Author + information
- Received February 20, 2013
- Revision received April 26, 2013
- Accepted May 9, 2013
- Published online March 1, 2014.
- Christos V. Bourantas, MD, PhD∗,
- Michail I. Papafaklis, MD, PhD†,
- Anna Kotsia, MD‡,
- Vasim Farooq, MB, ChB∗,
- Takashi Muramatsu, MD, PhD∗,
- Josep Gomez-Lara, MD, PhD∗,
- Yao-Jun Zhang, MD, PhD∗,
- Javaid Iqbal, PhD∗,
- Fanis G. Kalatzis, MD, PhD§,
- Katerina K. Naka, MD, PhD‡,
- Dimitrios I. Fotiadis, PhD§,
- Cecile Dorange, MSc‖,
- Jin Wang, PhD¶,
- Richard Rapoza, PhD¶,
- Hector M. Garcia-Garcia, MD, PhD∗,
- Yoshinobu Onuma, MD∗,
- Lampros K. Michalis, MD‡ and
- Patrick W. Serruys, MD, PhD∗∗ ()
- ∗Department of Interventional Cardiology, Erasmus University Medical Centre, Thoraxcenter, Rotterdam, the Netherlands
- †Cardiovascular Division, Brigham & Women's Hospital, Harvard Medical School, Boston, Massachusetts
- ‡Department of Cardiology, Medical School, University of Ioannina, Ioannina, Greece
- §Department of Materials Science and Engineering, University of Ioannina, Ioannina, Greece
- ‖Abbott Vascular, Diegem, Belgium
- ¶Abbott Vascular, Santa Clara, California
- ↵∗Reprint requests and correspondence:
Dr. Patrick W. Serruys, Interventional Cardiology Department, Erasmus Medical Center, ‘s-Gravendijkwal 230, 3015 CE Rotterdam, the Netherlands.
Objectives This study sought to investigate the effect of endothelial shear stress (ESS) on neointimal formation following an Absorb bioresorbable vascular scaffold (BVS) (Abbott Vascular, Santa Clara, California) implantation.
Background Cumulative evidence, derived from intravascular ultrasound–based studies, has demonstrated a strong association between local ESS patterns and neointimal formation in bare-metal stents, whereas in drug-eluting stents, there are contradictory data about the effect of ESS on the vessel wall healing process. The effect of ESS on neointimal development following a bioresorbable scaffold implantation remains unclear.
Methods Twelve patients with an obstructive lesion in a relatively straight arterial segment, who were treated with an Absorb BVS and had serial optical coherence tomographic examination at baseline and 1-year follow-up, were included in the current analysis. The optical coherence tomographic data acquired at follow-up were used to reconstruct the scaffolded segment. Blood flow simulation was performed on the luminal surface at baseline defined by the Absorb BVS struts, and the computed ESS was related to the neointima thickness measured at 1-year follow-up.
Results At baseline, the scaffolded segments were exposed to a predominantly low ESS environment (61% of the measured ESS was <1 Pa). At follow-up, the mean neointima thickness was 113 ± 45 μm, whereas the percentage scaffold volume obstruction was 13.1 ± 6.6%. A statistically significant inverse correlation was noted between baseline logarithmic transformed ESS and neointima thickness at 1-year follow-up in all studied segments (correlation coefficient range −0.140 to −0.662). Mixed linear regression analysis between baseline logarithmic transformed ESS and neointima thickness at follow-up yielded a slope of −31 μm/ln(Pa) and a y-intercept of 99 μm.
Conclusions The hemodynamic microenvironment appears to regulate neointimal response following an Absorb BVS implantation. These findings underline the role of the ESS patterns on vessel wall healing and should be taken into consideration in the design of bioresorbable devices.
Neointimal formation is modulated by several local factors, including the vessel wall trauma caused during stent deployment, the plaque burden, and the composition of the underlying plaque, as well as local endothelial shear stress (ESS) patterns (1–5). Several clinical and experimental studies have provided evidence that local hemodynamic factors, in particular low ESS, promote neointimal formation in bare-metal stents, whereas in drug-eluting stents, the association between ESS and neointimal proliferation is weak and appears to be affected by the mechanisms of action and probably the release kinetics of the eluted drug (6–8).
