Clinical application and image interpretation in intracoronary ultrasound

Introduction

Miniaturized flexible ultrasound catheters providedetailed information on the vessel wall. They are anadditional diagnostic evaluation technique, partly in therealm of coronary interventions in patients with coronary artery disease. Fluoroscopy and angiographicroad-mapping are indispensable elements of the intracoronary ultrasound examination and of the therapeuticprocedure. Thus intracoronary ultrasound should not beconsidered an alternative to angiography but rathera complementary diagnostic technique. The clinicaladvantages deriving from the use of intracoronary ultrasound have not yet been established in randomizedtrials. However, there is increasing evidence from largeprospective studies that ultrasound guidance improvesthe results of catheter-based intracoronary interventions in terms of immediate lumen enlargement,reduced procedure-related complications and long-termrestenosis. Although intracoronary ultrasound hasbecome a routinely applied diagnostic technique ininterventional cardiology, no attempts have beenmade to standardize the examination procedure, thedefinitions and the format of reporting qualitative andquantitative data.

The aim of this article is to propose guidelinesfor the acquisition, classification and analysis of intracoronary ultrasound images and to recommend indications for clinical application of intracoronary ultrasound based on recent experience.

Part I: Image acquisition

To acquire intracoronary ultrasound images requires theintracoronary insertion of a dedicated catheter probe.The cardiologist performing intracoronary ultrasoundshould be familiar with the selection and positioning ofguiding catheters, the steering and positioning ofguidewires, and the handling of possible complicationssuch as guidewire loop formation, spasm, anddissections[1–3]. Intracoronary ultrasound studies should,therefore, only be performed by, or under directsupervision of, experienced interventional cardiologists.

Selection and positioning of guiding catheters

Ultrasound catheters currently available for intracoronary application have an outer diameter of between 2·9and 3·5 French at the distal end. Although mechanicalcatheters are smaller at the distal end (2·9 and 3·2French models are available), their diameter increases to5 French in the shaft. Therefore, 7 French large lumenguiding catheters are required and 8 French guidingcatheters are preferred to facilitate positioning of intracoronary ultrasound probes. Intermittent injection ofcontrast medium also facilitates positioning and reducesthe risk of non-uniform rotation. Electronic catheters,although slightly larger in diameter at the distal end,maintain the same size along the entire catheterlength, allowing insertion through large lumen 6Frguiding catheters.

Steering and positioning of intracoronaryultrasound catheters

Although the handling of intracoronary ultrasoundprobes is similar to the handling of over-the-wire or

monorail PTCA catheters, additional caution isrequired. A stable guiding catheter position is desirablesince intracoronary ultrasound catheters still have lesstrackability and a larger profile than most ballooncatheters. The short monorail catheters are prone toprolapse, so that the tip of the guidewire must be placeddistal in the target vessel and the intracoronary ultrasound catheter should never be advanced over thefloppy end of the guidewire. For negotiating tortuousvessels, long monorail or over-the-wire intracoronaryultrasound catheters should be selected.

Damage to the vessel wall from modern flexibleintracoronary ultrasound catheters is rare, but to avoiddamage the catheters should not be advanced to thesmallest distal coronary vessels and caution must beused in crossing stents immediately after deployment.Intracoronary ultrasound catheters are larger and haveless tapering than PTCA balloons, so that disruption offragile coil stents can be induced by forceful pushing ofthe intracoronary ultrasound catheter. Occasionally,adjusting the guiding catheter and guidewire position toobtain a more central orientation of the intracoronaryultrasound catheter facilitates crossing of the stentedsegment.

Intracoronary ultrasound examination

A standard operating procedure is essential for bothinterpreting and reviewing the intracoronary ultrasoundexamination (Table 1). For mechanical systems, carefulpreparation of the catheter with generous forceful flushing is required to remove air bubbles around the transducer. The catheter should always be handled carefully,avoiding twists or kinks, especially during rotation ofthe ultrasound crystal. For electronic systems, the registration of an ultrasound image while the catheter is notin contact with the vessel wall is also required to subtractthe ring-down artifact. The best results are obtainedwhen the registration of the image is performed in theaorta, disengaging the guiding catheter from the coronary ostium. Before starting the insertion, demographic data and annotations of the type of vesselexamined and intervention performed should be enteredusing the alpha-numeric keyboard available in all ultrasound machines. Optimization of the machine settingshould be controlled during insertion and checked againbefore starting the pull-back. Inexperienced operatorsoften tend to reduce the near gain to eliminate artifactsor excessive backscatter from blood. If the near fieldsignals are almost extinguished, however, soft plaquesaround the catheter can be missed.

An intravenous injection of 5000 to 10 000 unitsheparin and an intracoronary bolus of 100–300ìg nitroglycerine or 1–3 mg isosorbide dinitrate should beadministered before the intracoronary ultrasound study.The role of aspirin and other antiplatelet agents inpreventing thrombotic complications during intracoronary ultrasound is not clear at present. The insertion of the intracoronary ultrasound catheter should be gentle but continuous and rapid, avoiding stopping orwithdrawal of the catheter to examine, compare ormeasure specific sites (stenosis, reference segments, etc).The Stanford Group has suggested that before enteringthe left anterior descending or left circumflex artery fromthe left main coronary artery, the image should beelectronically rotated so that, entering the left anteriordescending, the circumflex is coming o at 9 o’clock and,entering the circumflex, the left anterior descending ispositioned at 3 o’clock. With this orientation, duringinsertion into the left anterior descending the diagonalbranches will come o to the left, between 8 and12 o’clock, and the septal branches will come o at thebottom (between 2 and 8 o’clock). In the circumflex, theobtuse marginal branches will come o between 12 and6 o’clock. For the right coronary, the orientation is donewith the first right ventricular marginal branch, whichshould be rotated to the 9 o’clock position. Althoughnot essential for intracoronary ultrasound interpretation, the consistent use of this rotational orientation inserial intracoronary ultrasound examinations facilitatesthe comparison of corresponding cross-sections beforeand after interventions, especially when selective plaqueremoval with DCA is performed.

When the intracoronary ultrasound catheter hasbeen inserted distally to the segment of interest, acontinuous pull-back should be started. The use of amotorized pull-back device at constant speed (mostfrequently 0·5 mm . s"1) is highly recommended toincrease reproducibility and allow precise measurements of vessel length. For mechanical systems, recentadvances in catheter design allow the withdrawal only ofthe imaging cable within an external sheath, minimizingthe risk of malrotation or uneven speed during pull-backfrom friction of the catheter shaft against the vessel wall.The position from which the pull-back was started, andall the relevant sites explored, should be indicated by theoperator using a voice comment or written annotationon tape and showing, when possible, the correspondingfluoroscopic position of the ultrasonic catheter on asplit-screen. Side-branches, well visualized with bothangiography and ultrasound, are clear landmarks thatfacilitate interpretation and comparison of sequentialexaminations. Furthermore, the distance from a sidebranch can be used as a precise method to identify thesame arterial site in serial intracoronary ultrasoundexaminations (i.e. before and after interventions) when afixed speed is used for the pull-back.

