REVIEWS. Causes, assessment, and treatment of stent thrombosis intravascular imaging insights. Daniel S. Ong and Ik-Kyung Jang

Transcription

1 Causes, assessment, and treatment of stent thrombosis intravascular imaging insights Daniel S. Ong and Ik-Kyung Jang REVIEWS Abstract Stent thrombosis is a rare, but serious, complication of percutaneous coronary intervention and is associated with severe morbidity and mortality. In addition to clinical and pathological studies, intravascular imaging has advanced our understanding of the mechanisms underlying stent thrombosis. In particular, intravascular imaging has been used to study stent underexpansion, malapposition, uncovered struts, and neoatherosclerosis as risk factors for stent thrombosis. Intravascular ultrasonography and optical coherence tomography can be used to guide stent implantation and minimize the risk of stent thrombosis. Additionally, optical coherence tomography offers the unique potential to tailor treatment of stent thrombosis to address the specific mechanism underlying the thrombotic event. Bioresorbable stent technologies have been introduced with the goal of further reducing the incidence of stent thrombosis, and intravascular imaging has had an integral role in the development and assessment of these new devices. In this Review, we present insights gained through intravascular imaging into the causes of stent thrombosis, and the potential utility of intravascular imaging in the optimization of stent deployment and treatment of stent thrombosis events. Ong, D. S. & Jang I. K. Nat. Rev. Cardiol. 12, (2015); published online 17 March 2015; doi: /nrcardio Cardiology Division, Massachusetts General Hospital, Harvard Medical School GRB 800, 55 Fruit Street, Boston, MA 02114, USA (D.S.O., I. K.J.). Correspondence to: I. K.J. ijang@mgh.harvard.edu Introduction Over the past 3 decades, advances in percutaneous coronary intervention (PCI) have revolutionized the treatment of coronary artery disease. Coronary bare-metal stents (BMS) were first introduced in the 1980s with the goal of minimizing two major complications of balloon angioplasty abrupt vessel closure and restenosis. 1 Innovations in stent design, deployment techniques, and adjunctive therapies continue to focus on further reducing the risk of abrupt vessel closure mediated by stent thrombosis and in-stent restenosis. Stent thrombosis is a rare, but catastrophic, complication associated with severe morbidity and mortality and, therefore, has received substantial attention. In addition to clinical and pathological studies, intravascular imaging has advanced our understanding of the mechanisms underlying stent thrombosis. In this Review, we explore the insights gained through intravascular imaging into the causes of stent thrombosis, its usefulness in optimizing stent deployment to minimize risk, and its potential benefits in guiding the treatment of stent thrombosis events. Historical perspective The first human coronary artery stent was implanted in In an initial series of 105 patients undergoing placement of a self-expanding BMS, 24% suffered from Competing interests I. K.J. declares that he has received a research grant and honorarium from St. Jude Medical and research grants from Boston Scientific and Medtronic. D.S.O. declares no competing interests. complete stent occlusion after a mean follow-up of only 6 months. 3 Although the introduction of high-pressure balloon inflation and dual antiplatelet therapy helped to reduce the 6 month incidence of stent thrombosis to 1.6%, 4 the long-term benefit of BMS remained limited by a 20 30% prevalence of in-stent restenosis within 6 months of implantation. 5,6 Drug-eluting stents (DES) were developed with the aim of reducing the incidence of in-stent restenosis. 7,8 The first-generation of DES included a sirolimuseluting stent that was approved by the FDA in 2003, and a paclitaxel-eluting stent. Although initial studies did not detect an increased risk of stent thrombosis with these DES compared with BMS, 9,10 the report of hundreds of cases of subacute stent thrombosis and >60 associated deaths led the FDA to release an advisory warning about the potential risk of this complication. 11 Meta-analyses were performed with pooled trial data to clarify the risk of stent thrombosis and showed no significant difference between first-generation DES and BMS over 1 year of follow-up. 12,13 However, registry data reflecting real-world use of first-generation DES suggested a substantially higher rate of stent thrombosis compared with that observed in major clinical trials (1.3% versus % cumulative incidence at 9 months). 9,10,14 Furthermore, data from registries and trials suggested that, although the incidence of early stent thrombosis might be similar for BMS and firstgeneration DES, 9,10,15 17 the risk of late stent thrombosis seemed to be increased with the latter. 17,18 In one registry study with 4 year follow-up, stent thrombosis with first-generation DES occurred at a steady annual rate of NATURE REVIEWS CARDIOLOGY VOLUME 12 JUNE

2 Key points Stent thrombosis is a rare, but serious, complication of percutaneous coronary intervention and is associated with severe morbidity and mortality Intravascular imaging has furthered our understanding of the mechanisms underlying stent thrombosis Stent underexpansion and malapposition can be assessed and optimized at the time of stent implantation with the use of intravascular imaging During follow-up imaging, strut coverage and neoatherosclerosis are best evaluated in vivo using optical coherence tomography Optical coherence tomography might be particularly useful during evaluation of stent thrombosis events to determine specific underlying mechanisms and to guide therapy Intravascular imaging has been crucial to the development and assessment of new bioresorbable stent technologies, although the risk of stent thrombosis associated with these devices requires further investigation Box 1 Definite,* probable, and possible stent thrombosis 23 Definite stent thrombosis Angiographic confirmation of stent thrombosis. The presence of a thrombus that originates in the stent or in the segment 5 mm proximal or distal to the stent and presence of at least one of the following criteria within a 48 h time window: Acute onset of ischaemic symptoms at rest. New ischaemic electrocardiographic changes suggestive of acute ischaemia. Typical rise and fall in cardiac biomarkers. Nonocclusive thrombus: intracoronary thrombus is defined as a (spheric, ovoid, or irregular) noncalcified filling defect or lucency surrounded by contrast material (on three sides or within a coronary stenosis) seen in multiple projections, or persistence of contrast material within the lumen, or a visible embolization of intraluminal material downstream. Occlusive thrombus: TIMI 0 or TIMI 1 intrastent or proximal to a stent up to the most adjacent proximal side branch or main branch (if originates from the side branch). Pathological confirmation of stent thrombosis. Evidence of recent thrombus within the stent determined at autopsy or via examination of tissue retrieved following thrombectomy. Probable stent thrombosis Clinical definition of probable