In-Stent Restenosis and Thrombosis—Etiology, Prevalence, Pathophysiology

Restenosis after angioplasty can result from early vessel recoil, late constrictive remodeling (also called negative remodeling), and neointimal proliferation. Elastic recoil occurs nearly instantaneously, secondary to passive recoil of elastic media. Late constrictive remodeling occurs within 1 to 6 months and is secondary to increased collagen in the extracellular matrix and adventitial thickening.¹ Intracoronary stents prevent these mechanisms (acute recoil and constrictive remodeling) (Fig. 37-2).

FIGURE 37-2 Mechanisms contributing to coronary restenosis. Elastic recoil and negative remodeling contribute to stenosis but are controlled by coronary stenting. Stenting does, however, not reduce and indeed may promote neointimal proliferation.

(From Dobesh PP, Stacy ZA, Ansara AJ, et al. Drug-eluting stents: a mechanical and pharmacologic approach to coronary artery disease. Pharmacotherapy 2004; 24(11):1554-1577.)

Unfortunately, stent-induced vascular injury causes increased neointimal hyperplasia/proliferation. The pathophysiology includes proliferation and migration of smooth muscle cells and extracellular matrix formation.¹ In-stent restenosis, the principal problem after angioplasty, has a clinical incidence of 20% to 40% for bare-metal stents.^5,6 Although several characteristics of high-risk populations have been described as clinical predictors, the likelihood of restenosis in a particular patient remains largely unpredictable.^6,7

Drug-eluting stents (DES) have revolutionized the treatment of coronary artery disease through marked reduction of in-stent restenosis.⁸ The components of these stents can be divided into a platform (the stent), a carrier (usually a polymer), and an agent (a drug) to prevent restenosis (Fig. 37-3).² The advantage of using a stent as a delivery system is that it allows for local delivery of the drug and averts the need for higher systemic doses. The biology of vascular cells and the cell cycle is being used as a target to prevent restenosis and current agents used in drug-eluting stents all interfere with the cell cycle in some manner.¹ Numerous trials have established that sirolimus and paclitaxel drug-eluting stents markedly reduce the incidence of in-stent restenosis. DES have reduced restenosis rates to less than 10% at a 12-month follow-up.^6,9,10

FIGURE 37-3 Schematic diagram of a bare-metal stent and a drug-eluting stent. A drug-eluting stent is coated with a polymer that allows consistent controlled release of a drug into the arterial wall.

Although drug-eluting stents have significantly reduced restenosis events, its very effectiveness has led to increased rates of an uncommon but potentially very severe complication, in-stent thrombosis. To understand this phenomenon, it is important to be aware of risk factors for development of in-stent thrombosis. Persistent, slow coronary blood flow such as occurs with dissection or hypoperfusion, exposure of the blood to prothrombotic subendothelial constituents, such as tissue factor or to the stent itself, and failure to suppress platelet adhesion/aggregation at a time of prothrombotic risk predispose the patient to in-stent thrombosis.¹¹

A probable explanation for an increased risk of thrombosis in DES compared to BMS after cessation of antiplatelet therapy is delayed arterial healing.¹¹ Normally after vascular injury, the vessel wall undergoes a number of changes including migration and hypertrophy of vascular smooth muscle cells. During this period, endothelial cells colonize the surface of the stent and regain their normal function.¹² DES cause incomplete neointimal coverage and delay endothelialization of the stent. Subsequently, there is eventual propagation of thrombi over the surface of the stent.

The reported incidence of thrombosis of bare-metal stents 1 month postprocedure has ranged from 0.5% to 2.5% in clinical trials. Bare-metal stent thrombosis usually occurs within the first 24 to 48 hours or much less commonly within the first month after stent placement. Eighty percent of stent thrombosis events have been found to occur within the first 2 days and events occurring more than 1 week after the procedure were rare.¹³ Stent thrombosis within 1 year of implantation occurs with similar frequency between DES and BES; however beyond 1 year, there is a small but significant increased risk of very late stent thrombosis in DES.¹⁴ Because of this increased risk, extension of antiplatelet therapy is recommended.

The decision to place a drug-eluting rather than a bare-metal stent therefore rests on clinical grounds and evaluation of the benefits and risks of these stents. The lower rate of repeat target revascularization with DES must be weighed against the cost of longer term antiplatelet therapy to prevent stent thrombosis, the risk of bleeding, and the complications of noncompliance with drug therapy (Fig. 37-4).

