Coronary Intravascular Ultraso

Coronary Intravascular Ultrasound Imaging: Introduction

After more than a decade of continuous development, coronary intravascular ultrasound has achieved general acceptance as an essential element in all contemporary catheterization laboratories. Although angiography continues to serve as the primary imaging modality used to assess the anatomy of coronary artery disease, intravascular ultrasound represents an important alternative method for examination of the coronaries during diagnostic or interventional catheterization.1–8 Studies comparing angiography and intravascular ultrasound have demonstrated important differences in quantitative and qualitative findings.7–11 Unlike angiography, which portrays the vessel as a silhouette of the lumen, intravascular ultrasound provides tomographic images that depict not only the lumen but also the deeper intramural structures within the vessel wall. The ability of ultrasound to penetrate and image soft tissue enables direct visualization of the atheroma, providing insights into the pathophysiology of coronary disease not obtainable by any other technique. Accordingly, intraluminal ultrasound imaging is now commonly utilized to confirm, refute, or supplement angiographic data in patients with coronary disease.8 Recently, because of its ability to measure atherosclerosis progression and regression, intravascular ultrasound has been increasingly employed in clinical trials. Ongoing studies using this methodology offer the opportunity to develop new antiatherosclerotic therapies, supplementing traditional long-term morbidity and mortality trials.

Rationale for Intravascular Ultrasound

Limitations of Angiography Visual interpretation of angiograms is associated with significant observer variability, and necropsy examination is often discordant with the apparent angiographic severity of lesions.12–18 In comparison to postmortem evaluation, angiography often significantly underestimates the extent of atherosclerosis.13,18 Angiographic assessment of lesion severity is strikingly discordant with measurements of the physiologic effects of stenoses.19 Angiography depicts coronary anatomy from a planar two-dimensional silhouette of the contrast-filled lumen. However, coronary lesions are often complex, with markedly distorted or eccentric luminal shapes, and mechanical interventions (other than stenting) exaggerate luminal eccentricity by fracturing or dissecting the atheroma.9,20,21 The angiographic appearance of the postintervention vessel often reveals an enlarged but "hazy" lumen. This indistinct, broadened angiographic silhouette may overestimate actual vessel diameter and misrepresent the gain in luminal size.21 The traditional method for characterizing angiographic lesion severity depends on visual or computer measurements of the percentage stenosis. This process requires comparison of luminal dimensions within both the lesion and an adjacent, uninvolved "normal" reference segment. However, necropsy studies demonstrate that coronary disease is frequently diffuse and contains no truly normal reference segment.18 In the presence of diffuse disease, calculation of percent stenosis will predictably underestimate disease severity. Diffuse, concentric, and symmetrical disease affecting the entire vessel may result in the angiographic appearance of a small but normal artery.21 Angiography is also confounded by the phenomenon of coronary "remodeling," observed histologically as the outward displacement of the external vessel wall in segments with atherosclerosis.22 This adventitial enlargement attenuates lumen encroachment, thereby concealing the presence of the atheroma on angiography. Although such lesions do not restrict blood flow, clinical studies have demonstrated that these minimal, nonobstructive lesions represent an important cause of acute coronary syndromes.23 Angiographically unrecognized disease virtually always underlies an ergonovine-positive response in symptomatic patients with a "normal" coronary angiogram.24 Theoretical Advantages of Ultrasound Intravascular ultrasound has several unique properties of theoretical value in the detection and quantitation of coronary disease.25,26 The cross-sectional perspective of ultrasound permits visualization of the full 360-degree circumference of the vessel wall. Accordingly, measurement of luminal area can be determined by planimetry, independent of the radiographic projection or magnification.7,21,25,26 The tomographic perspective of ultrasound enables evaluation of vessels difficult to assess by angiographic techniques, including diffusely diseased segments and bifurcation or ostial lesions. The ability to directly image the atheroma within the vessel wall represents a truly unique capability not possible using any other commonly available imaging modality.

