Coronary Intravascular

Diagnostic Clinical Applications

Quantitative Luminal Measurements A broad spectrum of therapeutic decisions hinge on assessment of coronary luminal dimensions. Accordingly, in diagnostic and interventional practice, quantitation of vascular dimensions represents a common clinical application of intravascular ultrasound. Recently, a committee of the American College of Cardiology and the European Society of Cardiology defined standards for the acquisition, measurement, and reporting of intravascular ultrasound studies.60 This standards document precisely defines the terrminology and methodology for performing intravascular ultrasound measurements and constitutes the recognized standard for both clinical and research studies. The presence of standardized nomenclature and measurement methods represents a critical advance in the wider acceptance of intravascular ultrasound. Several studies have compared luminal measurements by intravascular ultrasound and quantitative angiography.6,7 For vessels without atherosclerosis, most studies document a relatively close correlation between angiographic and ultrasonic coronary dimensions, although a few studies suggest slightly larger measurements by ultrasound.7 However, in patients with atherosclerotic arteries, most investigators report only a moderate correlation between ultrasonic and angiographic dimensions, with the greatest disparities in vessel segments with a noncircular lumen shape.5,7,10 This reduced correlation is probably explained by the irregular, noncircular cross-sectional profile of diseased vessels, which cannot be adequately measured using angiography.10 Quantitation of Atherosclerosis Analysis of intravascular ultrasound images permits quantitative measurements of the extent and severity of coronary atherosclerosis (Fig. 18–7).26 However, the inherent properties of ultrasound require utilization of different anatomic landmarks than those employed in classical histology. In all ultrasound imaging, reflections at the leading edge of any interface are located precisely at the boundary where acoustic impedance abruptly changes. However, the position of the trailing edge of any anatomic structure is determined by multiple nonanatomic factors, including ultrasound beam properties, particularly the wavelength (frequency). Thus, leading-edge measurements accurately describe the location of a boundary, whereas trailing-edge measurements are unreliable. As previously noted, strong reflections are generally produced at two locations, the leading edge of the intima and the border between the media and the external elastic membrane. The position of the trailing edge of the intima is not accurately localized in intravascular ultrasound images. Accordingly, quantitative measurements must calculate the atheroma's cross-sectional area by subtracting the area bounded by the intimal leading edge from the area enclosed by the external elastic membrane. This approach results in a slight overestimation of atheroma area (in comparison to histology) by including the area of the media within the calculation.   Figure 18–7. Boundaries for intravascular ultrasound measurements. In these two identical images, an atherosclerotic plaque is well visualized. The right panel illustrates the planimetry typically employed to measure the extent of atherosclerotic disease. Both the lumen and external elastic membrane (EEM) are measured. The atheroma area represents the difference between the EEM and the lumen areas. The area reduction is calculated as the atheroma area divided by the EEM area multiplied by 100. Normal Intimal Thickness The threshold for abnormal intimal thickness by intravascular ultrasound is controversial, particularly since the categorical classification of a continuous variable intimal thickness as normal and abnormal is inherently arbitrary. In various histologic and ultrasound studies, normal intimal thickness ranges between 0.10 and 0.35 mm, and the normal medial thickness ranges from 0.15 to 0.25 mm. In a necropsy study, normal intimal thickness not including media was age-dependent, averaging 0.21 mm in 21- to 25-year-olds, 0.22 mm in 26- to 30-year-olds, and 0.25 mm in 36- to 40-year-olds.37 In a comparative study, intravascular ultrasound measurements of the intima plus media averaged 20 percent greater than histologic measurements.61,62 Considering the histologic and ultrasound data, most clinical studies have defined threshold for coronary disease by ultrasound as a measured intimal thickness 0.5 mm.63–67 Currently, there is no well-defined threshold for normal values for other measures of atherosclerosis, such as intimal cross-sectional area. Assessment of Atheroma Burden The tomographic orientation of intravascular ultrasound represents a problem in quantifying atherosclerosis. Since each image contains information from only a thin "slice" of the vessel, global measures of atheroma burden require the integration of multiple cross sections. One successful approach to this conundrum employs a motorized device to steadily and progressively withdraw the ultrasound catheter through the interrogated vessel, typically at 0.5 mm/s. Since motor speed is kept constant, the operator can obtain a series of cross sections separated by a constant, recurring time interval (fixed distance from each other). These slices are individually measured and then summated to calculate an approximate total atheroma burden using Simpson's rule. This method is increasingly employed in