement comparable to that of the peakbolus for a period which matches more closelythat of the clinical ultrasound examination.There is, then, a clear advantage to infusing

contrast agents, not least of which is that theimproved efficiency translates into an economicsaving on the cost of the agent. Against thismust be weighed a few practical considerationswhich vary from agent to agent.

First, if a slow or mechanically controlled injec-tion is to be used, the agent must remain stablein the syringe for the duration of the infusion.

Levovist, which is quite viscous after mixing,

Bolus: pros and cons

Pro Con

Easy to perform Contrast effect

short lived

Highest peak Contrast effect changing

enhancement during study

Wash-in and Timing of bolus difficult

wash-out visible

Agent is used quickly: Comparative contrast

no stability problems studies difficult

Fig. 16 Integrated fractional enhancement (definedas the area under the curves of Figure 13) plottedas a function of dose for Optison. A steadily increas-ing response is seen as the administered volume ofthe agent is increased.


17/47

provides this more easily than, say Optison, inwhich emulsion the bubbles tend to float.Agitating the preparation while it is in thesyringe is in practice difficult. On the otherhand, some perfluorocarbon agents such asDefinity are effective in very high dilution,

rendering them ideal for injection into a dripbag and drip infusion. This cannot be attempt-ed with Levovist because of the adverse effectof dilution on the bubble population. Finally,some agents which are effective at very lowvolumes require whole body doses of much less


Fig. 17 Bolus vs infusion. (a) A bolus injection of SonoVue gives a wash-out time of about 1 minute.

(b) Infused, the agent delivers uniform enhancement over a period of about 7 minutes.

a) b)

Fig. 18 The enhancement provided by a slowmanual injection of Levovist. Enhancement is seen

to last for approximately nine minutes.

Infusion: pros and cons

Pro Con

Extends time More complex to perform:

of enhancement may require pump

Provides consistent Titration of infusion

effect takes time and effort

Avoids blooming/ Stability of agent over

artifacts infusion period

can be a problem

Dose is optimised:

agent is used more

efficiently

Allows negative bolus

method for quantitation

of perfusion (see Ch4)


18/47

than 1ml, rendering infusion difficult and slowinjection virtually impossible. Others, whichare fundamentally unstable out of the body,such as Echogen, offer no obvious means for

infusion or slow injection. Clearly, the manu-facturers advice should be sought beforeattempting infusion in a clinical setting. TheTable below shows when an infusion issessential to contrast examination.

1.3 Mode of action

The interaction of an ultrasound beam with apopulation of bubbles is a process whose sub-tlety and complexity has only recently beenrecognised. Understanding what happens

when a microbubble contrast agent is exposedto an ultrasound beam is the key to understand-ing and using new clinical methods forcontrast imaging and thus the key to interpret-ing a clinical contrast echocardiographic study.

A sound field comprises a train of travellingwaves, much like ripples on a pond. The fluidpressure of the medium (in this case tissue)

changes as the sound propagates through it. A

gas bubble is highly compliant and hence issquashed when the pressure outside it is raisedand expanded when the pressure is lowered. Ata typical clinical frequency of 2 MHz, for

example, a bubble sitting in an acoustic fieldundergoes this oscillatory motion two milliontimes per second. As it moves in this way, thebubble becomes a source of sound that radiatesradially from its location in the body, as would

ripples on a pond from a small object movingat a point on its surface (Figure 19). The soundthat reaches the transducer from this bubble,combined with that from all of its neighbours,is what constitutes the scattered echo from a

contrast agent. Characterising this echo so that


When is an infusion necessary?

Infusion mandatory Bolus sufficient

Myocardial perfusion: Myocardial perfusion:

rest and stress studies, rest studies only,

quantitative analysis qualitative assessment

Coronary flow reserve LVO: rest or stress,

thrombus detection

Pulmonary vein

Doppler enhancement

Fig. 19 A bubble in an acoustic field responds tothe changes in pressure which constitute the sound

wave by changing in size. The radius oscillates atthe same rate as the incident sound, radiating anecho.


19/47


it can be differentiated from those of ordinarytissue such as the cardiac muscle is the basis ofcontrast specific imaging modes such as pulseinversion Doppler or harmonic imaging.

1.3.1 Bubble behaviour and incidentpressure

Unlike tissue, a bubble does not scatter in thesame way if it is exposed to weak (that is lowamplitude) sound, than to strong, high ampli-tude sound. Instead, there are three broad

regimes of scattering behaviour that depend onthe peak pressure of the incident sound fieldproduced by the scanner (see Table). These areused in different ways in contrast imaging ofthe heart.

Looking at the Table, we see that at low inci-dent pressures (corresponding to low transmitpower of the scanner), the agents producelinear backscatter enhancement, resulting in anaugmentation of the echo from blood. This isthe behaviour originally envisaged by the

contrast agent manufacturers for their firstintended clinical indication: Doppler signalenhancement. It is this for which they obtainedapproval from the regulatory authorities, an

indication which does not require any changeto ultrasound imaging and Doppler instru-mentation. As the transmit intensity controlof the scanner is increased and the pressureincident on a bubble goes beyond about50100 kPa, which is still below the level usedin most diagnostic scans, the contrast agentbackscatter begins to show nonlinear characte-ristics, such as the emission of harmonics. It isthe detection of these that forms the basis ofcontrast specific imaging modes such asharmonic and pulse inversion imaging andDoppler. Finally, as the peak pressure passesabout 1 MPa, near the maximum emitted by atypical echo imaging system, many agentsexhibit transient nonlinear scattering. Thisforms the basis of triggered imaging and moststrategies for detection of myocardial perfusion.The Table shows these three regimes together

with the imaging methods developed to exploitthem. It should be noted that in practice,

Three regimes of bubble behaviour in an ultrasound field

Peak pressure Mechanical Bubble Acoustic Clinical(approx) Index (MI) behaviour behaviour Application

@ 1MHz

< 100 kPa < 0.1 Linear Backscatter Coronary artery Doppler,oscillation enhancement PV Doppler, fundamental

B-mode LVO

100 kPa1 MPa 0.1 1.0 Nonlinear Harmonic Coronary artery Doppler,

oscillation backscatter harmonic B-mode LVO,

real-time perfusion imaging

> 1 MPa > 1.0 Disruption Transient Power Doppler LVO,

harmonic echoes intermittent perfusion

imaging


20/47

because of the different sizes present in a real-istic population of bubbles (19), the bordersbetween these behaviours are not sharp. Norwill they be the same for different agent types,

whose acoustic behaviour is strongly depend-ent on the gas and shell properties (20). In thenext section we examine the clinical imagingmethods which rely on each of these behav-iours in turn.

