Gray-Scale Intravascular Ultrasound

Gray-scale intravascular ultrasound (IVUS) utilizes one or more miniaturized transducers secured at the end of a flexible catheter that is inserted into arteries for in vivo tomographic visualization of vascular anatomy. Intracoronary ultrasound has become the most commonly used form of intravascular imaging and is regularly employed to delineate plaque morphology and distribution, to assess angiographically ambiguous lesions, to assess atherosclerosis progression and regression, to guide transcatheter coronary intervention, and to assess mechanisms of restenosis (Fig. 7-1).^1–2 The IVUS catheter is attached to either a portable or installed ultrasound machine through an interface that ensures sterility of the catheter and allows for motorized pullback. Motorized pullback of the IVUS catheter (or of the transducer within the catheter) at 0.5 mm/s or 1.0 mm/s allows calculation of lesion length and plaque volumes (necessary for progression/regression studies), ensures adequate time to visualize the lesion, and allows comparison between IVUS “runs” done at different times but with similar pullback speeds.

Figure 7-1 A, Pictorial representation of an intravascular ultrasound catheter inserted into a coronary artery after stent deployment. B, Actual intravascular ultrasound image of a well-deployed coronary stent.

There are two types of transducers used in IVUS catheters: mechanical and phased array. The mechanical catheter uses a single transducer that is mounted at the tip of the catheter and rotated quickly (at 1800 rotations per minute [rpm]). Because the IVUS transducer is oriented at a 90° angle to the length of the catheter, the images are displayed as cross-sectional views of the artery. The mechanical transducer has the advantage of a simple design and a greater signal-to-noise ratio. These properties account for its high image quality with an overall resolution (with current 40- to 45-MHz transducers) of approximately 100 to 120 µm. The primary disadvantage of mechanical transducers is that they require a central drive shaft to rotate the transducer on the tip of the catheter. This drive shaft decreases catheter flexibility. If a mechanical transducer is placed within a tortuous vessel, the drive shaft may be constricted, resulting in nonuniform rotation of the distal transducer, thus creating an artifact of the image termed nonuniform rotational distortion.

The phased array transducer uses multiple, tiny transducer elements. Each transducer element is permanently mounted around the circumference of the catheter tip and sends and receives ultrasound from a small sector. The multiple imaging sectors are then “added together” to produce a cross-sectional image of the artery. Because the phased array transducer is permanently fixed with no moving parts, it is more flexible and transverses torturous vessels more easily. These transducers do not produce nonuniform rotational distortion artifacts. However, the 20-MHz phased array transducer does require more complex programming to add the sectors together and thus has a lower temporal resolution (a function of computational speed) and lower spatial resolution.

Image Formation

An IVUS image is formed when ultrasound bounces off the multiple layers of the artery and returns to the transducer.³ The coronary artery is a muscular vessel composed of three basic layers: intima, media, and adventitia (Fig. 7-2). The intima, only a few cells thick at birth, is the site of atherosclerotic plaque accumulation. The intima is directly surrounded by the media, a layer of predominately homogeneous smooth muscle cells that provides the artery with its vascular tone. The outermost layer is the adventitia. Each of these layers tends to have a different acoustic property, allowing each to be visualized separately. When ultrasound encounters the intima, there is a large change in acoustic impedance; and much of it is bounced back to the transducer and displayed as a single concentric echo. Unless the intima becomes extremely hard, such as with the formation of a calcified plaque, some ultrasound penetrates through to the media. Because the media is composed of homogeneous smooth muscle cells, it passes through with minimal reflection; thus the media appears as a dark zone devoid of echoes. The adventitia has numerous layers of collagen fibers and therefore is a highly reflective structure that appears bright. As a result of the different acoustic properties of each layer, the normal coronary artery has a three-layered appearance, which includes (1) a bright echo from the intima, (2) a dark zone from the media, and (3) bright echoes from the adventitia. Because the resolution of IVUS is approximately 100 to 120 µm, the intima has to be at least this thick to be visualized as a distinct structure. In Western society, most patients presenting to the cardiac catheterization laboratory have some diffuse thickening of the intima. But if the intima is truly disease free, as in a newborn, child, or adolescent, then the intima will be much thinner and below the resolution of IVUS so that the artery will appear as a monolayer (see Fig. 7-2).

Figure 7-2 Center image was magnified and displayed at right and left. (Right) Typical “three-layered” appearance indicating intima, media (dark band), and adventitia. (Left) “Mono-layered” appearance indicating very thin intima below intravascular ultrasound resolution, media (not visible), and adventitia.