Electrocardiographic triggering of the pull-backusing a step-motor has also been proposed to obtainsmoother contours during three-dimensional reconstruction and increase the accuracy of the measurement[1].

Safety of intracoronary ultrasound and handling of complications

Guidewire looping

The relatively short monorail tip of some intracoronaryultrasound catheters easily bends or kinks, especially intortuous segments (sheperd’s crook origin of the rightcoronary artery or acute origin of the left circumflexfrom the left main coronary artery). If a wire loop ispresent proximal to its insertion in the intracoronaryultrasound catheter, the intracoronary ultrasoundprobe and guiding catheter should be simultaneouslyremoved.

Spasm

Spasm during intracoronary ultrasound can occurduring purely diagnostic procedures, such as after hearttransplantation, and was reported in approximately 3%of patients in a large multicentre survey including morethan 2000 intracoronary ultrasound examinations[1]. Incase of spasm, one or more intracoronary boluses ofnitrates should be given followed, if necessary, by a slowintracoronary injection of 1–1·5 mg verapamil[2,3]. If thespasm is located distal to the intracoronary ultrasoundprobe, the probe should be removed carefully, and notpulled out forcefully, as this will cause additional intimaldamage.

After the intracoronary ultrasound study, thepatient with spasm should be monitored carefully. Afinal contrast injection following removal of all intracoronary ultrasound hardware and the guidewire isrecommended to document the integrity of the vesselwall in all cases.

Dissections and acute closure

Coronary artery dissections and acute closure are a rare,but severe potential intracoronary ultrasound complication with an incidence of 0·4% of all patientsstudied[1,4]. These complications are very rare after diagnostic intracoronary ultrasound studies, particularly inpatients with mild or stable disease and mainly occurduring therapeutic procedures, when it is often di cultto decide whether the complication is related to theintervention or to the intracoronary ultrasound examination. The handling of such a complication should bethe same as recommended in patients with symptomaticdissections after invasive procedures (new balloon inflation with standard or perfusion balloons, stent implantation, bypass surgery or medical treatment alone,depending on the severity of the dissection and on theclinical condition of the patient)[2].

Part II: Image interpretation

Normal arterial morphology

The ultrasound appearance of normal human arteries invitro and in vivo has been studied extensively[5–10]. In thenormal intima, a superficial layer of endothelial cellscovers a very thin subendothelial layer of connectivetissue and smooth muscle cells. Its thickness increaseswith age, from a single cell layer at birth, to a meanof 60ìfrom infancy to 5 years and reaches 220ìat  30 years and 250ìat 40 years[11]. Further adaptive,physiological thickening of the intima occurs at pointswhere wall tension is increased, such as at arterialbifurcations and on the outer parts of bends, and may beeither eccentric or di use[12]. Di use intimal thickeningis common in older patients, and is a process which ishistologically distinct from atherosclerosis[13].

The muscular media of coronary arteries ispredominantly composed of smooth muscle cells withsmaller amounts of collagen, elastic tissue and proteoglycans. Fibrous degeneration of the media, particularlyof the inner third, is not uncommon in elderly patientsor in patients with concomitant atherosclerotic disease[14,15]. The thickness of the media ranges from 125ìto 350ì(mean 200ì) but medial thinning occurs in thepresence of atherosclerotic disease[16].

The adventitia is composed of loose collagen andelastic tissue that merges with the surrounding periadventitial tissue and is 300–500ìthick. Two sheets ofelastic tissue separate the media from the intima (internal elastic lamina) and the adventitia (external elasticmembrane).

The sudden change in acoustic impedancebetween adjacent tissue plays a particularly importantrole in the determination of the characteristics of theultrasound image of the vessel wall17,18]. The leadingedge of the intima and of the adventitia are two strongacoustic interfaces well visualized with ultrasound inmost instances. Although the internal elastic lamina iscomposed of strong echogenic elastic tissue, fibrouschanges in the inner third of the media decrease thedi erence in acoustic impedance between these adjacentlayers, making clear delineation of the internal elasticlamina and of the inner border of the media rare[19]thusprecluding reliable estimation of the media thickness.Therefore, only two layers are normally distinguishedwith intracoronary ultrasound, an internal wall layeroften described as intima or intimal plaque whichshould be more correctly defined ‘intima-media complex’ and an external or adventitial layer. The absenceof an acoustic interface between the adventitia andthe surrounding peri-adventitial tissue precludes theidentification of the adventitia as a discrete, quantifiableentity.

Atherosclerotic lesions

Atherosclerotic plaque has been studied extensively withintracoronary ultrasound in vitro and in vivo, both inperipheral and in coronary arteries[8,9,15,20–23]. Earlychanges that occur in the development of atherosclerosis, such as fatty streaks or duplication or fragmentation of the inner elastic lamina, do not changethe ultrasonic appearance of the vessel wall and cannotbe visualized with intracoronary ultrasound as longas they remain below the threshold of resolution ofintracoronary ultrasound.

With further progression of atherosclerosis anincrease in intimal thickness can be detected in theultrasound image. In advanced atherosclerosis, threebasic types of lesion are distinguished: (1) highly cellular fibromuscular lesions or lesions with di use lipidinfiltration which have a low echoreflectivity; (2) densefibrous lesions which produce bright, heterogeneousand sometimes speckled echoes, with an echoreflectivity equal or superior to the echoreflectivity of theadventitia; (3) calcified lesions which produce intenselybright reflections with acoustic shadowing (Fig. 1).Occasionally, areas of di use lipid deposition andnecrotic degeneration appear as dark areas (lowechodensity), often located within fibrous areas orcovered by a fibrous cap[5,15,22,23]. Small deposits,however, can be missed due to the limited resolutionand dynamic range of the currently available ultrasound systems. Dissections (false lumen) or broadecholucent areas between the intima and the adventitiadue to attenuation of the ultrasound signal by thickened fibrous intimal plaques are also often misinterpreted as lipids.

Thrombus appears as a bright heterogeneousspeckling reflection which cannot reliably be distinguished from other types of plaques. Intraluminalmasses having these ultrasound characteristics or the presence of multiple channels within the plaque communicating with the lumen are highly suggestive ofthrombi (Fig. 2). Sometimes mural thrombi generatelinear echoes within a thickened intima which representsan acoustic interface between thrombus and underlyingintima (wall layering)[24].

Qualitative classification

Normal artery/mild intimal thickening

The presence of a homogeneous vessel wall or of a thinintima is a rare finding in the intracoronary ultrasoundpopulation. Since mild intimal thickening is part of theageing process of the arterial system and does not inducelumen narrowing, a thickness of the intima-mediacomplex smaller than 0·3 mm is often suggested as anempirical arbitrary cut-o to distinguish betweenatherosclerotic plaque and mild ‘physiological’ intimalthickening (Table 2(a)). It should be stressed that a thinor minimally thickened intimal layer does not automatically indicate that a vessel wall is normal in terms ofreactivity to vasoactive stimuli. Angiographic studieshave shown that angiographically smooth and normalsegments may have an abnormal response to vasoactivestimuli in patients with coronary artery disease or riskfactors for coronary artery disease[25,26]. Ultrasound candetect atherosclerotic lesions in arterial segments thatappear normal on angiography, and recent studies haveshown that abnormal vasoconstrictive responses occurmore frequently and are more severe in segments showing di use plaque accumulation with ultrasound[27,28].

