Accuracy and Reproducibility of Stent-Strut Thickness Determined by Optical Coherence Tomography
- Volume 21 - Issue 11 - November, 2009
- Posted on: 11/6/09
- 0 Comments
- 10172 reads
ABSTRACT: Background. Optical coherence tomography (OCT) has been increasingly used to evaluate stent apposition following implantation. Since stent struts are visualized as linear structures with strong surface reflection and typical dorsal shadowing, apposition of struts is evaluated by measuring the distance between the strut surface reflection and adjacent vessel surface in consideration of strut thickness. However, there are no data available to validate the measurements of strut thickness by OCT. The aim of this in vitro study is to validate the accuracy of OCT measurement of stent-strut thickness of different commercially available stents in evaluating stent apposition. Methods. We performed the in vitro study after implantation of 5 commonly used stents in a phantom model artery. Stent-strut thickness was measured by a commercially available OCT system and was compared to the manufacturers’ nominal strut-thickness data for each stent. Intra- and interobserver variability were also assessed. Results. A total of 239 stent struts were evaluated. The differences in stent-strut measurements as compared to the manufacturers’ nominal strut thickness data were low. The intra- and interobserver measurement differences were low (6 ± 7 µm, and 6 ± 7 µm, respectively), with high correlation coefficients (r = 0.957 and r = 0.957, respectively; p < 0.0001). Conclusions. This in vitro study demonstrates that OCT analysis measuring stent-strut thickness provides accurate data with high reproducibility, suggesting that assessment of stent-strut apposition using OCT is feasible.
J INVASIVE CARDIOL 2009;21:602–605
Key words: accuracy, reproducibility, stent,
Optical coherence tomography (OCT) is being introduced as a new intravascular imaging modality and is gaining widespread use in the cardiac catheterization laboratory. OCT has a distinct advantage over intravascular ultrasound (IVUS) owing to its higher resolution power, with axial resolution of 10–20 µm and lateral resolution of 20–94 µm. OCT uses a near-infrared light source (1,310 nm), which detects the backscattering of light from superficial structures in the vessel wall. OCT has emerged as a very useful tool to evaluate stent-strut apposition, neointimal tissue quantification and to assess plaque morphology.1–8
Recently, incomplete stent apposition has been reported following usage of drug-eluting stents (DES), potentially contributing to associated clinical events such as stent thrombosis.9 OCT can provide detailed information regarding stent structure and its apposition to the underlying vessel wall; however, strong stent surface reflection with typical dorsal shadowing may make it difficult to evaluate stent apposition (Figure 1). Therefore, stent apposition can be evaluated by measuring the distance between the stent-strut surface reflection and the adjacent visible vessel surface.3,4,6 During these analyses, the stent-strut thickness information provided by manufacturers must be taken into consideration to precisely quantify true separation of the stent strut from the vessel wall; however, no validation studies have been conducted thus far to evaluate the manufacturers’ stated stent-strut thickness and the accuracy of OCT measurements for various commercially available stent platforms.
The aim of this in vitro study was to validate the accuracy of OCT measurement of stent-strut thickness of various commercially available stent platforms in the evaluation of stent apposition.
Materials and Methods
Five different types of commonly used coronary stents were used: 1) Bx Velocity™ (platform for Cypher™, Cordis Corp., Miami Lakes, Florida); 2) Cypher™; 3) Express2™ (platform for Taxus Express™, Boston Scientific Corp., Natick, Massachusetts); 4) Driver™ (platform for Endeavor™, Medtronic, Inc., Minneapolis, Minnesota); 5) Vision™ (platform for Xience™, Abbott Laboratories, Abbott Park, Illinois). All stents used were of 3.5 mm diameter and the stent length was 18 mm for the Bx Velocity, Cypher, Driver and Vision stents, and 16 mm for the Express2 stent. Each stent was implanted into a polyurethane artery phantom model with a diameter of 3.2 mm and a length of 50 mm at the nominal pressure, and was postdilated using a 3.5 x 20 mm noncompliant balloon at 12 atm as a nominal pressure (Quantum Maverick™, Boston Scientific) to achieve complete stent apposition.
OCT imaging procedure. The OCT images were acquired using the LightLab OCT imaging system (LightLab Imaging, Inc., Westford, Massachusetts). An OCT image wire emits 1,310 nm of near-infrared light source, with an axial resolution of 10–20 µm and a lateral resolution of 20–94 µm. This image wire is 0.014 inches at the tip and 0.016 inches at the lens.
The stent-implanted polyurethane tubes were glued on the metal board and immersed in the container filled with normal saline. The image wire was placed in the center of the stent implanted in the polyurethane tube using a plastic introducer and was held in place (Figure 2). Motorized pullback OCT imaging was performed at a rate of 1.0 mm/second for a length of 30 mm. Images were acquired at 15.4 frames/second and were digitally archived. The images were saved in the OCT image system console and then saved on a compact disk for off-line analysis.