FIGURE 37-4 Illustration demonstrating potential complications of coronary stenting: Restenosis in a traditional bare-metal stent (A) and late thrombosis in a drug-eluting stent (B).

(Reprinted with permission from Curfman GD, Morrisey S, Jarcho JA, et al. Drug-eluting coronary stents—promise and uncertainty. N Engl J Med 2007, 8;356 (10):1059-1060.

Clinical Presentation

Restenosis following angioplasty and stent implantation is classically considered a benign process in which the typical patient presents with exertional angina.¹⁴ Importantly, however, an appreciable proportion of these patients present with an acute coronary syndrome. One study from the Cleveland Clinic demonstrated that 9.5% of patients with restenosis presented with acute myocardial infarction and 26.4% presented with unstable angina.¹⁵

Stent thrombosis is particularly concerning because of its potential catastrophic consequences. As stents are often placed in proximal segments of major coronary arteries, thrombotic occlusion usually manifests as severe ischemia or myocardial infarction. One study demonstrated that 70% of cases of stent thrombosis manifested as acute myocardial infarction.^8,16 Mortality rates are very high ranging from 11% to 15% for BMS thromboses and 25% to 45% with DES thrombosis. The higher mortality with DES thrombosis has been suggested to be secondary to a combination of abrupt thrombotic events and decreased collateral formation.¹⁴ Again, the risk of coronary artery stent thrombosis and its frequent consequences of myocardial infarction or death are minimized by the use of dual antiplatelet therapy. Patient compliance is also an important issue when determining the type of stent to use.

Imaging Techniques and Findings

CT

Noninvasive assessment using computed tomography to assess in-stent stenosis has been attempted since the era of electron beam CT (EBCT). Early studies with EBCT could not directly visualize in-stent restenosis and quantification was not possible.^24,25

Multidetector CT (MDCT) has several advantages over EBCT; of note, MBCT has increased spatial resolution and improved signal-to-noise ratio.¹⁷ Early MDCT scanning with four-detector scanners were inadequate for stent interpretation with inability to visualize the majority of the stent lumen.^17,26 Subsequent studies using 16-, 40-, and 64-detector scanners have had mixed results. Most studies demonstrated high sensitivity and specificity but in general ignored nonevaluable segments from analysis. The number of stents found nonevaluable (up to 54% of 232 stents in a study by Gilard and colleagues), is a major limitation of this analysis. The negative predictive value was generally greater than 95%; however, the positive predictive value ranged from 29% to 78%.^6,27–34

Hamon and colleagues performed a meta-analysis of 15 studies analyzing the diagnostic capabilities of 16-, 40-, and 64-detector CT scanners in comparison with invasive coronary angiography for detection of in-stent stenosis. 13% of 1175 stents in these studies were nonevaluable. After exclusion of these stents from analysis, they found a sensitivity of 84% and a specificity of 91%.³⁵ The positive-predictive value (PPV) was almost uniformly low secondary to nonevaluable stents being classified as in-stent restenosis (ISR) for statistical analysis. An additional meta-analysis of 16- to 64-detector row CT scanners performed by Sun and colleagues demonstrated similar results. Although there was the hope that 64-detector row CT scanners would be more accurate than slower 16-detector row CT scanners, both meta-analyses demonstrated equivalent sensitivity and specificity.^35,36 Kumbani and coworkers recently performed a meta-analysis of 14 studies solely using 64-detector row CT scanners for detection of ISR. Overall sensitivity was 91%, specificity was 91%, PPV was 68%, and the negative-predictive value (NPV) was 98%. Nine percent of 1447 stents were deemed nonevaluable with decreased performance with inclusion of these stents. Only five studies in the analysis included stents less than 3 mm.³⁷ Given that a large number of stents in the general population are less than 3 mm, this further decreases the utility of MDCT in nonselected stent patients.