Atherosclerotic disease

Severity of intimal thickening

The maximal thickness of the intima-media complex or,more appropriately, the percentage of the total vesselarea occupied by plaque, are the most common quantitative indices used to define the severity of atherosclerotic involvement. Atherosclerotic lesions may bepresent in segments which are angiographically normalbecause compensatory total vessel enlargement in theearly phases of atherosclerosis tends to keep the lumenconstant[29[. Lumen reduction does not occur, accordingto these pathology studies, until the plaque occupiesmore than 40% of the total cross-sectional vessel area.However, atherosclerotic lesions occupying less than20% and 40% of the total vessel area can still beconsidered as lesions with a minimal and moderateatherosclerotic burden, respectively (Table 2(a)). Abovethis threshold (plaque area greater than 40% of the totalvessel area), atherosclerotic lesions may reduce thelumen area and can be classified as lesions with a large or massive atherosclerotic burden. Although manyultrasound studies have confirmed the pathology studiesof Glagovet al.[30–33], a reduction of the total vessel areain the stenotic segment has been described in restenoticlesions and in primary lesions (reversed Glagove ect)[34–36]. These observations suggest that cautionmust be taken in the evaluation of the functional severityof an atherosclerotic lesion based on intracoronaryultrasound measurements of plaque area.

Plaque eccentricity

This feature is based on either the presence of a nonthickened portion of the arterial circumference (diseasefree wall), or on a low ratio of the thinnest and thethickest part of the circumference (eccentricity index). Aplaque is often defined as eccentric in the presence of aneccentricity index smaller than 0·5.

Eccentric plaques are not uncommonly seen onintracoronary ultrasound despite the angiographicappearance of concentric narrowing[37]. Conversely,angiographically eccentric plaques rarely have a segmentof completely normal vessel wall (intima-medial thickness smaller than 0·3 mm). Both these observations arerelevant to guide selective plaque removal interventions.

Plaque composition

The echointensity of the di erent plaque components inthe image changes according to the system settings and ultrasound system used. In order to define a standardintensity which takes into account this variability, theechointensity of the intima can be compared to theechointensity of the adventitia[38]. Thus intimal thickening with less echointensity than the adventitia is oftenindicated as ‘soft’ material, whereas ‘hard’ plaques arecharacterized by equal or greater intensity than theadventitia. Low and high echoreflectivity should bepreferred to common denominations such as soft andhard since these common terms are not indicators of themechanical characteristics of the plaque. Intracoronaryultrasound definitions of soft plaque are misleading asmany plaques classified as soft will show high resistanceto dilatation[39].

The presence of acoustic shadowing and reverberations are specific landmarks of the presence ofcalcification. With the exception of multiple scatteredmicrocalcification, ultrasound can be considered highlyspecific and sensitive for the detection of calcium in aplaque. Extremely bright echoes can be induced bydensely fibrotic plaques and the extreme attenuation ofthe echo-signal can be misinterpreted as shadowing. Theabrupt disappearance of echoes and the presence ofduplicate echoes can be used to distinguish attenuationfrom true shadowing. Because of the importance ofcalcium for the selection of coronary interventionaldevices, it is important to define: (a) the presence ofsingle or multiple calcium deposits; (b) their depth(‘superficial’, defined as no tissue between the calciumdeposit and the lumen and ‘deep’, defined as all otherdepths), (c) their circumferential extent, measured indegrees or hours, or defined semi-quantitatively as calcium occupying less/more of 1, 2 or 3 quadrants of thevessel; (d) their axial distribution (length in mm).

Histologically, atherosclerotic plaques are rarelyhomogeneous and contain a mixture of plaque components with di erent echoreflectivity. Using theclassification proposed in Table 2(b), most of the atherosclerotic plaques are described as mixed and the majorityof the homogeneous plaques are described as soft orcalcific. Although a group of investigators found a largerprevalence of soft plaques in unstable lesions, they wereunable to define features that were truly pathognomic ofunstable angina[38]. Using this classification, no di erences in composition of the culprit lesion were observedbetween patients with stable and unstable syndromes,despite the confirmation with angioscopy of large di erences in terms of superficial plaque disruption andthrombosis[40].

These limitations and the relatively large interobserver variability of this qualitative classification suggest the need of quantitative techniques for analysis ofthe ultrasound characteristics of plaque components(densitometry[41], computer assisted gray level textureanalysis[42]or, more promisingly, backscatter analysis).

Plaque disruption

With improvements in image quality and increasedoperator experience, spontaneous plaque ruptures orfissures are increasingly observed, mainly in unstableischaemic syndromes[43]. These ultrasound observationshave clarified the pathological changes underlying manyof the angiographic ‘pseudoaneurysms’ by showing thepresence of niches within the atherosclerotic plaquelikely due to emptying of the lipid-rich necrotic plaquecore into the lumen.

Wall rupture or dissection are frequently theconsequence of percutaneous interventions directed toenlarge the arterial lumen and displace or remove theplaque. Two main types of wall disruptions should beconsidered (Fig. 3, Table 2(c)). Rupture of the vesselwall is defined as a radial tear, i.e. perpendicular to thevessel wall layers. The troughs produced by atherectomydevices, that may be classified as a specific type of wallrupture, are more readily appreciated when the matchedpre-interventional images are available for comparison.

Dissection of the vessel wall is defined as tear parallel to the vessel wall. The diagnosis of wall dissection orfracture is based on the visualization of blood flow in thenewly created lumen, if necessary confirmed by saline orcontrast injection. Pulsatility of an echolucent areawithin or behind a plaque is also suggestive of a falselumen. The following characteristics of disruptions mustbe noted: (1) location relative to the narrowest point(proximal, distal or at the narrowest point); (2) axiallength if a motorized pullback is available; (3) circumferential arc in hours, measured in the cross-section withthe largest circumferential extension of the dissection;(4) maximal depth, classified as partial (some plaqueremaining intact between the rupture and the underlyingadventitia) or complete (extending through the plaqueup to the adventitia).

A peculiar type of dissection is the superficialintimal flap, characterized by a slight, thickening ofthe dissected intima (< 0·20 mm), but still very visiblebecause of its great motility.

Various classifications of the e ects of coronaryinterventions have been proposed, combining thepresence and axial/circumferential extent of wall disruption[44,45]. The advantage of using these complex classi-fications over the more descriptive approach proposedabove is questionable, especially in the absence of a clearpredictive value in terms of risk of acute complicationsand late restenosis.