clinical trials designed to assess the effect of pharmacologic agents on atherosclerosis progression or regression. The first large scale trial, the REVERSAL study (Regression of Atherosclerosis with Aggressive Lipid Lowering), will compare the effects of atorvastatin 80 mg and pravastatin 40 mg on total atheroma burden in a single coronary artery.68 This study, which randomized 655 patients to the two treatment cohorts, showed enhanced plaque stabilization in the high-dose statin group as reported by Nissen (AHA scientific session 2003). Many additional clinical trials of antiatherosclerotic therapies are ongoing. If such studies demonstrate the ability of intravascular ultrasound to determine the relative effectiveness of antiatherosclerotic therapies, the application is likely to grow rapidly in use during the coming years. Atheroma Distribution The circumferential distribution of the atheroma varies from nearly symmetrical plaques to very eccentric lesions in which the entire atheroma is located on one side of the artery. Assessed by ultrasound, the majority of plaques are eccentric, with a maximum atheroma thickness more than twice the minimum plaque thickness.69 Studies have demonstrated a poor correlation between the apparent circumferential pattern by angiography and the actual plaque distribution revealed by ultrasound examination.69 Such studies demonstrate the inaccuracy inherent in determining plaque distribution from the projected two-dimensional silhouette of the lumen (angiography). Although previously advocated as means to guide directional atherectomy, the decreased popularity of this interventional method has reduced the value of eccentricity measurement as a clinical application for intravascular ultrasound. Angiographically Unrecognized Disease In patients undergoing angiography for clinically suspected coronary artery disease, no angiographic evidence of narrowing is present in 10 to 15 percent of cases. In these patients, intravascular ultrasound commonly detects atherosclerosis at angiographically normal sites.21,25,26,70,71 Using intravascular ultrasound, atherosclerotic abnormalities were documented in 21 of 44 patients (48 percent) with suspected coronary artery disease and normal coronary angiograms.70 Combining ultrasound and functional assessment (coronary flow reserve and endothelium-mediated vasodilator response), only 36 percent of patients in this cohort were completely normal. Other studies demonstrate that, if any luminal irregularity is present by angiography, ultrasound will usually demonstrate atherosclerosis at nearly all other examined sites.21 The prevalence of atherosclerosis at angiographically normal sites confirms the finding, previously reported from necropsy studies, that coronary involvement is frequently underestimated using angiographic evaluation methods (Fig. 18–8)12,13.   Figure 18–8. Underestimation of coronary atherosclerosis by angiography. In the angiogram in the left panel, a relatively minor lesion of the left anterior descending coronary is seen (arrow). In the right panel, this lesion is depicted by intravascular ultrasound and consists of a large eccentric atherosclerotic plaque that appears much more extensive than would be suspected from the angiogram. There are several mechanisms by which angiography may underestimate the presence, extent, or severity of atherosclerosis.21 First, to detect focal narrowing, angiography relies on comparison of the interrogated site to an adjacent uninvolved segment. However, the involved vessel is often reduced in caliber along its entire length, containing no truly normal segment for comparison. The angiographer may erroneously conclude that the vessel is simply "small in caliber." Overlapping structures and mechanical limits in x-ray positioning may prevent the angiographer from obtaining optimal radiographic projections (orthogonal to the lesion). Accordingly, eccentric plaques that occupy only a portion of the vessel circumference represent an important source of false-negative angiography. At atherosclerotic sites, compensatory enlargement of the vessel wall overlying the plaque often preserves lumen diameter, resulting in false-negative angiography, because the lumen size in the involved segment is identical to that of adjacent, uninvolved segments. Finally, radiographic foreshortening can conceal short "napkin ring" lesions. For each of these mechanisms of false-negative angiography, intravascular ultrasound has been employed to confirm the presence and estimate the extent of atherosclerosis.21 However, the long-term clinical implications of angiographically unrecognized atherosclerosis remain uncertain, since no outcomes-based research has demonstrated a worse prognosis for patients with atherosclerosis detected only by ultrasound. However, several investigators have demonstrated that plaques with minimal to moderate angiographic narrowing are the most likely lead to acute coronary syndromes. Accordingly, the presence of angiographically occult coronary disease may have prognostic significance. Studies are currently under way to determine the value of ultrasound in predicting the clinical outcome in patients with angiographically unrecognized coronary disease. Prevalence of Coronary Atherosclerosis Recent intravascular ultrasound studies have demonstrated an extraordinary prevalence of coronary atherosclerosis in the general population, beginning at a relatively young age.71 In the most thorough study, intravascular ultrasound was performed