Bioresorbable scaffold is a new technology introduced to overcome the long-term implications of metallic caging, because these devices have the unique ability to disappear after implantation, allowing restoration of vessel physiology (9). The first clinical studies provided evidence about the safety and efficacy of these devices and revealed a gradual increase of the neointima tissue, which, however, was accommodated by the expanding scaffold and did not appear to affect luminal dimensions (10). We have recently demonstrated that neointima tissue has an asymmetric distribution around the circumference of the vessel wall, indicating that local factors (i.e., vessel wall trauma, increased plaque inflammation, and local hemodynamics) are likely to be involved and regulate this process (11).
The aim of the present analysis was to investigate the impact of ESS on neointima proliferation following an Absorb bioresorbable vascular scaffold (BVS) (Abbott Vascular, Santa Clara, California) implantation. In contrast to previous reports, we utilized optical coherence tomographic (OCT) data to reconstruct the surface of the scaffolded segment at baseline, simulate blood flow, and assess vessel wall healing. The high resolution of this imaging technique allows more detailed reconstruction of luminal morphology and evaluation of the hemodynamic microenvironment and its impact on neointimal growth.
Included patients and study design
We analyzed data from the patients recruited in the second group of the ABSORB Cohort B Trial (A Clinical Evaluation of the Everolimus Eluting Bioresorbable Vascular Scaffold System in the Treatment of Patients With De Novo Native Coronary Artery Lesions; NCT00856856). The study design has already been described in detail by Serruys et al. (12). In brief, 101 patients with single- or 2-vessel de novo coronary disease implanted with an Absorb BVS (dimensions 3.0 × 18 mm) were included in this prospective multicenter trial. The studied population was divided into 2 groups. Both groups had serial angiographic, grayscale intravascular ultrasound (IVUS), IVUS virtual histology, and OCT evaluation at 3 time points. The first group (B1) had these tests at baseline post-device implantation and at 6 months and 2 years follow-up, whereas the second group (B2) had these investigations at baseline, 1-year, and 3-year follow-up. Thus the current analysis included only the patients from Cohort B2 who had received an Absorb BVS in a relatively straight coronary segment and had undergone OCT evaluation at baseline and 1-year follow-up.
The ABSORB Cohort B study was approved by the human research committee of the institutions that participated. Informed consent was obtained from all patients.
Computation of vessel angulation
Biplane coronary angiographic imaging that would allow accurate reconstruction of coronary geometry was available only in 1 patient at 1-year follow-up. Thus, coronary reconstruction was performed using only the OCT data neglecting the vessel curvature. To minimize the error introduced by this approximation, we analyzed relatively straight arterial segments. The angulation of the treated segments was assessed using a previously described methodology (13). In brief, an expert observer reviewed the angiographic images acquired at 1-year follow-up and selected an end-diastolic projection where there was minimal foreshortening and overlapping of the scaffolded segment. The angulation of the scaffolded segment was defined as the angle formed by the tangents of the centerlines of the 5-mm proximal and distal parts of the segment. To classify the scaffolded segments as straight or curved, we estimated the curvature of the segments treated in the Absorb Cohort B study during balloon inflation, assuming that the balloon straightened the treated artery. The mean ± 2 SD (29°) of the angulations measured during balloon inflation was used as a cutoff value (13).
Optical coherence tomography
OCT image acquisition was performed at baseline (immediately after scaffold implantation) and at 1-year follow-up using a C7XR Fourier Domain system (LightLab Imaging, Westford, Massachusetts). The data acquired at baseline were analyzed by 2 expert observers blinded to patients' procedural and clinical characteristics, who reviewed the scaffolded segment and identified in each frame the embedded, the protruded, and the malapposed struts. Embedded struts were struts where the surface was >50% impacted in the vessel wall; struts were considered protruded when they were in contact with the vessel wall but with <50% of their surface being impacted in the vessel wall; and malapposed when the struts' abluminal surface was not in contact with the vessel wall. The feasibility of this classification has previously been tested in our department, and the reported results have shown a high interobserver agreement (kappa index 0.75) (14).
The data acquired at 1-year follow-up were used to reconstruct 2 surfaces: 1) the luminal surface at baseline post-scaffold implantation; and 2) the luminal surface at 1-year follow-up. In each follow-up OCT examination, the observers identified the frames portraying the scaffolded segment and analyzed 1 frame at every 0.4-mm interval in the non-scaffolded segment and 1 frame at 0.2-mm intervals in the scaffolded segment. The luminal morphology at baseline was approximated by the luminal borders in the non-scaffolded segment at 1-year follow-up. In the scaffolded segment, the observers identified corresponding struts between baseline and follow-up OCT examinations, and in the case where the baseline struts were embedded, the baseline luminal borders were defined by splines connecting the adluminal sides of the struts (portrayed at the frames acquired at 1-year follow-up), whereas in the case where the baseline struts were protruded or malapposed, the baseline luminal borders were defined by the adluminal sides of the struts (portrayed at the frames acquired at 1-year follow-up) and between struts by the segment connecting the abluminal sides of adjacent struts (Fig. 1). For the struts where it was not possible to identify correspondence between baseline and follow-up examinations, we assumed that these struts were protruded because the majority of the struts were protruded at the baseline examination (14).