stent thrombosis is considered to have occurred after intracoronary stenting in the following cases: Any unexplained death within the first 30 days. Irrespective of the time after the index procedure, any myocardial infarction that is related to documented acute ischaemia in the territory of the implanted stent without angiographic confirmation of stent thrombosis and in the absence of any other obvious cause. Possible stent thrombosis Clinical definition of possible stent thrombosis is considered to have occurred with any unexplained death from 30 days after intracoronary stenting until end of trial follow-up. *Definite stent thrombosis is considered to have occurred by either angiographic or pathological confirmation. The incidental angiographic documentation of stent occlusion in the absence of clinical signs or symptoms is not considered a confirmed stent thrombosis (silent occlusion). Intracoronary thrombus. For studies with ST-segment elevation myocardial infarction population, one may consider the exclusion of unexplained death within 30 days as evidence of probable stent thrombosis. Reprinted from Cutlip, D. E. et al. Clinical end points in coronary stent trials: a case for standardized definitions. Circulation 115 (17), (2007) %. 16 Moreover, extended use of clopidogrel was associated with a reduced risk of death or myocardial infarction (MI) in patients with DES, but not in those who received BMS, which supports the concern about late thrombogenic risk with DES. 19 Thereafter, the ACC/ AHA guideline-recommended duration of dual antiplatelet therapy after DES implantation increased from 3 6 months (as used in initial trials) to 1 year. 20 European guidelines similarly advised 6 12 months of dual antiplatelet therapy after DES implantation, 21 although this recommendation was shortened to 6 months in the 2014 ESC guidelines. 22 Importantly, the definition of stent thrombosis has varied considerably between studies, with some requiring angiographic proof of in-stent thrombus and others broadly including any unexplained cardiac death. 23 To facilitate uniformity in clinical end points for future trials, the Academic Research Consortium (ARC) convened in 2006 to develop standardized definitions for definite, probable, and possible stent thrombosis (Box 1), and to specify the timing of events as acute (0 24 h after stent implantation), subacute (>24 h to 30 days), late (>30 days to 1 year), and very late (>1 year). 23 Of note, stented segments that require t arget-lesion revascularization and subsequently develop secondary stent t hrombosis are included as late and very late events. Using the ARC definitions, data from eight randomized trials to compare first-generation DES with BMS were reanalysed. 24 Target-lesion revascularization was significantly more frequent in patients who received BMS than in those who received either sirolimus-eluting or paclitaxel-eluting stents (15 30% versus 8%), and the prevalence of late and very late stent thrombosis events observed in the BMS groups was greater when defined using the ARC criteria than when using the original study protocol definitions. Mauri and colleagues proposed that the combined category of definite or probable stent thrombosis provided the best approximation of the true incidence of this complication. 24 Using this definition, no significant difference existed in the rate of stent thrombosis between BMS and either s irolimuseluting or paclitaxel-eluting stents over 4 years of f ollow-up ( % for all stent types). 24 A subsequent network meta-analysis of 38 trials with up to 4 years of follow-up similarly found no significant difference in the risk of ARC definite stent thrombosis between BMS, s irolimus eluting stents, and paclitaxel-eluting stents. 25 The second generation of DES includes everolimuseluting and zotarolimus-eluting stents, which were introduced in Europe in 2005 and approved by the FDA for use in the USA from In a large registry study comparing the risk of ARC definite stent thrombosis between BMS, first-generation DES, and second-generatio n DES, no significant difference was found in the incidence of early stent thrombosis. 26 However, the risk of very late stent thrombosis was higher with first-generation DES than with BMS, but was similar with secondgeneration DES and BMS. 26 Over 3 years of follow-up, the cumulative incidence of definite stent thrombosis was 1.5% with BMS, 2.2% with first-generation DES, and 1.0% with second-generation DES. 26 Despite becoming increasingly rare, stent thrombosis remains a serious complication of PCI. An analysis of BMS trials and registries conducted in the 1990s identified a 60% rate of associated major MI and 20% mortality at 6 months. 27 In a subsequent large registry study, the 2 year mortality in patients with definite stent thrombosis 326 JUNE 2015 VOLUME 12

3 was 15.6%. 16 Data from multiple trials and registries have been used to explore the risk factors for this potentially catastrophic complication. Risk factors can be broadly grouped into four major categories: patient-related factors, lesion-related factors, procedure-related and stent-related factors, and p harmacotherapy related factors (Box 2). Causes of stent thrombosis Intravascular imaging has improved understanding of several mechanisms underlying stent thrombosis. Studies suggest that early (including acute and subacute 23 ), late, and very late stent thrombosis have different underlying mechanisms, risk factors, and outcomes. 28,29 For example, in one study, early stent thrombosis was associated with a 2 year mortality of 20%, compared with 10% after late or very late stent thrombosis. 16 Stent underexpansion and malapposition have been identified as the most prevalent intravascular ultrasonography (IVUS) abnormalities in patients with early stent thrombosis. 30 These procedure-related mechanisms might explain the similar incidence of early stent thrombosis with different stent types. 24,26 Late malapposition and delayed endothelialization (manifesting as uncovered struts) have been studied with intravascular imaging in relation to late and very late stent thrombosis. Lastly, neoatherosclerosis has been identified using intravascular imaging as an important and novel mechanism for very late stent thrombosis. Stent underexpansion Stent underexpansion is identified by comparing the minimum luminal area achieved within an implanted stent with that of neighbouring reference segments. The most widely used IVUS definition of stent underexpansion is in-stent minimum luminal area <90% of the distal reference luminal area, or <80% of the average of proximal and distal reference segment luminal areas In a retrospective registry of 53 patients with early stent thrombosis after BMS placement, IVUS revealed stent underexpansion in 49% of individuals. 30 In a study of 7,484 patients who underwent BMS placement guided by IVUS, 0.4% of patients developed angiographicallyproven stent thrombosis within 7 days of the procedure. Among the 23 patients with early stent thrombosis and adequate IVUS image quality for complete analysis, stent underexpansion was identified in 78%, compared with in only 33% of 69 patients in the matched control group. 