A post-hoc analysis of 75 stents in 52 patients from the Core-64 trial, the first multicenter international, single-blinded study to determinate the accuracy of 64-detector row CT MDCT, was performed to assess the accuracy of MDCT in detecting in-stent restensosis. This analysis demonstrated less favorable results than the previous meta-analyses. Only 48 of 75 stents (64%) were considered evaluable and there was an overall accuracy of 77.1% for detection of 50% in-stent restenosis of evaluable stents. The PPV and NPV was 57.1% and 80.5%, respectively. A quantitative approach using in-stent and peri-stent attenuation didn’t improve accuracy. The researchers attribute the poor performance of MDCT to the fact that 80% of stents were <3 mm with secondary problems of calcifications, motion artifact, and blooming artifact.³⁸

In general, there is increased evaluability and improved detection of in-stent restenosis with larger stents and unfavorable results with stents less than 3 mm diameter. High heart rates, calcification, and increased body mass index decreased the accuracy of MDCT.³² When considered in the evaluation, thicker stent struts also reduced the evaluability of the stents.³⁶

Dual-source CT scanners have improved temporal resolution compared with single source scanners and promise to potentially minimize motion artifacts that limit in-stent evaluation. Pugliese and associates evaluated a dual-source CT scanner for evaluation of instant restenosis in 100 patients with 178 stents. Nine of the stents were uninterpretable (all of which were smaller than 2.75 mm in diameter) secondary to high-density artifacts obscuring the stent lumen. Sensitivity, specificity, PPV, and NPV in detecting >50% restenosis was 94%, 92%, 77%, and 98%, respectively. The diagnostic performance of the dual source CT scanner at heart rates less than 70 bpm did not differ significantly from its performance with faster heart rates. The stent diameter was important as the sensitivity was 100% in >3 mm diameter stents but dropped to 84% in <2.75 mm stents. Similarly the specificity and the PPV dropped precipitously as the stent diameter decreased. This study demonstrated that the dual-source CT scanner was able to negate the effect of heart rate on evaluation of stent stenosis. The NPV was very good (97% to 98%), even in patients with fast heart rates. Stent diameter was the most important feature in this study.³⁹

In the evaluation of the coronary arteries in a patient with history of percutaneous coronary intervention and stent placement, a clinical history with specific locations of stent placement is of utmost importance. Given that concentric calcification can mimic a stent and not all stents are uniformly recognized on MDCT;⁴⁰ a history of stent placement allows for a more accurate interpretation—understanding that even with the most current scanners, evaluation of in-stent restenosis remains challenging.

CT scanners with 64 or more detectors are, in general, currently considered the standard for coronary CTA because the reduced breath-holding time afforded by these scanners is better tolerated by patients owing to minimization of motion artifacts.^7,41 Metallic stents cause beam hardening on CT imaging resulting in stent struts appearing thicker (or “bloomed”) than they really are with secondary overlap of the vessel lumen and underestimation of the in-stent luminal diameter. In addition, this artifact can cause streaky dark bands which can simulate stenosis (Fig. 37-5).⁴² Calcification of the vessel wall near or at the outer surface of an implanted stent also contributes to beam hardening. Partial volume averaging is another artifact inherent in cross-sectional imaging and particularly problematic given the small caliber of the coronary arteries.^7,43

FIGURE 37-5 In-stent restenosis in the first diagonal coronary artery cannot be excluded secondary to significant blooming artifact. This artifact results from beam hardening and causes stent struts to appear thicker than they are with secondary underestimation of in-stent luminal diameter.

Sharp filters and thin slices (0.5 to 0.6 mm) reduce blooming and partial volume averaging for improved assessment of stent patency.⁴⁴ Studies have demonstrated that dedicated edge-enhancing convolution kernels decrease severity of blooming artifacts, resulting in superior depiction of the stent lumen.⁷ The disadvantages of these filters are an increase in image noise. The most appropriate filter must be chosen to account for this trade-off.

As with all cardiac CT examinations, reduction of cardiac motion, which exacerbates metal-related artifacts, as well as optimization of the contrast agent (fast speed of injection and high iodine concentration of contrast), improves in-stent evaluability. After the scan is performed, electrocardiographic (ECG) editing should be implemented in order to remove aberrant beats from the data set.

Analysis

When these technical factors are optimized and the data are obtained, the stent lumen should be evaluated using multiplanar reformation (MPR) and cross-sectional/short axis images. Appropriate display window settings must be considered to most accurately evaluate these images. Wide window settings (e.g., width 1500 HU) and center level of 300 HU have been recommended to further decrease image interpretation difficulties with blooming artifact.⁷

The first step in analyses is the determination of stent evaluability. If there is excessive blurring of the stent contour, motion artifact, or inadequate delineation between the stent and the surrounding fatty tissue or the in-stent lumen, the stents should be considered nonevaluable and should be reported to the clinician as such. As numerous papers have demonstrated, a significant number of stents are not evaluable and, therefore, cannot be commented on with any degree of accuracy.