1.3.1.1 The Mechanical Index (MI)For reasons unrelated to contrast imaging,ultrasound scanners marketed in the US arerequired by the Food and Drugs Adminis-tration (FDA) to carry an on-screen label ofthe estimated peak negative pressure to whichtissue is exposed. Of course, this pressurechanges according to the tissue through whichthe sound travels as well as the amplitude andgeometry of the ultrasound beam: the higherthe attenuation, the less the peak pressure intissue will be. A scanner cannot know whattissue it is being used on, so the definition of

an index has been arrived at which reflects theapproximate exposure to ultrasound pressure atthe focus of the beam in an average tissue. TheMechanical Index (or MI) is defined as thepeak rarefactional (that is, negative) pressure,divided by the square root of the ultrasoundfrequency. This quantity is related to theamount of mechanical work that can be per-formed on a bubble during a single negative

half cycle of sound (21). In clinical ultrasoundsystems, this index usually lies somewherebetween 0.1 and 2.0. Although a single value isdisplayed for each image, in practice the actualMI varies throughout the image. In theabsence of attenuation, the MI is maximal atthe focus of the beam. Attenuation shifts thismaximum towards the transducer. In a phasedarray, steering reduces the intensity of the

ultrasound beam so that the MI is also less at

the edges of the sector. Furthermore, because itis a somewhat complex procedure to calculatethe index, which is itself only an estimate ofthe actual quantity within the body, the indices

displayed by different machines are notprecisely comparable. Thus, for example, morebubble disruption might be observed at a dis-played MI of 1.0 using one machine than onthe same patient using another. For this reason,recommendations of machine settings for aspecific examination will include MI valuespeculiar to a given ultrasound manufacturersinstrument. Nonetheless, the MI is one of themost important machine parameters in acontrast echo study. It is usually controlled bymeans of the output power setting of the

scanner.

1.3.2 I Linear backscatter: Dopplerenhancement

Although bubbles of a typical contrast agentare smaller than red blood cells, and theirvolume concentration in the blood followingintravenous injection is a fraction of a percent,the amplitude of the echo from the micro-

bubble agent eclipses that of the blood itself.


The Mechanical Index (MI)

Defined by MI = , where Pneg is the peak

negative ultrasound pressure,fthe ultrasound

frequency

Reflects the normalised energy to which a target

(such as a bubble) is exposed in an ultrasound

field

Is defined for the focus of the ultrasound beam

Varies with depth in the image (lessens with

increasing depth)

Varies with lateral location in the image (lessens

towards the sector edges)


21/47


This is because the weakness of the echofrom blood originates from the cells them-selves, which are poor scatterers of ultrasound.Their acoustic impedance is almost identical

to that of the surrounding plasma. A bubblecontaining compressible gas, on the otherhand, presents a strong discontinuity inacoustic impedance and hence acts as a strongreflector. Size for size, a bubble is about onehundred million million times more effectiveat scattering ultrasound. Thus the injection ofa relatively sparse population of bubbles intothe bloodstream results in a substantial enhance-ment of the blood echo. In a Doppler examina-tion, the arrival of the contrast agent, someseconds after peripheral venous injection, inthe portion of the systemic vasculature underinterrogation is marked by a dramatic increasein signal strength. In spectral Doppler this isseen as an intensifying of the greyscale of thespectrum (Figure 2), whose enhancement isrelated to dose and whose duration depends onthe method of administration. A 10 ml bolus

of Levovist at a concentration of 400 mg/ml,for example, provides a peak enhancement ofabout 25 dB with a usable wash-out period ofabout 24 minutes. Spectral Doppler examina-tions of, for example, the pulmonary vein, thatfail because of lack of signal strength, can berescued by the contrast examination, detect-ing signals that would be otherwise obscuredby noise.

1.3.2.1 Enhancement studies withconventional imaging

In the initial studies which were carried out toobtain regulatory approval of ultrasoundcontrast agents, the sole indication was tosalvage a nondiagnostic Doppler examination.These demonstrated the capacity of contrastagents to increase the technical success rate of

Doppler assessments of aortic stenosis, mitral

regurgitation and pulmonary venous flow, inwhich latter case it rose from 27 percent to 80percent (22).

Echo enhancement caused by the agent may beconsidered in other ways too. For example,given a satisfactory transthoracic colourDoppler study, one use of the agent might besimply to enable a higher ultrasound frequencyto be used, exploiting the agent to counter thehigher tissue attenuation. In such a case, thecontrast enhancement translates into higherspatial resolution (Figure 3). Alternatively, thecolour system may be set to use fewer pulsesper scan line (that is a lower ensemble length)while still achieving the same sensitivity toblood flow by means of the contrast enhance-ment. The agent will then provide the userwith a higher frame rate.