35 Stent underexpansion is also an independent predictor of early stent thrombosis in patients receiving first generation DES. 36 Malapposition Malapposition, also called incomplete stent apposition, is the separation of at least one stent strut from the arterial wall that does not occur at a side branch. Autopsy studies have shown an association between malapposition and stent thrombosis. In one study of patients with acute coronary syndrome who died within 30 days of coronary intervention, malapposed struts were more frequently Box 2 Risk factors for stent thrombosis Patient-related factors Malignancy 122 Current smoking 123 Diabetes mellitus 14,16 Low ejection fraction 14,122,124 Renal failure 14,125 Presentation with acute coronary syndrome 16,54,125 Lesion-related factors Long lesions 33 Bifurcation lesions 14,15,122 Inflow or outflow disease 122 Vein-graft lesions 123 Procedure-related and stent-related factors Off-label use 126 Small stent size 29 Long stent length 12,123,125 Multiple stents 33 Overlapping stents 33,123 Substantial residual reference segment stenosis 36 Stent underexpansion 35,36 Malapposition 30 Edge tears or dissection 30,124 In-stent thrombus 30 Pharmacotherapy-related factors Choice of antiplatelet therapy 127,128 Hyporesponsiveness to antiplatelet therapy 93,129,130 Nonadherence or discontinuation of antiplatelet therapy 14,125,131 identified in thrombosed stent segments than in patent stent segments (34% versus 18%; P = 0.008). 37 Intravascular imaging has enabled the in vivo identification of malapposition at various time points after stent implantation. Malapposition is defined as acute when present on final imaging performed during the index procedure and late when identified on follow-up imaging. In studies with serial intravascular imaging, late malapposition can be specified as persistent or late-acquired on the basis of its presence or absence on final imaging during the index procedure. IVUS and optical coherence tomography (OCT) have been used to investigate stent malapposition. With IVUS, the definition of malapposition requires visualization of blood flow speckling behind the malapposed stent strut to confirm separation of the strut from the arterial wall. 38 Although cross-sectional analysis is possible, the resolution of IVUS is inadequate for strut-level quantification. By contrast, malapposition of individual struts can be detected with OCT imaging and is defined as a separation distance between the endoluminal surface of the strut and the vessel wall greater than the sum of the metal and polymer thickness. 39,40 The high resolution of OCT imaging facilitates moresensitive detection of malapposition than is possible using IVUS (Figure 1). 41,42 Although studies have suggested an association between malapposition and increased risk of stent thrombosis, a cause-and-effect relationship remains unproven. Mechanistically, greater malapposition distance has been shown to create greater flow disturbance and increased shear rates, both of which can predispose to thrombus NATURE REVIEWS CARDIOLOGY VOLUME 12 JUNE

4 a b c d * * ** * * * * * * * Figure 1 Malapposition evaluated with intravascular ultrasonography Nature Reviews and Cardiology optical coherence tomography. 132 a Intravascular ultrasonography and b optical coherence tomography cross-sectional images show stent struts (arrows) with malapposition (asterisks) in the same co-registered region of interest. c Intravascular ultrasonography does not clearly show malapposition, whereas d optical coherence tomography of the same co-registered region clearly depicts stent struts (arrows) with malapposition (asterisks). The blue lines in the longitudinal views indicate the region of the pullback from which each cross-section was obtained. Reprinted from Attizzani, G. F. et al. Mechanisms, pathophysiology, and clinical aspects of incomplete stent apposition. J. Am. Coll. Cardiol. 63 (14), (2014), with permission from Elsevier. formation. 43 In one study, thrombus was visualized using OCT in association with 20.6% of mal apposed struts, compared with only 2.0% of well-apposed struts. Importantly, however, no cases of malapposition were associated with clinically apparent stent thrombosis. 42 The clinical importance of acute m alapposition, th erefore, remains unclear. 44 Late-acquired malapposition develops when struts separate from the vessel wall after initial well-apposed implantation, and might occur through a variety of potential mechanisms, including chronic stent recoil, reduction in plaque size via plaque regression, apoptosis, or clot lysis, and increased external elastic membrane area (positive remodelling). 45 Chronic stent recoil is an uncommon mechanism for late-acquired malapposition, 45,46 although it has been observed after stent implantation for highly calcified lesions. 47 Similarly, plaque regression and apoptosis are not common mechanisms for late-acquired malapposition. 48,49 Clot lysis might be an important mechanism, particularly when stents are implanted in patients with acute coronary syndrome. 47,50 Overall, however, most studies suggest that positive remodelling is the prevailing mechanism for late-acquired malapposition. 44,45,48,50,51 Late-acquired malapposition is more common with first-generation DES than with BMS. 52,53 At 6 months of follow-up, the incidence of late-acquired stent malapposition is 5% with BMS 45,54 and 12% with firstgeneration DES. 50 Pathological examination of autopsy 55 and aspiration thrombectomy 56 samples from patients with very late stent thrombosis of a first-generation DES has shown a neutrophil-predominant inflammatory cell infiltrate and eosinophils. This finding suggests that lateacquired stent malapposition and positive remodelling after implantation of first-generation DES is a marker for underlying vascular inflammation or toxic effects related to hypersensitivity reaction. 57 Rates of late malapposition are lower with secondgeneratio n DES than with first-generation DES. In an autopsy study of stents implanted >30 days but 3 years previously, the frequency of late malapposition was lower with cobalt chromium everolimus-eluting stents than with sirolimus-eluting stents and paclitaxel-eluting stents (4% versus 16% and 18%). 58 No hypersensitivity reactions were observed with cobalt chromium everolimus-eluting stents. 58 Multiple OCT studies have similarly demonstrated smaller percentages of malapposed struts with second-generation DES than with first-generation DES at around 10 months of follow-up ( % versus %) As with acute malapposition, the clinical importance of late malapposition remains unclear. Although pathological and imaging studies have identified malapposition at the time of stent thrombosis events, 33,57 no clear association between late malapposition burden and subsequent adverse clinical events has been proven. 44,50,51,54,62,63 Uncovered struts Vascular healing after stent implantation is an important determinant of stent thrombosis risk, because the endothelium has a critical role in preventing thrombus formation. An early autopsy series of 13 patients with BMS-related late stent thrombosis identified incomplete neointimal coverage of stent struts as the underlying mechanism in all but one case. 64 In pathological studies, endothelial coverage was the strongest histological predictor of late or very late stent thrombosis. 