When deemed evaluable, the in-stent lumen should be check for in-stent restenosis. The level and quality of enhancement should be ascertained and any in-stent filling defects should be noted. Homogeneous high attenuation in the stent similar to the attenuation in the proximal or distal reference vessel implies normal flow (Figs. 37-6 and 37-7). Different grades of restenosis similar to those used by Gasper and colleagues and Ehara and associates^27,34 can be used (Figs. 37-8 and 37-9), including mild neointimal proliferation (<50% stenosis), significant (>50% restenosis), severe (>75% stenosis)/possible occlusion, and definitive occlusion (100% occlusion).

FIGURE 37-6 Homogeneous contrast attenuation within LAD stent demonstrates lack of significant in-stent restenosis.

(Courtesy of Dr. Sablayrolles, Saint-Denis France.)

FIGURE 37-7 RCA stent is fully patent with homogeneous enhancement throughout the stent and minimal blooming artifact. Furthermore, contrast enhancement in the RCA proximal and distal to the coronary stent is similar in attenuation, suggesting lack of significant in-stent restenosis.

FIGURE 37-8 A 56-year-old woman with chest pain after LIMA graft, s/p stent placement in the left main coronary artery. Axial CT (A) demonstrates a stent in the left main coronary artery (brackets). Long axis (B) and short axis images (C) demonstrate significant lack of contrast enhancement of the proximal portion measuring >75% of the vessel lumen (white arrow) consistent with severe in-stent restenosis or total occlusion. Distal flow does not exclude the possibility of total occlusion because collateral vasculature can supply distal segments. Compare this appearance to the normal distal in-stent appearance (D) with homogeneous enhancement. Finding was confirmed on conventional catheter angiography with 90% restenosis noted.

FIGURE 37-9 Blooming artifact somewhat limits evaluation of this stent in the second diagonal coronary artery. Given this limitation, there is apparent decreased attenuation within the stent (large arrow) and in-stent restenosis (ISR) cannot be excluded. More concerning are foci of marked decreased attenuation seen on curved MPR (A) and computer-generated straightened MPR images (B) in the immediate persistent vasculature (arrows). There is also global decreased attenuation in the distal coronary vasculature (bracket). These findings are concerning for ISR and persistent occlusion with collateral vasculature reconstituting distal vasculature. Angiogram (not shown) demonstrated severe ISR and high-grade persistent stenosis but no occlusion.

The perigraft vasculature should also be analyzed as restenosis at stent borders and is reported to occur frequently.⁴⁵ No contrast distal to the stent indicates definitive occlusion.⁴⁶ The presence of contrast distal to the stent does not, however, exclude occlusion secondary to the possibility of collateral vasculature feeding the vessel segments distal to the occluded stent in a retrograde direction (Fig. 37-10). It has also been noted that with stent restenosis, the CT attenuation of vessels distal to the stent is decreased compared with the attenuation of the artery proximal to the stent; however, no specific cutoff point has been ascertained.³¹

FIGURE 37-10 Multiple stents in the left circumflex artery in this 70-year-old man with recurrent angina pectoralis. There is complete lack of contrast enhancement seen on curved MPR images (A) as well as computer generated “straightened” MPR images (B). Short axis views (C), perpendicular to the center long axis of the stent, confirms lack of contrast enhancement filling the entire lumen consistent with high-grade stenosis or total occlusion. Lack of contrast enhancement in the coronary artery distal to the stents confirms total occlusion.

The diagnosis of in-stent thrombosis is much rarer and requires a collaborative clinical history. This is a highly morbid event that results in Q-wave infarction and death in the majority of cases. It is likely to show complete occlusion on cardiac CTA and the angiographic definition is no flow or faint flow beyond the occlusion.⁴⁷

Background

Surgical revascularization is well recognized as a modern achievement in therapy for advanced coronary artery disease. Although the concept of coronary bypass grafting for occlusive disease was originally proposed by Carrel in 1910,⁵² Drs. Michael DeBakey and Edward Garrett are credited with the first successful saphenous vein graft (SVG) bypass in 1964. Performed as a salvage for a complicated left anterior descending coronary endarterectomy, the patient did well postoperatively and demonstrated continued patency of the SVG graft when restudied 8 years later.⁵³ That same year, Kolesov performed the first sutured mammary-artery-coronary bypass.⁵⁴ It wasn’t until the end of the decade that broader acceptance and use of internal mammary grafting took place.