Image Interpretation

The interpretation of an IVUS image involves only a few simple steps: (1) identify the IVUS catheter in the center of the screen, keeping in mind that the IVUS catheter is traveling down the blood-filled lumen; (2) identify the dark stripe of the media that tells the size of the artery in the absence atherosclerotic disease; (3) remember that all of the echoes within the media stripe represent intima or intimal (atherosclerotic) disease and (4) all of the echoes outside the media are adventitia (see Fig. 7-2).

The amount of ultrasound that reflects off of tissue depends on the acoustic impedance (density) of the tissue. This property provides IVUS with some ability to differentiate plaque of different compositions. Hard material (calcium or fibrotic plaque) will reflect more ultrasound and appear brighter, whereas soft plaque (fat) will not reflect the ultrasound and will appear dark. Calcium is so dense that no ultrasound penetrates to the deeper tissues and, thus, produces acoustic shadows; shadowing, along with the bright echoes is the hallmark of calcification. A conventional approach is to compare the overall brightness of the plaque to the surrounding adventitia. Plaque is typically characterized as hypoechoic if less dense than the adventitia, hyperechoic if more dense than the adventitia, calcified if bright with acoustic shadowing, or mixed. IVUS can detect dense calcium with high sensitivity and specificity,^4–6 but its accuracy in identifying high-risk lipid rich plaque or even thrombus formation is poor.⁷ As shown in Figure 7-3, gray-scale IVUS can also detect two other types of high-risk plaques: (1) noncalcified attenuated plaque (which correlates with fibroatheroma) and (2) calcium nodules (calcium with a convex shape and irregular surface).^8–10

Figure 7-3 Both intravascular ultrasound (IVUS) and corresponding histology (Russell-Movat’s pentachrome staining) are shown. (Top) IVUS image shows typical “attenuated plaque”—non-hyperechoic plaque with attenuation. Corresponding histology shows advanced fibroatheroma (black arrows). Intravascular ultrasound image (bottom) shows “calcified nodule,” which has convex irregular surface with acoustic shadow. Corresponding histology shows cluster of calcified nodules (black arrows).

Radiofrequency Ivus

Standard gray-scale IVUS is limited, in part, because it uses only the amplitude of reflected ultrasound to formulate the image. In an effort to improve on the qualitative assessment of the reflected ultrasound signal, Kawasaki and coworkers developed a plaque characterization algorithm called integrated backscatter IVUS using time domain information directly from the radiofrequency signal. This process has resulted in improved plaque characterization with a reported in vitro sensitivity of 90% and specificity or 92% for lipid-rich plaque.¹¹

In a similar effort to improve plaque characterization, spectral analysis (virtual histology [VH]-IVUS) combines signal frequency and amplitude analysis to create an algorithm that detects fibrous tissue (dark green), fibrofatty (light green), necrotic core (red), and dense calcium (white). Reported sensitivity and specificity of VH-IVUS are 91.7% and 96.6% for identification of the lipid-rich necrotic core.^12–14 However, VH-IVUS cannot detect thrombus (in fact, thrombus appears as either fibrotic or fibrofatty plaque depending on the age of the thrombus) and has not been validated for assessment of stent metal or intimal hyperplasia. VH-IVUS lesion phenotype has been classified as (1) VH thin-cap fibroatheroma, (2) thick-cap fibroatheroma, (3) pathologic intimal thickening, (4) fibrotic plaque, and (5) fibrocalcific plaque (Fig. 7-4). Fibrotic plaque has mainly fibrous tissue with less than 10% confluent necrotic core, less than 10% confluent dense calcium, less than 15% fibrofatty. Fibrocalcific plaque has mainly fibrous tissue with greater than 10% confluent dense calcium, but less than 10% confluent necrotic core. Pathologic intimal thickening has a mixture of all plaque components, but dominantly fibrofatty with less than 10% confluent necrotic core and less than 10% confluent dense calcium. Fibroatheroma (both VH thin-cap fibroatheroma and thick-cap fibroatheroma) has greater than 10% confluent necrotic core. Because the resolution of VH-IVUS is 150 to 250 µm, it is not possible to detect fibrous cap thickness less than 65 µm (the typical pathologic definition of a thin fibrous cap). Therefore, if there is greater than 30° of necrotic core abutting to the lumen in three consecutive slices, the fibroatheroma is classified as VH thin-cap fibroatheroma; otherwise, it is classified as thick-cap fibroatheroma.¹⁵

Figure 7-4 Virtual histology (VH) IVUS classification.

(red, necrotic core; white, dense calcium; light green, fibrofatty; dark green, fibrous tissue) VH-TCFA, VH thin-cap fibroatheroma; ThCFA, thick-cap fibroatheroma; PIT, pathologic intimal thickening; fibrotic plaque; and fibrocalcific plaque.