Imaging Technology

Catheter Design Intracoronary ultrasound equipment consists of two major components: a catheter incorporating a miniaturized transducer and a console containing the electronics necessary to reconstruct the image. High frequencies (20 to 50 MHz) are employed, resulting in excellent theoretical resolution (axially <100 m and laterally <250 m). Two dissimilar technical approaches to transducer design exist: mechanically rotated devices and multielement electronic arrays.1–5 Each design has yielded small intravascular devices suitable for coronary imaging, typically ranging in size from approximately 2.6 to 3.5F (diameter of 0.86 to 1.17 mm). To facilitate subselective coronary cannulation and catheter exchanges, ultrasound catheters provide a lumen for a movable guidewire. Most systems generate images at a temporal frequency of 30 frames per second for recording on videotape. Limitations and Artifacts Intravascular ultrasound devices generate artifacts that may adversely affect image quality, alter interpretation, or reduce quantitative accuracy (Fig. 18–1).27 Ring-down artifact arises from acoustic oscillations in the piezoelectric transducer, resulting in high-amplitude signals that preclude imaging close to the transducer surface. Accordingly, the "acoustic" size of catheters is slightly larger than their physical size. Since the minimum size of current devices is approximately 0.9 mm, some severe stenoses cannot be imaged prior to intervention. Geometric distortion can result from imaging in an oblique plane (not perpendicular to the long axis of the vessel), resulting in an elliptical rather than circular imaging plane.28   Figure 18–1. Intravascular ultrasound artifacts. In the left panel, there is an example of nonuniform rotational distortion (NURD) with circumferential "stretching" of the image from 8 to 10 o'clock (arrow). The right panel shows an example of ring-down artifact. Mechanical, but not electronic, transducers may exhibit cyclical oscillations in rotational speed, resulting in an artifact known as nonuniform rotational distortion (NURD).27 This artifact arises from mechanical friction within the catheter drive shaft during the portions of its rotational cycle. This speed variation produces readily visible distortion, often observed as circumferential stretching of a portion of the image with compression of the contralateral vessel wall (Fig. 18–1). NURD is most evident when the drive shaft is bent into a small radius of curvature by a tortuous vessel. Improvements in the mechanical precision of ultrasound devices have reduced the impact of the artifact, but it still remains troublesome during some examinations.

Coronary Imaging

Examination Technique Standard interventional techniques for intracoronary catheter delivery are used for intraluminal ultrasound examination. Intravenous heparin [to maintain activated clotting time (ACT) >200 to 250 s] and intracoronary nitroglycerin (100 to 300 g) are routinely administered, although there are no controlled studies documenting the necessity for anticoagulation. Using a 6- or 7F guiding catheter, the operator advances a steerable guidewire to subselectively cannulate the vessel. A stable guiding catheter position with good support is desirable, since current ultrasound catheters have less trackability and a larger profile than modern angioplasty equipment. The operator carefully advances or retracts the imaging catheter over the wire to examine the vessel in real time, recording images on videotape for subsequent quantitative or qualitative analysis. Side branches, visualized with both angiography and ultrasound, are often used as landmarks to facilitate interpretation. Some practitioners advocate use of a uniform format, electronically rotating the ultrasound image so that branches appear in a standardized orientation. For example, imaging of the left anterior descending is often performed with the septal branches at 6 o'clock and the left circumflex appearing at about 9 o'clock. Some centers use a motorized pullback device to withdraw the catheter at a constant speed (between 0.25 and 1 mm/s, but most often 0.5 mm/s). In clinical practice, motorized pullback is often used to survey the coronary prior to more prolonged and thorough examination of sites of interest. However, in studies of atherosclerosis progression versus regression, the motorized pullback is an integral feature of the investigative procedure. In this application, during analysis of the motorized pulback, "slices" are selected at regular intervals (typically every 1 mm) and subsequently analyzed in a core laboratory. By comparing atheroma burden at a baseline examination with a