in 262 transplant recipients within 31 days of transplantation to determine the prevalence of atherosclerotic disease in the donor hearts. These heart transplant donors (116 women and 146 men) had a mean age of 33 years and no known coronary artery disease. Imaging of multiple coronary segments was performed to determine the greatest and least intimal thickness in each segment for an average of 2.3 coronary arteries per patient. Assessment of 2014 sites in 1477 segments of 574 coronary arteries showed that atherosclerotic lesions, defined as an intimal thickness 0.5 mm, were present in 51.9 percent of donor hearts. Intimal thickness correlated with donor age with the prevalence of disease ranging from 17 percent for donors aged <20 years to 85 percent in those aged 50 years (Fig. 18–9). For all age groups, average intimal thickness was greater in male than female donors, although similar proportions (52 and 51.7 percent) had atherosclerosis. Coronary angiography was completely normal in 92 percent of these subjects and none of the donors aged <30 years had angiographic evidence of atherosclerosis. This study confirms previously reported necropsy findings showing a high prevalence of atherosclerosis in young adults. However, unlike earlier postmortem studies, this intravascular ultrasound report demonstrates the extent of angiographic underestimation of the phenomenon.   Figure 18–9. Atherosclerosis in a heart transplant donor. The heart was harvested from a 32-year-old woman who suffered brain death following a motor vehicle accident. In both the left circumflex coronary (left panel) and the ramus branch (right panel), there is extensive atherosclerosis. Lesions of Uncertain Severity Angiographers commonly encounter lesions that elude accurate characterization despite thorough examination using multiple radiographic projections. Difficult-to-assess sites include ostial or bifurcation lesions and moderate stenoses (angiographic severity ranging from 40 to 75 percent) in patients whose symptomatic status is difficult to evaluate. For ambiguous lesions, ultrasound provides a tomographic perspective, independent of the radiographic projection, that may permit quantification of the lesion. In two prospective series, intracoronary ultrasound changed the management strategy in approximately 20 percent of the examinations performed immediately prior to coronary intervention.73,74 In both studies, however, operator selection of patients for ultrasound examination may have resulted in an overestimation of the true impact of ultrasound imaging on clinical decision making. Angiographic assessment of left main coronary artery (LMCA) obstruction represents a particularly vexing clinical problem.14 Radiographic contrast in the aortic cusp can obscure the ostium, and "streaming" of contrast from the injection vortex can result in a false impression of luminal narrowing. The LMCA is often short in length, leaving no normal reference segment. The bifurcation or trifurcation of the LMCA into daughter branches may produce vessel overlap, thereby concealing a stenosis. Intravascular ultrasound is commonly used to quantify LMCA lesions when angiographic interpretation is uncertain.75 The technique for examination consists of subselective placement of the ultrasound transducer in the circumflex or anterior descending, followed by slow pullback to the aorta with the guiding catheter disengaged. There is no consensus regarding the threshold for critical LMCA obstruction. However, an area stenosis >50 percent or an absolute area <9 mm2 has been proposed as a threshold.75 Cardiac Allograft Disease Transplant coronary artery disease is the leading cause of death beyond the first year after cardiac transplantation, with a reported incidence of 15 to 20 percent per year.76 Although most transplant centers perform arteriograms annually for screening, these surveillance studies often fail to detect atherosclerosis prior to a clinical event.77,78 Necropsy studies have demonstrated that angiography systematically underestimates coronary atherosclerosis in transplant recipients.78 Patients may have diffuse vessel involvement that, for reasons already enumerated, conceals the atherosclerosis from the angiographer. Many active transplant centers now routinely perform intravascular ultrasound at the time of annual catheterization in all recipients. Investigations using ultrasound to detect transplant vasculopathy report a very high incidence of abnormal intimal thickening, involving 80 percent of patients at 1 year and more than 92 percent studied 4 or more years after transplantation.65–67,79–83 Recent studies have revealed two pathways to transplant-associated atherosclerosis. Some patients receive atherosclerotic plaques transmitted via the donor heart, while others develop an immune-mediated vasculopathy.67,72,81 The natural history of donor lesions after transplantation is largely unknown. Since angiography is relatively insensitive, ultrasound remains the most important method used to study the early atherosclerotic lesion. In the first year after transplantation, progression occurred in 42 percent of patients.82

Interventional Clinical Applications