Ivus Versus Angiography

Angiography permits the measurement of only luminal diameters and typically in only two or three views. Because of its tomographic presentation and high resolution, IVUS is far more accurate than angiography in assessing atherosclerotic plaque extent, morphology, and distribution.

To appreciate and apply IVUS in clinical practice, one needs to have a full understanding of how IVUS and angiography differ. Angiography is often referred to as a “shadowogram” of the lumen because a silhouette of the arterial lumen is obtained by injecting radiopaque dye into the lumen and projecting the shadow from the dye on to cineangiographic film. The shadow-o-gram is then used to extrapolate a luminal narrowing to a plaque accumulation. It is common to detect atherosclerosis in an artery that appears “normal” by angiography. Although it may appear to be a discrepancy at first, it is not, and both the angiography and the IVUS studies are providing complementary and accurate information. The primary difference is that angiography is visualizing just the lumen, whereas IVUS visualizes both the lumen and arterial walls. If the assumption is that all angiographic stenoses are from plaque, then it would make sense to conclude that an angiogram displaying a lumen without any narrowing is free of plaque. However, that may not be true (and usually is not). Conversely, an angiogram of an artery can be free of any stenosis, but that artery can still have many areas of plaque. This is true because of three phenomena: (1) atherosclerotic plaque is often diffusely distributed, (2) arteries can undergo vascular remodeling, and (3) complex atherosclerotic plaques are not appreciated by the two-dimensional silhouette.

It has been verified in autopsy studies and other IVUS studies of healthy donors that most people have diffuse atherosclerotic disease throughout their arteries by midlife, and this disease often remains angiographically silent because of its diffuse nature. In addition to diffuse plaque, vascular remodeling is also responsible for the high prevalence of angiographically silent disease. Both pathology and IVUS studies have documented that coronary arteries will enlarge to accommodate focal plaque deposition in an attempt to maintain luminal integrity. Compensatory dilation of the arterial wall occurs in direct response to the accumulation of atherosclerotic plaque. An absolute reduction in lumen dimensions typically does not occur until the lesion occupies approximately 40% to 50% of the area within the internal elastic membrane (40% to 50% cross-sectional narrowing).¹⁶ As a result, most of the atherosclerotic burden is contained within angiographically normal reference segments.¹⁷

Complex luminal shapes can result from irregular plaques or the disruption of a plaque (from balloon angioplasty). Although such a lesion may appear hazy, it may not demonstrate a luminal stenosis when viewed from only one or two views and projected as a silhouette. This may be similar to an eccentric lesion, which can fool inexperienced angiographers, unless a large number and variety of views are obtained, each displaying a different severity. However, a lumen with highly irregular borders may not be fully appreciated, even in multiple views. The tomographic images of IVUS are necessary to appreciate the true luminal shape. Therefore, diffuse plaque, vascular remodeling, and irregular plaque can all lead to an inability of angiography to appreciate the plaque that is seen on IVUS. This is predominately because angiography only visualizes the lumen and does so in limited views using silhouettes of the dye.

Finally, angiography may contain areas that are hazy or lesions that are ambiguous. These hazy or ambiguous lesions may result from irregular plaque, a focal area of calcium, a dissection, or an intraluminal thrombus. IVUS is particularly useful for ambiguous lesions because each of the causes (plaque irregularities, calcium, dissection, or thrombus) appear different on IVUS.

Integration Into Practice

Many practical aspects of IVUS imaging are facilitated by designating specific IVUS technologists who become responsible for all practical aspects of IVUS imaging: (1) equipment and catheter setup, (2) image optimization, (3) proper recording of IVUS imaging runs, (4) accurate voice and onscreen alphanumeric documentation, (5) patient and procedure logs, and (6) equipment maintenance. With time, technologists can be trained to interpret images accurately, to provide the iterative feedback necessary for IVUS-based decision making, and to answer questions posed by the primary operators. Measurements should be made offline (from hard drive after the imaging run is complete), not when the catheter is in the vessel; this saves procedure time and minimizes patient ischemia.

The angiographic (i.e., “road-map”) monitors can be “wired” to display IVUS images. Angiographic monitors offer superior resolution, are convenient and readily visible to the operator, and allow the IVUS machine to be placed in a position away from the patient table and out of the way of the nurses providing patient care.

In a busy laboratory with multiple operators, it is important to standardize image acquisition and analysis. The use of a motorized transducer pullback device aids (in fact, enforces) discipline and acquisition standards; there is no question whether the transducer is being advanced or withdrawn. The preferred pullback speed is 0.5 mm/s; this is the fastest rate at which the trained eye can assimilate the information. Standardization of image acquisition facilitates offline image analysis and comparison of serial (preintervention versus postintervention or postintervention versus follow-up) studies; it is essential for multicenter studies.