Within a decade from these initial advances, coronary bypass operations had a tremendous impact on therapy for atherosclerotic disease. The National Center for Health Statistics report 450,000 coronary revascularization procedures performed in 2006.³ Continued improvements in operative techniques have allowed increasingly challenging patients to be operated on with success.

Selective conventional catheter coronary angiography is considered to be the reference standard for the assessment of CABG. However, recent advances in competing cross-sectional modalities—specifically, MDCT and MRI—represent minimally invasive alternatives with promising results for bypass patency evaluation.

Bypass Graft Conduits

The choice of conduits is highly dependent on the particular surgeon and institution. Primarily, the various conduits used for surgical revascularization can be divided into arterial and venous grafts.

Saphenous Vein Grafts

Saphenous venous grafts were first successfully used in a CABG operation in 1964.⁵⁵ Although resistant to spasm versus their arterial counterparts, the use of SVGs is limited by a higher occurrence of intimal hyperplasia and atherosclerotic changes after exposure to systemic blood pressure, resulting in lower graft patency rates.^56,57 Saphenous vein grafts are attached proximally from the ascending aorta to the coronary artery distal to the diseased coronary lesion(s). The SVG can be sutured directly to the anterior portion of the ascending aorta or can be attached with an anastomotic device.

Occlusive failures of saphenous vein grafts are well documented and have been extensively investigated.^58–63 A large angiographic study of bypass graft patency (n = 5065, 91% venous, 9% arterial) found 88% of grafts to be patent perioperatively, 81% to be patent at 1 year, and 75% of grafts to be patent at 5 years.⁶¹ Further declines in patency to 50% were observed in 353 grafts examined more than 15 years after revascularization. Nevertheless, continued improvements in surgical techniques, combined with concomitant antiplatelet or anticoagulant agents, and lipid-lowering therapy have allowed SVGs to remain a main choice for surgical bypass.

Arterial Grafts

Despite early success with internal mammary grafts, widespread acceptance came nearly a decade later. Compared with the saphenous vein, the internal mammary artery (IMA) has unique biologic characteristics that enable it to resist atherosclerosis and maintain high patency rates. The IMA has a nonfenestrated internal elastic lamina and lacks vaso vasorum inside the vessel wall, which tends to protect against intimal hyperplasia and cellular migration.⁶³ In addition, the medial layer of IMA is thin and is limited in muscle cells, resulting in a decreased tendency for maladaptive vasoconstriction.⁶⁴

Advantages of IMA conduits over SVGs include decreased postoperative mortality, improved cardiac event-free survival rates, and improved long-term patency rates well above 90% at 10 years.^62,63 Because of its proximity to the left anterior descending (LAD) artery and favorable patency rates, the left IMA (LIMA) is most commonly used as an in situ graft to revascularize the LAD or diagonal artery, supplying the anterior or anterolateral cardiac wall.

Other Arteries

The profound benefits afforded by IMA grafting have given foundation to utilization of arterial conduits as coronary bypass grafts including the right IMA, right gastroepiploic artery, radial artery, and inferior epigastric arteries. The use of right gastroepiploic and inferior epigastric arteries in CABG procedures has been limited because of the need to extend the median sternotomy to expose the abdominal cavity. Although the use of these arteries increases surgical time and technical difficulty of the surgery, these arteries can be used as a free graft to perform total arterial revascularization.^65,66 In rare instances, the right gastroepiploic artery may be used in situ for revascularization to the posterior descending artery.

Imaging Techniques and Findings

Catheter coronary angiography is still considered to be the gold standard for evaluation of severity of coronary artery disease. The procedure is not without risk of complications including pseudoaneurysm (1.2%), arrhythmia (1%), cerebrovascular accidents (0.3%), and myocardial infarction (0.2%). Noninvasive coronary imaging holds great potential utility in the pre-, peri-, and postoperative evaluation of bypass patients.