The luminal surface at follow-up was constructed by the luminal borders in the scaffolded segment defined by the endoluminal borders of the neointima. Malapposed struts at 1-year follow-up were not taken into consideration, and in these segments, the baseline luminal border was approximated by the endoluminal border of the neointima (Fig. 1). Segments exhibiting extensive malapposition (malapposition seen in >25% of the frames portraying the scaffolded segment) were excluded from the analysis.
Reconstruction of coronary anatomy
The vessel centerline was approximated by a straight line with a length equal to the length of the studied coronary artery. The center of mass of the luminal borders in native segments, and of the scaffold borders (defined by the abluminal side of the scaffold) in scaffolded segments, were estimated, and these borders were then placed perpendicularly onto the centerline in their corresponding locations with their center of mass being positioned on the vessel centerline. This approach is likely to distort the anatomy of eccentric lesions located proximally or distally to the scaffold; however, it is expected to provide a more accurate representation of the coronary morphology in the scaffolded segment and eliminate the wave-like artifacts introduced by the movement of the catheter (during the cardiac circle) that can affect the ESS measurements (Fig. 2).
The final outcome of this process is 2 non-uniform, rational B-spline surfaces: the first representing the luminal surface at baseline (constructed by the luminal borders in the non-scaffolded segment and by the scaffold borders in the scaffolded segment), and the second representing the luminal surface at 1-year follow-up (constructed by the luminal borders in the scaffolded segment).
Blood flow simulation
The obtained geometries were further processed with computational fluid dynamics techniques for the generation of a finite volume mesh. Anisotropic meshes with unstructured tetrahedral elements were generated for each baseline surface using an automated mesh generation program (ICEM CFD version 11, ANSYS, Canonsburg, Pennsylvania). To capture the detailed characteristics of the hemodynamic microenvironment, mesh density was increased around the stent struts and within the boundary layer of the flow field, and had a maximum element edge equal to approximately one-fourth of the BVS strut thickness (i.e., ∼40 μm). Blood flow simulation was performed by solving the 3-dimensional (3D) Navier-Stokes equations (CFX 11, ANSYS). Blood was considered as a homogeneous, Newtonian fluid with a dynamic viscosity of 0.0035 Pa·s and a density of 1,050 kg/m3. Blood flow was considered to be steady, laminar, and incompressible, and a flat velocity profile was imposed at the inlet of the entire reconstructed arterial segment. Coronary blood flow for each artery was estimated by measuring in 2 angiographic projections, obtained at baseline post-scaffold implantation, the number of frames required for the contrast agent to pass from the inlet to the outlet of the reconstructed segment, the volume of the segment at baseline, and the cine frame rate (6). The arterial wall was considered to be rigid, and no-slip conditions were applied at the baseline luminal surface, whereas zero-pressure conditions were imposed at the outlet. ESS at the baseline luminal surface was calculated as the product of blood viscosity and the gradient of blood velocity at the wall.
Data analysis and statistics
In the scaffolded segment, the lumen volume at baseline and 1-year follow-up was estimated, and the neointimal volume was defined as: the lumen volume at baseline minus the lumen volume at follow-up, whereas the percentage volume obstruction was defined as: 100 times the neointima volume divided by the lumen volume at baseline. Neointima thickness (NT) was calculated as the distance between the luminal surface at baseline and the luminal surface at follow-up using an in-house–developed algorithm implemented in Visual Fortran 6.5 (Compaq Computer Corporation, Houston, Texas) (15). Positive values corresponded to areas of neointimal hyperplasia. NT was measured along the axial direction per 0.2 mm and around the vessel circumference per 5° for each cross section, and was associated with ESS at the corresponding location. A part of the scaffolded segment at the entrance and exit of the scaffold, covering a length of 2.5 mm, was excluded from the analysis. Segments within the scaffold located at the origin of side branches and their adjacent segments with length equal to the diameter of the side branch were also excluded from the analysis; the branches were not included in the 3D reconstruction, as the ESS assessment in these areas is not considered to be reliable.