Quantitative assessment

Normal range of coronary artery dimensions

The normal range of coronary artery diameters in adultshas been established in autopsy studies[14]. The left maincoronary artery ranges between 2·5 and 5·5 mm (mean4·0 mm); the proximal left anterior descending arterybetween 2·0 and 5·0 mm (mean 3·6 mm); the proximalleft circumflex artery between 1·5 and 5·5 mm (mean3·0 mm) and the right coronary artery between 1·5 and5·5 mm (mean 3·2 mm). These measurements are largerthan the angiographic measurements of corresponding segments in apparently normal coronary arteries,especially in older patients, but were confirmed byultrasonic measurements in arteries with no plaque[46].The left anterior descending and left circumflex arteriestaper along their length, but the calibre of the rightcoronary artery remains constant up to the cruxcordis[47].

Calibration and ultrasound artefacts

Unlike quantitative coronary angiography, intracoronary ultrasound quantitation does not requireroutine calibration. The accuracy of measurementdepends on the incorporation of the correct o set andestimated average speed of sound in blood and vasculartissue into the scan-converting algorithm. However,correct system calibration should not be taken forgranted, and must be confirmed in in-vitro phantomsprior to the use of a new scanner.

Catheter malalignment

Non-coaxial alignment of the transducer within theartery results in an epilliptic rather than circular crosssectional imaging plane, leading to overestimation ofboth areas and diameters. In a comparative study ofintracoronary ultrasound and quantitative coronary angiography in normal arteries, no significant inaccuracywas caused by non-axial alignment of the catheter. Thiswas probably because the small size of the coronaryartery lumen relative to the length of the intracoronarysegment of the imaging catheter prevents significantmalalignment[48]. Nevertheless, this potential source oferror should be considered in aorta-ostial lesions, intortuous segments, in large or ectatic vessels, and whenmeasuring close to acute bends.

Non-uniform rotation

Image distortion may occur as a result of rotation angleartefacts induced by non-uniform rotation of the driveshaft in mechanical imaging catheters. When nonuniform rotation is obvious visually the degree of distortion may be significant and measurements are notreliable.

Systolo-diastolic changes

The lumen area varies in relation to changes in distending pressure during the cardiac cycle. The maximumarea is present in mid-systole and the minimum area inlate diastole except in the presence of a ‘tunnelled’ arterywhich runs under a muscle bridge causing systolicexternal compression of the artery[49]. The pulsatilevariation in lumen area of normal coronary arteries is onaverage 8% in native coronary arteries, whereas in thepresence of plaque this variation is reduced dependingupon the thickness, eccentricity and composition of theplaque[50]. By convention, the end-diastolic frames ofangiograms are used quantitatively when cardiac motionand contrast streaming are at a minimum. A number ofarguments can be made for employing the oppositestrategy and measuring the intracoronary ultrasoundimages at end-systole. The maximum lumen dimensionsduring the cardiac cycle are clinically the more relevantmeasurements for the purpose of sizing percutaneousinterventional devices and the movement of the ultrasound catheter within the artery is minimal at endsystole, making the ultrasound image more easilyinterpretable and the area measurements more reliableand reproducible. Furthermore, after interventions theminimum lumen area may vary considerably during thecardiac cycle as tissue flaps move to-and-fro into thelumen. The most practical approach is to measure lumendimensions in systole, when the size of any false lumen isminimized by the maximal distending pressure duringthe cardiac cycle.

Area measurements

Table 3 summarizes the most frequently used measurements with intracoronary ultrasound. All measurements(diameters and areas) should be performed in thestenosis at the site of the minimum lumen area and in thereference segments, proximal and distal to the stenosis.While the stenosis location is unequivocal, the positionof the reference cross-section is highly subjective.Specific indications (5 mm from both ends of the stentedsegment) have been proposed for stent implantation butrecommendations are more di cult for other types ofinterventions.

Lumen area

The lumen area is measured by tracing the leading edgeof the circumferential blood/intima interface signals.Edge detection may be facilitated on the real-timeimages by observing the dynamic alteration in specklepattern characteristic of flowing blood compared to themore static pattern of adjacent tissue. A bolus injectionof contrast dye or saline (at body temperature) into thevessel temporarily clears the bright blood signals andfacilitates edge detection. These images should not beused for quantitative assessment because of the di erentpropagation speed of sound waves in water or contrastand in blood.

The complex morphology of the lumen aftercoronary interventions, and especially after balloonangioplasty, raises questions concerning the appropriateness of also considering small fissures behind plaqueas part of the vessel lumen, since their functional importance for blood passage is questionable and unpredictable. A commonly applied solution in the presence ofextensive wall disruption is to distinguish between the‘true’ lumen area (in general the lumen in which theultrasound catheter is positioned), and the ‘dissectionarea’, separated from the true lumen by the dissectionflap. This distinction, straightforward in the example ofFig. 3, can become very subjective when a broaderjunction between true lumen and dissection lumen ispresent.

Total vessel area

As the adventitia imperceptibly merges with thesurrounding perivascular tissue, for the purpose ofintracoronary ultrasound measurements the total crosssectional area of the vessel is taken to mean the areaenclosed by the outermost definable interface, i.e. thewell delimited interface between the media and adventitia coinciding with the position of the external elasticlamina. This is also referred to as the external elasticlamina area. Measurement of the area within the internal elastic lamina theoretically allows the calculation oftrue plaque or intimal area, but, as previously discussed,the internal elastic lamina is not well delineated in mostcases. Vessel area cannot be measured when calciuminduced acoustic shadowing obscures more than90 degrees of the vessel circumference. When lesserdegrees of shadowing are present, the vessel border isextrapolated from the closest identifiable segments ofthe media/adventitia interface. The variable degree ofacoustic shadowing cast by the stent struts and theblurring of the vessel layers deep to the stent may alsogive rise to di culties in quantitating the vessel area.

Plaque area

The plaque area should be more accurately termed the‘plaque+ media’ area and is calculated as the di erencebetween the total vessel area and the lumen area. As thismeasurement is derived from the vessel and lumen areasproblems applicable to total vessel and lumen area measurement are also applicable to the measurement ofplaque area. In the case of plaque dissection, it has beenproposed that the plaque area between true lumen anddissection lumen be planimetered as a means of quantitatively expressing the severity of coronary dissections(dissection arm).

Percentage plaque area

The percentage of the vessel area occupied by plaque iscalculated using the formula: (total vessel area"lumenarea)/vessel area#100. This parameter has been referredto as ‘percent plaque area’, or ‘percent plaque burden’,‘percent cross-sectional area narrowing or stenosis orobstruction’. As the last terms are also used to describethe ratio of the lumen area at the site of stenosis relativeto the lumen area in the reference segment they mustbe avoided. A number of investigators have takenthe opposite approach and calculated the proportion ofthe vessel area occupied by the lumen, termed thepercentage lumen cross-sectional area.