With conventional greyscale imaging, enhance-ment might be seen in lumina of the ventriclesor large vessels if the concentration of the

bubbles is sufficiently high, as it is for per-fluorocarbon gas agents such as Optison andDefinity. The contrast is not, however, normal-ly seen in the small vessels within the muscle ofthe myocardium itself. This is because the1025 dB of enhancement provided by theagent still leaves the blood echo some1020 dB below that of the echogenic tissue ofthe heart wall. In order to enhance the visible

grey level, either higher concentrations ofbubbles must be achieved or specific, newimaging strategies employed. A greater numberof bubbles, however, inevitably leads to higherattenuation of the sound beam as it traversesthe cavities, so perfusion imaging cannot beeffective using conventional imaging (23).Thus new, bubble specific imaging methodsare necessary.


22/47

1.3.3 II Nonlinear backscatter:harmonic imaging

1.3.3.1 The need for bubble-specific imaging

In many of the potential applications of contrastagents, one might ask whether it is possible tocontinue to increase the amount of agentsinjected and obtain progressively strongerechoes from blood; to the point, for example,where the myocardium becomes visible on agreyscale image. Unfortunately, attenuation ofthe ultrasound beam by the agent in the cavityalso increases with bubble concentration, with

the result that shadowing occurs distal to theagent and the myocardium disappears alto-gether. Because of this limitation of the useableconcentration of the agent, we are generally leftwith small enhancements in the myocardiumecho that must be identified against the strongbackground echo from the solid tissue itself.X-ray angiography, which is faced with a similarproblem after dye is injected into the blood-stream, deals with these clutter components ofthe image by simple subtraction of a pre-injec-tion image. What is left behind reveals flow inindividual vessels or the blush of perfusion atthe tissue level. If, however, we subtract twoconsecutive ultrasound images of a solid organ,we get a third ultrasound image of the sameorgan, produced by the decorrelation of thespeckle pattern between acquisitions. In order toshow parenchymal enhancement due to the

agent, speckle variance must first be reduced byfiltering, with a consequent loss of spatial ortemporal resolution. Even if the speckle pro-blem could be overcome, subtraction would stillbe poorly suited to the dynamic and interactivenature of cardiac ultrasound imaging. InDoppler modes, the problem of the moving wallinterference (the clutter) prevents the smallerecho from the blood itself being detected, as

discussed in 1.1.3.

How then might contrast agents be used toimprove the visibility of blood in movingvascular structures such as the myocardium?Clearly, a method that could identify the echo

from the contrast agent and suppress the echofrom solid tissue would provide both a realtime subtraction mode for contrast-enhancedB-mode imaging, and a means of suppressingDoppler clutter without the use of a velocity-dependent filter in spectral and colour modes.Nonlinear imaging aims to provide such amethod, and hence the means for detection offlow in smaller vessels than is currently poss-ible.

1.3.3.2 Harmonic imagingExamining the behaviour of contrast-enhancedultrasound studies reveals two important piecesof evidence. First, the size of the echo enhance-ment at very high dilution following a smallperipheral injection (7 dB from as li tt le as0.01 ml/kg of Levovist, for example (24)) ismuch larger than would be expected from suchsparse scatterers of this size in blood. Second,investigations of the acoustic characteristics ofseveral agents (25) have demonstrated anapproximately linear dependence of back-scattered coefficient on numerical density ofthe agent at low concentrations, as expected,but a dependence of attenuation on ultrasoundfrequency different to that predicted by theRayleigh law, which describes how the echo-

genicity of normal tissue changes with fre-quency. Instead, peaks exist which are depen-dent on both ultrasound frequency and the sizeof the microbubbles, suggestive of resonancephenomena. This important observationsuggests that the bubbles resonate in the ultra-sound field. As the ultrasound wave whichcomprises alternate compressions and rarefac-tions propagates over the bubbles, they ex-

perience a periodic change in their radius in



23/47


sympathy with the oscillations of the incidentsound. Like vibrations in other structures,these radial oscillations have a natural orresonant frequency of oscillation at which

they will both absorb and scatter ultrasoundwith a peculiarly high efficiency. Consideringthe linear oscillation of a free bubble of air inwater, we can use a simple theory (1) to predictthe resonant frequency of radial oscillation of abubble of 3 m diameter, the median diameterof a typical transpulmonary microbubbleagent. As Figure 20 shows, it is about 3 MHz,

approximately the centre frequency of ultra-sound used in a typical echocardiographyscan. This extraordinary and fortunate coincidence explains why ultrasound contrastagents are so efficient and can be administeredin such small quantities. It also predicts thatbubbles undergoing resonant oscillation inan ultrasound field can be induced to emit

harmonics, the basis of harmonic imaging.

One consequence of this extraordinary coinci-dence is that bubbles undergoing resonantoscillation in an ultrasound field can be inducedto nonlinear motion. It has long been recog-

nised (26) that if bubbles are driven by theultrasound field at sufficiently high acousticpressures, the oscillatory excursions of thebubble reach a point where the alternateexpansions and contractions of the bubblessize are not equal. Lord Rayleigh, the origina-tor of the theoretical understanding of soundupon which ultrasound imaging is based, wasfirst led in 1917 to investigate this by hiscuriosity over the creaking noises that histeakettle made as the water came to the boil.The consequence of such nonlinear motion is

Fig. 20 Microbubbles resonate in a diagnosticultrasound field. This graph shows that the reson-ant or natural frequency of oscillation of abubble of air in an ultrasound field depends on itssize. For a 3.5 m diameter, the size needed for anintravenously injectible contrast agent, the reson-ant frequency is about 3 MHz.

Fig. 21 Higher incident pressures induce non-linear behaviour in a bubble. Asymmetric oscillationof a resonant bubble in an ultrasound field gives

rise to an echo with even harmonics.


