Tissue Characterization After Drug-Eluting Stent Implantation Using Optical Coherence TomographySignificance
Objective—To validate optical coherence tomography (OCT) imaging for assessment of vascular healing in a preclinical animal model and human autopsy cases and to translate the findings to the assessment of vascular healing after drug-eluting stent implantation in clinical practice.
Approach and Results—Drug-eluting and bare metal stents were imaged 28 and 42 days after implantation in atherosclerotic rabbits using OCT and simultaneously evaluated by histology. After coregistration with histology, gray-scale signal intensity (GSI) was measured for identified mature or immature neointimal tissue. Autopsy specimens were imaged with OCT and GSI values correlated with histology. Finally, prospective OCT imaging and GSI measurements were acquired in 10 patients undergoing follow-up 6 months after stenting with drug-eluting stents. Histopathologic and OCT morphometric analysis of implanted stents showed excellent correlation. Neointimal growth and vessel healing at 28 days in the animal model best correlated with human stented arteries at 6 months. In animal and human autopsy specimens, mature neointimal tissue consistently showed higher GSI values. Receiver operating characteristic curve analysis displayed high sensitivity and specificity for detection of mature neointima in animal (96% and 79%, respectively) and human autopsy (89% and 71%, respectively) data. In patients undergoing OCT follow-up 6 months after drug-eluting stent implantation, prospective GSI analysis revealed that a minimum of 27.7% of areas above stent struts represented mature neointima.
Conclusions—Novel GSI analysis of OCT imaging data allows distinction between mature and immature neointimal tissue in animal models, autopsy specimens, and patients undergoing invasive surveillance in simple atherosclerotic lesions.
- drug-eluting stents
- optical coherence tomography
- signal intensity
- vascular healing
Stent strut coverage is an important predictor of stent thrombosis in human autopsy studies particularly in patients treated with drug-eluting stents (DES).1,2 Optical coherence tomography (OCT) is a high-resolution imaging technology that permits a precise assessment of strut coverage in preclinical and clinical settings and may have an important role in the stratification of patients at risk for stent thrombosis.3,4 However, estimation of quantitative and qualitative strut tissue coverage requires fundamental validation in preclinical models, and currently available study data are associated with several important limitations, which hinder translation to human disease states.
First, fundamental differences in the temporal course of vascular healing between animal models and humans mean that the relevance of preclinical research data are limited by the inability to find chronological correlates in human disease.5 Second, although vascular healing after stenting is dependent on underlying disease morphology and plaque composition, existing OCT histopathology correlation studies have mostly been performed in healthy animal arteries.6,7 Third, lack of strut coverage by mature neointimal tissue, including smooth muscle and endothelial cells with interspersed extracellular matrix, has been reported to be an important risk factor for late stent thrombosis.2 However, although OCT imaging of stented coronary arteries has improved our capability to distinguish covered from uncovered stent struts, not all covered struts are covered by mature neointimal tissue, and OCT tissue characterization studies are lacking to date. Indeed, the development and validation of quantitative OCT tissue characterization methods to differentiate between mature and immature tissue coverage may have important implications for clinical practice.
In the current study, we aimed to compare OCT morphometric parameters with preclinical histopathologic data at 2 different time points after stent implantation (28 and 42 days) in a rabbit iliac model of atherosclerotic disease. Furthermore, we sought to perform tissue characterization by gray-scale signal intensity (GSI) analysis to classify tissue areas overlying stent struts as either mature or immature. We then aimed to apply these findings to the assessment of human autopsy specimens and to a cohort of patients treated with DES undergoing OCT surveillance at 6 months after intervention.
Materials and Methods
Materials and Methods are available in the online-only Supplement.