Continuous variables are presented as mean ± SD, whereas categorical variables are presented as counts and percentages. Pearson correlation coefficient and linear regression analyses were implemented to investigate the association between the baseline ESS or the logarithmic transformed ESS (lnESS) and the estimated NT at follow-up. To control for patient effect, a mixed model with random intercept and slope was used to estimate the overall association between lnESS and NT. In this mixed model, the autoregressive covariance structure was utilized to take into consideration the nested structure of cross sections within subjects.
The interobserver agreement in the identification of corresponding struts between baseline and follow-up examination was evaluated with the use of the Cohen's kappa test. A p value <0.05 was considered statistically significant. Statistical analysis was performed with the SPSS statistical software package (version 18.0 for Windows, SPSS, Chicago, Illinois) and SAS software (version 9.2, SAS Institute, Cary, North Carolina).
Twenty-one patients (22 lesions) from Cohort B2 were investigated with OCT at baseline and 1-year follow-up. In 14 patients (14 arteries), the treated segments had an angulation <29°. Two cases were excluded from the analysis: the first because of extensive malapposition and the second because of poor OCT image quality. Thus, 12 segments were reconstructed. The baseline characteristics of the studied population are shown in Table 1. Most of the patients had hypercholesterolemia (58%) and hypertension (67%), and 83% were treated for stable angina.
Coronary reconstruction analysis
The length of the reconstructed segments was 47.6 ± 10.5 mm, whereas the length of the scaffolded segment was 19.6 ± 0.8 mm. Neointimal proliferation was evident in all treated segments, which resulted in a percentage scaffold volume obstruction of 13.1 ± 6.6% (Table 2). The mean NT derived from the measurements in the 12 patients (scaffold level analysis) was 113 ± 45 μm (range 72 to 200 μm).
Shear stress analysis
In the 12 patients (scaffold-level analysis), the mean ESS measured in the scaffolded segments at baseline was 1.10 ± 0.35 Pa. Sixty-one percent of the ESS estimations in the 12 scaffolded segments were <1 Pa. In native segments, the ESS values were higher (2.53 ± 1.24 Pa, p < 0.0001); only 21% of the ESS estimations were <1 Pa. As shown in Figure 3, the protruded struts created a rough luminal surface at baseline that affected the local hemodynamic microenvironment. Flow recirculation zones were noted proximally and distally to the protruded struts, resulting in low ESS in these regions and relatively high ESS on top of the struts (Fig. 4). The low ESS noted in these areas appeared to have affected neointima formation with increased tissue formation proximally and distally to the scaffold struts and with limited neointima formation on top of the struts (Figs. 3 and 4).
A negative correlation coefficient (average −0.385, statistically different from 0; p < 0.001) was noted between the baseline ESS and the NT at 1-year follow-up in all of the studied scaffolded segments. The correlation coefficient was higher after the logarithmic transformation of the baseline ESS (average −0.451; p < 0.001). Table 3 provides the correlation coefficients between baseline lnESS and NT for each subject, as well as the estimated slopes and y-intercepts after applying linear regression analysis for each subject (Fig. 5). The linear mixed effects model yielded an overall slope of −31 μm/ln(Pa) and y-intercept of 99 μm.
Interobserver agreement in identifying corresponding struts in OCT examinations
The reliability of the 2 observers in identifying corresponding struts between baseline and the follow-up examinations was examined in 100 frames randomly selected from 8 patients. The Kappa index was 0.79, indicating a good agreement.
In this study, we used for the first time serial OCT data and computational fluid dynamics techniques to evaluate the effect of the hemodynamic microenvironment on neointimal formation post-Absorb BVS implantation. We found that: 1) the thick protruded struts (strut thickness 156 μm) of the implanted scaffold create a rough surface that causes flow disturbance and recirculation zones resulting in low ESS; and 2) the reported low ESS contributed to neointimal formation because there was a negative correlation in all of the scaffolded segments between ESS and NT.
Several computational experimental and clinical reports in the past have examined the effect of stent implantation on local hemodynamics and the association between ESS and neointimal formation. In silico blood flow simulation studies have underscored the effect of stent design on local hemodynamic patterns: the stents that have thick, rectangular-shaped struts cause flow disruption and lead to a low ESS environment, whereas the devices with thin, circular-shaped struts appear to minimally influence the local flow and ESS (16–19). The connection of the struts seems also to affect ESS patterns: strut connectors that are parallel to the flow direction have a minimal effect on ESS, whereas strut connectors that have a perpendicular arrangement to the flow direction influence considerably the ESS distribution and increase the in-stent areas exposed to low ESS (20).
