Author + information
- Received April 13, 2015
- Revision received August 4, 2015
- Accepted August 6, 2015
- Published online November 1, 2015.
- Francesco Prati, MD, PhD∗,†∗ (, )
- Enrico Romagnoli, MD, PhD†,
- Francesco Burzotta, MD, PhD‡,
- Ugo Limbruno, MD§,
- Laura Gatto, MD∗,†,
- Alessio La Manna, MD‖,
- Francesco Versaci, MD¶,
- Valeria Marco, RN†,
- Luca Di Vito, MD, PhD∗,
- Fabrizio Imola, MD∗,
- Giulia Paoletti, RN†,
- Carlo Trani, MD‡,
- Corrado Tamburino, MD‖,
- Luigi Tavazzi, MD# and
- Gary S. Mintz, MD∗∗
- ∗San Giovanni Addolorata Hospital, Rome, Italy
- †Centro per la Lotta contro l’Infarto–CLI Foundation, Rome, Italy
- ‡Università Cattolica Del Sacro Cuore, Rome, Italy
- §Misericordia Hospital, Grosseto, Italy
- ‖Division of Cardiology, University of Catania, Catania, Italy
- ¶Ospedale Civile Ferdinando Veneziale, Isernia, Italy
- #GVM Care and Research, E.S. Health Science Foundation, Cotignola, Italy
- ∗∗Cardiovascular Research Foundation, New York, New York
- ↵∗Reprint requests and correspondence:
Dr. Francesco Prati, Cardiology Unit, San Giovanni–Addolorata Hospital, Via Amba Aradam 9, 00184 Rome, Italy.
Objectives The goal of this study was to assess the clinical impact of optical coherence tomography (OCT) findings during percutaneous coronary intervention (PCI).
Background OCT provides unprecedented high-definition visualization of plaque/stent structures during PCI; however, the impact of OCT findings on outcome remains undefined.
Methods In the context of the multicenter CLI-OPCI (Centro per la Lotta contro l’Infarto–Optimisation of Percutaneous Coronary Intervention) registry, we retrospectively analyzed patients undergoing end-procedural OCT assessment and compared the findings with clinical outcomes.
Results A total of 1,002 lesions (832 patients) were assessed. Appropriate OCT assessment was obtained in 98.2% of cases and revealed suboptimal stent implantation in 31.0% of lesions, with increased incidence in patients experiencing major adverse cardiac events (MACE) during follow-up (59.2% vs. 26.9%; p < 0.001). In particular, in-stent minimum lumen area <4.5 mm2 (hazards ratio [HR]: 1.64; p = 0.040), dissection >200 μm at the distal stent edge (HR: 2.54; p = 0.004), and reference lumen area <4.5 mm2 at either distal (HR: 4.65; p < 0.001) or proximal (HR: 5.73; p < 0.001) stent edges were independent predictors of MACE. Conversely, in-stent minimum lumen area/mean reference lumen area <70% (HR: 1.21; p = 0.45), stent malapposition >200 μm (HR: 1.15; p = 0.52), intrastent plaque/thrombus protrusion >500 μm (HR: 1.00; p = 0.99), and dissection >200 μm at the proximal stent edge (HR: 0.83; p = 0.65) were not associated with worse outcomes. Using multivariable Cox hazard analysis, the presence of at least 1 significant criterion for suboptimal OCT stent deployment was confirmed as an independent predictor of MACE (HR: 3.53; 95% confidence interval: 2.2 to 5.8; p < 0.001).
Conclusions Suboptimal stent deployment defined according to specific quantitative OCT criteria was associated with an increased risk of MACE during follow-up.
Optical coherence tomography (OCT) is the newest intracoronary imaging technique designed for better definition of coronary atherosclerosis and its functional consequences (1–4). The current OCT systems are rapid, with unprecedented spatial resolution allowing high-definition visualization of intraluminal and endothelial structures. Despite the high quality of OCT images, the clinical utility of this technique to improve percutaneous coronary interventions (PCIs) and clinical outcomes remains to be defined (1–3).
Recently, the CLI-OPCI (Centro per la Lotta contro l’Infarto–Optimisation of Percutaneous Coronary Intervention) study compared angiography alone versus angiographic guidance plus OCT guidance for routine PCI (5). The researchers found that OCT can identify nonoptimal stent deployment in approximately one-third of cases, thus providing preliminary evidence of the technique’s clinical utility. Importantly, for the first time, the CLIO-PCI study addressed the question of how to interpret OCT findings by setting specific quantitative criteria to identify suboptimal stent deployment (6).
