Background and Limitations of Angiography
Contrast angiography has been the gold standard of coronary artery imaging for over six decades. However, it is important to understand that angiograms only delineate the coronary lumen with no direct imaging or examination of the arterial wall. With the growing interest in vascular biology and understanding of the metabolically active atherosclerotic plaque, intravascular ultrasound (IVUS) imaging fills a gap in our understanding of coronary disease, as it allows in vivo examination of both the arterial wall and the lumen of the coronary arteries with high resolution. Although newer intravascular imaging modalities have evolved in the past decade, IVUS imaging remains the most mature adjunctive intravascular imaging modality and the one technique with a large body of literature to support its application.
Conceptually, it is easy to understand how angiography of the coronary arteries may not be the perfect imaging modality. The coronary arteries are constantly moving structures with complex, three-dimensional lumen shapes and atherosclerotic lesions of various distributions and compositions that need to be depicted on a two-dimensional display. Therefore, it is critically important for the angiographer to understand the limitations of the technique. This will allow a more realistic processing of the information, which is necessary when making diagnoses and in clinical decision making. One of the most basic limitations is that fluoroscopy and cine angiography have limited resolution that is not adequate to delineate all the information needed from a coronary angiogram, even with near universal adoption of digital imaging and flat panel technology. Because of these complexities and limitations, interpretation of coronary angiograms has traditionally been marked by significant interobserver and intraobserver variability.
As stated, and because angiograms are essentially contrast silhouettes of the lumen, it is not possible to directly discern whether the arterial wall is thickened (i.e., diseased). This is inferred by a reduction in the diameter of the luminogram in one segment compared to an adjacent one. Thus, the segmental construct in interpretation of angiograms is based on the assumption that narrower segments are the sites of disease, while adjacent larger segments represent the normal size of the artery or at least of its lumen. Yet autopsy studies have consistently demonstrated that atherosclerosis is diffuse, affecting almost all segments of an artery, albeit with varying severity. Therefore, the “focal” lesions frequently described on angiography are the more diseased sites, but the “normal” or “mildly irregular” segments are almost always diseased as well. This leads to the angiographic underestimation of disease severity and degree of narrowing at the worst lesion sites ( Figure 16-1 ).
Arterial remodeling, first described by Glagov et al., is another important phenomenon that can significantly affect angiographic interpretation. Arterial remodeling describes the outward displacement of the arterial wall with increasing plaque size. This compensatory enlargement allows the artery to accommodate a certain amount of plaque volume without affecting the size of the lumen. Stenoses develop only when this mechanism is overwhelmed by increasing plaque size. Therefore, by definition, interpretation of contrast luminograms based on differences in lumen size from one segment to the other is not helpful in recognizing early, “well-compensated” stages of disease. Remodeled arterial segments with significant plaque burden can thus be described as “angiographically normal.”
Another phenomenon, “negative remodeling,” can also contribute to development of luminal narrowing in a reverse manner. In these cases, the reduction in the size of the artery itself contributes more to the reduction in lumen size than what the size of the plaque would indicate ( Figure 16-2A ). The remodeling index is a metric relating the size of the artery at the lesion site to that at the proximal reference segment. An index exceeding 1.0 indicates compensatory enlargement at the lesion site and a value less than 1.0 indicates shrinkage or negative remodeling ( Figure 16-2B ). There is some evidence that positive remodeling is associated with acute presentations, while negative remodeling is more commonly seen in patients with stable angina.
The complexity of coronary anatomy is another factor that limits the accuracy of angiographic depiction. The constantly moving tortuous arteries give off branches in different planes in a three-dimensional space. The lesions may be eccentric in distribution, at bifurcation points, ostial in location, calcified, or have complex luminal topography. It takes a perfect orthogonal angiographic projection to delineate accurate information about one or more of those findings. Given that projections are frequently selected arbitrarily, it may be difficult for angiographers to visualize the lesion, define the percentage of stenosis, and proceed with a management decision ( Figures 16-3 and 16-4 ).