Computed Tomography

The ability of CT to characterize the patency of bypass grafts has been discussed since the 1990s^60,67–69 with first-generation (four- and eight-detector) MDCT scanners. A recent meta-analysis demonstrated that obstructive bypass graft disease can be detected by using at least a 16-detector row CT with a high diagnostic accuracy, with a sensitivity of 98%, a specificity of 97%, a positive-predictive value of 93%, and a negative-predictive value of 99%.⁷⁰

The advent of ECG gating and improved subsecond data acquisition has given rise to near complete suppression of cardiac motion with superior subcentimeter spatial resolution. In particular, 64-detector row CT devices, available since 2005, have not only shown the ability to acquire high-quality images, but also have demonstrated impressive accuracy in evaluating coronary artery stenosis and bypass graft patency in a larger cohort of patients.⁷¹

The increased number of detectors lead to increased temporal resolution (e.g., 83 msec) and spatial resolution (e.g., 0.4 × 0.4 × 0.4 mm³) and reduction of both cardiac and respiratory motion artifacts, leading to improved ability to assess graft stenosis and occlusion.¹³ Newer three-dimensional (3D) image processing algorithms and advanced volumetric visualization techniques can provide improved evaluation of grafts in multiple planes using various projections (Fig. 37-11).

FIGURE 37-11 A 57-year-old male with diffuse disease of the right coronary artery. A, Multiplanar view of the native RCA demonstrated diffuse mixed calcified and soft coronary plaqu (arrows). B, Multiplanar image of a saphenous vein graft is observed arising from the ascending aorta with end-to-side anastomosis with the posterior descending artery.

Ropers and colleagues were among the first to compare bypass graft imaging with 64-detector MDCT compared to conventional catheter angiography and reported sensitivity of 97% and specificity of 98% of MDCT for detecting graft occlusion.⁷² Other subsequent studies using 64-detector MDCT have reported sensitivity and specificity values of 93.3% and 100% and 91.4% and 100%, respectively, for graft occlusion and significant luminal stenosis (>50% luminal narrowing).^73–76

The protocol for CT bypass imaging is similar to that for coronary CTA. One important difference is that the scan should be extended superiorly to include the origins of the internal mammary arteries. As is routine, heart rate control and proper breathing instructions are critical to performing a diagnostic study. Development of a new generation of scanner technology such as cardiac freeze-frame technique, dual-source CT, and flat-panel CT promise further improvements in temporal resolution and the ability to acquire slices with a gantry rotation time of about 100 ms, potentially reducing the problems of breath-holding, motion artifact, and artifacts related to variations of heart rate during the scan.⁷⁷ The role of bypass assessment with MDCT holds great potential to replace conventional catheter angiography in the near future as MDCT continues to evolve.

Magnetic Resonance Imaging

With the potential of multiplanar imaging, lack of radiation, and ability to evaluate physiologic parameters, MRI is an important adjunct to coronary imaging.

Coronary imaging with MRA can be performed using two-dimensional (2D) or 3D techniques. Cardiac synchronization with ECG gating is necessary to minimize degradation of images owing to cardiac motion. Motion of the diaphragm also presents a limitation. Different techniques, such as a breath-hold technique versus a free breathing with navigator echo-based technique, have been used to overcome these issues.⁷⁸ For a 2D coronary MRA using the breath-hold technique, the patient may be required to do many breath-holds for 16 to 20 seconds (and even longer for a 3D coronary MRA).^79,80

Few studies have focused on the ability of MR to assess coronary bypass grafts. Brenner and coworkers evaluated 85 patients after an average of 7 days post CABG and were able to assess saphenous and arterial grafts with 94% specificity and 90% sensitivity.⁸¹ Langerak and colleagues focused only on venous grafts and demonstrated a sensitivity and specificity of 65% to 82% and 82%, respectively, for the detection of a graft stenosis >50% and 73% and 80% to 87% for a graft stenosis >70%.⁸²

Cardiac MRA is particularly limited in visualizing IMA bypass grafts. This is primarily due to the smaller diameters of arterial versus venous grafts (1 to 3 mm compared to the 3 to 6 mm), as well as the imaging artifacts arising from surgical metal clips. Several early studies failed to evaluate IMA grafts simply due to the suboptimal image quality.^83,84 Notwithstanding these results, Knoll and colleagues⁸⁵ and Wintersberger and associates⁸⁶ compared the assessment of IMA grafts with MRA versus conventional angiography. Both studies were able to determine IMA graft patency with MRI with a sensitivity of 94%

Despite these initial investigations, MR imaging of bypass grafts remains difficult in patients with atrial fibrillation and tachycardia. Furthermore, extremely agitated or claustrophobic patients cannot be examined. Whole heart SSFP coronary magnetic resonance angiography has emerged as another promising MR technique that has already shown promise in evaluating native coronary arteries.^87–90 However, its value in assessment of CABG is yet to be determined.