24/47

that the sound emitted by the bubble, anddetected by the transducer, contains harmonics,

just as the resonant strings of a musical instru-ment, if plucked too vigorously, will produce

a harsh timbre containing overtones(themusical term for harmonics). The origin of thisphenomenon is the asymmetry which beginsto affect resonant oscillation as the amplitudebecomes large. As a bubble is compressed bythe ultrasound pressure wave, it becomes stifferand hence resists further reduction in itsradius. Conversely, in the rarefaction phase ofthe ultrasound pulse, the bubble becomes lessstiff, and, therefore, enlarges much more(Figure 21). Figure 22 shows the frequencyspectrum of an echo produced by a micro-bubble contrast agent following a 3.75 MHzburst. The particular agent is Levovist, though

many microbubble agents behave in a similarway. Ultrasound frequency is on the horizontalaxis, with the relative amplitude on the verticalaxis. A strong echo, at -13 dB with respect to

the fundamental, is seen at twice the transmit-ted frequency, the second harmonic. Peaks inthe echo spectrum at sub- and ultraharmonicsare also seen. Here, then, is one potentialmethod to distinguish bubbles from tissue:excite them so as to produce harmonics anddetect these in preference to the fundamentalecho from tissue. Key factors in the harmonicresponse of an agent, which varies frommaterial to material, are the incident pressureof the ultrasound field, the frequency, as well asthe size distribution of the bubbles and themechanical properties of the bubble capsule.A stiff capsule, for example, will dampen the


Fig. 22 Harmonic emission from Levovist. A sample of a contrast agent is insonated at 3.75 MHz and theecho analyzed for its frequency content. It is seen that most of the energy in the echo is at 3.75 MHz, butthat there is a clear second peak in the spectrum at 7.5 MHz, as well as a third at 1.875 MHz. The secondharmonic echo is only 13 dB less than that of the main, or fundamental echo. Harmonic imaging and

Doppler aims to separate and process this signal alone. The smaller peak is the first subharmonic.


25/47


oscillations and attenuate the nonlinear re-sponse.

Harmonic B-mode imaging

A new real-time imaging and Doppler methodbased on this principle, called harmonic imag-ing (27), is now widely available on most

modern echocardiography ultrasound scanners.In harmonic mode, the system transmitsnormally at one frequency, but when in har-monic mode is tuned to receive echoes pre-

ferentially at double that frequency, where theechoes from the bubbles lie. Typically, thetransmit frequency lies between 1.5 and3 MHz and the receive frequency is selected bymeans of a bandpass filter whose centrefrequency is at the second harmonic between 3and 6 MHz (Figure 23). Harmonic imaginguses the same array transducers as conventionalimaging and for in most of todays ultrasoundsystems involves only software changes. Echoesfrom solid tissue, as well as red blood cellsthemselves, are suppressed. Real-time har-monic spectral Doppler and colour Dopplermodes have also been implemented (some-times experimentally) on a number of com-mercially available systems. Clearly, an excep-tional transducer bandwidth is needed tooperate over such a large range of frequencies.Fortunately much effort has been directed in

recent years towards increasing the bandwidthof transducer arrays because of its significantbearing on conventional imaging performance,so harmonic imaging modes do not require theadditional expense of dedicated transducers.

Harmonic Doppler imaging

In harmonic images, the echo from tissuemimicking material is reduced but not elim-

inated, reversing the contrast between theagent and its surrounding (Figure 24). Thevalue of this effect is to increase the conspicuityof the agent when it is in blood vessels normal-ly hidden by the strong echoes from tissue. Inspectral Doppler, one would expect the sup-pression of the tissue echo to reduce the tissuemotion thump that is familiar to all echo-cardiographers. Figure 25 shows spectral

Doppler applied to a region of the aorta in

Fig. 23 The principle of harmonic imaging andDoppler. A conventional phased array is used, withthe receiver tuned to double the transmittingfrequency. The tissue and blood give an echo at thefundamental frequency, but the contrast agentundergoing nonlinear oscillation in the sound fieldemits the harmonic which is detected by the har-

monic system.


26/47

which there is wall motion as well as bloodflow within the sample volume. The conven-tional Doppler image of Figure 25 a shows thethump artifact due to clutter, which is almostcompletely absent in the harmonic Dopplerimage of Figure 25 b. All instrument settings,including the filters, are identical in theseimages; we have merely switched between

modes. In vivo measurements from spectral

Doppler show that the signal-to-clutter ratio isimproved by a combination of harmonic imag-ing and the contrast agent by as much as 35 dB(28). Figure 26 shows harmonic colour imagesof an aorta, this time with flash artifact fromrespiratory motion. The harmonic imagedemonstrates the flow without the flashartifact. The potential application of this diag-

nostic method is to detect blood flow in small


Fig. 24 Harmonic imaging. An vitro phantom shows a finger-like void in tissue equivalent material, filledwith dilute contrast agent, barely visible in conventional mode (a). Harmonic imaging is seen to reversethe contrast between the microbubble agent and the surrounding material (b).

Fig. 25 Clutter rejection with harmonic spectral Doppler. (a) The abdominal aorta of an animal is ex-amined with harmonic spectral Doppler. In conventional mode, clutter from the moving wall causes thefamiliar artifact which also obscures diastolic flow. (b) In harmonic mode, the clutter is almost completelysuppressed, so that flow can be resolved. The settings of the filter and other relevant instrument para-meters are identical.

a) b)

a) b)


27/47


Fig. 26 Clutter rejection with harmonic colour Doppler. (a) Flow in the abdominal aorta is superimposedon respiratory motion, producing severe flash artifact. (b) In harmonic mode at the same point in therespiratory cycle, the flash artifact disappears. Flow from a smaller vessel (the cranial mesenteric artery)

is visualised. All instrument settings are the same.

vessels surrounded by tissue which is moving:in the branches of the coronary arteries (29) orthe myocardium itself (30), as well as in theparenchyma of abdominal organs (31).