OCT data analysis. Each stent-strut thickness was measured on the three cross-sectional OCT images at 5 mm inside the stent both from the proximal and distal edges and at the middle of the stent. Representative cross-sectional OCT images of each stent are shown in Figure 3. On each cross-sectional image, the distance between the center of the stent-strut surface reflection and the adjacent vessel surface was measured as stent-strut thickness. This experiment assumes that the stent is fully apposed to the artificial artery wall. This procedure was repeated in similar fashion for all five stents. In total, the thickness of 239 stent struts of the five different stents was measured. All measured stent-strut thicknesses using the above-mentioned method were compared with the manufacturers’ provided nominal stent-strut thickness information, including the polymer thickness.
To assess the reproducibility of the measurements, each OCT image was analyzed by two independent observers. These measurements were then compared to examine the interobserver variability. To determine the intraobserver variability, the images were measured again by the first observer at least 4 weeks after the initial measurement, and those two measurements were compared.
Statistical analysis. To assess intra- and interobserver variability, linear regression and the Bland-Altman test were performed. The intra- and interobserver agreement (reproducibility) was assessed by determining the mean and the standard deviation of the between-observation and between-observer differences, respectively. Values of p < 0.05 were considered statistically significant.
Total measured stent-strut thicknesses and their comparison with the manufacturers’ provided strut thickness are shown in Figure 4 and Table 1. Total of 42 struts, 42 struts, 50 struts, 60 struts and 45 struts were measured for the Bx Velocity, Cypher, Express, Driver, and Vision stents, respectively. The mean difference between the measured and the manufacturers’ nominal strut thickness in each stent was less than the OCT resolution (10–20 µm). The strut thickness was significantly larger in Cypher stent as compared to Bx-Velocity stent.
The intra- and interobserver correlation coefficients were high (r = 0.957, p < 0.001 and r = 0.951, p < 0.001, respectively; Figures 5 and 6). Bland-Altman plots shows low intra- and interobserver variability. The intra- and interobserver measurement differences were 6 ± 7 µm and 6 ± 7 µm, respectively.
This is the first validation study to evaluate the intra- and interobserver variability and the correlation of stent-strut thickness measurement using OCT and comparing this with the manufacturers’ provided nominal stent-strut thickness in commonly used stents.
Apposition of the stent struts to the underlying vessel wall is likely very important for short- and long-term outcomes. Intravascular ultrasound (IVUS) has been used as a gold standard for the assessment of stent-strut apposition. In the drug-eluting stent era, several studies have shown conflicting results in relation to the adverse events and incomplete stent apposition.10–16 This could be potentially explained by the low resolution of IVUS and incorrect assessment of incomplete stent apposition. Because of its high resolution, OCT can provide more accurate information in identifying incomplete stent apposition.
Since OCT uses an infrared light source and it is unable to penetrate the metal struts, stent struts are visualized as linear structures with strong reflection from the endoluminal strut surface and typical dorsal shadowing behind the strut surface, as shown in Figure 1. Apposition of the stent strut to the vessel wall is indirectly evaluated by measuring the distance between the stent-strut surface reflection and the adjacent vessel surface. This measurement requires actual stent-strut thickness in defining stent apposition.3,4,6 Therefore, it is important for the evaluation of stent apposition with OCT to confirm this measurement of stent-strut thickness with high accuracy and reproducibility.
As shown in this study, the measured stent-strut thickness of each stent using OCT closely correlated with the manufacturers’ provided nominal stent-strut thickness information. Assessment of stent-strut apposition needs to be tailored according to the type of the stent used. Since a reliable cut-off value remains to be established,3,4,6 based on the measurements (upper 95% confidence interval, Table 1) and resolution of OCT (10–20 µm), we propose that a stent strut is classified as incompletely apposed when the distance between its surface reflection and the vessel wall is > 150 µm for the Bx Velocity, > 160 µm for the Cypher, > 130 µm for the Express2, > 90 µm for the Driver, and > 90 µm for the Vision stents.
The other important observation revealed in this study was the comparison of Bx Velocity stent and Cypher stents. The measured strut thickness of the Cypher stent was greater than that of its platform (Bx Velocity) stent. This may be due to the polymer-based drug coating on the stent struts. Therefore, in evaluating drug-eluting stent apposition, polymer thickness should also be taken into account.
We have also shown high reproducibility in measuring stent-strut thickness using commercially available OCT equipment. The intra- and interobserver measurement differences in stent-strut thickness measurements were 6 ± 7 µm and 6 ± 7 µm, respectively, which are relatively small differences made possible by OCT’s high-resolution capabilities.
This knowledge of low intra- and interobserver variability and accuracy of measurement with OCT would be helpful in designing larger studies and as a research tool following intracoronary stent implantation.
Study limitations. Any bench-top testing is intrinsically limited because it can never precisely reproduce in-vivo conditions. This study did not evaluate several inherent limitations of in vivo OCT analysis including the motion artifacts during the cardiac cycle and the position artifact of the OCT imaging wire due to vessel curvature. Moreover, we have not compared this with IVUS, thus the incremental benefit with the use of OCT could not be demonstrated in this study.