Measurement of the plaque burden relates tohistological practice, and is therefore recommended.A simple but noteworthy distinction must be madebetween intracoronary ultrasound percent plaque areaand angiographically assessed percent stenosis. Compensatory vessel expansion and disease in the proximal‘reference’ segment accounts for the poor correlationnoted between intracoronary ultrasound and angiographic percent stenosis with, in general, intracoronary ultrasound percent stenosis more severe thanangiographic percent stenosis[9,10].

Intracoronary ultrasound derived diameters

Lumen diameter measurements remain central in everyday clinical practice, as the appropriate size of interventional devices is chosen on the basis of the estimateddiameter of the reference segments adjacent to astenosis. In addition, the severity of a stenosis andthe outcome of coronary interventions continue to beroutinely evaluated in terms of the percentage diameterstenosis of the vessel.

With intracoronary ultrasound, direct measurement of maximum and minimum diameter is the mostwidely applied method. Whereas the maximum lumendiameter is usually readily identified, selection of theminimum diameter may be di cult in cases in which theborders of the lumen are irregular and incorporatesections that protrude into the lumen. The minimumdiameter is normally drawn as the smallest diameter inany direction passing through the mid-point of themaximum diameter.

The ratio between the maximum and minimumlumen diameter can be used to define the symmetry ofthe lumen, with ratios lower than 1·0 indicating increasing lumen asymmetry. The mean diameter can also bederived from the lumen area assuming a circular area.Once the lumen boundary has been traced, automatedmethods can determine the maximum and the minimumdiameter through the geometric centre of the lumen.

Part III: Clinical applications

Angiographically normal coronary arteries

Normal angiograms are present in 10%–15% of patientsundergoing coronary angiography because of suspectedcoronary artery disease. Plaque formation can often bedemonstrated with intracoronary ultrasound in thesepatients. Erbelet al. observed atherosclerotic changesin 21/44 patients (48%) with suspected coronary arterydisease and a normal coronary angiogram[51]. Iffunctional parameters are also considered (coronaryflow reserve and endothelium-mediated vasodilatoryresponse) only 36% of patients were confirmed to befully normal.

These findings suggest a revision and a newclassification for patients with syndrome X or chest painwithout significant angiographic changes. Before recommending a routine intracoronary ultrasound assessmentin these patients, however, it is necessary to demonstratethe clinical relevance of these findings and in particulardi erences in prognosis in patients with and withoutintracoronary ultrasound-detectable atheroscleroticchanges.

Intracoronary ultrasound can also be used toevaluate other vessel abnormalities such as myocardialbridging[49], spontaneous coronary dissection[52]andcontrast inhomogeneties within the vessel lumen.

Evaluation of intermediate stenoses andambiguous lesions

Suboptimal angiographic visualization impairs accurateassessment of stenosis severity. Ostial stenoses, at theorigin of the left and right coronary arteries from theaorta, at the bifurcation of the left main coronary arteryor at the origin of large side branches, especially in theproximal left anterior descending coronary artery, areoften poorly visualized because of guiding catheterwedging, vessel overlap or foreshortening. Out-of-planeprojections do not always solve the problem.

Occasionally, insu cient quality of the angiogram results from extreme obesity, emphysema orchest deformities. Extreme lumen eccentricity (slit-likeorifices) are certainly more rare than estimated frompathology studies in non-pressurized arteries. However,a discrepancy between measurements in orthogonal projections (lesions significant in one projection and moderate in another projection) is not rare in clinical practiceand complicates clinical decision-making. Intracoronaryultrasound does not su er from these limitations andmeasurements of lumen cross-sectional area in lumensof non-circular morphology are straightforward. It iscommon experience that ultrasound can often solvethe problem of angiographically intermediate or ambiguous lesions by showing obviously normal or severelydiseased vessels.

In two large prospective series, in more than 20%of the examinations before coronary interventions,intracoronary ultrasound changed the managementstrategy (treatment of angiographically non-significantlesions after intracoronary ultrasound examinationand vice-versa)[53,54]. In both studies, however, the selection of the patients scheduled for an intracoronaryultrasound examination before intervention may haveresulted in an overestimation of the real impact ofultrasound for clinical decision making. Furthermore,the criteria used to define the severity of the stenosis withultrasound were not objectively defined.

In most instances, intracoronary ultrasound mayhelp to solve the clinical dilemma proposed by angiographically ambiguous or intermediate stenoses on apurely visual analysis of the angiogram. In particular,the minimal lumen diameter derived from intracoronaryultrasound area measurement is well correlated withphysiological parameters such as measurements of coronary flow (Kernet al., personal communication). Notinfrequently, however, the analysis of the severity ofintermediate stenoses in the catheterization laboratoryrequires functional investigations such as intracoronaryDoppler and post-stenotic pressure measurements.

Coronary artery disease after hearttransplantation

Accelerated transplant coronary artery diseaserepresents the most important cause of morbidity andmortality in cardiac transplant recipients beyond thefirst year after transplantation[55–57]. Because cardiacallografts are functionally denervated, major clinicalevents due to advanced coronary atherosclerosis, including myocardial infarction, congestive heart failure, andsudden death, usually occur without prodromal angina.Thus, repeated coronary angiography is performed forsurveillance of coronary artery disease progression[58].

The pathology of transient vasculopathy is distinctive in that it initially consists of concentric intimalproliferation throughout the coronary tree which thenprogresses to di use arterial obliteration. This is characterized angiographically by longitudinal narrowing ofthe arteries with pruning of distal vessels[58]. The limitations of standard coronary arteriography to accurately measure the severity of transplant coronary arterydisease has been highlighted by angiographic–pathologycorrelation studies[59].Intracoronary ultrasound is an e ective andreproducible method of measuring intimal proliferationin cardiac transplant recipients. One or more years aftercardiac transplantation the majority of patients haveintracoronary ultrasound evidence of silent intimalthickening not apparent by angiography[60,61].

Intracoronary ultrasound o ers early detection and quantitation of transplant coronary disease, and providescharacterization of vessel wall morphology. The studiesperformed in patients early after transplantation serve asa reference for the ultrasound appearance of young,morphologically normal coronary arteries[62]. A subsetof these patients, however, studied early after cardiactransplantation, has provided ultrasound evidence ofdonor-related atherosclerotic changes.

The ultrasound images obtained in patients ayear or more after transplantation show a broadspectrum of morphological abnormalities and a highincidence of angiographically silent intimal thickening[60,62,63]. Furthermore, in contrast with the histological paradigm of coronary artery disease after hearttransplant, focal lesions are often observed[63,64].

Preliminary longitudinal studies comparingthe sensitivity of intracoronary ultrasound to coronaryangiography for detecting and monitoring progressionof atherosclerotic disease have shown that progressionof intimal proliferation identified with intracoronaryultrasound in cardiac transplant recipients, occurs inapproximately 40% of the serially studied sites[65]. It hasbeen seen that intimal proliferation occurs mostly duringthe first 2 years after transplantation, whereas calcifi-cation of plaques occurs only later in the process[61,66].The presence of a mean intimal thickness greater orequal to 0·3 mm was shown to be an independentpredictor of overall and cardiac survival as well as offreedom from retransplantation[67].