65 Among morphometric parameters, the ratio of uncovered to total struts per section correlated most strongly with histological endothelialization, which suggests that it would be a useful surrogate end point for delayed healing and risk of stent thrombosis. A ratio of >30% was associated with an odds ratio of 9.0 for late or very late stent thrombosis. 65 In vivo assessment of endothelialization requires intravascular imaging. In a study of patients who underwent follow-up angioscopy 3 6 months after stent implantation, all BMS showed complete neointimal coverage, compared with only 13.3% of sirolimus-eluting stents. Additionally, incomplete neointimal coverage was associated with formation of subclinical thrombus. 66 OCT has been used to study vascular healing at an individual strut level and in the context of strut apposition (Figure 2). In one study, OCT performed 3 months after sirolimuseluting stent implantation showed that 15% of struts remained uncovered. 67 A serial OCT study with longterm follow-up showed that, after 2 years, only 5% of sirolimus-eluting stent struts were uncovered, 68 although notably, a small percentage (0.9%) of struts remained uncovered even 4 years after stent implantation. 69 Second-generation DES show favourable strut coverage compared with first-generation DES. In an autopsy comparison of first-generation and second-generation DES implanted >30 days but 3 years previously, the frequency of late and very late stent thrombosis was 21% with sirolimus-eluting stents, 26% with paclitaxel-eluting stents, and 4% with cobalt chromium e verolimuseluting stents. 58 The mean durations of implantation were similar for all three stent types. In patients with 328 JUNE 2015 VOLUME 12

5 cobalt chromium everolimus-eluting stents, the percentage of uncovered struts was significantly lower than in patients with sirolimus-eluting or paclitaxel-eluting stents (2.6% versus 18.0% and 18.7%), and ratios of uncovered to total struts per section >30% were seen in 20% versus 60% and 67% of lesions. 58 Additionally, cobalt chromium everolimus-eluting stents were associate d with less inflammation and fibrin deposition. 58 OCT studies support the difference in stent coverage between first-generation and second-generation DES (Table 1). 59,61 In a comparison of sirolimus-elutin g and paclitaxel-eluting first-generation DES with everolimus-eluting and zotarolimus-eluting secondgeneration DES, the prevalence of uncovered struts at around 12 months after implantation was significantly lower with the second-generation DES (3.8% versus 7.5%; P <0.001). 60 Moreover, within-stent thrombi were seen less frequently in the second-generation DES (8% versus 20%; P = 0.04). 60 Although all patients were free from major adverse events at the time of follow-up OCT, the improved stent coverage and reduced withinstent thrombi observed with the second-generation DES suggest a benefit in terms of stent thrombosis risk. Importantly, neointimal hyperplasia was suppressed similarly in both groups, which indicates retention of antiproliferative benefit in addition to improved vascular healing with second-generation DES. 60 An important consideration in the use of OCT for evaluating vascular healing is that, despite its high resolution, this imaging technique cannot be used to identify the specific type of tissue covering a strut. Therefore, a strut covered with functional neointimal tissue might appear similarly to one covered with fibrin or inflammatory cells. A correlation study of OCT images with histology samples, however, introduced the possibility of using OCT greyscale signal intensity to differentiate between mature and immature neointimal tissue. 70 Neoatherosclerosis In 2002, we reported the potential for patients with in-stent restenosis to present with acute coronary syndrome. 71 Pathological studies provided further insight a c Figure 2 Strut apposition and coverage. Optical coherence tomography images of stents with a well-apposed and covered struts (arrows), b well-apposed and uncovered struts (arrows), c malapposed and covered struts (arrows), and d malapposed and uncovered struts (arrows). into the potential for such lesions to contribute to un stable presentations and identified neoathero sclerosis as a novel mechanism for late and very late stent thrombosis. Neoatherosclerosis is defined as the presence of lipid-laden foamy macrophage infiltrates within the neointima above a stent, with or without formation of a necrotic core. 72 In an autopsy study of 143 stented segments, neoatherosclerosis was significantly more prevalent within sirolimus-eluting stents than in BMS (35% versus 10%; P <0.001). Additionally, these atherosclerotic changes were seen earlier with DES, with foamy macrophage infiltration observed as early as 4 months after implantation compared with >2 years after implantation of BMS. 72 Features associated with plaque vulnerability in de novo coronary lesions have been identified in neoatherosclerotic lesions. In autopsy cases with neoathero sclerosis, unstable features such as thin-cap fibroatheroma and ruptured plaque with thrombosis b d Table 1 Prevalence of uncovered struts on OCT Type of stent Prevalence at mean follow-up interval to OCT imaging 3 months 67,133 9 months 59,61, months 60,138 >12 months years years 69 5 years 140 BMS 0.1 NA NA NA NA NA NA Sirolimuseluting DES Paclitaxeleluting DES Zotarolimuseluting DES* NA NA NA NA NA NA Zotarolimuseluting NA NA NA NA DES Everolimuseluting DES NA NA NA NA *Resolute (Medtronic, USA). Endeavor (Medtronic, USA). Abbreviations: BMS, bare-metal stent; DES, drug-eluting stent; NA, not available; OCT, optical coherence tomography. NATURE REVIEWS CARDIOLOGY VOLUME 12 JUNE

6 a b c d e Fibrotic tissue Fibrofatty tissue Dense calcium Necrotic core Figure 3 Neoatherosclerosis, assessed with VH IVUS. 74 Top panels show Cardiology VH IVUS images of neointimal composition at sites of maximum percent intimal hyperplasia, and bottom panels show intravascular ultrasound images for comparison. Follow-up after implantation of a paclitaxel-eluting stent at a 6 months shows necrotic core 10% and dense calcium 2%, b 9 months shows necrotic core 28% and dense calcium 8%, and c 22 months shows necrotic core 39% and dense calcium 20%. Follow-up after implantation of a bare-metal stent at d 48 months shows necrotic core 40% and dense calcium 25%, and e 57 months shows necrotic core 57% and dense calcium 15%. Abbreviation: VH IVUS, virtual histology intravascular ultrasonography. Reprinted from Kang, S. J. et al. Tissue characterization of in-stent neointima using intravascular ultrasound radiofrequency data analysis. Am. J. Cardiol. 106 (11), (2010), with permission from Elsevier. were identified within 2 years of implantation of firstgeneration DES, compared with an average 6 years after implantation of BMS. 73 Identification of neoatherosclerosis in vivo requires intravascular imaging. Greyscale IVUS is of limited usefulness, but a virtual histology IVUS study of patients with in-stent restenosis revealed that the risk of neoatherosclerotic tissue components, such as necrotic core and dense calcium, increases with time after implantation (Figure 3). 