Preinterventional Imaging Several studies have demonstrated that ultrasound imaging of interventional target lesions may influence the approach to therapy. In one study (313 lesions), the intended revascularization strategy before ultrasound imaging was compared with the treatment actually performed.42 In 40 percent of cases, the intended strategy was altered based on the ultrasound findings, most often ultrasound assessment of lesion composition or eccentricity (26 percent). Although there was a relatively close correlation between angiographic and ultrasonic lumen diameter (r = 0.83), a disagreement between the two methods was cited as the reason for altering the procedure in 13 percent of lesions. In another small, non-randomized study (n = 56) of ultrasound guidance of balloon angioplasty or directional atherectomy, operators reclassified lesion characteristics after ultrasound in 68 percent of patients and the therapeutic approach was modified in 48 percent. Ultrasound measurements revealed a smaller lumen diameter than expected from angiography, leading to balloon "upsizing" in 34 percent of angioplasty cases.84 Several studies have purported to show benefits of ultrasound imaging prior to implantation of coronary stents.85,86 The preliminary results of one prospective study identified vessel calcification as one of the predictors of "inadequate" stent expansion.85 For ostial lesions, ultrasound imaging is sometimes used to determine whether the lesion involves the "true" ostium or spares the most proximal few millimeters, which may assist optimal stent positioning. When stents are used to treat dissections, ultrasound may reveal involvement of a longer segment than can be appreciated by angiography. This may be particularly relevant in bailout stenting for threatened abrupt closure, where it may be preferred to cover the full length of the dissection.86 Despite promising data, reports on preinterventional imaging must be interpreted with caution. There are no prospective controlled trials demonstrating a superior outcome using an preinterventional ultrasound guidance. In most studies, patients were not randomized, allowing for bias in selection of more complex cases for ultrasound guidance, which would likely emphasize the contributions of imaging. Furthermore, cases in which the operators were unable to advance the ultrasound catheter through the lesion were systematically excluded. Finally, previous studies suggesting a benefit for preinterventional imaging were performed in an era when techniques such as directional atherectomy or rotational ablation were commonly performed. The preeminence of coronary stenting has reduced the likelihood that one of these alternative revascularization approached will be performed. Accordingly, device selection using intravascular ultrasound has diminished in importance in recent years. Imaging during Specific Interventions Balloon Angioplasty Ultrasound guidance of balloon sizing has been proposed as a means to improve procedural result and late clinical outcome for percutaneous transluminal coronary angioplasty (PTCA).87 In a study of 104 lesions, ultrasound was performed after obtaining a "satisfactory" angiographic result and revealed remodeling at the lesion or extensive plaque within the reference segment in 73 percent of the cases. In this subset, the balloon-to-artery ratio was increased from 1.12 ± 0.15 to 1.30 ± 0.17 (p <0.0001) and the resulting angiographic minimal lumen diameter increased from 1.95 ± 0.5 to 2.21 ± 0.5 mm. Ultrasound lumen area improved from 3.16 ± 1.0 to 4.52 ± 1.1 mm2 (p > 0.0001). Following ultrasound-guided balloon upsizing, the incidence of angiographic dissection was not increased (37 vs. 40 percent, p = 0.67). However, the study was too small to demonstrate any effect upon intermediate or long-term clinical restenosis rates. Intravascular ultrasound studies have evaluated the mechanisms of luminal enlargement following balloon angioplasty. Prior necropsy studies in patients who expired shortly after balloon angioplasty have described plaque fracture or disruption as the most common mechanism of dilatation.20 Most ultrasound studies have confirmed that dissection is an important mechanism of luminal enlargement, occurring in 40 to 80 percent of patients.9,41,88–92 Identification of dissection or fracture is based on the visualization of blood flow in the newly created lumen, sometimes aided by injection of saline or iodinated contrast to opacify the lumen via microbubbles. Wall disruptions can be further defined by measuring the circumferential extent, length, and/or maximal depth of the dissection. One small study reported that calcified lesions had a higher incidence of dissection (67 vs. 25 percent, p = 0.03) with a trend toward restenosis in lesions with no dissection.41 Following iliac artery angioplasty, ultrasound evidence of dissection was noted in all 40 cases, accounting for 72 percent of the total lumen gain.89 Several alternative mechanisms for luminal enlargement not discernible by angiography have been identified using ultrasound, including arterial wall stretching and plaque compression, or "axial redistribution."93–95 The contribution of vessel stretch to lumen gain following balloon angioplasty has been validated in experimental and clinical investigations. A peripheral angioplasty study reported that plaque area was reduced by 33 percent, accounting for only 20 percent of luminal gain.89 However, studies using automatic pullback devices have shown that "compression" actually represents redistribution of plaque along the long axis of the vessel.95 The prognostic significance of different mechanisms of luminal enlargement remains uncertain. Directional Atherectomy Directional coronary atherectomy (DCA) is currently performed relatively uncommonly. DCA devices