Contemporary, well-trained angiographers should be able to utilize adjunctive technologies that can resolve the clinical dilemmas caused by limitations of angiography and reach accurate conclusions. Currently, the most commonly utilized technologies in the cardiac catheterization laboratory are IVUS, optical coherence tomographic (OCT) imaging, and fractional flow reserve (FFR) measurement. While IVUS and OCT imaging address the limitations of angiography related to the resolution and complexities of severity and distribution of disease, FFR measurement addresses the discrepancy between the anatomical and functional significance of coronary lesions.
IVUS Imaging, Basic Image, and Measurements
Imaging is performed using a miniaturized ultrasound transducer placed into the center of the vessel lumen on the tip of a catheter. Contemporary imaging catheters are 3 to 3.5 Fr in diameter and can be used within 5 or 6 Fr guiding catheters. Coronary imaging catheters are typically advanced over standard angioplasty guidewires, 0.014 inch in diameter. By emitting and receiving reflected ultrasound beams, sector images can then be generated. Two different technologic approaches are used to generate a circumferential tomographic image of the entire cross section of the vessel: (a) mechanical systems in which the transducer is composed of a single large piezoelectric crystal that is rotated at high speed to acquire images from all sectors of the circumference, and (b) phased array systems in which transducers are made of multiple small crystals that are sequentially activated to image adjacent sectors of the arterial cross section. The imaging frequency is 40 MHz to 45 MHz for the mechanical systems and 20 MHz for the phased array system. In both cases, the reflected ultrasound waveforms are processed into grayscale images and the sectors are reconstructed into the full tomographic cross section of the artery.
Definitions and methodology of acquiring IVUS images and measurements are outlined in the American College of Cardiology and the European Society of Cardiology expert consensus documents on the standards of IVUS imaging. Tomographic images obtained by the IVUS transducer show the reflection of the intima as an echodense layer, followed by the media as an echolucent stripe. The adventitia is the outer echodense layer that represents the outer boundary of the artery and the adherent connective tissue and gives the arterial wall a trilaminar appearance. Less commonly, and especially in very young individuals, the intima is so thin (<300 µm in thickness) that it leads to signal dropout and the traditional trilaminar appearance is replaced by a monolayer ( Figure 16-5 ). Intimal thickening is the hallmark of arterial wall disease. When the intima thickens, it becomes more echodense and easier to visualize on display.
The basic IVUS measurements performed on a coronary artery image are shown in Figure 16-6 . The echodense intimal leading edge defines the boundaries of the lumen, while the leading edge of the adventitia (the external elastic membrane [EEM]) defines the vessel area. The area between both tracings represents the plaque plus media or atheroma area. Lumen and vessel diameters can be measured as well. The plaque or atheroma area can be normalized to the size of the vessel by calculating a percentage of the cross-sectional narrowing or plaque burden (plaque area ÷ vessel area × 100).
The IVUS catheter can be mounted on an automatic pullback device that moves it along the arterial segment at a known and constant velocity (0.5 or 1.0 mm/sec), thus allowing calculation of the length of the segment. Lumen and plaque areas, in addition to length, can then be used to calculate lumen and plaque volume. These are typically more accurate measures of disease burden and are commonly used in research studies, particularly those examining small degrees of progression or regression of plaque size. However, volumetric calculations are time consuming and of less value for everyday clinical applications of IVUS imaging.
Qualitatively, the echodensity of the plaque on grayscale displays is somewhat related to its tissue content. Using the adjacent adventitia as a visual reference, echodensity indicates a plaque is “bright” or “brighter” than adventitia (closer to the white end of the grayscale), while echolucency refers to plaques that appear “darker” than adventitia (closer to the black end of the grayscale). Most plaques are heterogeneous with varying densities, even on the same still frame image ( Figure 16-7 ). Plaques rich in lipid are typically echolucent, whereas echodense ones are typically rich in fibrous tissue and calcification. Calcified lesions are usually very dense and have a back shadow due to the complete absorption of the ultrasound beam. The association between echodensity and tissue content is not robust due to the overlap between various levels of grayscale and heterogeneous tissue content. As discussed later, advanced analysis of ultrasound backscatter (such as virtual histology technology) attempts to overcome the limitations of visual analysis of the reconstructed grayscale images.