similar study at follow-up, the extent of disease progression or regression can be precisely characterized. Safety of Coronary Ultrasound Although intravascular ultrasound requires intracoronary instrumentation, studies have demonstrated few serious untoward effects.29–31 The most frequently encountered complication is focal coronary spasm, which usually responds rapidly to intracoronary nitroglycerin. Data from European centers report a 1.1 percent complication rate in 718 ultrasound examinations.30 Another report from 28 centers (2207 studies) documents spasm in 2.9 percent and major complications, such as occlusion or dissection, judged to have a "certain relation" to instrumentation in 0.4 percent.29 In both studies, complications (spasm, vessel dissection, or guidewire entrapment) occurred in patients undergoing angioplasty rather than diagnostic imaging. In 170 cardiac transplant recipients (240 studies), there was no morbidity, but spasm occurred in 20 patients (8.3 percent) despite pretreatment with nitroglycerin.31 Any intracoronary instrumentation carries the potential risk of intimal injury or vessel dissection. Accordingly, most laboratories limit credentialing for this procedure to personnel with interventional training. Normal Coronary Anatomy Studies performed either in vivo or using excised, pressure-distended vessels have characterized the appearance of normal coronaries by intravascular ultrasound.32–36 Important determinants of vessel wall appearance include both the normal arterial structure and the inherent properties of ultrasound. An ultrasound reflection occurs at a tissue boundary whenever there is an abrupt change in acoustic impedance. Normally, two strong acoustic interfaces are visualized by ultrasound, the leading edge of the intima (at the interface between the blood-filled lumen and the endothelium) and the outer border of the media (at the junction of media and external elastic membrane). Underlying the trailing edge of the intima, a middle sonolucent layer is usually evident, which is composed principally of the tunica media. The echodense intima and adventitia with a sonolucent medial layer often give the wall a trilaminar appearance. However, this pattern is not a universal finding; in 30 to 50 percent of normal segments, a thin intimal layer reflects ultrasound poorly, which results in a monolayer appearance (Fig. 18–2).35 In a necropsy study, the ultrasound-derived intimal thickness in segments with three layers was significantly greater than for monolayered sites (0.24 ± 0.1 vs. 0.11 ± 0.06 mm, p <0.001). The mean age in the three-layered group was greater, 42.8 ± 9.8 versus 27.1 ± 8.5 years (p <0.001).37 Other studies demonstrate that a trilaminar appearance is dependent not only on the age but also on the histologic characteristics of the vessel. A three-layered appearance is consistently observed if an internal elastic membrane is present.38 However, if an internal elastic membrane is absent, a trilaminar appearance is observed only when the collagen content of the media is low. In older "normal" subjects, intimal thickening usually results in a pattern of two distinct echogenic layers "sandwiching" a sonolucent intermediate layer. In nearly all cases, the deepest arterial layers exhibit a characteristic onionskin pattern, representing the adventitia and periadventitial tissues with an indistinct outer vessel border (Fig. 18–2). In both normal and abnormal arteries, the lumen exhibits faint, finely textured, swirling echoes that arise from acoustic reflections from circulating blood elements. This blood "speckle" may assist image interpretation by providing a means to confirm the communication between dissection planes and the lumen. The pattern of blood speckle is dependent on the velocity of flow, showing increased intensity and a more coarse appearance when flow is reduced. In some cases, the coarse blood speckle can mimic the appearance of tissue, complicating image interpretation. The physical presence of the ultrasound catheter may exacerbate this problem, particularly if there exists a stenosis with relatively severe narrowing proximal to the imaged site. In such cases, the reduction of blood flow produced by partial obstruction of the proximal narrowing, may result in reduced blood flow at the imaged site, and the resulting coarse blood speckle may be misinterpreted at tissue protruding into the lumen.   Figure 18–2. Two variants of normal coronary anatomy by intravascular ultrasound. In both images, a magnified view of the area contained within the rectangle is shown at the top. In the left panel, there is a monolayered artery; in the right panel, the artery has a trilaminar structure.