The aim of the present study was to assess the impact of these pre-specified OCT quantitative criteria on clinical outcomes after PCI. For this purpose, the end-procedural OCT data in a large retrospective study (CLI-OCPI II) were evaluated. The study included 1,002 lesions in 832 patients with a follow-up length of at least 1 year.
This retrospective multicenter PCI registry included cases with frequency domain OCT assessment of stent positioning. All case subjects had at least 1 OCT assessment of the treated vessel, performed at the end of the procedure, with a sufficient acquisition length to address the whole length of the stented segments plus the proximal and distal reference segments (2,6–8). Indications for periprocedural OCT assessment and its practical utilization were left to the operator’s discretion; no formal selection criteria or treatment strategies (e.g., routinely stent post-dilation) were prospectively adopted in the enrolled case samples. For the purposes of this study, only OCT findings obtained at the end of the procedure were considered.
All patients provided written informed consent for the index procedure and for telephone/direct visit follow-up. Ethical approval was waived because of the study’s observational retrospective design.
As the primary objective of the study, the impact of the presence of an OCT-based suboptimal stent deployment on clinical outcome was explored; the impact of the individual OCT findings on outcome was also appraised. For this purpose, the incidence of major adverse cardiac events (MACE) was a composite of all-cause mortality, myocardial infarction (MI) not clearly attributable to a nontarget vessel (including periprocedural MI defined as creatine kinase-myocardial band level >3 times the upper limit of normal), and target lesion revascularization. All outcomes were defined according to the recommendations of the Academic Research Consortium (9).
Endpoint adjudication was performed by a central clinical event committee in a blinded fashion. No extramural funding was used to support this work, and the authors were solely responsible for the design, conduct, and final contents of the study.
Patients and procedures
Overall, 832 consecutive patients in Italy undergoing OCT guidance at 5 experienced and high-volume OCT centers entered the study. Given the retrospective design, treatment choices (including stenting technique, drug-eluting stent [DES] utilization, and additional pharmacological therapy) were according to local practice. In particular, OCT guidance during the procedure was not codified but left to the operator’s discretion.
PCIs were performed with standard techniques and catheters by using a femoral or radial approach. All patients received unfractionated heparin (a bolus of 70 IU/kg with additional doses aimed at achieving an intraprocedural activated clotting time of 250 to 300 s). All patients were pretreated with 325 mg of aspirin and a loading dose of clopidogrel 600 mg, prasugrel 60 mg, or ticagrelor 180 mg, if the patient was not already on a maintenance dose. Unless contraindicated, dual antiplatelet therapy was recommended for at least 12 months. During the first year after discharge, patients were followed up by means of scheduled direct visits (generally at 1 and 6 months) and telephone contacts. In case of any adverse event or new hospitalization, source documents were obtained and examined in detail.
OCT measurements and definitions
OCT was acquired by means of the frequency domain C7-XR system or the OPTIS system (both St. Jude Medical, St. Paul, Minnesota) with a nonocclusive technique according to a well-standardized method (2,8). OCT assessment of stent implantation was on the basis of conventional definitions reported in expert consensus OCT documents (2,4,10). The value with maximal predictive accuracy for outcome was used as a cutoff point for each variable (Figure 1). In particular, the following factors were considered significant findings:
1. Edge dissection: the presence of a linear rim of tissue with a width ≥200 μm and a clear separation from the vessel wall or underlying plaque that was adjacent (<5 mm) to a stent edge (2,6).
2. Reference lumen narrowing: lumen area <4.5 mm2 in the presence of significant residual plaque adjacent to stent endings (6);
3. Malapposition: stent-adjacent vessel lumen distance >200 μm (6,10,11);
4. In-stent minimum lumen area (MLA) <4.5 mm2 (6);
5. In-stent MLA <70% of the average reference lumen area;
6. Intrastent plaque/thrombus protrusion: tissue prolapsing between stent struts extending inside a circular arc connecting adjacent struts or intraluminal mass ≥500 μm in thickness, with no direct continuity with the surface of the vessel wall or highly backscattered luminal protrusion in continuity with the vessel wall and resulting in signal-free shadowing (2,10,12).