Guidelines for Use and Appropriate Indications
Over the past two decades, numerous clinical studies have established a number of acceptable and appropriate applications of IVUS imaging in the cardiac catheterization laboratory. The updated clinical practice guideline for coronary interventions published by the ACC, AHA, and SCAI in 2010 outlines the recommendations for coronary IVUS imaging ( Table 16-1 ).
| Class I |
|
| Class IIa |
|
| Class IIb |
|
| Class III |
|
The following section reviews the current diagnostic and interventional applications outlined in the guideline statement. In addition, a few more advanced or highly specialized as well as research applications are discussed in detail.
Diagnostic Applications
Ambiguous Angiograms and Indeterminate Coronary Lesions
Vessel overlap, vessel tortuosity, eccentric lesions, ostial or bifurcation lesions, and severe calcification constitute the major reasons for suboptimal angiographic visualization of the lumen (see Figures 16-3 and 16-4 ).
This is particularly important when the lesions are of intermediate (40%-70%) severity in patients with mild or atypical symptoms. In these settings, IVUS imaging provides a tomographic perspective that is independent of the radiographic projection, which allows accurate quantification of lumen size, plaque burden, plaque distribution in relation to branch points, and distribution of calcified plaque.
When an angiographically intermediate lesion (50%-70% diameter stenosis) is encountered, interobserver and intraobserver variability is quite high. Further evaluation can be accomplished either by functional assessment (FFR measurement) or by a more accurate definition of the lumen size using IVUS or OCT imaging. Measurement of FFR accurately defines hemodynamically significant lesions. Multiple prospective randomized trials have validated the use of FFR measurement in clinical decision making on the need for revascularization versus conservative management.
IVUS measures that are used to define hemodynamically significant lesions have mostly been benchmarked against established FFR cutoff thresholds. In a small study of 51 lesions, a lesion was considered hemodynamically significant when FFR was <0.75. The IVUS measurements that identified such a lesion were minimum lumen area (MLA) <3.0 mm 2 (sensitivity, 83.0%; specificity, 92.3%) and area stenosis >60% (sensitivity, 92.0%; specificity, 88.5%) ( Figure 16-8A ). The combination of both criteria (MLA <3.0 mm 2 and area stenosis <60%) had 100% sensitivity and specificity. In another study of 53 lesions, a minimal luminal diameter (MLD) of <1.8 mm, MLA of ≤4 mm 2 , and the cross-sectional area stenosis of >70% were the best indicators of hemodynamic significance, as determined by an FFR <0.75. Methodological differences in the measurements of FFR (such as intracoronary vs. intravenous administration and use of adenosine vs. papaverine for generation of hyperemia) in these two studies may explain the discrepant cutoff values.
As the FFR threshold of <0.8 has been adopted more recently to define ischemia-producing lesions and to address discrepancies in reference vessel size, a larger retrospective analysis of 205 angiographically intermediate lesions was performed using both IVUS and FFR measurements. FFR was <0.8 in 26% of lesions. Overall, there was moderate correlation between FFR and IVUS measurements, including MLA (r = 0.36, p < 0.001), lesion length (r = −0.43, p < 0.001), and area stenosis (r = 0.33, p = 0.01). For the whole sample, an MLA >4.0 mm 2 had an excellent negative predictive value (>94%), and MLA <3.09 mm 2 was the best determinant of lesions with FFR <0.8 (sensitivity, 69.2%; specificity, 79.5%) ( Figure 16-8A ). The correlation between FFR and IVUS was better for large vessels compared to small vessels. Depending on the reference vessel diameter, threshold MLA values for ischemia-producing lesions (FFR <0.8) are as follows: MLA <2.4 mm 2 in small vessels, MLA <2.7 mm 2 in medium-sized vessels, and MLA <3.6 mm 2 in large vessels ( Figure 16-8B ).