incorporate a rotating circular blade to remove atherosclerotic plaque from the luminal surface.96 Because angiographic and/or ultrasound calcification is a well-documented predictor of failure of directional atherectomy, ultrasound imaging was advocated for guidance of atherectomy, particularly for preintervention lesion selection.97,98 Some investigators have proposed that achieving a larger lumen after atherectomy using ultrasound guidance would result in a lower restenosis rate. This hypothesis was tested in a multicenter registry (the Optimal Atherectomy Restenosis Study, or OARS), in which residual stenosis was reduced from 64 to 7 percent with ultrasound guidance.98 The angiographic restenosis rate at 6 months was 28.9 percent and the 1-year target lesion revascularization rate was 17.8 percent. However, in the larger CAVEAT trial, DCA failed to reduce late events as compared with PTCA, with or without ultrasound guidance.99 The success of coronary stenting in reducing restenosis has largely rendered directional atherectomy obsolete. Rotational Ablation Rotational ablation employs a high-speed (up to 200,000 rpm) diamond-coated burr to debulk atheroma. Theoretically, this device minimizes injury to the normal arterial wall by "differential cutting," in which normal elastic tissue is deflected away from the burr while relatively inelastic atheroma is not displaced and is therefore abraded by rotation of the burr. Clinical indications for rotational ablation include calcified segments or lesions that resist balloon dilatation. Rotational ablation is also sometimes used in long lesions, ostial lesions, and in-stent restenosis.100–104 Demonstration of a heavily calcified vessel by angiography or intravascular ultrasound is often cited by operators as an indication for rotational ablation. As previously noted, there is a poor correlation between ultrasound and fluoroscopy in assessing the presence and extent of calcification. Accordingly, ultrasound is sometimes employed prior to rotational ablation to confirm or refute the presence of calcification. Vessels revascularized using rotational ablation are often diffusely diseased, and the "normal" dimension can be difficult to determine by angiography. Therefore, ultrasound is sometimes used to size the vessel and determine the largest burr that can be safely employed. Intravascular ultrasound studies have confirmed the principle of selective plaque removal or differential cutting. In 48 lesions treated with rotational ablation, atheroma area decreased from 15.7 ± 4 to 13.0 ± 5 mm2 and the arc of calcium decreased slightly from 227 ± 107 to 209 ± 107 degrees, p <0.05.104 Vessel expansion or dissection was noted in a minority of cases and did not contribute significantly to lumen gain. The residual narrowing of the cross-sectional area measured by ultrasound averaged 74 percent. Following rotational ablation, the residual lumen is usually round or ellipsoid and may have a 15 to 20 percent greater area than the largest burr used, presumably due to lateral movement of the burr during the procedure. Luminal Measurements Postintervention A poor correlation has been reported for comparisons of ultrasound and angiography in assessment of residual stenosis following balloon angioplasty, with measurements that are usually smaller by ultrasound than by angiography.9,41,88–92 Two factors probably influence the overly optimistic tendency of angiographic imaging.21 At the reference site, angiography tends to underestimate the diameter of the normal reference vessel because of the frequent presence of unrecognized atherosclerosis. At the target site, angiography tends to overestimate the actual gain in luminal diameter because contrast material penetrates into complex cracks and fissures produced by the intervention, giving the appearance of an enlarged lumen. To calculate a postprocedure percent diameter stenosis, the diameter at the target site (an overestimate) is divided by the reference diameter (an underestimate), resulting in a more favorable impression of the actual gain in luminal dimensions. Quantitative angiography showing a residual stenosis of 10 to 15 percent is commonly associated with a 60 to 80 percent atheroma burden by intravascular ultrasound. Coronary Stent Deployment Initial Studies of Ultrasound Guidance The use of stents in percutaneous revascularization has increased exponentially over the last few years. Intravascular ultrasound imaging has played a pivotal role in understanding and optimizing the benefits of stent therapy.105–107 In initial trials leading to approval by the U.S. Food and Drug Administration (FDA), articulated slotted-tube stents were deployed using moderate balloon pressures (6 to 10 atm).108,109 To reduce subacute thrombosis, patients received aggressive anticoagulation with both antiplatelet and antithrombotic agents, including warfarin, for 3 to 6 months. Initial studies demonstrated a reduction in the restenosis rate compared with balloon angioplasty but reported a high incidence of hemorrhagic complications and longer hospital stays. A pioneering report detailing the intravascular ultrasound experience of Colombo et al. in Milan, Italy, significantly altered the understanding of optimal stent deployment and prevention of subacute thrombosis.107 Ultrasound examination revealed a mean residual stenosis of 51 percent following angiographically guided stent deployment and a high prevalance of incomplete stent apposition (Fig. 18–10).   