Standard Ivus Acquisition Protocol

IVUS imaging should include careful uninterrupted imaging of (1) at least 10 mm of distal reference, (2) the lesion site(s), and (3) the entire proximal reference back to the aorta. Advantages of motorized pullback are (1) controlled catheter withdrawal so no segment of the vessel is skipped or imaged too quickly by pulling the catheter back too rapidly, (2) the ability to concentrate on the images without having to simultaneously pay attention to catheter manipulation, (3) the ability to make length and volumetric measurements, and (4) uniform and reproducible image acquisition for multicenter and serial studies. Disadvantages of motorized pullback are the following: (1) even at slow pullback speeds, it is possible to skip over focal lesions; (2) not enough attention may be paid to important regions of interest; and (3) it is not possible to have the transducer “sit” at one specific site in the vessel.

However, motorized transducer pullback does not preclude the addition of careful, manually controlled interrogation of the lesion. Manual transducer pullback should be at a slow rate similar to motorized pullback. Advantages are that it is possible to concentrate on specific regions of interest by having the transducer sit at a specific site in the vessel. Disadvantages include the following: (1) it is easy to skip over significant pathology by pulling the transducer back too quickly or unevenly, (2) length and volume measurements cannot be performed, and (3) antegrade and retrograde manual catheter movement can be confusing when the study is reviewed at a later date.

When interrogating ostial lesions, it is important that the guiding catheter be disengaged from the ostium. If not, the true ostial lumen dimensions may not be identified. In all cases, at least one continuous uninterrupted imaging run should be performed from approximately 10 mm beyond the target lesion to the ostial junction of the ostium with the aorta.

Safety

Three large studies have evaluated the safety of IVUS. Other than transient spasm, complications appear to be rare (Table 7-1).^18–20

Assessment Of Stenosis Severity

Coronary angiography underestimates stenosis severity most markedly in vessels with 50% to 75% plaque burden (plaque area divided by arterial area) at necropsy and in patients with multivessel disease.^21–25 Clinical events are determined by lumen size, not by the amount of plaque. Therefore, it is important to focus on accurate measurement of lumen dimensions and not to be distracted by the plaque burden. Plaque burden in patients with atherosclerosis tends to be large even in the absence of lumen compromise.

Studies performed more than 10 years ago suggested that an IVUS minimum lumen area (MLA) greater than 4 mm² was a valid criterion for deferring an intervention based both on comparisons to physiologic measures and on follow-up data.^26–29 This cutoff only applied to major epicardial vessels excluding the LMCA and excluding saphenous vein grafts. However, it was necessary to interrogate the lesion carefully to identify the image slice with the smallest lumen, especially in very focal stenoses. Once the smallest lumen is identified, careful measurement was required. Recently this cutoff has come under criticism for a number of reasons including the size of the vessels (studies were done mostly in large 3.5-mm vessels), the average MLA of the intermediate lesions studied (4 mm²), and the fact that these studies have been misinterpreted to suggest that lesions with an MLA below 4 mm² justify stent implantation. For this reason, it is recommended that for major epicardial vessels excluding the LMCA, physiologic lesion assessment (using an intracoronary pressure wire and maximum hyperemia with intravenous adenosine to measure the fractional flow reserve) is a complementary approach to assess lesion severity.³⁰

Assessment Of Lmca Disease

A high percentage of patients with angiographically “normal” LMCA with only nonsignificant stenosis have substantial disease by IVUS.^31–34 Reasons for the discrepancy between angiography and necropsy or IVUS assessment of LMCA disease include the following: (1) diffuse atherosclerotic involvement affects the angiographic diameter stenosis (DS) calculation because of the lack of a normal reference segment; (2) a short LMCA also makes identification of a normal reference segment difficult; (3) unique geometric issues exist in LMCA disease (the correlation between angiography and necropsy or IVUS appears to be somewhat better in non-LMCA stenoses); and (4) there is significant interobserver and intraobserver variability in the angiographic assessment of LMCA disease.^35–42 In fact, the LMCA is the coronary arterial segment with the greatest variability in angiographic assessment.

We reported 122 patients who underwent angiographic and IVUS assessment of the severity of LMCA disease, who did not have subsequent catheter or surgical intervention, and who were followed for 1 year.⁴³ There were three distinct predictors of these cardiac events: diabetes; a major epicardial vessel or bypass graft with a quantitative angiography (QCA) DS greater than 50% that was left untreated; and LMCA lesion site minimal lumen diameter measured by IVUS. An example is shown in Figure 7-5.