74 The best intravascular imaging tool for the assessment of neoatherosclerosis is OCT. Rupture of neoatherosclerotic plaques (Figure 4) has been identified using OCT in patients presenting with stable angina as well as those with ST segment elevation MI related to very late stent thrombosis. 75,76 The high resolution of OCT imaging also enables in vivo identification of morphological features associated with plaque vulnerability (Figure 4). 77 In an OCT study of 60 BMS in 39 patients who presented with recurrent ischaemia on average 6.5 years after stent placement, 33% of stents showed instent neoatherosclerosis with lipid-rich plaque and thin fibrous caps (mean 56.7 μm). 78 Of these lesions, 30% showed evidence of recent in-stent fibrous-cap rupture and were found in patients who presented with unstable angina. Taken together, these findings suggest that thin-cap neoatheroma plaque rupture is an important mechanism for very late stent thrombosis. 78 Further insights into the development of neoatherosclerosis were gained from an OCT study in which BMS implanted 5 years previously (late group, n = 21) were compared with those implanted <6 months previously (early group, n = 20). 79 Lipid-laden neointima was identified in 67% of cases in the late group, compared with none in the early group, where only normal homogenous neointimal proliferation was seen. 79 Further confirmation of progressive neoatherosclerosis within neo intimal tissue is provided by a serial OCT study including both first-generation and second-generation DES. 80 On 9 month follow-up imaging, lipid-laden neointima was identified in 14.5% and thin-cap neoatheroma in 3.9% of stents. 80 At 2 years, the prevalence of both had increased, with lipid-laden neointima identified in 27.6% and t hin cap neoatheroma in 13.2% of stents. 80 In a study of BMS and first-generation and secondgeneration DES, the prevalence of neoatherosclerosis increased with stent age. Neoatherosclerosis was identified in 8% of BMS and 37% of DES placed within the preceding 9 months, 28% and 63% placed 9 48 months previously, and 77% and 75% placed 48 months previously. Of note, although neoatherosclerosis developed later in BMS than in DES, most restenotic lesions in the oldest stents of both types contained neoatherosclerosis. 81 OCT has also been performed in patients with BMS and DES at the time of presentation with stent thrombosis and has confirmed that most cases of very late stent thrombosis are attributable to neoatherosclerosis. 82,83 These findings highlight the importance of long-term follow-up in studies to explore stent thrombosis risk across different stent platforms. Stent assessment Several studies have explored the usefulness of intravascular imaging to guide stent implantation. In a metaanalysis of three randomized, controlled trials and nine high-quality observational cohort studies to compare DES implantation performed with or without IVUS guidance, the use of IVUS was associated with significantly larger minimum luminal diameters and lower rates of major adverse cardiac events, MI, and death. Five of these 12 studies included data on rates of early and late stent thrombosis and two included data on rates of very late stent thrombosis. Overall, IVUS guidance was associated with a 50% risk reduction in the rate of stent thrombosis (P = 0.07). 84 A subsequent nonrandomized, prospective study included 3,349 patients in whom IVUS was used to guide DES implantation. In 74% of cases, the operator changed the PCI strategy on the basis of the intravascular imaging result, including use of a larger stent or balloon (38%), higher inflation pressure (23%), longer stent (22%), additional dilatation to correct stent underexpansion (13%) or malapposition (7%), and placement of additional stents (8%). 85 Compared with patients undergoing stent implantation using angiography alone, IVUS guidance resulted in significantly reduced rates of MI (2.5% versus 3.7%; P = 0.004), major adverse cardiac events (3.1% versus 4.7%; P = 0.002), and ARC definite or probable stent thrombosis (0.6% versus 1.0%; P = 0.003) over 1 year of follow-up JUNE 2015 VOLUME 12

7 a b c d * * * * e f g Figure 4 Neointimal features identified with optical coherence tomography. a Lipid-laden (neoatherosclerotic) neointima within a stent (solid arrow), appearing as a signal-poor region with diffuse borders (asterisks), with fibrous-cap disruption (dashed arrow). b Homogenous neointimal proliferation without evidence of neoatherosclerosis. c Neointima with a diffusely-bordered signal-poor region and an overlying signal-rich homogenous band (arrows), indicating lipid-rich neoatherosclerosis with an overlying fibrous cap. d Neointima with small vesicular or tubular structures, indicating in-stent neovascularization (arrow). e Neointima with a linear series of signal-rich spots with high signal attenuation, indicating in stent macrophage infiltration (arrow). f Neointima with a sharply demarcated area of heterogeneous reflectivity with low signal attenuation, indicating in-stent calcification (arrows). g Neointima with thin, linear, signal-rich structures with low signal attenuation, indicating in-stent cholesterol crystals (arrows). Panel a reprinted from Vergallo, R. et al. Correlation between degree of neointimal hyperplasia and incidence and characteristics of neoatherosclerosis as assessed by optical coherence tomography. Am. J. Cardiol. 112 (9), (2013), with permission from Elsevier. Panels c e reprinted from Yonetsu, T. et al. Comparison of incidence and time course of neoatherosclerosis between bare metal stents and drugeluting stents using optical coherence tomography. Am. J. Cardiol. 110 (7), (2012), with permission from Elsevier. Few studies have been designed to explore the benefit of OCT-guided stent implantation compared with angiography alone. In the CLI OPCI study, patients undergoing PCI guided by angiography alone were compared with 335 matched patients undergoing PCI with the addition of OCT guidance. In the OCT group, intravascular imaging revealed abnormalities that required further intervention in 34.7% of patients. In a multivariate analysis, OCT guidance was associated with a significantly lower risk of cardiac death or MI than angiography alone (P = 0.037). The rate of ARC definite stent thrombosis did not differ between groups, although importantly, only three stent thrombosis events occurred in the entire study population over 1 year of follow-up. 86 Although a paucity of data for the use of OCT to guide stent implantation remains, the high resolution of this imaging technique and the capacity to perform strut-level evaluations suggest that it is a promising method that warrants further study. Additionally, proposed criteria for IVUS guidance to achieve optimum stent implantation, such as the AVIO algorithm, 87,88 might be extrapolated for use with OCT, although it is important to note that lumen dimensions measured using OCT might be smaller than those measure d with IVUS. 89,90 OCT is currently the best intravascular imaging tool for the study of stent coverage and neoatherosclerosis and, therefore, provides unique potential for use during follow-up imaging in risk assessment for late and very late stent thrombosis. The prevalence of uncovered struts might be helpful when weighing the risks and benefits of discontinuing or extending dual antiplatelet therapy in individual patients. Additionally, vascular healing and endothelialization with new stent platforms might be more easily studied with OCT than with pathological techniques. Neoatherosclerotic lesions can be assessed on follow-up imaging for features such as lipid-rich plaque, thin fibrous cap, neovascularization, macrophage infiltration, calcification, and cholesterol crystals, which have been associated with plaque vulnerability in de novo coronary lesions. 