Characterization of Atherosclerosis

Atheroma Composition The subtle changes that occur early in the development of atherosclerosis, such as fatty streaks, are not visible using current ultrasound devices. Atherosclerotic arteries exhibit a variety of features that reflect the distribution, severity, and composition of the atheroma.32–34 Sites with limited disease exhibit generalized or focal thickening of the intimal leading edge, while advanced lesions appear as large echogenic masses encroaching on the lumen. A comparative study of ultrasound and histology in 1100 fresh necropsy sections demonstrated that lipid-laden lesions are usually hypoechoic.32 Soft, low-intensity echoes most often represent fibromuscular lesions and very bright echoes are characteristic of dense fibrous or calcified tissues (Fig. 18–3). In highly echogenic plaques, areas of calcification are recognized by obstruction or severe attenuation of ultrasound penetration, which obscures deeper layers, a phenomenon known as acoustic shadowing. In lipid-laden or fibromuscular lesions, a prominent echogenic overlying "fibrous cap" may be observed.   Figure 18–3. Atheroma morphology by intravascular ultrasound. In the left panel, a large, "soft," lipid-laden atheroma with a thin fibrous cap is seen (arrows). It is eccentric, involving only about 50 percent of the vessel wall. The right panel shows a circumferential atheroma with an area of focal calcification is evident (arrow). The echogenicity of the plaque components is dependent not only on the acoustic properties of tissue, but also on the acquisition settings (gain, compression, etc.). Accordingly, most morphologic classification schemes compare the echogenicity of the plaque to the surrounding adventitia to adjust for differences in ultrasound technique. However, in plaques containing a zone of reduced echogenicity, it is not possible to determine whether these represent areas of lipid deposition, thrombus, or necrotic degeneration, all of which can appear as zones of low density. Plaque composition was accurately predicted by ultrasound imaging in 96 percent of 112 quadrants from 21 freshly explanted human coronary arteries.38 Fibrous and calcified plaque quadrants were correctly identified in almost all cases (100 of 103, or 97 percent), but only 7 of 9 quadrants (78 percent) with predominantly lipid deposits were correctly identified. Accordingly, some caution is warranted in the intravascular ultrasound classification of atheroma composition. Although currently available devices produce detailed views of the vessel wall, interpretation employs visual inspection of acoustic reflections to impute morphology. Different histologic features may exhibit comparable acoustic properties, and well-validated methods for objective or automated classification of atheromatous lesions do not yet exist. Thus, intravascular ultrasound can delineate the thickness and echogenicity of vessel wall structures, but this technique does not provide actual histology. More recently, sophisticated image processing techniques have been employed to classify the composition of atherosclerotic plaques. The most promising methods employ analysis of raw radiofrequency data from ultrasound studies to assess the morphology of the atheroma.39 Using necropsy specimens, some recent investigations have demonstrated that automated methods for characterization of atheroma morphology are more accurate and reproducible than methods based on simple visual inspection.40 By plotting morphologic characteristics as a color overlay superimposed on the ultrasound image, such methods facilitate relatively facile image interpretation. Accordingly, such methods are likely to achieve broader acceptance within the next few years. Detection of Calcification Ultrasound imaging is more sensitive than fluoroscopy or angiography for the detection of coronary calcification. In a series of 110 patients undergoing intervention, target lesion calcification was detected by ultrasound and fluoroscopy in 84 and 50 patients, respectively (76 vs. 14 percent, p <0.001).41 Another retrospective study analyzed calcification by angiography and ultrasound in 183 interventional patients.42 Assessment by the two techniques was concordant in 92 and discordant in 91 cases. Calcification was detected in 138 patients by ultrasound and 63 by fluoroscopy, showing a sensitivity and specificity for angiography of 46 and 82 percent, respectively. When calcium was detected angiographically, calcification by ultrasound often subtended >90 degrees and was superficial to the lumen in location. If no calcification could be visualized on the angiogram, the chance of detecting a large superficial arc of calcium by