Figure 18–10. Underdeployed coronary stent. In this example, intravascular ultrasound images show several stent struts (arrows) that are not in full contact with the underlying vessel wall. This process is referred to as incomplete stent apposition. Because stents are porous structures, angiographic contrast can flow outside of a partially deployed stent, resulting in the angiographic appearance of full deployment despite the presence of incomplete apposition. In the Milan study, the operators performed additional balloon inflations at higher pressures (typically 18 to 20 atm) or used a larger balloon (or both), reducing final ultrasound residual stenosis to 34 percent and with a subacute thrombosis rate of only 0.3 percent using no systemic antithrombotic agents (antiplatelet therapy only). It is now widely accepted that high-pressure deployment of the stents dramatically reduces the incidence of subacute thrombosis and obviates the need for acute and chronic administration of antithrombotic agents.110,111 Subsequently, routine high-pressure deployment without ultrasound imaging became the standard of therapy. Routine versus Nonroutine Ultrasound Following the widespread acceptance of high-pressure post-dilatation with antiplatelet rather than anticoagulation regimens, the further benefit of ultrasound imaging has been debated.86,110–114 Some investigators have suggested that despite routine use of high-pressure postdilatation, ultrasound-guided therapy could improve procedural results.86,114 In a retrospective analysis of 315 lesions treated by high-pressure stenting, ultrasound was defined as "beneficial" if imaging resulted in further interventions that increased stent area by >25 percent or identified other lesions that required treatment.115 Prior to ultrasound examination, the mean inflation pressure was 14.7 ± 3.2 atm, and only 47 percent of stents were considered "optimally" deployed. Additional ultrasound-triggered inflations improved in-stent lumen area by more than 25 percent in 83 lesions (26 percent of patients). Additional procedures were performed in other lesions identified by ultrasound in 51 (16 percent). Final in-stent area improved from 6.9 ± 2.2 to 8.0 ± 1.93 mm2 (p <0.001). Procedural results were "improved" in 39 percent of the cases following ultrasound imaging.115 It is now generally accepted that after high-pressure coronary stenting, ultrasound imaging results in additional procedures in approximately 20 to 40 percent of cases. Since in-stent restenosis is predominantly determined by the degree of intimal hyperplasia, a larger lumen could theoretically accommodate more tissue growth without flow-limiting obstruction.116 However, it remains uncertain whether ultrasound-guided "optimal" expansion translates into better clinical outcome. A randomized trial in 164 patients of ultrasound-guided stenting demonstrated a 6.3 percent absolute reduction in restenosis rate, which was not statistically significant, because the study was powered to detect a 50 percent reduction of the restenosis rate.117 A nonrandomized substudy of 538 patients from the Stent Anticoagulation Regimen Study (STARS) compared the outcome of ultrasound and angiographically guided stenting. The ultrasound arm achieved a significantly larger lumen area and a 39 percent relative reduction in clinical restenosis.118 However, the impact of the more aggressive dilatation on restenosis rates has not been adequately examined by properly designed, prospective randomized clinical trials. Ultrasound Imaging of Peri-Stent Segments Ultrasound imaging of reference segments following stenting may be useful in identifying reference segment disease or dissections that require additional interventions. The presence of significant peri-stent flow-limiting lesions or dissections has been linked to higher likelihood of stent thrombosis.119 These findings are often angiographically occult or appear as areas of indistinct haziness at the vessel border. In 201 stent patients, 31 segments with peri-stent angiographic haziness were detected. Ultrasound imaging revealed an angiographically inapparent obstructive lesion in 15, a peri-stent wall injury in 14, and mild intimal thickening in the remaining 2 segments.120 The extent of neointimal hyperplasia at the stent margins has been linked to preexisting disease in the reference segment.121 In stenting as a bailout for dissection, intravascular ultrasound is more sensitive in detecting the extent of dissection, often revealing a greater true length than is evident from angiography, which may be helpful in guiding vessel salvage. Optimal Procedural Goals of Ultrasound Guided Stenting Although ultrasound guidance of stenting has been practiced for several years, there is no consensus regarding optimal procedural endpoints. Colombo initially recommended achieving 60 percent of the average proximal and distal reference areas but later altered the definition to 100 percent of the distal reference lumen area.105–107 Other definitions of optimal expansion include 90 percent of the distal reference area, 80 percent or 90 percent of the average reference area, a "lumen symmetry index" >0.7, and/or full coverage of reference-segment disease or dissections.122,123 In most clinical trials, procedural endpoints are not achieved in the majority of cases. In the Optimal Stent Implantation Trial, the target of >90 percent of the average reference or >100 percent of the smaller reference area were not achieved in half the patients at an inflation pressure of 15 