Identification of factors predisposing to intimalproliferation is an important contribution to the understanding of the pathogenesis of transplant vasculopathyand to develop preventive and therapeutic strategies.Recent studies have correlated multiple immunologicaland metabolic factors with intimal thickness by univariate analysis, suggesting a multifactorial aetiology fortransplant vasculopathy[68].

Guidance during interventions

Table 4 summarizes the indications for intracoronaryultrasound in association with coronary interventions.

Lesion assessment before coronary interventions:selection of treatment

The intracoronary ultrasound examination o ers potential advantages over angiography for deciding whichspecific treatment modality is most appropriate for agiven lesion. Despite the extreme miniaturization ofthe ultrasound catheters, before interventions theprobe occludes the lesion in most cases, precluding aprolonged assessment because of the rapid developmentof symptoms and signs of myocardial ischaemia andcomplicating the image interpretation because of bloodstagnation.

Since the ultrasound catheter must be advancedinto the lesion, in the examination of ostial stenosesof the two main coronary arteries particular attentionmust be paid to avoid the complications of the severeischaemia induced by the partial or complete occlusionof flow.

Despite these limitations, the additional information provided by ultrasound on lesion composition,eccentricity and length modified the treatment strategy in almost 20% of the cases in the large experience of theWashington Heart Center[54]. Intracoronary ultrasoundmade an even greater impact on clinical decisionmaking, as reported by Leeet al. They showed thatincreased confidence with the technique and theincreased experience of the operators in the interpretation of the images led to a progressive increase overtime of decisions based on intracoronary ultrasoundfindings[53].

CalciumPresence, depth and circumferential extent ofthe calcification is of great importance for selecting thetype of interventional device and for estimating the riskof complications. Fluoroscopy detects calcium depositswhich occupy more than 180 degrees of the vesselcircumference with insu cient sensitivity (60%) and iscompletely unreliable in the presence of smaller calciumdeposits[69,70]. In 1155 coronary lesions examined withintracoronary ultrasound, Mintzet al.[69]detected lesioncalcium in 73% of the lesions (38% by angiography),showing that calcium is more often subendothelial (72%)and located at the maximal thickness of the plaque.Subendothelial calcium is an important factor limitingtissue retrieval and increasing the incidence of procedural complications after directional atherectomy[71–73].The presence and circumferential extent of calcium isalso predictive of the development of dissection orfracture after PTCA, almost always present when the arcof vessel calcium is greater than 90 degrees[74,75]. Dissections frequently occur at the junction between soft tissueand calcium due to the di erent stress modulus of thesetwo plaque components[76,77].

Rotational ablation can successfully remove subendothelial calcium and create a smooth channel whichcan be further enlarged with balloon angioplasty, directional atherectomy, or stent implantation[78,79]. There isgeneral agreement among operators from high volumecentres used to perform routine pre-intervention intracoronary ultrasound, that rotational atherectomy isindicated in the presence of an area of superficial calcium greater than 180 degrees in multiple cross-sectionsalong the stenotic segment. Although the removal ofplaque calcification with excimer laser is more di cult,shattering of the calcific deposits was observed withultrasound, facilitating the subsequent angioplasty[80].

Plaque eccentricityLesion eccentricity and location ofmaximal plaque accumulation is another element ofgreat importance to guide the interventional procedure,which is only indirectly assessed with angiography. Witha direct measurement of the maximal and minimalplaque thickness, eccentricity is recognized much morefrequently than appreciated from the angiographicappearance[9,37,81]. For highly eccentric plaques, in theabsence of subendothelial calcium, directional atherectomy appears a logical choice, but the advantage of thisprocedure over balloon angioplasty and stenting in thissetting remains to be confirmed. In eccentric plaques, theorigin of side-branches from the diseased part of thevessel wall is an element predictive of occlusion afterballoon dilatation or stent implantation.

Diffuse atherosclerotic disease in vein graftsDegeneratedvein grafts are a challenge for the interventionalist since percutaneous treatment can avoid the increased risk ofsurgical reintervention but the immediate complications(distal emboli, myocardial infarction) and the long-termrestenosis are both high in this setting. Vein grafts, asnative coronary arteries, undergo a process of remodelling and compensatory enlargement which leads to anunderestimation of the di useness of the disease withangiography[82]. Intracoronary ultrasound examinationleads to extension of treatment to longer graft segments,treatment of lesions at high risk of rapid progression atthe medium term and to referral of patients to surgery inthe presence of di use vein degeneration with friableplaque (low echoreflectivity with irregular borders)[54].Various types of interventions (PTCA, extraction,directional or laser atherectomy) can also be guided withintracoronary ultrasound[83]. This is particularly usefulfor appropriate sizing of stents in these large conduits,with only 9% of the stents optimally expanded with angiography matching the reference cross-sectional area[84].

Type of vessel remodellingAs discussed in the previoussections, lumen reduction can be induced either byplaque accumulation that has exceeded the capacity ofthe vessel to remodel, or by failure to remodel in thepresence of a small or moderate plaque burden. Thesetwo conditions cannot be recognized by angiographybut are readily distinguished with ultrasound and mayrequire a di erent therapeutic approach. Pasterkampet al. observed a di erence in the mechanism of lumenenlargement after balloon angioplasty in these two typesof lesions, with a similar final increase in lumen areaafter balloon angioplasty[35]. Although other studies areneeded to establish the optimal treatment for these twolesion types, it is conceivable that stenoses due tonegative remodelling might mainly require expansion ofthe total vessel area, possibly using a stent to avoid acuteor chronic vessel recoil, while in lesions with a largeplaque burden partial plaque removal would facilitatethe lumen expansion with adjunctive balloon dilatationor stent implantation.

Intracoronary ultrasound during balloon angioplasty

Modifications of the dilatation strategy based on intracoronary ultrasound results include changes in balloonsize and inflation pressure. Occasionally a lesion lengthgreater than that expected from the angiographic imagemay suggest the use of a long balloon. Di use lesioncalcification certainly requires higher inflation pressuresand carries a higher risk of dissection[73–75,86], but thelength and circumferential extent of calcification atwhich the risk of an unsatisfactory result after plainballoon angioplasty is so high as to prompt the use ofalternative techniques such as rotational atherectomy isnot yet defined[85,86]. Plaque area reduction is the majorcause of lumen gain in unstable angina and acutemyocardial infarction suggesting that compression,redistribution or dislodgement of mural thrombusoccurs in acute coronary syndromes[24,86]. With intracoronary ultrasound the selection of the balloon size canbe based on measurements of the total diameter of thevessel. The CLOUT Study Group has proposed upsizingthe diameter of the balloon based on the calculation ofthe mid-wall diameter and has reported initial favourable results in terms of increase in lumen area without ahigher incidence of complications[87]. The e cacy of thisstrategy has been confirmed in a recent preliminaryreport, showing a large lumen gain and low targetlesion revascularization (17%) using a balloon diameterequal to the total vessel diameter and high pressuredilatation[88].