ultrasound was low (12 percent). Ultrasound calcification is an important determinant of the arterial response to intervention, portending a greater risk of dissection following balloon angioplasty, less tissue retrieval with directional atherectomy, and potentially greater benefit with the use of rotational atherectomy.43,44 Most classification schemes quantify the extent of calcification, usually by measuring the circumferential angle subtended by calcified plaque.41 Commonly, the axial length of the calcified portion of the lesion is also reported. The depth of calcification is also assessed, described as superficial when the calcium remains in contact with the luminal surface and deep if no portion of the calcium deposit is superficial. During initial development of intravascular ultrasound, the extent of calcification was often utilized in the selection of interventional devices, particularly as a means to select vessels suitable for directional coronary atherectomy. However, in recent years, the reduced importance of directional atherectomy and the nearly universal application of coronary stenting has lessened the importance of calcification as a determinant of the interventional approach. Accordingly, in current practice, intravascular ultrasound is uncommonly performed solely as a means to detect coronary calcification. Arterial Remodeling The term arterial remodeling refers to a change in arterial dimensions associated with the development of atherosclerosis. In a necropsy study of 136 human left main coronary arteries, Glagov et al. originally described focal arterial enlargement at atherosclerotic sites, reporting a positive correlation between external elastic membrane (EEM) area and the area occupied by atheroma (r = 0.44, p <0.001).22 At sites with area stenosis less than 40 percent, the increase in arterial size "overcompensated" for the plaque deposition, leading to an increase in absolute lumen area. With more advanced lesions (area stenosis >40 percent), the degree of arterial enlargement or remodeling was blunted, resulting in a smaller lumen area. The authors hypothesized that this phenomenon represented a compensatory mechanism to preserve lumen size. The findings of Glagov et al. were later confirmed in vivo by intravascular ultrasound imaging (Fig. 18–4).45,46 In 80 ultrasound cross sections obtained from 44 patients undergoing coronary interventions, EEM area correlated closely with plaque area (r = 0.79, p = 0.0001). In this study, lumen area increased with early atherosclerosis, confirming the phenomenon of overcompensation in early stages of the disease. With more advanced atherosclerosis, there was a correlation between increasing area stenosis and decreasing lumen area (r = 0.58, p = 0.0001).45 Compensatory enlargement has also been demonstrated by ultrasound in superficial femoral arteries; however, there was no difference between lesions less than, and greater than 40 percent stenosis.47   Figure 18–4. Example of coronary remodeling. The left upper panel shows a normal segment of the circumflex coronary. In the right upper panel, an atherosclerotic segment of the coronary a few millimeters proximal to the normal segment is shown. In the bottom two panels, measurements taken at each of the sites show very similar cross-sectional areas. The preservation of luminal area results in a coronary angiogram that is normal despite the presence of a large atherosclerotic plaque in the involved segment. In recent years, ultrasound studies have demonstrated a new dimension to arterial remodeling, the phenomenon of "negative" remodeling.48,49 At diseased sites, the EEM area may actually be reduced in size, contributing to luminal narrowing, rather than compensating for it. In 51 femoral arteries, EEM area was smaller at lesions than adjacent reference sites, with a negative correlation between stenosis severity and EEM area reduction (r = 0.62 by histology and 0.66 by ultrasound, p <0.001 for both).48"Inadequate" remodeling, defined as an EEM area within the lesion less than 78 percent of a proximal reference site, has also been described in the coronaries of patients with stable angina.49 Although 91 of 603 lesions (15 percent) fit this definition, there was a highly variable response among lesions within the same patient. However, when remodeling is defined in this fashion, there is an assumption that the reference EEM area represents the original vessel size, which may not be correct, since angiographic reference sites are frequently diseased by ultrasound. Although the exact mechanisms of compensatory or negative remodeling remain unclear, these phenomena have important clinical implications. Compensatory remodeling represents an