atm and only 60 percent of patients at 18 atm.122 In the Angiography Versus Intravascular Ultrasound Directed Stent Placement (AVID) trial, the target endpoint of 90 percent of the distal reference area was not achieved in >70 percent of 225 patients.124 Other reports have questioned the clinical relevance of using the stent-to-reference ratios as target for ultrasound-guided stenting. In 165 patients, target vessel revascularization was predicted by final in-stent lumen area (OR 1.4, 95 percent CI 1.1 to 1.9) and not the ratio of stent-to-reference area (OR 1.1, 95 percent CI 0.85 to 1.6).112 Repeat revascularization was required in 30 percent of patients with a minimum in-stent lumen area <5 mm2 but only 3 percent of cases with an area exceeding 9 mm2. In another large cohort undergoing ultrasound-guided stenting, restenosis was inversely related to the minimum in-stent area.125 An area of 9 mm2 was achieved in 23 percent, but the incidence of restenosis in this subgroup was only 8 percent, compared with 29 percent in the remaining patients, p < 0.0001. Thus, commonly employed ultrasound endpoints based upon a predefined stent-to-reference ratio are both difficult to achieve and correlate weakly with clinical outcome.125–127 Ultrasound studies have demonstrated that the degree of in-stent neointimal hyperplasia is independent of final lumen size, which may explain the higher restenosis rates in smaller vessels and poorly expanded stents.116 If acute lumen gain is not adequate to accommodate subsequent tissue proliferation, there is significant late loss and restenosis. Intravascular Ultrasound and Restenosis A more complete understanding of restenosis has evolved from serial ultrasound measurements of plaque and lumen areas following balloon angioplasty and directional atherectomy.128,129 In some studies, serial ultrasound examinations have shown that a late reduction in total vessel area (chronic negative remodeling) is an important mechanism of restenosis after interventional procedures.129 These observations suggest that mechanical interventions to prevent chronic recoil (such as stenting) may be more important in preventing restenosis than interventions designed to prevent intimal hyperplasia. This concept likely explains the lower restenosis rate observed in randomized multicenter studies comparing balloon angioplasty and stent implantation.108,109 In 212 native coronary lesions in 209 patients following intervention, the ultrasound cross-sectional area with the smallest lumen area at late follow-up was compared with the matching site obtained immediately following the intervention.129 At follow-up examination, there was a significant decrease in EEM area and an increase in plaque area (p < 0.0001 for both) that combined to reduce lumen area. More than 70 percent of lumen loss was attributable to the decrease in EEM area, whereas the neointimal area accounted for only 23 percent of the decrease in lumen area. The change in lumen area correlated more strongly with the change in EEM area (r = 0.75, p < 0.0001) than with the change in plaque area (r = 0.28, p < 0.0001). At lesions that demonstrated an increase in EEM area at follow-up (47 percent), there was no change or an actual gain in lumen area and a reduction in angiographic restenosis (26 versus 62 percent for lesions with a decrease in EEM area at follow-up, p < 0.0001). Other investigators have suggested a bidirectional remodeling response following percutaneous coronary interventions: early adaptive enlargement and late shrinkage of the vessel. In a unique study, 61 lesions in 57 patients who underwent balloon angioplasty or atherectomy were examined by intravascular ultrasound in a serial manner before and immediately after the intervention and after 24 h, 1 month, and 6 months.48 The lumen area significantly improved during the first month following the intervention, but significantly decreased at 6 months. Simultaneously, the EEM area increased in the first month, but later decreased at 6 months. However, plaque area steadily increased from immediately postintervention to the 6-month follow-up. Thus the changes in lumen size closely tracked the changes in EEM area (r = 0.72, p = 0.0001). Although the increase in plaque area correlated with lumen loss, the correlation was not as strong (r = 0.34, p = 0.0008). The lumen gain observed during the first month was solely due to the compensatory vessel enlargement, whereas the late lumen loss was mostly caused by vessel shrinkage but also by progressive neointimal hyperplasia. Investigations employing quantitative angiography have demonstrated that late lumen loss is significantly greater with stents than with balloon angioplasty. This, however, is offset by the much larger acute lumen gain, such that the net gain at follow-up is significantly greater with stenting.108,109 Intravascular ultrasound has been employed to examine the mechanism of stent restenosis. Unlike the restenotic response to other percutaneous devices, which is a mixture of arterial remodeling and neointimal growth, stent restenosis is almost exclusively due to the neointimal proliferation.129 In a serial study using intravascular ultrasound of stented coronary segments, there was no significant change in the area bound by stent struts, indicating that stents can withstand and resist the arterial remodeling process.130 In some cases, restenosis develops at the margins of the stent. Predictors of stent restenosis have been identified by multivariate