91,92 Guidance of treatment Few data are available regarding the optimal treatment strategy for stent thrombosis events. Certainly, confirmation of adherence to prescribed antiplatelet therapy is of paramount importance. Although the usefulness of platelet-function testing is beyond the scope of this Review, modification of antiplatelet therapy (for example, the replacement of clopidogrel with prasugrel or ticagrelor) might be advisable, either empirically or in response to documented high residual platelet activity. 93 PCI strategies vary widely. In a large study to examine the management of 7,315 stent thrombosis events NATURE REVIEWS CARDIOLOGY VOLUME 12 JUNE

8 identified in the CathPCI registry, aspiration thrombectomy was performed in around 30% of cases, and a new stent was implanted in over half of patients, and more frequently in cases of very late than of early stent thrombosis (69.9% versus 51.2%). 28 In a Japanese registry study of 611 patients with sirolimus-eluting stent thrombosis, aspiration thrombectomy was performed in 77% of cases, and a new stent was implanted in 36% of cases. 94 Intravascular imaging might be particularly useful in guiding the management of stent thrombosis by enabling treatment to be tailored to the underlying mechanism of the event. OCT is superior to IVUS for determining the underlying mechanism of acute coronary syndromes in native coronary arteries, 95 which translates to superiority in the evaluation of stent thrombosis events. OCT has been used successfully to guide treatment of stent thrombosis. 96 In particular, OCT can be used to identify stent underexpansion and malapposition, which can be treated with additional dilatation by balloon angioplasty; uncovered struts, which can be treated with aspiration thrombectomy and more-aggressive antiplatelet therapy; and in-stent neoatherosclerosis, which might require additional stent placement. Further insights into the causes, assessment, and treatment of stent thrombosis are anticipated from the PRESTIGE study, 97 which is a multicentre collaboration that began enrolment in December The investigators aim to enrol at least 500 patients with stent thrombosis to explore molecular and cellular events triggering late stent thrombosis, devise and validate novel stent thrombosis prevention strategies, develop imaging technologies to study vascular healing and dysfunction after stent placement, and create a registry to enable multimodal characterization of patients with late stent thrombosis. Bioresorbable polymers and scaffolds Bioresorbable technologies might be the next revolutionary advance in PCI. DES that use bioresorbable polymers on metal scaffolds have been developed with the aim of retaining antiproliferative benefits while minimizing or eliminating the polymer-stimulated chronic inflammation that might contribute to late and very late stent thrombosis. 55,98 A meta-analysis of data from three randomized trials found significantly lower rates of target-lesion revascularization and ARC definite stent thrombosis (driven by a reduced risk of very late stent thrombosis) with bioresorbable-polymer DES than with first-generation DES over 4 years of followup. 99 An OCT substudy of one trial found significantly fewer uncovered struts at 9 months of follow-up with a biolimus-eluting bioresorbable-polymer stent than with a first-generation sirolimus-eluting stent (1.8% versus 6.3%). 100 Of note, however, the prevalence of mal apposed struts was similar in the two groups. 100 Network metaanalyses have suggested similar or worse outcomes with bioresorbable-polymer DES than with second-generation DES, including a higher rate of MI and ARC definite stent thrombosis, particularly when compared with cobalt chromium everolimus-eluting stents Fully bioresorbable coronary scaffolds take this technology one step further and capitalize on the idea that resorption of the entire stent structure might reduce the risks of late and very late stent thrombosis related to intracoronary foreign material, and avoid the long-term restrictions on vasomotion, late luminal enlargement, and expansive remodelling associated with the persistently rigid structure of metallic stents. 104 Although not currently available in the USA, two bioresorbable scaffolds have been approved for use in Europe: the Absorb e verolimus-eluting Bioresorbable Vascular Scaffold System (Abbott Vascular, USA) and the DESolve Novolimus -eluting Bioresorbable Coronary Scaffold System (Elixir Medical Corporation, USA). 105 Intravascular imaging is crucial for the assessment of bioresorbable coronary scaffolds. New definitions and descriptors have had be to developed to account for their unique appearance in comparison with metallic stents. 104 Bioresorbable coronary scaffold struts appear as hyperechogenic (white) structures on greyscale IVUS and as dense calcium and necrotic core on colour-coded virtual histology IVUS. 106 The quantitative change in dense calcium and necrotic core content over time has been used as a surrogate end point for assessment of scaffold resorption. 107 On OCT imaging, the struts of bioresorbable coronary scaffolds initially appear as black squares bordered by a signal-rich frame and, in contrast to metallic struts, cast no optical shadow. 108 As the scaffold resorbs and the strut cores become replaced with connective tissue, their appearance becomes white or signal-rich on OCT. 109 The first version of the Absorb Bioresorbable Vascular Scaffold System was investigated in 30 patients in the ABSORB cohort A trial. 110 Angiography and intravascular imaging at 6 months of follow-up was notable for late scaffold recoil causing an 11.8% reduction in scaffold area on IVUS. 110 Although 5 year follow-up confirmed complete scaffold resorption, restoration of vasomotion, and late luminal enlargement beyond 6 months, 106,111,112 the second version of the scaffold was modified to increase its radial strength and mechanical durability. A total of 101 patients received this version in the ABSORB cohort B trial. 113 Follow-up imaging was performed with angiography and multimodality intravascular imaging at 6 and 24 months in 45 patients and at 12 and 36 months in 56 patients. OCT imaging confirmed no substantial change in the scaffold area from baseline to 6 months of follow-up. 108 Intravascular imaging, in particular OCT, has been used to explore bioresorbable scaffold expansion, malapposition and strut coverage (Figure 5), and neoplaque characteristics. In a series of 19 patients in whom 29 Absorb Bioresorbable Vascular Scaffolds were implanted, OCT imaging was performed after optimum angiographic results were achieved. The OCT findings prompted further intervention to 28% of scaffolds owing to underexpansion or malapposition. 114 In ABSORB cohort B patients who had serial OCT imaging, the prevalence of malapposed struts decreased from 3.5% at baseline to 0.8% at 6 months of follow-up. 115 The prevalence of uncovered 332 JUNE 2015 VOLUME 12