The most important information obtained withintracoronary ultrasound concerns the results of theprocedure. After angioplasty, intracoronary ultrasoundcan detect circumferential and longitudinal extension ofplaque fracture or dissection[44,45,89](Fig. 3). Althoughthe angiographic presence of dissections increases therisk of in-hospital complications after angioplasty, theyoccur in only 5% of the stenoses with angiographic signsof dissection after balloon angioplasty. Intracoronaryultrasound has the potential to more accurately detectdissections at risk which require immediate further treatment. Although depth and circumferential extension ofthe dissection appear the most relevant parameters tobe considered, firm intracoronary ultrasound predictorsof complications have not yet been established andprobably require the integration of longitudinal andcircumferential measurements with three-dimensionalintracoronary ultrasound[90].

Since the angiographic parameters, includingquantitative angiographic measurements are poor predictors of the long-term result after balloon angioplasty[91,92], the best application of intracoronaryultrasound after balloon angioplasty is the detection oflesions at high-risk of development of restenosis at thetime of the initial procedure. This would allow theoperator immediately to perform further interventionsto improve long-term outcome. Preliminary studies haveshown conflicting data concerning the factors predictive of restenosis after balloon angioplasty, indicatingthat absence of plaque fracture or conversely largedissections are prognostic markers of restenosis[44,93].

In 200 patients studied with intracoronary ultrasound after the final balloon inflation in the PICTUREstudy (Post Intra-Coronary Treatment UltrasoundRestenosis Evaluation)[94], no correlation was foundbetween ultrasonically identified lesion composition,fracture or dissection after dilatation and quantitativemeasurements of the lumen and plaque after intervention and clinical and angiographic results. More encouraging results have been reported by Mintzet al. whohave studied lesions after transcatheter interventionsand tested the predictive value of multiple angiographicand ultrasonic parameters for restenosis[95]. With multivariate analysis, these authors found that the residualplaque burden measured with intracoronary ultrasoundwas an independent predictor of restenosis. Plotting theresidual plaque burden vs the probability of restenosis at6 months, a curvilinear relationship was observed, withrestenosis occurring in more than 50% of the lesionswith more than 70% residual plaque burden immediately after intervention[96]. The GUIDE II trial (Guidance byUltrasound Imaging for Decision Endpoints) is a multicentre study assessing the factors predictive of restenosisafter balloon angioplasty and directional atherectomybased on a final intracoronary ultrasound examinationwith the operator blinded to the intracoronary ultrasound results[97]. The first interim analysis performedwithout discriminating between balloon angioplasty anddirectional coronary atherectomy has shown that twointracoronary ultrasound parameters were predictive oflong-term recurrence of symptoms: residual plaque burden and luminal cross-sectional area after intervention.Similar intracoronary ultrasound parameters (residualplaque burden and minimal luminal diameter measuredby intracoronary ultrasound but not by angiography)were found to be predictive of restenosis after PTCA ina large single-centre study including 89 patients[98].

The final analysis of these trials will showwhether intracoronary ultrasound provides sensitive andspecific predictors of restenosis after balloon angioplasty, justifying its more widespread application duringPTCA. There is little doubt, however, that intracoronaryultrasound unmasks pseudosuccessful angiographicresults in which the lumen area increase is only due tocircumferential dissections filled with contrast. Di usehaziness or intraluminal defects in the treated segmentare suggestive of the presence of these suboptimal resultswhich are likely to be associated with a high risk ofpersistence or recurrence of symptoms. In these selectedcases, especially in large vessels, it is the routine practiceof many centres to use intracoronary ultrasound toconfirm the need for further interventions.

Directional atherectomy

The direct visualization of the quadrants of maximalplaque accumulation has great potential for guidanceof interventions directed to selective plaque removal.Unfortunately, until now combined ultrasound–atherectomy devices are not in current clinical use andonly prototypes have been tested, providing imageslimited to the quadrant towards which the cutter isoriented[99].

The orientation of the atherectomy cutter basedon images obtained in a separate preliminary insertionof the ultrasound catheter is cumbersome. An angiographically visible side-branch close to the lesion mustbe identified and the arc between this branch and theradiant of maximal plaque accumulation defined. Afterwards, using angiography the cutter is positioned pointing to the side-branch identified and appropriaterotation of the cutting catheter is performed[100]. Morerecently, a technique which allows both a proper orientation of the atherectomy cutter and a complete plaqueremoval has been proposed[101]. The initial ‘referencecut’ is performed based on the angiographic image andthen imaged by intracoronary ultrasound. The orientation of the cutter is maintained within the hemisphereof the ‘reference cut’ until complete plaque removal hasbeen achieved. If necessary (concentric plaques), thecutter is then turned 180 degrees from its initial positionand cutting is performed until also the opposite hemisphere is appropriately treated. During atherectomy,serial ultrasound examinations are performed: beforeintervention to confirm the appropriateness of theindication (absent or deep calcification, short stenosis,ideally soft plaques); between subsequent atherectomypasses to assess the completeness of plaque removal andavoid deep cuts in the periadventitial tissue; afteratherectomy to determine the need and e ect of adjunctive balloon dilatation or stent implantation to be usedto tackle the flaps and smooth the irregular wall contours often induced by the atherectomy cuts. It iscommon experience that the use of intracoronary ultrasound during atherectomy results in a more aggressivestrategy and leads to greater plaque removal and a largerlumen diameter[102]. These experiences suggest thatintracoronary ultrasound helps to overcome some of thelimitations of the angiographically-guided directionalatherectomy, reflected by its inability to provide aclinically relevant reduction in restenosis as documentedin randomized multicentre comparisons with balloondilatation[103,104]. The results of the OARS trial (OptimalAtherectomy Restenosis Study)[105]confirm the possibility of achieving an improved immediate result(residual angiographic diameter stenosis 8&11%) and ahigh procedural success with aggressive ultrasoundguided atherectomy. Despite these ‘stent-like’ immediateresults, however, the angiographic restenosis rate was29%, with subsequent target vessel revascularization ormajor ischaemic cardiac events in 20% of the patients.More encouraging are the results reported in a Japanesemulticenter trial (ABACAS) showing that plaque removal, more complete than in the OARS trial, (45% vs57% residual plaque burden) results in a reduction inangiographic restenosis rate (21%)[106].

Stent implantation

Serial intracoronary ultrasound examinations haveshown, with slight di erences among the di erentinvestigators[36,107–110], that a late reduction in totalvessel area (chronic negative remodelling) is an important mechanism of restenosis after PTCA and DCA.These observations explain why coronary stenting isable to reduce the restenosis rate in comparison withPTCA[111,112]. In these large multicentre trials, however,stent implantation was associated with a high incidenceof subacute thrombosis, especially for implantation asbail-out after PTCA.