analysis, including the smaller reference vessel and lumen size, the larger plaque burden at the reference segments, and the smaller achieved in-stent lumen area at the stent margins.121 Intravascular Ultrasound and Brachytherapy Ultrasound has proven useful in clarifying the mechanisms of benefit and refining the techniques used for brachytherapy.130,131 Intravascular ultrasound studies demonstrate that radiation inhibits neointimal proliferation within a stent. In a randomized study of 70 lesions with in-stent restenosis, at follow-up, 79 percent of stents in patients who received radiation had no measureable intimal proliferation, compared to 27 percent of those randomized to the "no radiation" cohort.130 In the nonstented segments, some studies suggest that radiation initiates a process of vessel expansion (a type of positive remodeling).131 These effects are strongly influenced by the dose delivered to the media or adventitia, which is dependent on the thickness and composition of the atheroma and the position of the catheter in the lumen. Current research is examining whether an ultrasound image-based dosing algorithm will be required to optimize therapeutic benefit. Ultrasound has already demonstrated the potential for radiation to accelerate restenosis at the edges of the treatment region where the dosing falls off ("candy-wrapper" effect).131 Intravascular Ultrasound and Drug-Eluting Stents Intravascular ultrasound has proven useful in assessing the mechanism of benefit of drug-eluting stents.132,133 Studies consistently demonstrate that drug-eluting stents have the potential to markedly reduce neointimal proliferation within the stent at 6 to 9 months follow-up. In the RAVEL trial, a subset of 95 patients underwent intravascular ultrasound examination 6 months following stent inplantation. The patients who received a sirolimus-eluting stent exhibited an average of only 2 mm3 of intimal hyperplasia compared to 37 mm3 for the control group (p <0001).132 More recently, similar although less pronounced results were reported for a paclitaxel-eluting stent.133 Using paclitaxel, investigators reported a dose-dependent reduction in intimal hyperplasia at 4 to 6 months from 31 mm3 in the control arm to 18 mm3 in a low-dose cohort to 13 mm3 in the highest dose paclitaxel group.

Future Directions

Intravascular ultrasound is commonly but not routinely performed in the United States during coronary interventions. Approximately 5 to 8 percent of interventional procedures are currently performed with ultrasound guidance. Usage in Europe is considerably less than in the United States; in Japan it is considerably higher, reflecting differing reimbursement rates and practice patterns. Technical developments in both catheters and systems are continuing, but at a relatively slow pace.134 Additional technologic advances in intravascular imaging are anticipated, including further reductions in the size of imaging catheters and higher-frequency ultrasound catheters, yielding significantly better spatial resolution.135 Although high-frequency probes enable better axial and lateral resolution, there are significant trade-offs in moving beyond the current 40-MHz frequency. For example, penetration is likely to be impaired in comparison with more conventional devices, and greater backscatter from blood cells at high frequencies may interfere with discrimination of the interface between lumen and vessel wall. As previously discussed, analysis of backscattered ultrasound signals has been employed by several investigators to perform "tissue characterization."39,40 Intrinsic properties of the backscattered ultrasound signals—including the amplitude distribution, frequency response, and power spectrum of the signal—may convey specific information about tissue types.39 However, the ability of computer-based analysis of the unprocessed radiofrequency backscatter to differentiate the histologic layers of the normal vessel wall remains investigational. The promise of this research is the potential to identify "vulnerable" atherosclerotic plaques, defined as lesions at high risk for plaque rupture leading to thrombosis and acute coronary syndromes. Three-dimensional reconstruction of intravascular ultrasound has been proposed as a means to facilitate understanding of the spatial relationship between the structures within different tomographic cross sections.136,137 Despite the promise of these methods, many unresolved problems remain. The algorithms applied for three-dimensional reconstruction do not consider the presence of curvatures of the vessel and assume that the catheter passes in a straight line through the center of consecutive cross sections. The systolic expansion of the coronary vessel and the movements of the catheter within the vessel during the cardiac cycle also generate artifacts. Accordingly, the reconstructed images should not be considered faithful representations of the vessel and should not be used for volumetric plaque determination.

Summary

The equipment, technique, and applications for intravascular ultrasound imaging continue to evolve. The insights provided by the unique ability of intravascular ultrasound to directly image coronary plaques have contributed greatly to our understanding of the nature of atherosclerosis and the effects of interventional devices. Current studies using intravascular ultrasound to measure atherosclerosis progression and regression have the potential to further expand the utility of this important diagnostic method.