9 a b c d e f g h Figure 5 Bioresorbable vascular scaffold apposition and strut coverage. 115 Optical coherence tomography images of the second version of the Absorb Bioresorable Vascular Scaffold System (Abbott Vascular, USA) at baseline and 6 month follow-up. At baseline, struts can appear a protruding, b embedded, c malapposed, or d located at a side branch. On follow-up imaging, struts can appear e apposed and covered, f apposed and uncovered, g malapposed and covered, or h malapposed and uncovered. Reprinted from Gomez-Lara, J. et al. Serial analysis of the malapposed and uncovered struts of the new generation of everolimus-eluting bioresorbable scaffold with optical coherence tomography. JACC Cardiovasc. Interv. 4 (9), (2011), with permission from Elsevier. struts was 3.2% at 6 months, % at 12 months, % at 24 months, 117 and 1.7% at 36 months. 109 In addition to favourable rates of malapposition and strut coverage, OCT imaging has revealed favourable changes in plaque morphology after implantation of the second version of the Absorb Bioresorbable Vascular Scaffold. In one study, 92% of thin-cap fibroatheroma detected at baseline were sealed by neointimal tissue and transformed to thick-cap fibroatheroma by 6 12 months of follow-up. 118 Over 2 3 years of follow-up, this growth was accommodated by scaffold expansion, helping to preserve luminal dimensions. 118 No scaffold thrombosis was seen in the entire ABSORB cohort B over 3 years of clinical follow-up, although three non Q-wave MIs and seven cases of ischaemia-driven target-lesion revascularizations resulted in a 3 year major adverse cardiac event rate of 10%. 109 Over long-term intravascular imaging follow-up, scaffold resorption makes it impossible to distinguish between underlying plaque and bioresorbed struts and neointima. However, in a 5 year follow-up OCT study of eight patients implanted with the first version of the bioresorbable vascular scaffold, high-risk features such as new necrotic core formation were absent from the most superficial signal-rich neoplaque layer, which comprised pre-existing fibrous tissue, resorbed struts, and neointima. 112 Although late strut discontinuity is an expected finding related to scaffold resorption, acute scaffold disruption can be caused by overexpansion of the Absorb Bioresorbable Vascular Scaffold and has been associated with recurrent angina and target-lesion revascularization. 119 The DESolve Bioresorbable Coronary Scaffold was designed to accommodate a wide safety margin for expansion without strut fracture, to self-correct minor acute malapposition, and to resorb in about 1 year. In a first-in-man study of 16 patients, no scaffold thrombosis was seen over 12 months of follow-up. 120 OCT imaging performed at 6 months revealed a low prevalence of malapposed struts (0.04%) and uncovered struts (1.32%). 120 In the large, multicentre GHOST EU registry of realworld outcomes after Absorb Bioresorbable Vascular Scaffold implantation, the cumulative incidence of ARC definite or probable scaffold thrombosis was 1.5% at 30 days and 2.1% at 6 months. 121 These rates were noted to exceed the incidence typically observed in registries and trials of patients implanted with second-generation DES. 121 Further studies in both trial and real-world populations with long-term follow-up are necessary to clarify the risk of stent thrombosis with bioresorbable coronary scaffolds. Conclusions Stent thrombosis is a rare, but serious, complication with multifactorial causes. Intravascular imaging has improved our understanding of several mechanisms underlying stent thrombosis and can be used to optimize stent placement. Although PCI guided by IVUS has been shown to improve outcomes compared with PCI guided by angiography alone in de novo coronary lesions, OCT might be uniquely useful for guiding the treatment of stent thrombosis events. Further studies are needed to develop intravascular-imaging-based guidelines for optimal stent implantation and stent thrombosis treatment. Additionally, although intravascular imaging has been integral to the development and assessment of bioresorbable stent technologies, the benefit of these novel devices in terms of reducing the risk of stent thrombosis remains to be proven through large studies with long-term follow-up. NATURE REVIEWS CARDIOLOGY VOLUME 12 JUNE

10 1. Sigwart, U., Puel, J., Mirkovitch, V., Joffre, F. & Kappenberger, L. Intravascular stents to prevent occlusion and restenosis after transluminal angioplasty. N. Engl. J. Med. 316, (1987). 2. Roguin, A. Stent: the man and word behind the coronary metal prosthesis. Circ. Cardiovasc. Interv. 4, (2011). 3. Serruys, P. W. et al. Angiographic follow-up after placement of a self-expanding coronary-artery stent. N. Engl. J. Med. 324, (1991). 4. Colombo, A. et al. Intracoronary stenting without anticoagulation accomplished with intravascular ultrasound guidance. Circulation 91, (1995). 5. Serruys, P. W. et al. A comparison of balloon expandable stent implantation with balloon angioplasty in patients with coronary artery disease. Benestent Study Group. N. Engl. J. Med. 331, (1994). 6. Fischman, D. L. et al. A randomized comparison of coronary-stent placement and balloon angioplasty in the treatment of coronary artery disease. Stent Restenosis Study Investigators. N. Engl. J. Med. 331, (1994). 7. Morice, M. C. et al. A randomized comparison of a sirolimus-eluting stent with a standard stent for coronary revascularization. N. Engl. J. Med. 346, (2002). 8. Grube, E. et al. TAXUS I: six- and twelve-month results from a randomized, double-blind trial on a slow-release paclitaxel-eluting stent for de novo coronary lesions. Circulation 107, (2003). 9. Moses, J. W. et al. Sirolimus-eluting stents versus standard stents in patients with stenosis in a native coronary artery. N. Engl. J. Med. 349, (2003). 10. Stone, G. W. et al. A polymer-based, paclitaxeleluting stent in patients with coronary artery disease. N. Engl. J. Med. 350, (2004). 11. Muni, N. I. & Gross, T. P. Problems with drugeluting coronary stents the FDA perspective. N. Engl. J. Med. 351, (2004). 12. Moreno, R. et al. Drug-eluting stent thrombosis: results from a pooled analysis including 10 randomized studies. J. Am. Coll. Cardiol. 45, (2005). 13. Bavry, A. A., Kumbhani, D. J., Helton, T. J. & Bhatt, D. L. What is the risk of stent thrombosis associated with the use of paclitaxel-eluting stents for percutaneous coronary intervention? A meta-analysis. J. Am. Coll. Cardiol. 45, (2005). 14. Iakovou, I. et al. Incidence, predictors, and outcome of thrombosis after successful implantation of drug-eluting stents. JAMA 293, (2005). 