Evaluation of the Left Main Coronary Artery
Angiographic severity of left main coronary artery (LMCA) lesions is almost always difficult to quantify. Visualizing the ostium and most proximal part of the vessel depends on the “reflux” of contrast into the aortic cusp. “Streaming” of contrast from the injection vortex can give a false impression of luminal narrowing near the tip of the injecting catheter. When the whole length of the LMCA is diseased, the absence of a near normal reference segment complicates the visual or automated quantification of stenosis severity. Furthermore, disease in the bifurcation or trifurcation into daughter branches is often complex in topography and is frequently concealed by the overlap of the branches in different projections ( Figure 16-9A ). Thus, reproducibility of quantitative angiography of the LMCA is worse than that of all other coronary arterial segments. For these reasons, IVUS imaging is commonly used for better delineation of LMCA stenoses.
Similar to non–left main coronary lesions, a few studies were based on correlating cutoff values of IVUS measurements with FFR measures of hemodynamic significance of LMCA stenosis. In a study of 55 patients with angiographically ambiguous LMCA, both FFR and IVUS measurements were performed. The two IVUS measurements that correlated best with hemodynamically significant lesions as determined by FFR were MLD <2.8 mm (sensitivity and specificity of 93% and 98%, respectively) and MLA <5.9 mm 2 (sensitivity, 93%; specificity, 95%) ( Figure 16-9B ). In another series of 122 patients with intermediate LMCA stenosis, patients were followed for 1 year following a clinical decision to defer revascularization. Similar to the FFR-based studies, the MLD was the most important predictor of adverse cardiac events. A threshold MLD of 3 mm appeared to provide the best cutoff between those who developed clinical events and those who did not. Other patient series demonstrated the safety of deferring revascularization in patients with LMCA stenosis when the minimum lumen area was >7.5 mm 2 .
Evaluation of Transplant Vasculopathy
Cardiac allograft vasculopathy (CAV) is a disease of unclear etiology that affects the coronary arteries of the transplanted heart and is characterized (at least in part) by progressive intimal proliferation of coronary arteries. Due to its diffuse nature, the sensitivity of coronary angiography for detection of CAV is appreciably lower than its sensitivity in detection of the more segmental atherosclerotic lesions. IVUS imaging provides a very useful and safe tool to study the early development and progression of CAV. CAV is commonly defined as a site where the intimal thickness is ≥0.5 mm, although a cutoff value of 0.3 mm is also used as a cutoff to diagnose earlier lesions.
IVUS imaging of coronary arteries soon after transplantation has demonstrated that arteries that appear angiographically normal may contain evidence of early donor atherosclerosis. In one series with a mean donor age of 32 years, atherosclerotic lesions (maximal intimal thickness ≥0.5 mm) were detectable in >50% of patients. Donor atherosclerotic lesions are focal, noncircumferential, maybe calcified, and more commonly involve the proximal segments. Presence of donor atherosclerosis does not seem to predispose to development of CAV.
On serial IVUS studies with matched segments, CAV progression initially manifests as increasing intimal thickening associated with compensatory positive remodeling. Lumen compromise over a 5-year follow-up period can also be the result of a negative remodeling response. Several studies have demonstrated an association between disease severity as assessed by IVUS and the clinical outcome in heart transplant recipients. Serial studies allow the evaluation of disease progression, which is another important predictor of outcome. In a serial study of 143 patients followed for an average of 5.9 years, IVUS evidence of rapid progression of CAV (defined as ≥0.5-mm increase in intimal thickness in the first year after transplantation) was a powerful predictor of all-cause mortality and MI. Patients with rapidly progressive CAV had a higher mortality rate compared to those without (26% vs. 11%, p = 0.03). Death and nonfatal MI were also more frequent among those with rapid progression (51% vs. 16%, p < 0.0001).
IVUS imaging has also been used in the evaluation of therapies directed toward control of CAV. The beneficial effects of pravastatin and everolimus in delaying the progression have been shown using IVUS imaging.
Interventional Applications
In the 1990s, IVUS coronary imaging was rapidly adopted, as it was seen to provide a wealth of information about atherosclerotic disease patterns that overcame the limitations of angiography and favorably impacted interventional techniques. Hence, it was instrumental in better understanding the pathophysiologic mechanisms of action of various interventional devices and the arterial responses to interventions. This information resulted in improvement in device design and refinement of procedure techniques. Subsequently, the knowledge gained by operators from IVUS imaging became assimilated in the technical approaches. Widespread and routine use of IVUS imaging during interventional procedures tapered down to a more selective approach: situations when specific questions cannot be accurately answered on the basis of angiography alone.