Definition of suboptimal OCT stent deployment required the presence of at least 1 of the OCT findings significantly associated with MACE.
By study design, only final OCT images performed at the end of the procedures were analyzed off-line at a certified central core laboratory (Rome Heart Research, Rome, Italy) whose operators were blinded to procedural characteristics and outcomes.
Continuous variables are reported as mean ± SD or median (1st to 3rd quartile) in case of normal or skewed distribution; discrete variables are reported as percentages. The Student t test, Mann-Whitney U test, chi-square test, and Fisher exact test were applied for bivariate analyses when appropriate. The receiver-operating characteristic curve was used to evaluate the predictive accuracy of each OCT parameter for outcome; the highest Youden index (J statistic) representing the maximum potential effectiveness was used to determine the optimal cutoff (13,14). Combined adverse events were evaluated on a per-patient hierarchical basis; thus, only 1 hard event per patient per event type was summarized as Kaplan-Meier estimates.
All study variables were tested for bivariate association with MACE; if nominally significant (p < 0.05), they were simultaneously forced into a Cox regression model to identify independent outcome predictors and to calculate their adjusted hazard ratios (HRs) with associated 95% confidence intervals (CIs). The Cox regression model included the following variables: left ventricular ejection fraction, diabetes mellitus, family history of coronary artery disease, non–ST-segment elevation myocardial infarction (NSTEMI) diagnosis, multivessel disease, left main disease, previous MI, angiographically ambiguous lesion (i.e., intermediate lesion with irregular contour and/or haziness), in-stent restenosis lesion, bare-metal stent (BMS) usage, ostial lesion treatment, and suboptimal final OCT result.
A score quantifying the propensity to incur MACE was computed to adjust for potential confounding factors inherent to the observational nature of the study (15,16). Specifically, the individual score, defined as the conditional probability of experiencing MACE, was estimated with a nonparsimonious logistic regression model, including all available co-variables but excluding those that were OCT related (C-statistic: 0.78; 95% CI: 0.71 to 0.85). Adjusted effect estimates were estimated from models in which the score was entered as covariates. In addition, as a sensitivity analysis, a matched pair analysis was performed on the basis of the propensity to develop MACE.
A 2-tailed, p value <0.05 was established as the level of statistical significance for all tests. All statistical analyses were conducted by using SPSS-PASW version 22.0 (IBM SPSS Statistics, IBM Corporation, Armonk, New York).
Between 2008 and 2013, a total of 832 patients with 1,002 lesions undergoing post-stenting OCT assessment were included in the registry. Clinical and procedural features of the study population are summarized in Tables 1 and 2, respectively. The patients’ median age was 64 years (interquartile range [IQR]: 56 to 72 years), and the study included 29.2% female subjects. Diagnosis at admission was acute coronary syndrome in 56.4% of patients, including acute ST-segment elevation MI in 31.0%. Most of the patients had a complex lesion profile (Ellis class B2/C 74.8%), with multivessel disease involvement in 52.6%.
Treated lesion location was as follows: left main, 4.8%; left anterior descending artery, 50.7%; left circumflex artery, 21.4%; right coronary artery, 22.7%; and graft conduit, 0.4%. DES implantation occurred in 71.4% of the lesions, and multiple overlapping stents were implanted in 21.4% of cases. Direct stenting and high-pressure stent post-dilation rates were 27.0% and 48.0%, respectively.
All OCT acquisitions were successfully performed; however, during off-line analysis, 1.8% of cases were discarded due to insufficient quality images (e.g., improper acquisition technique) (6). Therefore, OCT assessment was analyzed in 984 stented lesions, and suboptimal stent implantation was noted in 31.0% of cases (Table 3). In particular, OCT disclosed in-stent MLA <4.5 mm2 in 23.4% of the stented lesions, edge dissection in 12.7%, in-stent lumen underexpansion in 23.7%, malapposition in 49.3%, intrastent plaque/thrombus protrusion in 29.4%, and reference lumen narrowing in 7.5%.
The immediate angiographic success rate (residual stenosis <30% with Thrombolysis In Myocardial Infarction flow grade 3) was 97.6% with a periprocedural MI prevalence of 2.6%. The cumulative MACE rate at a median follow-up of 319 days (IQR: 123 to 576 days) was 12.6%, with 2.9% all-cause mortality, 7.7% nonfatal MI, and 6.7% target lesion revascularization (Table 4). Notably, 82% of adverse events occurred within the first 12 months after the procedure with a mean time-to-MACE of 26 days (IQR: 1 to 216 days).