Harmonic power Doppler imaging

In colour Doppler studies using a contrastagent, the effect of the arrival of the agent in acolour region of interest is often to produceblooming of the colour image, wherebysignals from major vascular targets spread outto occupy the entire region. Thus, althoughflow from smaller vessels might be detectable,the colour images can be swamped by artifac-tual flow signals. The origin of this artifact isthe amplitude thresholding that governs most

colour displays in conventional (or velocity)mode imaging. Increasing the backscatteredsignal power simply has the effect of displayingthe velocity estimate, at full intensity, over awider range of pixels around the detected loca-tion. A display in which the parameter mappedto colour is related directly to the backscatteredsignal power, on the other hand, has the advan-tage that such thresholding is unnecessary and

that lower amplitude Doppler shifts, such as

those which result from sidelobe interference,are displayed at a lower visual amplitude,rendering them less conspicuous. Echo enhanc-ed flow signals, in contrast, will be displayedat a higher level. This is the basis of the powerimaging map (also known as colour powerangiography, or colour Doppler energy map-ping). Power Doppler can help eliminate someother limitations of small vessel flow detectionwith colour Doppler. Low velocity detectionrequires lowering the Doppler pulse repetitionfrequency (PRF), which results in multiplealiasing and loss of directional resolution. Adisplay method that does not use the velocityestimate is not prone to the aliasing artifact,and therefore allows the PRF to be lowered and

hence increase the likelihood of detection ofthe lower velocity flow from smaller vessels.

Because it maps a parameter directly related tothe acoustic quantity that is enhanced by thecontrast agent, the power map is a naturalchoice for contrast enhanced colour Dopplerstudies. However, the advantages of the powermap for contrast enhanced detection of small

vessel flow are balanced by a potentially devas-

a) b)


28/47


tating shortcoming: its increased susceptibilityto interference from clutter. Clutter is bothdetected more readily, because of the powermodes increased sensitivity, and displayedmore prominently, because of the high intensitydisplay of high amplitude signals. Further-more, frame averaging has the additional effectof sustaining and blurring the flash over thecardiac cycle, so exacerbating its effect on theimage. This is the reason that conventionalpower mode, while quite popular in radiologi-cal and peripheral vascular ultrasound imaging,has almost no role in echocardiography.

At the small expense of some sensitivity, amply

compensated by the enhancement caused bythe agent, harmonic mode effectively over-comes this clutter problem (Figure 27).Combining the harmonic method with powerDoppler produces an especially effective toolfor the detection of flow in the small vessels ofthe organs of the abdomen which may bemoving with cardiac pulsation or respiration.In a study in which flow imaged on contrast

enhanced power harmonic images was

compared with histologically sized arterioles inthe corresponding regions of the renal cortex(24), it was concluded that the method iscapable of demonstrating flow in vessels of lessthan 40m diameter; about ten times smallerthan the corresponding imaging resolutionlimit, even as the organ was moving withnormal respiration. Recent studies of thispower mode method in the heart (32) showthat flow can be imaged in the myocardiumwith an agent such as Levovist (33).

1.3.3.3 The impact of harmonic imagingHarmonic imaging succeeds in identifyingmicrobubble contrast agent in the tissue vascu-

lature by means of its echo signature. In doingso it helps tackle some old problems in ultra-sound, such as the rejection of the tissue echoin Doppler modes designed to image movingblood, and creating a subtraction mode with-out sacrificing the real time nature of theexamination. Figure 28 summarises graphicallythe effect of contrast agents and harmonicdetection on the Doppler process. In conven-

tional Doppler, the signal from blood is larger

Fig. 27 Reduction of the flash artifact in harmonic power Doppler. The harmonic contrast method helpsovercome one of the principal shortcomings of power Doppler, its increased susceptibility to tissue motion.

(a) Aortic flow in power mode with flash artifact from cardiac motion of wall. (b) In harmonic mode at thesame point in the cardiac cycle, the flash is largely suppressed. All instrument settings are the same.

a) b)


29/47


than the clutter signal from tissue. In contrastenhanced Doppler, the signal from blood israised, sometimes to near that of tissue. Withharmonic mode, the signal from blood is raisedbut that from the tissue reduced, so reversingthe contrast between tissue and blood. Another

way of looking at the harmonic method is that,because of its greater sensitivity to small quan-tities of agent, a given level of enhancementdue to a bolus will last longer (Figure 29). This

is the reason that a number of investigatorshave found that cardiac contrast imaging inany mode is generally more effective with theharmonic method (29, 30). Harmonic imaginghas been shown to render possible the detec-tion of microvessels containing contrast agentseven in the presence of highly echogenic tissue,such as the liver (31) and myocardium (34).

Harmonic imaging demands exceptional per-formance from the transducer array and systembeamformer. Its implementation forces impli-cit compromises between, for example, imageresolution and rejection of the non-contrastagent echo. It places unusual demands on thebandwidth performance of transducers, as wellas the flexibility of the architecture of the imag-ing system. The ease with which it has beendeveloped on modern instruments, however,

Fig. 28 Quantifying the impact of harmonicDoppler. Clutter and spectral Doppler signal levelsmeasured in conventional, contrast enhanced, andharmonic contrast enhanced modes.

Fig. 29 The improved sensitivity of the harmonic method translates into increased useful imaging timefrom a contrast agent bolus. Time-enhancement curve caused by circulation of contrast agent bolus shows

how harmonic imaging, by reducing the detection threshold, increases the imaging lifetime of the agent.


30/47


reflects this flexibility and augurs well for thefuture of the method. More significantly,contrast agents are now being developed speci-fically with nonlinear response as a designcriterion. Entirely new agents present oppor-tunities for entirely new detection strategies(17). Our laboratory measurements show thatsome new contrast agents are capable of creat-ing an echo with more energy in the secondharmonic than at the fundamental: that is,

they are more efficient in harmonic than con-ventional mode. With such agents, nonlinearimaging is the preferred clinical method fortheir detection.