With the predominance of coronary stenting, data supporting the use of IVUS imaging to aid in PCI device selection have become less relevant. In the selected cases where atherectomy is considered, IVUS evaluation may provide useful data about lesion characteristics that lead to the use of one particular device to “prepare” the lesion for stenting. In the following sections, the role of IVUS in guiding coronary stenting is discussed in more detail.
Vessel Sizing
The use of properly sized devices (such as balloons, stents, atherectomy burrs, etc.) is essential to achieve optimal procedural outcomes. The underestimation of reference vessel size and the use of undersized devices increase the risk of suboptimal outcomes such as early recoil, restenosis due to inadequate procedural lumen gain, or stent complications such as stent underexpansion and/or strut malapposition. On the other hand, an oversized device increases the risk of dissection or rupture.
As previously discussed, IVUS imaging frequently reveals significant disease burden in vessels that appear angiographically normal or mildly diseased; thus the true size of the reference vessel can be underestimated (see Figures 16-1 and 16-3 ). Certain clinical scenarios are associated with significant underestimation of angiographic reference vessel size. Examples include PCI in setting of acute myocardial infarction (higher degree of coronary vasomotor tone) and chronic total occlusion (distal segments are underfilled and diffusely diseased). In these situations, adequate coronary vasodilators and IVUS imaging can accurately delineate the reference vessel size, which is the most important determinant of device size selection. Often, after IVUS imaging and quick quantitative analysis, a larger balloon or stent size is accurately and safely selected for use.
It is important for operators to use IVUS imaging for vessel sizing judiciously. The reference vessel size is determined by the lumen diameter in the reference segments, that is, those segments that have no or minimal disease in the vicinity of the lesion to be treated. Using the EEM dimension at the reference or at the site of the lesion will result in overestimation of the vessel size, since EEM diameters reflect a degree of positive remodeling in most cases. This can result in device oversizing and should be avoided. The operator should focus on lumen size, with some adjustment depending on plaque burden.
IVUS Imaging in Coronary Stenting
Historically, IVUS imaging has been instrumental in optimizing the technique of coronary stenting as it is currently practiced. Initial IVUS observations in the early 1990s revealed that the technique of stents was frequently inadequately expanded and not fully opposed to the arterial wall despite satisfactory angiographic results. These observations led to major refinements in the technique of stenting, most notably higher pressure postdilatations and the use of upsized balloons. Coupled with the introduction of dual antiplatelet therapy, these technical improvements resulted in a reduction in the incidence of stent thrombosis.
A few terms commonly used in the discussion of IVUS imaging and stenting need to be defined. Stent expansion refers to the size of the stent lumen achieved after the stent is deployed and postdilated. That can be expressed in terms of minimum stent lumen diameter (MLD) or more frequently minimum lumen area (MLA). Expansion is also indexed to the size of the reference segments, where MLD or MLA is divided by the mean diameter or area of the reference vessel to express a percentage of the reference. Stent MLA is an important predictor of restenosis, as is discussed later. Stent or strut apposition describes the contact between the stent struts and the underlying arterial wall. This is a qualitative metric of stenting that is detected with advanced imaging such as IVUS or OCT. It is related but not synonymous with stent expansion. It is common to see a well-opposed but underexpanded stent, for example, when the stent is slightly undersized or if a rigid lesion is not adequately predilated ( Figure 16-10 ). IVUS-guided stenting refers to the strategy of using IVUS imaging to optimize all aspects of the stenting procedure, that is, using IVUS to assess vessel size, facilitate lesion preparation, assess adequate expansion and apposition, confirm results after postdilation, and ensure the absence of complications in the reference segments ( Figure 16-11 ).
Specific criteria to define optimal stent expansion were developed out of multiple observation studies that correlated final IVUS measurements with angiographic and/or clinical outcome. The general consensus among those studies is that minimum in-stent lumen dimensions are the most important predictors of restenosis.