Compared with patients with event-free survival, patients with MACE during follow-up had a lower left ventricular ejection fraction (52% [IQR: 43% to 60%] vs. 55% [IQR: 48% to 60%]; p = 0.002), more frequent NSTEMI diagnosis (19.1% vs. 7.7%; p < 0.001), more prior MI (32.4% vs. 17.9%; p < 0.001), and more multivessel disease (65.7% vs. 50.8%; p = 0.024) (Table 1). Regarding the procedural aspects, patients with MACE were characterized by higher BMS use (34.4% vs. 20.0%; p = 0.002) and more frequent treatment of a left main (9.6% vs. 4.1%; p = 0.012), ostial (8.8% vs. 4.8%; p = 0.048), angiographically ambiguous (13.6% vs. 8.0%; p = 0.023), or in-stent (7.2% vs. 3.3%; p = 0.045) restenosis lesion (Table 2).
OCT analyses revealed a significantly higher incidence of suboptimal stent deployment in lesions associated with any adverse event during follow-up (59.2% vs. 26.9%; p < 0.001). In particular, patients with lesions and MACE reported more frequent in-stent MLA <4.5 mm2 (40.8% vs. 20.8%; p < 0.001), dissection >200 μm at the distal stent edge (16.0% vs. 5.7%; p < 0.001), and reference lumen area <4.5 mm2 in the presence of residual significant plaque at either the distal (22.4% vs. 3.4%; p < 0.001) or proximal (11.2% vs. 1.2%; p < 0.001) stent edges. Conversely, in-stent MLA <70% of the average reference lumen area (30.4% vs. 22.7%; p = 0.07), dissection at the proximal stent edge (6.4% vs. 6.6%; p = 0.92), malapposition (50.4% vs. 49.1%; p = 0.85), or in-stent plaque/thrombus prolapse (30.4% vs. 29.2%; p = 0.83) were not associated with an increased MACE rate (Table 3).
In the multivariable Cox hazard analysis, suboptimal OCT stent deployment was confirmed as an independent predictor of MACE (HR: 3.53; 95% CI: 2.2 to 5.8; p < 0.001), together with impaired left ventricular ejection fraction (HR: 2.12; 95% CI: 1.3 to 3.5; p = 0.003), NSTEMI diagnosis (HR: 1.99; 95% CI: 1.1 to 3.6; p = 0.021), and left main disease (HR: 2.79; 95% CI: 1.3 to 6.2; p = 0.012). Figure 2 presents the relative Kaplan-Meier curves, and Table 5 displays the predictive value of the individual OCT criteria of suboptimal stent deployment. Sensitivity analysis on the basis of 146 patients matched for the propensity to incur MACE confirmed the results stemming from the main analysis in terms of both statistical direction and magnitude.
The main finding provided by this large multicenter registry was that patients exhibiting suboptimal stent deployment on the basis of specific OCT criteria experienced a higher rate of MACE during follow-up. Indeed, suboptimal stent deployment was significantly more common in the MACE group (59.2% vs. 26.9%; p < 0.001) and was found to be an independent predictor of MACE.
The new angle of view
Recent intravascular ultrasound (IVUS) data derived from observational studies and large meta-analyses have proven the efficacy of an IVUS-guided approach for reducing adverse clinical outcomes (including death, MI, and stent thrombosis) after PCI (17,18). OCT represents a new angle of view to address the adequacy of stent deployment. Besides enabling measurement of IVUS-validated predictors of MACE (including MLA and inflow/outflow disease), the high resolution of the OCT technique permits detection of features that may be missed by IVUS, such as malapposition, intrastent plaque/thrombus protrusion, or dissections at the stent edges and inside the stents.
The multicenter CLI-OPCI registry (5) showed that OCT could potentially improve the clinical outcomes after coronary intervention in a real-world population. In fact, the 1-year composite of cardiac death or nonfatal MI was significantly lower in the OCT-guided intervention arm. However, the promising conclusions reached by the CLI-OPCI registry should be approached with caution because of its nonrandomized design and relatively small population size (335 patients in the OCT group).