Figure 7-5 Normal left main coronary artery (LMCA) compared to LMCA with severe disease.

Jasti and colleagues reported that an IVUS MLA of 5.9 mm² or a minimal lumen diameter of 2.8 mm corresponded to physiologic ischemia (fractional flow reserve less than 0.75) in the LMCA.⁴⁴ Hernandez and co-workers evaluated 354 angiographically intermediate (diameter of stenosis 25% to 60%) unprotected LMCA lesions by IVUS. Of the patients evaluated, 186 had an MLA of at least 6 mm², and only 7 patients were revascularized.⁴⁵ In 2 years, percentages for survival free of cardiac death (97.7%) and revascularization (96.5%) were very high, suggesting a MLA of 6 mm² or greater as a safety cutoff to defer the revascularization of LMCA disease.

Lesion Length And Reference Dimensions

The two most important measurements for stent size selection are lesion length and reference dimension. Motorized transducer pullback is essential for length measurements and has been validated in vivo.⁴⁶ A comparison of IVUS and QCA reference lumen measurements in 3311 nonostial native coronary arteries or saphenous vein graft lesions showed a QCA proximal reference lumen diameter of 3.05 ± 0.68 mm and an IVUS proximal reference lumen diameter of 3.41 ± 0.62 mm (p less than 0.0001 versus QCA). The difference between the IVUS and QCA measurements was 0.36 ± 0.64 mm or 15 ± 24%. This difference was greatest in the smallest vessels, decreased asymptotically with increasing vessel size, and was zero at a QCA reference of approximately 4 mm shown in Figure 7-6. This analysis suggests that, especially in the smaller arteries (which have higher restenosis rates), (1) the angiographic assessment of reference lumen dimensions is flawed and (2) an IVUS measurement would indicate that a larger device could be used that could result in larger final lumen dimensions without having to resort to midwall or media-to-media device/vessel sizing.^47,48 Alternatively, some authorities have advocated the use of IVUS “true vessel,” “media-to-media,” or midwall dimensions for device sizing; however, this should only be done by highly experienced IVUS users because it carries a greater risk of acute complications such as stent edge dissection.

Figure 7-6 Correlation between reference diameter by quantitative angiography (QCA) and intravascular ultrasound (IVUS).

At smaller vessel sizes, IVUS dimensions are larger than QCA; at larger vessel sizes, the difference is small.

Unusual Lesion Morphology

IVUS is useful in understanding lesions with an unusual angiographic appearance including filling defects, haziness, aneurysms, ulceration, myocardial bridge, and spontaneous dissection.^49–53 Examples are summarized in Figure 7-7 and shown in Figures 7-8 through 7-11.

Figure 7-7 Summary of preintervention qualitative assessment by intravascular ultrasound (IVUS). If the angiogram shows a filling defect, IVUS will show either thrombus or calcium nodule. If the angiogram shows haziness, IVUS will show thrombus, calcified plaque, myocardial bridge, normal site, or spontaneous dissection. Causes of an angiographic aneurysm include true aneurysm, false aneurysm, plaque rupture, or just a normal site that appears aneurysmal.

Figure 7-8 Angiogram shows filling defect in the proximal right coronary artery.

Intravascular ultrasound shows calcified nodule.

Figure 7-9 All angiograms have the appearance of an aneurysm.

Intravascular ultrasound can differentiate true (three-layered) or pseudoaneurysm, plaque rupture, or just normal segments with adjacent stenoses.

Figure 7-10 Angiogram shows diffuse stenosis with haziness in mid to distal left descending artery.

Proximal segment of haziness appears as calcified plaque, and distal end appears as myocardial bridge by intravascular ultrasound (surrounded by low echoic band).

Figure 7-11 Angiogram shows diffuse mild stenosis in mid to distal left descending artery (A-D). Corresponding intravascular ultrasound images shows spontaneous dissection (with intramural hematoma).

An IVUS classification of angiographic coronary artery aneurysms has also been created.⁵⁰ Of 77 aneurysm-containing lesions, 21 (27%) were classified as true aneurysm; 3 (4%) were classified as pseudoaneurysms (all having previous intervention); 12 (16%) were complex plaques; and the other 41 (53%) were normal arterial segments adjacent to one or more stenoses. Examples are shown in Figure 7-9. In a study of 300 consecutive IVUS-defined plaque ruptures, 91% were categorized as complex by angiography: ulceration in 81%, intimal flap in 40%, thrombus in 7%, and coronary aneurysm in 7%.⁵¹ Approximately one quarter of patients had IVUS-detectable myocardial bridges in the LAD coronary artery; only 3% (11 of 331) were diagnosed by angiography. In 13% of cases, IVUS-detectable myocardial bridges appeared as mild (diameter of stenosis less than 25%) fixed lesions and/or angiographic haziness (see Fig. 7-10).⁵² Spontaneous dissections are an intramural process that appears as a medial dissection with an intramural hematoma occupying the dissected false lumen without intimal tears and without a communication between the true and false lumens.⁵³ An example is shown in Figure 7-11.