Intracoronary ultrasound had an essential role indeveloping an optimal strategy for stent deployment.The demonstration that incomplete apposition of thestent struts to the vessel wall, residual lumen narrowingor irregular eccentric lumen in the stented segment werestill present in 88% of the cases with an optimal angiographic result suggested that the poor technique ofimplantation rather than the inherent stent thrombogenicity was responsible[113](Fig. 4). This promptedoperators to develop a more aggressive stent implantation strategy based on high-pressure balloon dilatationinside the stent[114–117]

The guidelines of the Milano group[116]havebeen modified and simplified by the MUSIC Investigators (Multicenter Ultrasound guided Stent Implantation in the Coronaries) (Table 5). These guidelinesare based on a comparison between lumen inside thestent and lumen of the proximal and distal referencesegment. Although the strict criteria proposed cannotbe achieved in all cases of stent implantation andthey do not need to be completely fulfilled to prevent subacute thrombosis, they should be consideredas an ideal goal to be reached. With this approach,the low incidence of subacute thrombosis of theMilan experience has been confirmed (personalcommunication).

With the consistent use of a strategy of highpressure dilatation but without intracoronary ultrasound guidance, a subacute thrombosis rate of 1·6% hasbeen reported in 1156 patients treated with a combination of aspirin and ticlopidine[118]. The improveddiagnostic accuracy of angiography after high pressuredilatation is explained by a recent report showing grossover-estimation of lumen diameter with quantitativeangiography when the stent is deployed at low pressurebut a progressive improvement in the correlationbetween intracoronary ultrasound and angiographicallymeasured diameters after high pressure dilatation[119].Similar conclusions have been reported by the Essengroup in a retrospective analysis of the comparisonbetween intracoronary ultrasound and angiography inthe early experience (deployment at low pressure) and inthe most recent cases treated with high pressure balloondilatation[120]. Although the subacute thrombosis rate ofthe French Registry and of other groups not usingintracoronary ultrasound guidance for stenting isbetween 1 and 2%[118,121], these data must be comparedto the even lower percentages of subacute thrombosis incentres using intracoronary ultrasound-guided implantation (Columbus Milan 0·9%[117], Cleveland ClinicRegistry 0%[122], APLAUSE trial in Washington HeartCenter (Anti PLAtelet treatment After UltrasoundGuided Stent Evaluation) 0·4%[123]). These di erences,although small, appear clinically relevant, especially ifthe extreme complexity of the lesions treated in thesetertiary referral centres is considered, suggesting thatthere is a price to pay in terms of additional immediatecomplications when the intracoronary ultrasoundguidance is abandoned. Unfortunately, it will be verydi cult to organize a trial su ciently large to demonstrate statistically the significance of this small di erence. In the French Registry small arteries, oftencorresponding to di use disease with ultrasound, bailout scenting, unstable syndromes and low operatorexperience were independent predictors of complicationsand subacute thrombosis[124]. These data suggest thatintracoronary ultrasound guidance can be more important and cost-e ective in these specific clinical andanatomical conditions. In practice, a possible compromise weighting risk of subacute thrombosis andcost of intracoronary ultrasound might be that intracoronary ultrasound is not applied to elective stentimplantation in short discrete stenoses in angiographically normal vessels when an optimal angiographicresult is obtained with high pressure dilatation. On thecontrary, in complex procedures for long stenoses ordissections, if the implantation of multiple stents isrequired or when there are still doubts angiographically after high pressure dilatation concerningadequacy of stent expansion and presence of edgelesions, intracoronary ultrasound is recommended. Thecombination of ultrasound probe and balloon in thesame catheter[125]or the immediate assessment ofthe appropriate and symmetrical expansion ofthe balloon using an intracoronary ultrasoundguidewire[126]are innovative approaches to facilitatethe application of intracoronary ultrasound duringroutine procedures of stent implantation.

Having learned from intracoronary ultrasoundhow a stent must be implanted to avoid subacutethrombosis, the new challenge is represented by thereduction of restenosis after stenting. Ultrasound canfacilitate the optimization of stent expansion since it canhelp to:

(1) identify the length of the diseased segment toavoid significant residual stenosis or dissection atthe edges of the stent after high-pressure balloondilatation; these axial measurements are highly facilitated using three-dimensional intracoronary ultrasound[127,128];

(2) detect presence and extension of plaque calcification,an important factor limiting stent expansion[129]whichcan be treated with rotational atherectomy (Fig. 5);

(3) guide selective interventions of plaque removal(DCA) before stent implantation to avoid plaqueprolapse or shift and reduce vessel stretch;

(4) guide and confirm the achievement of an optimallumen gain in the stented segment.

Despite the improvement in immediate results,the risk of stent restenosis still remains clinicallyrelevant, with a higher risk in long stenoses and in smallvessels[130]. In the Milan experience, despite the highincidence of restenotic lesions, long lesions, small vesselswith di use disease and total occlusions, the overallangiographic restenosis rate was low (21% using the 50%diameter stenosis criterion at 6 months) and repeatedinterventions were required in 13% of the patients[116]. Ina matched comparison of 346 lesions treated withangiographic guided and ultrasound guided stent implantation, both performed with high-pressure balloonexpansion, a 57% reduction in target lesion revascularization was observed in the ultrasound group[131]. Asignificant increase in post-stent cross-sectional lumenarea and a 40% decrease in target lesion revascularization has been recently reported in the ultrasoundguided group of the CRUISE trial. Randomized comparative studies (see Table 6) are targeted to a reductionof the restenosis rate in the high-risk group (small vessels, long lesions) in which restenosis still represents amajor factor limiting the application of percutaneousinterventional treatment modalities.

Intracoronary ultrasound will also become theultimate method to test the e cacy of di erent strategiesof prevention of intimal hyperplasia (local or systemicpharmacological treatment, radioactivity, etc), measuring the volumetric plaque increase inside thestent[133,134]. Ultrasound has shown that the mechanismof restenosis after stent implantation is not a chronicmechanical recoil as in other types of interventions butthat intimal hyperplasia or plaque protrusion explain thelumen reduction within the stent[135,136].

Conclusions

Although thousands of patients have successfullyundergone intracoronary ultrasound studies forguidance of coronary interventions (Table 6), theabsence of randomized studies precludes a definitiveidentification of role and clinical usefulness of thistechnique during interventions. Intracoronary ultrasound can be used to solve selected diagnosticproblems and is a standard method for in vivo quanti-fication of intimal proliferation in cardiac allograftrecipients and to follow progression of coronary disease.

The importance of intracoronary ultrasoundfor guidance of ballon selection during PTCA requiresfurther studies and its application is currently limited tothe identification of suboptimal results and complications. Intracoronary ultrasound can be used to optimizethe results of alternative treatment modalities, guidingselective plaque removal during directional coronaryatherectomy and assessing the adequacy of stent expansion. For these indications and in selected subsetsof patients, data obtained in experienced intracoronaryultrasound centres suggest that this technique canimprove the immediate result of the intervention andreduce in-hospital complications and late restenosis.

The support of the European Community to some of theMeetings required to write these consensus guidelines is gratefullyacknowledged (BIOMED II). The expert secretarial assistance ofMrs Ornella Tramontano and Miss Elena Rocca are gratefullyacknowledged.