Clinical studies in 3D models obtained from patient data have shown an inverse correlation between ESS and NT in bare-metal stents, whereas in drug-eluting stents, there are conflicting reports (6,7,21,22). Gijsen et al. (22) demonstrated an inverse relation between ESS and neointimal formation in patients treated with sirolimus-eluting stents, whereas Suzuki et al. (21) found that ESS had no impact on neointimal development in diabetic patients implanted with the same type of stent. Papafaklis et al. (6) evaluated the association between neointimal formation and ESS in 10 patients implanted with paclitaxel-eluting stents and 10 patients treated with sirolimus-eluting stents, and found an inverse correlation between ESS and NT in the paclitaxel arm, but no significant association in the group of patients implanted with sirolimus-eluting stents, advocating that these findings are due to the different pathophysiological effect of each drug on vessel wall healing and the pro-restenotic ESS-related biological pathway. However, a significant limitation of the aforementioned studies is the fact that neointimal formation was measured assuming that stent struts were well apposed post-device implantation and that coronary reconstruction was based on IVUS, an imaging modality with a limited radial resolution that cannot accurately evaluate the lumen surface irregularities caused by the stent struts and the modest neointima proliferation in drug-eluting stents.
The present study overcomes the aforementioned pitfalls because it uses for the first time OCT to reconstruct the coronary lumen. OCT, with its high radial resolution, allows reliable evaluation of luminal morphology, detailed assessment of strut apposition, and accurate estimation of NT. In addition, the heterogeneous appearance of the struts of the Absorb BVS in OCT permits us to find corresponding struts at baseline and follow-up examinations and use this information for more reliable reconstruction of the baseline luminal surface (23). The high resolution of OCT allowed us to confirm for the first time in vivo the findings of computational flow dynamic studies on theoretical models and demonstrate that the thick, protruded struts disrupt flow, resulting in recirculation zones in front of and especially behind the struts. These regions are exposed to low ESS (ESS were <1 Pa in 61% of the scaffolds' surfaces), which appears to promote neointimal growth (24).
In contrast to previous reports that showed that rapamycin derivatives abrogate the effect of ESS on neointimal formation, we found a statistically significant correlation between ESS and NT in all studied segments implanted with an everolimus-eluting BVS (6,8,21). This discrepancy should be attributed to the more reliable evaluation of NT thickness and baseline ESS patterns in the reconstructed by OCT data segments, as well as to the longer follow-up period. In previous studies, the association between NT and ESS was evaluated at 6 and 9 months follow-up—only 3 or 6 months after the complete elution of the antiproliferative drug (25). In Absorb BVS, 80% of the everolimus elution is released within the first month and is completed 3 months after device implantation (26). It can be speculated that after the first months, the low ESS promoted neointimal formation, resulting in a negative association between ESS at baseline and neointimal formation at 1-year follow-up.
In this analysis, we focused on the association between ESS and NT at 1-year follow-up and not at other time points (i.e., at 3 years follow-up or at 6 months or 2 years follow-up using the data of the ABSORB Cohort B1 group) for the following reasons: 1) one-half of the baseline OCT examinations in the Cohort B1 group were performed by an M-3 Time Domain system, which has different radial and axial resolution and image quality from the C7XR Fourier Domain system, and thus, it is not feasible to find corresponding struts between baseline and follow-up examinations; 2) our objective was not to compare our findings with the results of the IVUS-based reconstruction studies in metallic drug-eluting stents, but rather to examine the midterm effect of ESS on NT where the antiproliferative drug has been eluted and the effect of vessel wall injury has dissipated; and 3) although it would be interesting to compare the association between ESS and NT at 2 different follow-up time points (i.e., at 1- and 3-year follow-up), because it has been shown that neointimal tissue continues to develop, the proposed analysis cannot be applied at 2- or 3-year follow-up because the scaffold loses its structural integrity after 1 year and expands; thus we cannot approximate the baseline luminal surface in the scaffolded segment based on the location of the struts at 2- or 3-year follow-up (10,27). To our knowledge, this is the first study that examines the effect of ESS on neointimal proliferation post-bioresorbable scaffold implantation. We found that the early depolymerization of the scaffold's struts does not apparently affect the mechanotransduction pathways that regulate the response of the free (i.e., noncovered by struts) vessel wall to the local hemodynamic microenvironment (28–30). Hence, the current design (in-phase zigzag hoops linked with bridges; strut thickness of 156 μm) and composition (poly-l-lactic struts, covered by a thin layer of an amorphous matrix of poly-d,l-lactide that controls the release of the antiproliferative drug everolimus) of the Absorb BVS seems to provide a template for the formation of a potentially protective thin layer of neointimal tissue that without restricting the luminal dimensions covers the underlying plaque and the scaffold's struts, minimizing the risk for late scaffold thrombosis. Furthermore, the neointima tissue that developed in the areas between the struts appears to smooth the luminal morphology and normalize the ESS, creating in the long term an athero-protecting environment (31).