Assessment Of Vulnerable Plaque

PROSPECT (Providing Regional Observations to Study Predictors of Events in the Coronary Tree) was a prospective study of the natural history of coronary atherosclerosis using angiography and VH-IVUS to identify the clinical and lesion-related factors that predicted future coronary events.¹⁵ It was performed at 37 sites in the United States and Europe. Patients with acute coronary syndromes were enrolled after undergoing successful percutaneous coronary intervention of all culprit lesions responsible for the acute event. Angiography was performed of the entire coronary tree followed by gray-scale and VH-IVUS of the proximal 6 to 8 cm of all epicardial arteries. All baseline angiograms and gray-scale and VH-IVUS images were prospectively analyzed without knowledge of subsequent events using prespecified definitions and methodology. Plaque burden and luminal area as prospectively measured by gray-scale IVUS and plaque composition (thin-cap fibroatheroma) as assessed by VH-IVUS were determined to be independent predictors of future adverse cardiovascular events (Fig. 7-12).

Figure 7-12 Event rates for lesions with and without thin-cap fibroatheromas (TCFAs) at a median follow-up of 3.4 years. MLA, minimum lumen area; PB, plaque burden.

(From Stone GW, Maehara A, Lansky AJ, et al: A prospective natural-history study of coronary atherosclerosis. N Eng J Med 364:226-235, 2011.)

Prediction Of No-Reflow During Interventional Procedures

IVUS guidance may also be useful to identify high-risk lesions before intervention, such as those that cause no-reflow or periprocedural myocardial infarction, particularly noncalcified attenuated plaque.^54–57 Noncalcified attenuated plaque may be a gray-scale indication of a larger necrotic core and/or fibroatheroma (96% of attenuated plaques contained a fibroatheroma vs. 52.3% on nonattenuated plaques, p <0.0001).⁹ In support of this, multiple VH-IVUS studies have shown that a larger necrotic core is related to distal embolization during percutaneous coronary intervention in both acute coronary syndromes and stable angina patients.^58–61

Assessment Of Interventional Complications

IVUS detects poststent dissections or other complications more often than angiography.

One type of dissection, intramural hematoma, appears as an angiographic dissection in 60%, as newly developed stenosis that does not disappear after nitroglycerin (i.e., not a spasm) in 11%, and as no angiographic abnormality in 29%.⁶² Similarly, perivascular trauma appears as an angiographic perforation in 24%, as angiographic dissection in 33%, as new stenosis in 5%, and as no abnormality in 38%. Angiographic “haziness” at the stent edge may demonstrate an edge dissection or residual plaque, especially a calcified eccentric plaque, when studied by IVUS.

Restenosis, Late Malapposition, Stent Fracture, And Stent Thrombosis

In the bare metal stent (BMS) era, serial IVUS studies in humans showed (1) that tubular-slotted stent designs exhibit almost no chronic recoil, (2) that in-stent restenosis was almost entirely neointimal hyperplasia, (3) that there was no predilection for tissue accumulation within any one segment, and (4) that neointimal thickness was independent of stent size.^63,64 It is for these reasons that many studies have reported that the absolute minimum stent area (MSA) by IVUS was the strongest quantitative predictor of clinical and angiographic in-stent restenosis.^65–67

Because drug-eluting stents (DES) have less neointimal hyperplasia, the MSA necessary to minimize restenosis can be smaller than with BMS. Using the data from SIRIUS, Sonoda and coworkers reported that the best cutoff to predict an MLA at follow-up of 4 mm² was an MSA of 5 mm² for sirolimus-eluting stents (SES) and 6.5 mm² for BMS.⁶⁸ A similar study by Hong and colleagues showed that the MSA that best separated angiographic restenosis from no restenosis was 5.5 mm² in SES.⁶⁹ We combined various TAXUS studies and reported that the best cutoff to predict angiographic binary restenosis for the paclitaxel-eluting stent (PES) was 5.7 mm².⁷⁰ However, “bigger is better” is still true; also, (1) a larger MSA is still associated with a lower restenosis rate, whereas (2) an MSA of 5.0 to 5.5 mm² may not be achievable in small vessels that are increasingly stented.