The present study broadens the previous experience of the CLIO-PCI registry by assessing the role of OCT findings after PCI in a much larger population (832 patients and 1,002 lesions; median follow-up 319 days). Consistent with previous data (5), CLI-OPCI II showed that OCT-defined suboptimal stent deployment was a relatively common finding (31.0% of cases), with a significantly higher prevalence in patients experiencing MACE in the first year of follow-up (59.2% vs. 26.9%; p < 0.001), and was an independent predictor of worse outcome (HR: 3.53; p < 0.001).
Specific OCT findings of suboptimal stenting
Conclusions reached by this study are in line with those emerging from the IVUS substudy of the ADAPT-DES (Assessment of Dual Antiplatelet Therapy with Drug Eluting Stents) trial registry (18). In particular, the CLIO-PCI II highlighted the role of residual reference segment disease. Stented segments exhibiting a narrowing at the reference lumen area <4.5 mm2 in the presence of significant plaque experienced a worse outcome, with the risk of MACE approximately 5 times higher regardless of the location (proximal or distal reference segment). These data were not unexpected: a large IVUS-identified plaque burden at the stent margins represents a well-known risk factor for late restenosis and thrombosis (19–21).
Dissections >200 μm at the distal stent edge also conveyed a higher risk of MACE (HR: 2.54; p = 0.004), whereas proximal dissections had no clinical impact. The relationship between distal dissection and worsened clinical outcome has already been noted in the CLI-OPCI registry (5). This finding was also in line with the ADAPT-DES study confirming the ominous role of distal dissections regardless of the amount of luminal narrowing. The negative impact of stent edge dissection, shown in the present study, was emphasized by the early occurrence of cardiac events. The majority of MACE occurred during the first 3 months after the procedure (Figure 2). Importantly, even at the applied 200-μm threshold, significant dissections may be missed by using IVUS (22).
There are 2 general approaches to assessing stent underexpansion. The first is using absolute dimensions that can be expressed as in-stent MLA or minimal stent area (MSA); the second is the relative minimal stent–to–mean reference lumen area percentage. Using IVUS, Ziada et al. (23) and Sonoda et al. (24) showed that the absolute dimension was the strongest predictor of freedom from adverse events after BMS or DES implantation. Similarly, an absolute in-stent MLA <4.5 mm2 according to OCT in the present study predicted MACE, whereas the relative criterion of stent–to–mean reference lumen area did not. Patients with subsequent MACE had a smaller MLA compared with those with no events; relative stent expansion was virtually identical in the 2 groups. OCT measurements are reportedly smaller than IVUS (22); in keeping with this observation, an in-stent MLA <4.5 mm2 according to OCT in the present study was consistent with the IVUS MSA criterion that has been reported after implantation of second-generation DES (25). Finally, the in-stent MLA reflects both actual stent underexpansion (i.e., the MSA) and in-stent plaque/thrombus prolapse. In the present study, 56.4% of patients presented with acute coronary syndrome, including acute ST-segment elevation MI in 31.0%; this is a patient population in whom in-stent plaque/thrombus prolapse has an important impact on in-stent lumen dimensions, whereas expansion of the metallic scaffold into a thrombus containing lesions is often easier than into a fibrotic lesion in patients with stable angina.
It has been suggested that acute malapposition may be associated with reduced re-endothelialization and increased formation of neointima. However, our data corroborate IVUS findings that failed to relate acute stent vessel wall malapposition with clinical outcome (26,27). Such conclusions were also in line with those recently reported from the CLI-THRO study (28) as well as a report from Im et al. (29). In the CLIO-THRO study (i.e., an OCT prospective registry designed to address the mechanism of stent thrombosis) (28), patients with subacute stent thrombosis more often exhibited stent underexpansion, stent edge dissection, reference lumen narrowing, and smaller MSA but no increased frequency of stent malapposition.
The main limitation of the present study was its nonrandomized, retrospective design. Thus, some evident clinical imbalances were present in patients experiencing MACE compared with patients with no events during follow-up. However, the presence of nonoptimal OCT criteria for stent deployment was an independent predictor of MACE in the multivariable Cox hazard analysis. This study included patients with different clinical conditions (i.e., patients in stable and acute condition) and treatment approach (i.e., BMS and DES); the role and importance of the described OCT findings could vary among these categories.
Although all the adopted definitions of suboptimal stent deployment were derived from previous IVUS experiences and OCT consensus documents, the proposed “clinical” cutoffs reflect efforts to delineate a practical approach to OCT guidance; these need to be validated in further studies, however. Finally, although some OCT findings are clearly associated with worse outcome, treatment and reformability remain to be investigated.