Bare-Metal Stents
In bare-metal stents (BMS), minimum stent area <6 mm 2 was associated with the highest rates of restenosis and need for repeat revascularization ( Figure 16-12 ). The lumen area within the stented segment is primarily determined by the target vessel size, explaining the relatively higher rates of restenosis in smaller vessels despite the achievement of excellent procedural results in the majority of cases ( Figure 16-13 ).
The strategy of IVUS-guided bare-metal stenting has been evaluated in several nonrandomized registries and some randomized trials. There is a general consensus that this did not alter the incidence of death and nonfatal MI, but it did consistently lead to a larger in-stent lumen dimension as a result of higher pressure and/or balloon upsizing. Whether that translates into a reduction in clinical restenosis and need for repeat revascularization remains controversial. Results have not been consistent and favorable results were observed more in the registries than in the randomized trials.
In the study by Schiele et al., 155 patients undergoing coronary stenting were randomized to IVUS guidance versus angiographic guidance. Stent MLA was 20% larger in the IVUS-guided stenting arm, and at 6 months, there was a 22% relative reduction in angiographic binary restenosis in favor of the IVUS guidance arm (22.5% vs. 28.8%). However, this did not reach statistical significance (p = 0.25). This result was influenced by the small sample size and the overestimation of the expected benefit attributed to the use of IVUS guidance when the study was being designed. OPTICUS, the largest reported randomized trial on the comparison of IVUS-guided versus angiography-guided stenting, showed similar binary restenosis rates (24.5% and 22.8%, p = 0.68) and similar target vessel revascularization rates at follow-up, although acute gain was significantly higher in the IVUS-guided lesions. OPTICUS specifically included patients with lesions ≤25 mm in length and vessels ≥2.5 mm in diameter. Another randomized study, the TULIP trial, investigated the possible role of IVUS in stenting long (≥20 mm) stenoses. Binary restenosis and ischemia-driven target lesion revascularization were less common with IVUS guidance compared with angiography guidance (23% vs. 46%, p = 0.008, and 4% and 14%, p = 0.037, respectively). These studies do suggest that while IVUS imaging may not be needed to improve outcomes in every case, it may be of value in subsets of cases in which the risk of restenosis is elevated or carries a higher risk. In these cases, operators need to achieve a perfect or near perfect procedural result, which can only be accomplished by IVUS guidance.
Two recently published metaanalyses have also reached different conclusions. The first included seven randomized trials of IVUS versus angiographic guidance of bare-metal stenting in over 2000 patients. Overall, IVUS guidance results in improved procedural outcomes and significantly reduces the risk of repeat revascularization (13% vs. 18%, odds ratio 0.66, 95% CI, 0.48 to 0.91, p = 0.004). Although this analysis does demonstrate an advantage of IVUS guidance, that conclusion was mostly driven by the results of two of the smaller studies that accounted for 25% of the weight of the analysis. Such a difference between IVUS and angiographic guidance could not be demonstrated in the larger included studies. Additionally, the included studies were not homogeneous in methodology or patient populations, which raises concerns about the generalizability of the conclusion. A second metaanalysis addressing the same topic included only five randomized controlled trials including 1754 patients. The investigators were not able to demonstrate any improvement in clinical outcomes with IVUS guidance. This metaanalysis can also be criticized for the process of selection that may have left out studies fairly similar in methodology to ones that were included.
Drug-Eluting Stents
With more widespread use of DES and the lower rates of restenosis, the potential advantage of IVUS guidance to reduce restenosis has become less significant. Early in the DES era, several IVUS studies and substudies were performed to better understand the predictors and mechanisms of DES restenosis. Subsequently, there was significant interest in understanding mechanisms of stent thrombosis and the insights that IVUS can provide in these settings.