A significant limitation of this analysis is the use of the follow-up OCT data to reconstruct both the baseline and follow-up luminal surface. This was done because there were significant differences in corresponding OCT frames, regarding the number of the portrayed struts and their circumferential distribution, between baseline and follow-up examinations (Fig. 1). Thus, it was not possible to accurately associate the luminal surface derived from the baseline OCT data with the luminal surface reconstructed at follow-up so as to quantify the NT and examine the effect of the baseline ESS on neointimal proliferation.
Another limitation of the current analysis is the lack of biplane angiographic data that would allow accurate 3D reconstruction of the coronary anatomy. To address this drawback, we included in this analysis relatively straight coronary arteries, assuming that they were totally straight segments and that the scaffold was symmetrically expanded. We have recently reported a high minimal luminal/maximal luminal diameter ratio (0.85 ± 0.08) in IVUS cross-sectional images of segments implanted with an Absorb BVS 1.1, a fact that indicates a symmetric expansion of the device (32). However, potential asymmetries in scaffold expansion that may have affected the estimated association between ESS and NT cannot be excluded. The effect of malapposed struts on local hemodynamics was not taken into account in this analysis. Considering the overall low reported prevalence of malapposed struts and the fact that we excluded segments with extensive malapposition, it is unlikely for this approximation to have affected our results (14).
In this study, we used serial OCT data to reconstruct the coronaries and evaluate the effect of the hemodynamic microenvironment on neointimal formation after Absorb BVS implantation. We found a statistically significant inverse association between ESS and NT in all of the studied segments. These findings underscore the role of local hemodynamic milieu on vessel wall healing and should be taken into consideration in the design of bioresorbable devices.
The ABSORB Cohort B study was sponsored and financially supported by Abbott Vascular. Dr. Bourantas is funded by the Hellenic Cardiological Society. Dr. Dorange is an employee of Abbott Vascular. Drs. Wang and Rapoza are employees of and hold stock in Abbott Vascular. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- endothelial shear stress
- intravascular ultrasound
- neointimal thickness
- optical coherence tomographic/tomography
- Received February 20, 2013.
- Revision received April 26, 2013.
- Accepted May 9, 2013.
- American College of Cardiology Foundation
- Hoffmann R.,
- Mintz G.S.,
- Mehran R.,
- et al.
- Farb A.,
- Weber D.K.,
- Kolodgie F.D.,
- Burke A.P.,
- Virmani R.
- Koskinas K.C.,
- Chatzizisis Y.S.,
- Antoniadis A.P.,
- Giannoglou G.D.
- Prati F.,
- Di Mario C.,
- Moussa I.,
- et al.
- Papafaklis M.I.,
- Bourantas C.V.,
- Theodorakis P.E.,
- et al.
- Wentzel J.J.,
- Krams R.,
- Schuurbiers J.C.,
- et al.
- Cheng C.,
- Tempel D.,
- Oostlander A.,
- et al.
- Serruys P.W.,
- Garcia-Garcia H.M.,
- Onuma Y.
- Ormiston J.A.,
- Serruys P.W.,
- Onuma Y.,
- et al.
- Serruys P.W.,
- Onuma Y.,
- Ormiston J.A.,
- et al.
- Gomez-Lara J.,
- Garcia-Garcia H.M.,
- Onuma Y.,
- et al.
- Gomez-Lara J.,
- Radu M.,
- Brugaletta S.,
- et al.
- LaDisa J.F. Jr..,
- Olson L.E.,
- Guler I.,
- et al.
- Stone P.H.,
- Saito S.,
- Takahashi S.,
- et al.
- Serruys P.W.,
- Onuma Y.,
- Dudek D.,
- et al.
- Chatzizisis Y.S.,
- Coskun A.U.,
- Jonas M.,
- Edelman E.R.,
- Feldman C.L.,
- Stone P.H.
- Bourantas C.V.,
- Papafaklis M.I.,
- Garcia-Garcia H.M.,
- et al.