1.3.3.4 Tissue harmonic imagingIn second harmonic imaging, an ultrasoundscanner transmits at one frequency and receivesat double this frequency. The resulting

improved detection of the microbubble echo is

due to the peculiar behaviour of a gas bubblein an ultrasound field. However, any source ofa received signal at the harmonic frequencywhich does not come from the bubble, will

clearly reduce the efficacy of this method.

Such unwanted signals can come from non-linearities in the transducer or its associatedelectronics, and these must be tackled effec-tively in a good harmonic imaging system.However, tissue itself can produce harmonicswhich will be received by the transducer. Theyare developed as a wave propagates throughtissue. Again, this is due to an asymmetry: thistime the fact that sound travels slightly fasterthrough tissue during the compressional partof the cycle (where it is denser and hence morestiff ) than during the rarefactional part .Although this effect is very small, it is sufficientto produce substantial harmonic componentsin the transmitted wave by the time it reachesdeep tissue, so that when it is scattered by alinear target such as the myocardium, there is a

harmonic component in the echo, which isdetected by the scanner along with the har-monic echo from the bubble (35). This is thereason that solid tissue is not completely darkin a typical harmonic image. The effect is toreduce the contrast between the bubble andtissue, rendering the problem of detecting per-fusion in the myocardium more difficult.

Tissue harmonics, though a foe to contrastimaging, are not necessarily a bad thing. Infact, an image formed from tissue harmonicswithout the presence of contrast agents hassome properties which recommend it over con-ventional imaging. These follow from the factthat tissue harmonics are developed as thebeam penetrates tissue, in contrast to the con-ventional beam, which is generated at the

transducer surface (36). Artifacts which accrue

Application

LVO; myocardial

perfusion (offline

subtraction necessary)

Coronary flow reserve;

pulmonary vein

Doppler enhancement

Coronary vessel

imaging

LVO; myocardial

perfusion

Usual Format

Greyscale display

of harmonic

echoes

Spectral Doppler

with tissue motion

suppression

Colour Doppler

with tissue sup-

pression, conven-

tional B-mode

background

Power Doppler

with tissue motion

suppression,

conventional B-

mode background

Method

Harmonic

B-mode

Harmonic

spectral

Doppler

Harmonic

colour

Doppler

Harmonic

power

Doppler

Harmonic imaging modalities


31/47


from the first few centimetres of tissue, such asreverberations, are reduced by using tissue har-monic imaging. Sidelobe and other low-levelinterference is also suppressed, making tissueharmonic imaging the routine modality ofchoice for many echocardiographers (Figure 30).

For contrast studies, the tissue harmonic limitsthe visibilty of bubbles within tissue in aB-mode harmonic image and therefore can beconsidered an artifact. In contemplating howto reduce it, it is instructive to bear in mind

differences between harmonics produced by

tissue propagation and by bubble echoes. First,tissue harmonics require a high peak pressure,so are only evident at high MI. Reducing theMI leaves only the bubble harmonics. Second,tissue harmonics become greater at greaterdepths, whereas harmonics from bubbles aredepth independent. Finally, harmonics fromtissue at high MI are continuous and sustained,whereas those from bubbles are transient innature as the bubble disrupts.

1.3.3.5 Pulse inversion imagingThe method we have described above for har-monic imaging imposes some fundamentallimitations on the imaging process which

restrict its clinical potential in organ imaging.First, in order to ensure that the higher fre-quencies are due only to harmonics emitted bythe bubbles, the transmitter must be restrictedto a band of frequencies around the funda-mental (Figure 31 a). Similarly, the receivedband of frequencies must be restricted to thoselying around the second harmonic. If these tworegions overlap (Figure 31 b), the result will be

that the harmonic filter will receive echoes

Potential origin of echoesat harmonic frequency

Nonlinear scattering from bubbles

Nonlinear propagation of echoes from solid tissue

Inadvertent transmission of harmonics by non-

linearities in transmitter or transducer

Inadvertent production of harmonics by non-

linearities in receiver or transducer Deliberate transmission at second harmonic

frequency to improve bandwidth of image

Fig. 30 Tissue harmonic imaging of the heart. Note that in comparison to the fundamental image (a), thetissue harmonic image (b) has a reduced level of artifacts associated with reverberation and sidelobe inter-ference.

a) b)


32/47


33/47


power, nondestructive, continuous imaging ofmicrobubbles in an organ such as the liver. Themethod also relies on the asymmetric oscillati-on of an ultrasound bubble in an acoustic field,but detects all nonlinear (or, even) compo-nents of the echo, over the entire bandwidth ofthe transducer. In pulse inversion imaging, twopulses are sent in rapid succession into thetissue. The second pulse is a mirror image of

the first (Figure 32). That is, it has undergonea 180 phase change. The scanner detects theecho from these two successive pulses andforms their sum. For ordinary tissue, whichbehaves in a linear manner, the sum of twoinverted pulses is simply zero. For an echo withnonlinear components, such as that from abubble, on the other hand, the echoes produ-ced from these two pulses will not be simple

mirror images of each other, because of the

asymmetric behaviour of the bubble radiuswith time. The result is that the sum of thesetwo echoes is not zero. Thus, a signal is detec-ted from a bubble but not from tissue. It canbe shown by mathematical analysis that thissummed echo contains the nonlinear evenharmonic components of the signal, includingthe second harmonic (38).

One advantage of pulse inversion over the filterapproach to detect harmonics from bubbles isthat it no longer suffers from the restriction ofbandwidth. The full frequency range of soundemitted from the transducer can be detected inthis way, and a full bandwidth, that is highresolution, image of the nonlinear echoes frombubbles may be formed in real time. The pricepaid, using current technology, is a reduction

of the effective frame rate by a factor of two.