15. Ong, A. T. et al. Thirty-day incidence and sixmonth clinical outcome of thrombotic stent occlusion after bare-metal, sirolimus, or paclitaxel stent implantation. J. Am. Coll. Cardiol. 45, (2005). 16. Wenaweser, P. et al. Incidence and correlates of drug-eluting stent thrombosis in routine clinical practice. 4 year results from a large 2 institutional cohort study. J. Am. Coll. Cardiol. 52, (2008). 17. Stone, G. W. et al. Safety and efficacy of sirolimus- and paclitaxel-eluting coronary stents. N. Engl. J. Med. 356, (2007). 18. Lagerqvist, B. et al. Long-term outcomes with drug-eluting stents versus bare-metal stents in Sweden. N. Engl. J. Med. 356, (2007). 19. Eisenstein, E. L. et al. Clopidogrel use and longterm clinical outcomes after drug-eluting stent implantation. JAMA 297, (2007). 20. King, S. B. et al focused update of the ACC/AHA/SCAI 2005 guideline update for percutaneous coronary intervention: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines: 2007 Writing Group to Review New Evidence and Update the ACC/AHA/SCAI 2005 guideline update for percutaneous coronary intervention, writing on behalf of the 2005 Writing Committee. Circulation 117, (2008). 21. Wijns, W. et al. Guidelines on myocardial revascularization. Eur. Heart J. 31, (2010). 22. Windecker, S. et al ESC/EACTS guidelines on myocardial revascularization: the Task Force on Myocardial Revascularization of the European Society of Cardiology (ESC) and the European Association for Cardio-Thoracic Surgery (EACTS): developed with the special contribution of the European Association of Percutaneous Cardiovascular Interventions (EAPCI). Eur. Heart J. 35, (2014). 23. Cutlip, D. E. et al. Clinical end points in coronary stent trials: a case for standardized definitions. Circulation 115, (2007). 24. Mauri, L. et al. Stent thrombosis in randomized clinical trials of drug-eluting stents. N. Engl. J. Med. 356, (2007). 25. Stettler, C. et al. Outcomes associated with drugeluting and bare-metal stents: a collaborative network meta-analysis. Lancet 370, (2007). 26. Tada, T. et al. Risk of stent thrombosis among bare-metal stents, first-generation drug-eluting stents, and second-generation drug-eluting stents: results from a registry of 18,334 patients. JACC Cardiovasc. Interv. 6, (2013). 27. Cutlip, D. E. et al. Stent thrombosis in the modern era: a pooled analysis of multicenter coronary stent clinical trials. Circulation 103, (2001). 28. Armstrong, E. J. et al. Clinical presentation, management, and outcomes of angiographically documented early, late, and very late stent thrombosis. JACC Cardiovasc. Interv. 5, (2012). 29. Brodie, B. et al. Predictors of early, late, and very late stent thrombosis after primary percutaneous coronary intervention with baremetal and drug-eluting stents for ST-segment elevation myocardial infarction. JACC Cardiovasc. Interv. 5, (2012). 30. Uren, N. G. et al. Predictors and outcomes of stent thrombosis: an intravascular ultrasound registry. Eur. Heart J. 23, (2002). 31. Schiele, F. et al. Impact of intravascular ultrasound guidance in stent deployment on 6 month restenosis rate: a multicenter, randomized study comparing two strategies with and without intravascular ultrasound guidance. RESIST Study Group. REStenosis after Ivus guided STenting. J. Am. Coll. Cardiol. 32, (1998). 32. Oemrawsingh, P. V. et al. Intravascular ultrasound guidance improves angiographic and clinical outcome of stent implantation for long coronary artery stenoses: final results of a randomized comparison with angiographic guidance (TULIP study). Circulation 107, (2003). 33. Cook, S. et al. Incomplete stent apposition and very late stent thrombosis after drug-eluting stent implantation. Circulation 115, (2007). 34. Russo, R. J. et al. A randomized controlled trial of angiography versus intravascular ultrasounddirected bare-metal coronary stent placement (the AVID trial). Circ. Cardiovasc. Interv. 2, (2009). 35. Cheneau, E. et al. Predictors of subacute stent thrombosis: results of a systematic intravascular ultrasound study. Circulation 108, (2003). 36. Fujii, K. et al. Stent underexpansion and residual reference segment stenosis are related to stent thrombosis after sirolimus-eluting stent implantation: an intravascular ultrasound study. J. Am. Coll. Cardiol. 45, (2005). 37. Nakano, M. et al. Causes of early stent thrombosis in patients presenting with acute coronary syndrome: an ex vivo human autopsy study. J. Am. Coll. Cardiol. 63, (2014). 38. Kobayashi, Y. et al. Long-term vessel response to a self-expanding coronary stent: a serial volumetric intravascular ultrasound analysis from the ASSURE Trial: A Stent vs. Stent Ultrasound Remodeling Evaluation. J. Am. Coll. Cardiol. 37, (2001). 39. Gonzalo, N. et al. Incomplete stent apposition and delayed tissue coverage are more frequent in drug-eluting stents implanted during primary percutaneous coronary intervention for STsegment elevation myocardial infarction than in drug-eluting stents implanted for stable/ unstable angina: insights from optical coherence tomography. JACC Cardiovasc. Interv. 2, (2009). 40. Tanigawa, J., Barlis, P. & Di Mario, C. Intravascular optical coherence tomography: optimisation of image acquisition and quantitative assessment of stent strut apposition. EuroIntervention 3, (2007). 41. Bezerra, H. G. et al. Optical coherence tomography versus intravascular ultrasound to evaluate coronary artery disease and percutaneous coronary intervention. JACC Cardiovasc. Interv. 6, (2013). 42. Ozaki, Y. et al. The fate of incomplete stent apposition with drug-eluting stents: an optical coherence tomography-based natural history study. Eur. Heart J. 31, (2010). 43. Foin, N. et al. Incomplete stent apposition causes high shear flow disturbances and delay in neointimal coverage as a function of strut to wall detachment distance: implications for the management of incomplete stent apposition. Circ. Cardiovasc. Interv. 7, (2014). 44. Tanabe, K. et al. Incomplete stent apposition after implantation of paclitaxel-eluting stents or bare metal stents: insights from the randomized TAXUS II trial. Circulation 111, (2005). 45. Shah, V. M., Mintz, G. S., Apple, S. & Weissman, N. J. Background incidence of late malapposition after bare-metal stent implantation. Circulation 106, (2002). 46. Hong, M. K. et al. Intravascular ultrasound comparison of chronic recoil among different stent designs. Am. J. Cardiol. 84, (1999). 47. Fujino, Y., Attizzani, G. F., Nakamura, S., Costa, M. A. & Bezerra, H. G. Frequency-domain optical coherence tomography assessment of stent constriction 9 months after sirolimus-eluting stent implantation in a highly calcified plaque. JACC Cardiovasc. Interv. 6, (2013). 48. Mintz, G. S., Shah, V. M. & Weissman, N. J. Regional remodeling as the cause of late stent malapposition. Circulation 107, (2003). 49. Ako, J. et al. Late incomplete stent apposition after sirolimus-eluting stent implantation: a serial intravascular ultrasound analysis. J. Am. Coll. Cardiol. 46, (2005). 50. Hong, M. K. et al. Late stent malapposition after drug-eluting stent implantation: an intravascular ultrasound analysis with long-term follow-up. Circulation 113, (2006). 334 JUNE 2015 VOLUME 12