Analysis

Postoperative Evaluation

Using either prospective or retrospective ECG gated techniques, CTA images are acquired in a limited field of view with axial images centered on the heart. Both 3D volume rendering and multiplanar reformatted images (MPR) are used to assess the bypass grafts, proximal and/or distal graft anastomoses, and the cardiac anatomy. In particular, curved planar images with center lines through the bypass grafts and native coronary arteries should be obtained.

It is important to follow the anatomic course of the bypass graft and note the patency of the graft at its expected target. The LIMA extends from its origin at the subclavian artery and courses through the anterior mediastinum along the right ventricular outflow tract after being separated surgically from its original position in the left parasternal region. Infrequently, sequential distal anastomoses, with side-to-side and end-to-side anastomoses to the diagonal and LAD arteries, respectively, or involving separate sections of the LAD artery, are performed (Figs. 37-12 and 37-13).

FIGURE 37-12 A 64-year-old male with left internal mammary to left anterior descending coronary bypass. Sequential axial slices demonstrate the course of the LIMA graft running along its usual parasternal location (A, arrow) anterior to the heart (B, arrow). Note the mixed calcified and soft coronary plaque in the proximal LAD (arrowhead). The LIMA graft then courses along the apex of the heart to subsequently anastomose with the native LAD (C, curved arrow, and D, arrowhead). Additional multiplanar view provides full depiction of the course of the graft (E).

FIGURE 37-13 A 63-year-old patient with three vessel bypass. Volume rendered (A) image of the heart demonstrates patent left internal mammary graft to LAD (small arrow), and saphenous venous graft (large arrow) arising from the ascending aorta to a posterolateral branch from the RCA. The radial artery graft is observed as a smaller caliber vessel arising from the ascending aorta and attached to a diagonal branch from the LAD (arrowhead). The entire course of the graft is best seen on the multiplanar reconstructions (B, C, D).

The right IMA (RIMA) is used less frequently than the LIMA. The RIMA can be placed as an in situ graft to revascularize the RCA or one of its branches. In cases in which both in situ IMAs are necessary to revascularize the left heart, either the LIMA is anastomosed to the left circumflex coronary artery (LCx) or side branches (i.e., obtuse marginal [OM] or diagnonal branches) and the RIMA is attached to the LAD artery, or the LIMA is connected to the LAD artery and the RIMA is joined to the LCx artery or OM branches by extension through the transverse sinus of the pericardium. The RIMA can also be used as a composite or free graft. As a composite or “Y” graft, the RIMA is anastomosed proximally to LIMA, allowing total arterial revascularization instead of using a venous graft with LIMA.

On CTA, the LIMA is usually visualized as a small vessel in the left anterior mediastinum. As with other grafts, the distal anastomosis is typically difficult to visualize. Surgical clips are used routinely to occlude branch vessels of the IMA, and metallic artifact may limit assessment in some instances. The CTA appearance of the RIMA is similar to that of the LIMA.

Complications

Graft Thrombosis and Occlusion

Bypass graft failures are classified either as early or late following CABG surgery. During the early phase, usually within 1 month after CABG surgery, the most common cause of graft failure is thrombosis from platelet dysfunction at the site of focal endothelial damage during surgical harvesting and anastomosis.⁹¹

Additionally, other factors (such as the hypercoagulability state of the patient and the high-pressure distention or stretching of the venous graft, with its intrinsically weaker antithrombotic features) further initiate early venous graft failure, resulting in a 3% to 12% occlusion rate within 1 month postoperatively.⁹²

Late-phase venous graft failure is due primarily to progressive changes related to systemic blood pressure exposure. One month after surgery; the venous graft starts to undergo neointimal hyperplasia.⁹² Although this process does not produce significant stenosis, it is the foundation for later development of graft atheroma. Beyond 1 year, atherosclerosis is the dominant process, resulting in graft stenosis and occlusion (Fig. 37-14).