Fig. 32 Principle of Pulse Inversion Imaging. A pulse of sound is transmitted into the body and echoesare received from agent and tissue. A second pulse, which is an inverted copy of the first pulse, is thentransmitted in the same direction and the two echoes are summed. Linear echoes from tissue will beinverted copies of each other and will cancel to zero. The microbubble echoes are distorted copies of eachother, and the nonlinear components of these echoes will reinforce each other when summed, producinga strong harmonic signal.


34/47


One application for this advantage is in tissueimaging without contrast agents. In conventio-nal imaging, reverberation, phase aberrationand sidelobe artifacts caused by the linearsound beam can reduce image contrast,especially in hypoechoic structures such as theventricular cavities. Tissue harmonic imagingcan reduce these artifacts and improve imagecontrast, but with a somewhat degraded resolu-tion. Figure 37 illustrates how pulse inversionimaging provides better suppression of linearechoes than harmonic imaging and is effective

over the full bandwidth of the transducer,showing improvement of image resolution overharmonic mode.

Because this detection method is a more effi-cient means of isolating the bubble echo,weaker echoes from bubbles insonated at low,nondestructive intensities, can be detected.Figure 33 shows a dramatic improvement in

contrast sensitivity obtained over both funda-

mental and harmonic modes by pulse inver-sion imaging in a phantom. For LVO, pulseinversion imaging provides improved resolu-tion of the endocardium (Figure 34). At highMI, bubbles are disrupted and their contents

Fig. 33 Demonstration of pulse inversion imaging. In vitro images of a vessel phantom containingstationary Optison surrounded by tissue equivalent material (biogel and graphite). a) Conventional image,MI = 0.2. b) Harmonic imaging, MI = 0.2, provides improved contrast between agent and tissue. c) Pulseinversion imaging, MI = 0.2. By suppressing linear echoes from stationary tissue, pulse inversion imagingprovides better contrast between agent and tissue than both conventional and harmonic imaging.

Fig. 34 Pulse inversion imaging with Levovistshowing LV opacification. Notice that perforatingbranches of the coronary vessels are visible.

Courtesy Eric Yu, University of Toronto, Canada

a) b) c)


35/47


Fig. 35 Pulse inversion imaging showing coronary vessels using Levovist at two frames per heart beat.(a) Agent enters cavity. (b) Perfusion, 6 seconds later. (c) One second later, vessels are seen entering themyocardium.

Fig. 36 The appearance of blood vessels in harmonic and pulse inversion imaging in a liver.In a harmonic image of a liver following an injection of Optison, large vessels have a punctate appearanceas the high MI ultrasound disrupts the bubbles as they enter the scan plane (a). In the pulse inversionimage of the same liver, a lower MI can be used so that continuous vessels are now seen. Improved resolu-tion of pulse inversion imaging demonstrates 4th order branches of the portal vein (b).

a) b) c)

a) b)


36/47


Fig. 37 Tissue harmonic imaging. In conventionalimage (a), reverberation, phase aberration andsidelobe artifacts degrade depiction of cavityborders. Tissue harmonic image (b) shows reducedartifacts, but with a somewhat degraded resolution.Pulse inversion image (c) gives better suppressionof artifacts than harmonic imaging and uses full

bandwidth of the transducer, thus improving resolu-tion.

dispersed following insonation (see 1.3.4).Perforating coronary vessels, however, whichhave a sufficiently high flow velocity to refillbetween frames, become visible on pulse inver-sion images with Levovist (Figure 3435).In the liver, low MI imaging allows long

lengths of vessels to become visible. Figure 36compares the appearance of large hepaticvessels imaged in harmonic and pulse inversionmodes following a peripheral venous injectionof Optison. The high MI harmonic image(Figure 36 a) shows the punctate appearance ofvessels in which the agent is destroyed; thelower MI pulse inversion image (Figure 36 b)shows fourth order branches of the portal vein

imaged with great clarity. Finally, because of its

extended bandwidth, pulse inversion imagingis an ideal method to use for high MI tissueharmonic imaging. Figure 37 compares thefundamental (Figure 37 a), harmonic (Figure37 b) and pulse inversion images (Figure 37 c)of the same view. Note that artifact suppressi-

on is achieved in both the harmonic and pulseinversion images, but that the resolution of thepulse inversion image is superior.

1.3.3.6 Power pulse inversion imaging

In spite of the improvements offered by pulseinversion over harmonic imaging for suppress-ing stationary tissue, the method is somewhatsensitive to echoes from moving tissue. This is

because movement of tissue causes linear

a) b)

c)


37/47


Fig. 38 The principle of pulse inversion Doppler (also known as power pulse inversion). The method issimilar to pulse inversion imaging, except that a sequence of more than two pulses are transmitted withalternating phase. The echoes from successive pulses are recombined in a way that eliminates the effectof steady displacement of the target due to tissue motion. The method allows suppression of movingtissue without the need to disrupt the bubble, hence achieving real time, low MI imaging. The exampleshown here of three pulses illustrates a principle that can be applied to arbitrarily long ensembles ofpulses, improving bubble-to-tissue sensitivity.

echoes to change slightly between pulses, sothat they do not cancel perfectly. Furthermore,at high MI, nonlinear propagation also causesharmonic echoes to appear in pulse inversionimages, even from linear scattering structuressuch as solid tissue. While tissue motion arti-facts can be minimised by using a short pulserepetition interval, nonlinear tissue echoes can

mask the echoes from bubbles, reducing theefficacy of microbubble contrast, especially ininterval delay imaging when a high MI is used.A recent development seeks to address theseproblems by means of a generalisation of thepulse inversion method, called pulse inversionDoppler (38). This technique which is alsoknown as power pulse inversion imaging combines the nonlinear detection performance

of pulse inversion imaging with the motion

discrimination capabilities of power Doppler.Multiple transmit pulses of alternating polarityare used and Doppler signal processing tech-niques are applied to distinguish betweenbubble echoes and echoes from moving tissueand/or tissue harmonics, as desired by theoperator. In a typical configuration, the echoesfrom a train of pulses are combined in such a

way that signals from moving tissue whichpose a problem to pulse inversion imaging are eliminated (Figure 38). This method offersimprovements in the agent to tissue contrastand signal to noise performance, though at thecost of a somewhat reduced framerate. Themost dramatic manifestation of this methodsability to detect very weak harmonic echoes hasbeen its first demonstration of real-time per-

fusion imaging of the myocardium (39)