IVUS imaging studies of the first generation sirolimus-eluting stents (SES) demonstrated that a postprocedural minimum stent area of <5 mm to 5.5 mm 2 was the most important predictor of angiographic binary restenosis at 6 to 9 months. A larger analysis of a number of IVUS substudies performed within the TAXUS trial program examined the outcomes of the paclitaxel-eluting stents (PES). This included 1580 PES and control group patients. Postprocedure, the stent MLA was similar in PES and BMS (6.6 mm ± 2.5 mm 2 vs. 6.7 mm ± 2.3 mm 2 , p = 0.92). At 9 months, in-stent restenosis was significantly lower in the PES group (10% vs. 31%, p < 0.0001). Postintervention stent MLA by IVUS was the independent predictor of subsequent restenosis in both the PES and control BMS groups (p = 0.0002 for PES and p = 0.0002 for BMS). The optimal thresholds of postprocedure IVUS MLA that best predicted stent patency at 9 months were 5.7 mm 2 for PES and 6.4 mm 2 for BMS.
One of only two prospective randomized trials of IVUS-guided DES stenting included 284 patients and focused on complex lesions defined as bifurcations, long lesions, chronic total occlusions, or small vessels. The primary endpoint was the postprocedure MLD. This showed a statistically significant difference in favor of the IVUS group (2.70 mm ± 0.46 mm vs. 2.51 ± 0.46 mm, p = 0.0002). At 24 months, there was no difference in death or MI. Target vessel revascularization was nonsignificantly lowered in the IVUS-guided group compared to the angiography-guided arm (9.8% vs. 15.5%). A similarly designed Korean randomized trial of 543 patients with long lesions treated with DES and randomized to IVUS guidance versus angiography guidance showed no difference in postprocedure stent length or MLD. The authors attributed that lack of improvement in the IVUS-guided arm to the high rate of crossover to IVUS utilization in the angiography-guided arm, which was allowed by protocol and occurred in 15% of patients.
A specific high-risk group of patients in whom IVUS guidance has a valuable role are those undergoing unprotected LMCA stenting. In the MAIN-COMPARE registry, data from approximately 1000 stable patients undergoing unprotected left main stenting was compiled. IVUS guidance was used in 80% of cases. The 3-year outcomes were compared between the two groups using propensity-score matching in the entire population and separately in those receiving DES. In 201 matched pairs of the overall population, there was a trend toward reduced mortality with IVUS guidance compared with angiography guidance, but that did not reach statistical significance (6.0% vs. 13.6%, Cox-model p = 0.061). In the 145 matched pairs of patients receiving DES, mortality was lower with IVUS guidance (4.7% vs. 16.0%, log-rank p = 0.048; hazard ratio, 0.39; 95% CI, 0.15 to 1.02; Cox model p = 0.055). While the statistical significance of the benefit appears marginal and the underlying mechanism is unclear, it may be hypothesized that optimal expansion and apposition achieved by IVUS guidance may reduce the risk of late and very late stent thrombosis with DES.
A large metaanalysis examined the role of IVUS guidance of DES procedures on clinical outcome. The metaanalysis included a very large number of patients (approximately 20,000), but it was weakened by the propensity of uncontrolled, nonrandomized patient registries (of the 20,000 patients included, there were only 105 randomized patients from one trial). That analysis demonstrated an improvement in death and rates of stent thrombosis with IVUS guidance compared to angiography alone but no difference in the need for repeat revascularization. While intriguing, these conclusions were primarily driven by results from two or three uncontrolled registries with unadjudicated outcomes and should be considered with caution.
Recent data suggest much more significant advantages to IVUS guidance of DES procedures, with evidence of reduced risk of stent thrombosis, myocardial infarction, and major adverse events. This was generated from the ADAPT-DES trial, which was a prospective, multicenter registry of 8583 patients receiving DES, of whom 3349 (39%) were enrolled in a prespecified IVUS substudy. The study aimed to determine the frequency, timing, and correlates of early and late stent thrombosis. Patients in the IVUS imaging arm were more likely to have ACS. IVUS guidance led to a modification of the intended procedure in 74% of cases (larger stent, longer stent, additional stent, additional postdilation, and/or higher pressures). After 1 year of follow-up, definite/probable stent thrombosis occurred in 18 (0.6%) patients in the IVUS group versus 53 (1.0%) in the non-IVUS group (HR 0.4, 95% CI 0.20, 0.7, p = 0.003). Other adverse clinical outcomes were significantly reduced in the IVUS-guided arm ( Table 16-2 ). After propensity matching, the use of IVUS remained an independent predictor of freedom from probable/definite stent thrombosis.