On the other hand, full lesion coverage may be more important in the DES era than in the BMS era. Costa and colleagues reported that longitudinal geographic miss (GM)—injured or diseased segment not covered by SES—was observed in 47.6%,; axial GM—balloon-artery size ratio less than 0.9 or greater than 1.3—was observed in 35.2%; and both GM in 16.5%.⁷¹ One-year target vessel revascularization rates were 6.1% in the longitudinal GM group versus 2.6% in the no-longitudinal GM group (p = 0.001); the GM was the independent predictor for target vessel revascularization. In support of this, Liu and co-workers reported that the stent edge plaque burden that best separated restenosis from no restenosis was approximately 50%.⁷²

Figure 7-13 summarizes the percentage of neointimal hyperplasia (%NIH = NIH volume divided by stent volume) and late acquired malapposition in various multi-center randomized trials.^73–83 In BMS, %NIH averaged (approximately) 30% of stent volume, and %NIH was consistently greater in diabetes than nondiabetes.⁸⁴ In the DES era, %NIH was reported as 2.7% to 3.1% in SES, 2.5% to 6.9% in everolimus-eluting stents, 7.4% to 12.2% in PES, and 16.1% to 17.6% in zotarolimus-eluting stents. Though NIH varied among DES, the clinical efficacy (target vessel revascularization) and safety (stent thrombosis) were similar in noncomplex lesions.

Figure 7-13 Summary of %NIH (neointimal hyperplasia volume divided by stent volume) in various drug-eluting stent randomized trials. BMS, bare metal stent (gray); ZES, zotarolimus-eluting stent (green); PES, paclitaxel-eluting stent (lavender); EES, everolimus-eluting stent (orange); SES, sirolimus-eluting stent (dark blue). The numeric values above each bar indicate the incidence of late-acquired malapposition.

IVUS predictors of either drug-eluting acute/subacute stent thrombosis or restenosis are stent underexpansion and residual edge “problems,” such as GM, adjacent secondary plaques, or significant edge dissections—but not acute malapposition.^85–87 Stent underexpansion was more diffuse in stent thrombosis than restenosis.⁸⁸

IVUS studies have documented late-acquired stent malapposition in approximately 5% of BMS in non-STEMI (ST elevation myocardial infarction) patients, about 5% to 15% of BMS in STEMI patients and DES in non-STEMI patients, and about 25% to 30% of DES in STEMI patients.73–83,89–92 Late-acquired stent malapposition is typically the result of positive vessel wall remodeling and/or thrombus dissolution behind the stent struts (especially in myocardial infarction patients). Cook and colleagues reported that 10 of 13 very late stent thrombosis patients (mean 1.8 years after stent implantation) had stent malapposition with a malapposition area twice as large as in control patients with malapposition, but without very late stent thrombosis (mean = 8.3 vs. 4.0 mm²).⁹³ Alfonso and co-workers reported 15 (1.3%) patients with an angiographic aneurysm 9 months post-DES implantation; subsequently, 3 patients (with significantly larger aneurysms than others) suffered late stent thrombosis.⁹⁴ One metaanalysis suggested that there was an increased frequency of very late stent thrombosis in the setting of large areas of late malapposition.⁹⁵ In a second report, Cook and colleagues correlated the presence and size of the malapposition area with inflammatory cells from intracoronary aspirates, suggesting that it was not just the malapposition, but malapposition in the setting of vessel wall inflammation, that contributed to very late stent thrombosis.⁹⁶ In addition, in-stent neoatherosclerosis and subsequent rupture causing very late stent thrombosis has been reported by IVUS, VH-IVUS, optical coherent tomography, and coronary angioscopy.^97–100

More recently, stent fracture is being implicated as a cause of DES failure, especially with SES. In a prior retrospective IVUS study of 17 non-STEMI patients with 20 stent strut fractures, 18 (90%) occurred in SES and 2 (10%) in BMS, but none in PES.¹⁰¹ Of note, 5 of 20 stent fractures were accompanied by coronary aneurysm formation with two very late stent thromboses (Fig. 7-14). It has been proposed that biologic reactions to the eluted drug cause positive vessel remodeling with aneurysm formation and malapposition that allow motion and/or kinking of the stent within the aneurysm, leading to strut fracture. Angiographic data from three PES trials in nonacute coronary syndromes revealed a 1.1% rate of stent fracture associated with restenosis and/or stent thrombosis.¹⁰² In the Nordic IVUS study, stent malapposition and stent fracture were found in 43% (37 of 87) and 16% (14 of 87) of DES thrombosis cases and 27% (10 of 37) and 27% (10 of 37) of BMS thrombosis cases.¹⁰³

Figure 7-14 An example of late-acquired aneurysm with stent fracture and very late stent thrombosis.