important factor in the underestimation of the severity of atherosclerosis by angiography. Thus, a vessel site may contain a very large atheroma but minimal stenosis if outward remodeling of the EEM has "compensated" for the plaque accumulation. Recent evidence suggest such sites may be particularly prone to plaque rupture. Because remodeling can affect both the lesion and adjacent reference segments, remodeling may also influence the estimation of the vessel size during coronary interventions. Recently, negative remodeling has been implicated in restenosis following atherectomy and balloon angioplasty.50 Unstable Plaque and Thrombi An emerging application of intracoronary ultrasound is the characterization of the atheroma associated with acute coronary syndromes (Fig. 18–5).51–56 The typical angiographic appearance of a ruptured plaque is a stenosis with an eccentric or ulcerated lumen, often with overhanging edges (Ambrose type II lesion). However, retrospective reviews of angiograms of patients performed before an episode of unstable angina often reveal minimal stenosis severity within the culprit lesion segment.53 Such studies highlight the inability of angiography to identify "rupture prone" atherosclerotic lesions. Histologic examination of unstable plaques after rupture usually reveals a lipid-laden plaque with a thin fibrous cap.51 Based on these observations, it has been postulated that the size of the lipid pool and the thickness of the fibrous cap are more important than severity of stenosis in predicting plaque rupture.54 Some intravascular ultrasound studies have suggested the presence of an echolucent atheroma within culprit lesions in patients with acute coronary syndromes. In a very limited study of 22 stable and 43 unstable angina patients, type II eccentric lesions were detected on the angiograms in 18 percent of stable and 40 percent of unstable angina patients. Echolucent plaques were more frequently observed in patients with unstable than in those with stable angina syndromes (74 vs. 41 percent, p <0.01).11 However, this finding has not been confirmed by other investigators.   Figure 18–5. Ruptured coronary plaque. In these two identical images, the anatomy of a ruptured coronary plaque is seen. There is a large lipid core with a fracture of the fibrous cap (right panel, arrow). This image was obtained a few days after hospitalization of this patient for a unstable coronary syndrome. Recent intravascular ultrasound studies have examined the relationship between remodeling and the type of clinical presentation, suggesting difference in the remodeling pattern for unstable versus stable patients.57 The culprit lesions in 76 patients with acute coronary syndromes were compared with lesions in 40 patients with stable angina. In the unstable patients, both EEM and plaque areas were significantly larger than the corresponding measurements in the stable patients (p = 0.02 for both). Positive remodeling was more prevalent in the unstable group (51 vs. 18 percent, p = 0.002) and negative remodeling more prevalent in the stable group (58 vs. 33 percent, p = 0.002). This finding provides further insight into the relationship between lesion severity and the likelihood of plaque rupture. Because angiographic studies suggested that minimal stenoses were associated with atheroma rupture, most investigators assumed that the culprit lesion represented a small plaque. However, the finding that rupture sites frequently exhibit positive remodeling suggests that such lesions are not particularly small plaques. Rather, the presence of remodeling enables the atheroma to reach a large size without compromising the lumen. The formation of intraluminal thrombi at a ruptured or fissured plaque is considered to be the hallmark of acute coronary syndromes.58 Angiographic criteria for diagnosis of a coronary thrombus, the presence of haziness, an intraluminal filling defect, and/or irregular lumen contour are not sensitive.56 Small observational studies have attempted to differentiate the ultrasound appearance of thrombus, defined as hypoechoic material projecting into the lumen with a slight synchronous pulsation and a distinct acoustic interface, from more echogenic plaque (Fig. 18–6).59 However, in vitro studies have revealed limitations in the reliability of intravascular ultrasound diagnosis of thrombi (sensitivity of 57 percent and specificity of 91 percent), considerably inferior to angioscopy (sensitivity and specificity of 100 percent).58   Figure 18–6. Thrombus within a coronary stent. In this intravascular ultrasound image, a stent is well visualized. There is a globular mass projecting into the lumen at 6 o'clock; it probably represents a large thrombus.