38/47

Handbook of Contrast Echocardiography36

(Figure 39). By lowering the MI to 0.1 or less,bubbles undergo stable, nonlinear oscillation,emitting continuous harmonic signals. Becauseof the low MI, very few bubbles are disrupted,

so that imaging can take place at real-timerates. Because sustained, stable nonlinear oscil-lation is required for this method, perfluoro-

carbon gas bubbles work best.

1.3.4 III - Transient disruption:intermittent imaging

As the incident pressure to which a resonatingbubble is exposed increases, so its oscillationbecomes more wild, with radius increasing insome bubbles by a factor of five or more duringthe rarefaction phase of the incident sound.Just as a resonating violin string, if bowed over-zealously, will break, so a microbubble, if drivenby intense ultrasound, will suffer irreversibledisruption of its shell. A physical picture ofprecisely what happens to a disrupted bubble isonly now emerging from high speed videostudies (Figure 40). It is certain, however, thatthe bubble disappears as an acoustic scatterer(not instantly, but over a period of time deter-mined by the bubble composition), and that asit does so it emits a strong, brief nonlinearecho. It is this echo whose detection is the basisof intermittent myocardial perfusion imaging

with ultrasound contrast agents.

1.3.4.1 Triggered imagingIt was discovered during the early days of har-monic imaging that by pressing the freezebutton on a scanner for a few moments, andhence interrupting the acquisition of ultra-sound images during a contrast study, it is pos-sible to increase the effectiveness of a contrast

agent. So dramatic is this effect that it can raise

Fig. 40 Fragmentation of contrast agent observedwith a high-speed camera. The frame images (ag)are captured over 50 nanoseconds. The streakimage (similar to an M-mode) shows the variations

in bubble diameter with a temporal resolution of10 nanoseconds. The bubble is insonified with2.4 MHz ultrasound with a peak negative pressureof 1.1 MPa (MI~0.7). (a) The bubble is initially 3 min diameter. (b) The first large expansion. (cd) Thebubble fragments during compression after the firstexpansion. (ef) Fragments are seen during expan-sion. Resulting bubble fragments are not seen afterinsonation, because they are either fully dissolvedor below the optical resolution.Courtesy of James Chomas, Paul Dayton, Kathy

Ferrara, University of California, Davis CA, USA

Fig. 39 Pulse inversion Doppler (also known aspower pulse inversion) showing real-time myo-

cardial perfusion. The contrast agent is Optison,the mechanical index is 0.15, the frame rate is 15 Hz.


39/47


Fig. 41 Bubble disruption demonstrated in a laboratory phantom. a) Bubbles are suspended in a waterbath containing a block of tissue mimicking material. b) The block is scanned at high mechanical indexusing a linear array scanner. A bright echo is seen from the bubbles as they experience high peak pres-sure insonation. c) An image taken a moment later reveals a swirl of echo-free fluid, corresponding to thescanplane which no longer bears intact bubbles. This explains the apical 'swirl' seen in many LVO studies

at high MI.

the visibility of a contrast agent in the myo-cardial circulation to the point that it can beseen on a harmonic B-mode image, above theecho level of the normal heart muscle (40).This is a consequence of the ability of theultrasound field, if its peak pressure is suf-ficiently high, to disrupt a bubbles shell and

hence destroy it (37, 41) A phantom studyillustrating this effect is shown in Figure 41. Asthe bubble is disrupted, it releases energy, socreating a strong, transient echo, which is richin harmonics. This process is sometimes mis-leadingly referred to as stimulated acousticemission. The fact that this echo is transient innature can be exploited for its detection. Thefirst, and simplest method, is to subtract from

a disruption image a baseline image obtainedeither before or (more usefully) immediatelyafter insonation. The residual echoes can thenbe attributed to the bubbles which weredisrupted by the imaging process. Such amethod requires offline processing of storedultrasound images, together with softwarewhich can align the ultrasound images beforesubtraction. Chapter 4 contains more discus-

sion and examples of this technique. A more

effective method is to compare the echoes fromtwo or more consecutive high intensity pulses,separated by a fraction of a millisecond. This isthe principle of intermittent harmonic powerDoppler.

1.3.4.2 Intermittent harmonic power Doppler

Power Doppler imaging (also known as colourpower angio, or energy imaging) is a techniquedesigned to detect the motion of blood or oftissue. It works by a simple, pulse-to-pulsesubtraction method (42), in which two ormore pulses are sent successively along eachscan line of the image. Pairs of received echotrains are compared for each line: if they areidentical, nothing is displayed, but if there is a

change (due to motion of the tissue betweenpulses), a colour is displayed whose saturationis related to the amplitude of the echo that haschanged. This method, though not designedfor the detection of bubble disruption, is ideallysuited for high MI destruction imaging. Thefirst pulse receives an echo from the bubble,the second receives none, so the comparisonyields a strong signal. In a sense, power

Doppler may be thought of as a line-by-line


40/47

subtraction procedure on the radiofrequencyecho detected by the transducer.

A crucial question is how long one needs to

wait between pulses. I f the two pulses are tooclose together in time, the bubbles gaseouscontents, which are dispersed after disruptionof the shell by a process of diffusion andfragmentation, will st

Get Chapter1 1

Documents