Treatment Of In-Stent Restenosis

The uses of IVUS in evaluating patients with in-stent restenosis include (1) determining whether in-stent restenosis is severe enough to treat, (2) identifying occult mechanical problems with the stent that (presumably) occurred at the time of implantation, (3) assessing the patterns of in-stent restenosis, and (4) assessing the acute results. Using integrated TAXUS data (331 treated with PES and 304 treated with BMS who did not require revascularization in the first 9 months postintervention), we reported that the optimal thresholds of MLA that best predicted subsequent target vessel revascularization-free survival during the subsequent 3 years were 4.2 mm² for PES and 4.0 mm² for BMS.¹⁰⁴

There are a number of technical and mechanical complications of stent implantation that can remain unrecognized until the patient presents with in-stent restenosis. These include (1) missing the lesion, (2) stent underexpansion, (3) stent “crush,” and (4) having the stent stripped off the balloon during the implantation procedure. Because most stents are radiolucent, some lesions, especially aortoostial lesions, can be missed. Incomplete stent expansion during implantation can be missed angiographically because stents are porous and contrast can flow through and around them. Chronic stent recoil is rare.^63,105 If the guidewire is accidentally removed and, in recrossing the freshly placed (and presumably not fully implanted) stent, it courses adjacent to the stent and enters the stent through one of the diamonds, subsequent adjunct balloon dilation can crush part of the stent against the vessel wall. These mechanical problems constitute a minority of in-stent restenoses. However, the treatment of in-stent restenosis must begin by excluding these problems and, if they are present, by correcting them. Examples are shown in Figures 7-15 and 7-16. In an analysis of more than 1000 consecutive in-stent restenosis lesions, the frequency of these mechanical complications was found to be 4% to 5%.¹⁰⁶

Figure 7-15 The implanted stent is crushed against the vessel wall.

Figure 7-16 Chronic stent underexpansion in a saphenous vein graft.

Bioabsorbable Stents

The first generation of bioabsorbable everolimus-eluting stents (BVS) implanted in 30 simple de novo lesions showed higher acute recoil than conventional metallic platform stents. At 6 months, the 11.8% reduction in scaffold area (chronic recoil) and a small neointimal response of 5.5% resulted in a total 24.3% decrease in MLA by IVUS.¹⁰⁷ However, at 2 years, the lumen area increased because of decrease in plaque size without change in vessel size.¹⁰⁸ The second-generation BVS had an improved scaffold that had only a 2% chronic recoil.¹⁰⁹

Does Ivus Guidance Reduce Stent Restenosis And Thrombosis?

A number of randomized and nonrandomized studies compared IVUS versus angiographic guidance of BMS implantation.^110–112 A metaanalysis of the seven randomized IVUS versus angiography trials showed that IVUS improved final BMS dimensions to reduce BMS restenosis and repeat revascularization.¹¹² In the DES era, Roy and co-workers reported that IVUS guidance reduced stent thrombosis both within 1 month and from 1 to 12 months (0.7% vs. 2.0% at 1 year, p = 0.014), supporting the use of IVUS during interventional procedures.¹¹³ The improvement of major adverse cardiac events, particularly mortality, has been supported by Park and colleagues in unprotected LMCA-treated arteries.¹¹⁴

Practical Recommendations

In practical day-to-day terms, how should IVUS be used during stenting procedures? The following is one suggested algorithm for routine use of IVUS. It begins with preintervention IVUS and concludes with iterative IVUS to optimize the final result. Perform preintervention IVUS to assess lesion characteristics and measure reference vessel size and lesion length. From the maximum reference lumen diameter, determine the maximum achievable stent dimension. Select stent length based on distance between proximal and distal references.

After the stent is implanted, repeat IVUS imaging to assess the MSA. If the MSA is adequate, stop. If the MSA is inadequate, perform additional high-pressure inflations, if necessary using a larger balloon. If additional balloon inflations are performed, recheck the IVUS results. If IVUS detects a complication such as a major dissection, an intramural hematoma, or perforation, treat appropriately and repeat IVUS to ensure that the complication is corrected. Conversely, if, despite the angiographic appearance, IVUS imaging shows a good acute result, stop.

If the interventional cardiologist wants to selectively used IVUS, it will be especially useful in the following three situations: (1) when you expect a good result, but flow is limited (IVUS can identify a correctable cause such as flow-limiting dissection or intramural hematoma); (2) any patient presenting with stent failure in which IVUS may identify a mechanical problem such as simple stent underexpansion or a crushed stent; and (3) with LMCA lesions or small-vessel stenting.