The present large multicenter registry showed that suboptimal stent deployment on the basis of specific OCT criteria was frequent in patients experiencing MACE in the first year of follow-up after stent implantation. In particular, suboptimal OCT stent deployment was an independent predictor of worse clinical outcome. These data seem to corroborate the rationale for an OCT-guided strategy during PCI.
COMPETENCY IN MEDICAL KNOWLEDGE 1: OCT-defined suboptimal stent deployment was a relatively common finding (31.0% of cases), with a significantly higher prevalence in patients experiencing MACE in the first year of follow-up (59.2% vs. 26.9%; p < 0.001).
COMPETENCY IN MEDICAL KNOWLEDGE 2: Suboptimal OCT stent deployment, defined according to specific quantitative OCT criteria, was an independent predictor of worse outcome (HR: 3.53; p < 0.001).
COMPETENCY IN PATIENT CARE AND PROCEDURAL SKILLS: OCT guidance during PCI allowed identification of patients at increased risk of MACE.
TRANSLATIONAL OUTLOOK 1: The management and reformability of OCT-defined suboptimal stent deployment require further investigation.
TRANSLATIONAL OUTLOOK 2: Randomized studies are needed to assess the clinical impact of OCT guidance during PCI.
This research was supported by a grant from the Centro per la Lotta contro l’Infarto–Fondazione Onlus. Dr. Prati has served as a consultant for St. Jude Medical. Dr. Burzotta has received speaker fees from St. Jude Medical; and been involved in advisory board meetings for Medtronic. Dr. Limbruno has received consulting fees from Abbott, Biotronik, AstraZeneca, Lilly, and Boston Scientific. Dr. Tavazzi has served as trial committee member and a member of the Speakers Bureau for Servier; and a trial committee member for Boston Scientific, Medtronic, Cardiorentis, CVIE Therapeutics, and St. Jude Medical. Dr. Mintz has received a research grant from St. Jude Medical. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- bare-metal stent(s)
- confidence interval
- drug-eluting stent(s)
- hazard ratio
- interquartile range
- intravascular ultrasound
- major adverse cardiac event(s)
- myocardial infarction
- minimum lumen area
- minimal stent area
- non–ST-segment elevation myocardial infarction
- optical coherence tomography
- percutaneous coronary intervention
- Received April 13, 2015.
- Revision received August 4, 2015.
- Accepted August 6, 2015.
- American College of Cardiology Foundation
- Suter M.J.,
- Nadkarni S.K.,
- Weisz G.,
- et al.
- Prati F.,
- Guagliumi G.,
- Mintz G.S.,
- et al.,
- Expert's OCT Review Document
- Gonzalo N.,
- Escaned J.,
- Alfonso F.,
- et al.
- Tearney G.J.,
- Regar E.,
- Akasaka T.,
- et al.,
- International Working Group for Intravascular Optical Coherence Tomography (IWG-IVOCT)
- Prati F.,
- Di Vito L.,
- Biondi-Zoccai G.,
- et al.
- Prati F.,
- Regar E.,
- Mintz G.S.,
- et al.,
- Expert’s OCT Review Document
- Cutlip D.E.,
- Windecker S.,
- Mehran R.,
- et al.,
- Academic Research Consortium
- Serruys P.W.,
- Onuma Y.,
- Ormiston J.A.,
- et al.
- Egan J.P.
- Rosenbaum P.R.,
- Rubin D.B.
- Ahn J.M.,
- Kang S.J.,
- Yoon S.H.,
- et al.
- Witzenbichler B.,
- Maehara A.,
- Weisz G.,
- et al.
- Fujii K.,
- Carlier S.G.,
- Mintz G.S.,
- et al.
- Kubo T.,
- Akasaka T.,
- Shite J.,
- et al.
- Ziada K.M.,
- Kapadia S.R.,
- Belli G.,
- et al.
- Sonoda S.,
- Morino Y.,
- Ako J.,
- et al.,
- SIRIUS Investigators
- Steinberg D.H.,
- Mintz G.S.,
- Mandinov L.,
- et al.
- Guo N.,
- Maehara A.,
- Mintz G.S.,
- et al.
- Im E.,
- Kim B.K.,
- Ko Y.G.,
- et al.