FIGURE 37-14 A 63-year-old man with occlusion of a saphenous graft to diagonal-obtuse marginal graft (A,B). Perioperative study shows a patent aortocoronary graft: a saphenous vein to diagonal graft and sequential to the obtuse marginal artery (arrow on volume rendered image [A], and arrow on CT scan [B]) as well as a part of the left internal mammary graft. C and D, Study obtained 13 months later shows flow only to the left internal mammary graft (curved arrow); the saphenous graft has occluded.

Arterial grafts, such as IMA grafts, are resistant to atheroma development. Late IMA graft failure is more commonly due to progression of atherosclerotic disease in the native coronary artery distal to the graft anastomosis (Fig. 37-15).⁹³

FIGURE 37-15 A 78-year-old man with occlusion of a radial artery graft to obtuse marginal artery. A and B, Perioperative study shows two patent aortocoronary grafts on volume rendered image: a saphenous vein to diagonal graft (thin arrow on volume rendering [VR] and arrowhead on CT scan) and a radial artery to obtuse marginal graft (thick arrow on VR and thick arrow on CT scan). C and D, Study obtained 11 months later shows flow only to the saphenous vein to diagonal graft (squiggly arrow on VR); the radial artery graft has occluded (arrowhead on CT scan).

CT angiography can delineate multiple findings associated with graft stenosis and occlusion. Calcified and noncalcified atherosclerotic plaque is readily identified, and the calculation of the extent of graft narrowing is straightforward. Occlusion can be determined by nonvisualization of a vessel known to have been used for surgical grafting. In many instances, the most proximal part of an occluded aortocoronary graft fills with contrast, creating a small outpouching or “nubbin” from the ascending aorta, allowing a diagnosis (Fig. 37-16). A residual low attenuation structure, or “ghost”, of part of the occluded graft may be visible. Acute or chronic graft occlusion can sometimes be differentiated by the diameter of the bypass graft (Fig. 37-17). In chronic occlusion, the diameter is usually reduced from scarring, as compared with acute occlusion in which the diameter is usually enlarged (Fig. 37-18).

FIGURE 37-16 A 56-year-old male with previous CABG. A and B, Volume rendered (curved arrow) and MPR demonstrate a small focal outpouching of contrast, which is immediately truncated. We have noted this appearance as a “nubbin” sign related to graft occlusion.

FIGURE 37-17 A 67-year-old patient with graft occlusion. A and B, Straight and curved MPRs of a saphenous vein graft demonstrating long segment decreased attenuation, later confirmed to be occlusion of the saphenous to diagonal branch conduit. We have noted this appearance as a “ghost” of the previous patent graft. C, Volume rendered image shows cutoff of the conduit (arrow).

FIGURE 37-18 A, Volume rendered image demonstrating three-vessel bypass: saphenous graft to first diagonal branch, saphenous graft to first obtuse marginal branch, and saphenous graft to distal RCA. Axial CT scan (B) focuses on the origin of the saphenous to diagonal branch. The volume rendered image from a follow-up study performed 12 months later (C) demonstrates slightly decreased caliber of OM and RCA grafts with occlusion of SVG to diagonal branch (arrow). Note the decreased attenuation (“ghosting”) observed on axial CT (D) in the region of the previous graft (arrow).

Summary

Increasingly, the diagnosis of bypass graft patency and occlusion is feasible using noninvasive techniques. Although it is difficult to visualize smaller grafts, continued improvements in image quality promise greater spatial resolution and suppression of cardiac motion. With such improvements and continued research investigations, these noninvasive techniques may replace conventional angiography as a primary tool to visualize these conduits.

KEY POINTS

Two major complications of coronary stent placements are in-stent thrombosis and restenosis.

There is decreased rate of restenosis but potential for increased in-stent thrombosis with DES.

The newer generations of CT scanners have demonstrated increased accuracy for detection of in-stent restenosis, however artifacts still prevent it from being routinely used except in select cases.

Long-term patency of saphenous vein grafts is considerably less than with arterial grafts. The most common cause of early saphenous graft failure is thrombosis. Beyond 1 year, atherosclerosis is the dominant process, resulting in graft stenosis and occlusion.

Arterial grafts such as IMA grafts, are resistant to atheroma development. Late IMA graft failure is more commonly due to progression of atherosclerotic disease in the native coronary artery distal to the graft anastomosis.

Occlusion of bypass grafts may be visualized as diffuse low attenuation (ghosting), abrupt truncation of contrast in the proximal portion of the occluded graft (nubbin sign), or complete nonvisualization of the conduit.