Histological Data After Stent Implantation in Atherosclerotic Rabbits and Correlation to OCT
Data from a total of 14 atherosclerotic rabbits treated with 2 types of DES or control bare metal stents (BMS) were available. Neointimal area increased between 28 and 42 days and was significantly higher in BMS at 28 days. A similar trend was observed for neointimal thickness above stent struts. In DES, a high degree of stent struts remained uncovered at 28 days, whereas almost complete coverage was observed at 42 days. In contrast, BMS showed almost complete strut coverage at 28 days (Table I in the online-only Data Supplement). With regard to vascular healing, BMS consistently showed lower fibrin deposition in the neointimal tissue and higher percent luminal reendothelialization (Table II in the online-only Data Supplement).
In terms of pairwise correlations between OCT and histology-based morphometric parameters in atherosclerotic animals (Figure 1), there was a significant correlation (P<0.0001) throughout the different parameters assessed, although numeric values were consistently lower for histology (Table 1). The greatest absolute partial correlation coefficients were observed for neointimal area, lumen, and stent area, with somewhat lower correlation for the percentage of uncovered stent struts.
Baseline Characteristics and Procedural Data of Treated Patients
Ten patients were treated with either everolimus-eluting (n=5) or zotarolimus-eluting (n=5) stents. Mean age was 65 years, 80% were men, and 40% had diagnosed diabetes mellitus. Baseline angiographic characteristics of patients and lesions are shown in Table III in the online-only Data Supplement.
Comparison of OCT Data Between Animal and Humans
Lesion level–based percent volume obstruction showed comparable values for atherosclerotic rabbits at 28 days and humans (Table 2). Frame level–based analysis of neointimal area was 0.48 (0.26-0.71) and 0.98 (0.69-1.28) mm2 in rabbits 28 and 42 days after stent implantation, respectively, whereas it was 0.60 (0.42-0.77) mm2 in humans at 6-month follow-up (P=ns for humans versus 28-day rabbits; Table 3). Strut level analysis revealed 42.1% uncovered struts at 28 days versus 7.7% uncovered stent struts at 42 days in animals, whereas the percentage of uncovered stent struts was 28.3% in humans at 6-month follow-up (P=ns for humans versus 28-day rabbits; Table 4; Figure 2).
Validation of Tissue Characterization in the Preclinical Setting
Histologically characterized regions of mature tissue showed significantly higher smooth muscle cell and, in turn, lower proteoglycan/collagen, fibrin, and inflammatory cell content as compared with immature tissue regions (Table 5; Figure 3). By immunohistochemistry, mature neointimal tissue was predominantly composed of α-actin–positive smooth muscle cells with minimal presence of RAM-11–positive macrophages. In contrast, immature neointimal tissue was predominantly composed of RAM-11–positive macrophages with minimal interspersed smooth muscle cells (Figure 3). There was a significant statistical interaction between the presence of endothelialization above stent struts and mature tissue defined by histology (P=0.0005), whereas there was no statistically significant interaction for the presence of fibrin. Most regions of mature tissue were derived from vessels stented with BMS (118 of 136 mature areas).
GSI analysis of mature as compared with immature neointimal tissue above struts showed higher GSI values (115.6±14.6 versus 75.1±20.2 U) when retrieving data from converted OCT files normalized on the brightest strut. In line with these findings, GSI analysis of raw data obtained from the C7-XR imaging system showed consistent differences between mature and immature neointima (430.7±107.6 versus 205.2±87.8 U). Receiver operating characteristic curve analysis based on OCT image files revealed high sensitivity and specificity for detecting mature neointima (96% and 79%, respectively) at a GSI cutoff value of 91.6. GSI analysis of the raw data was also able to discriminate between tissue types with receiver operating characteristic curve analysis showing a cutoff value at 347.8 at a sensitivity of 84% and specificity of 77%. The area under the curve for converted (frame) data and raw data were 0.95 and 0.94, respectively (Figure 4). Mean pixel brightness of the struts was found to be higher than that of the guide wire. Therefore, calibrating on the guide wire resulted in higher GSI units with receiver operating characteristic curve analysis showing a cutoff at 109.7 at a sensitivity of 87% and specificity of 82% for converted (frame) data. In sensitivity analysis, similar area under the curve was found in frames with eccentric (area under the curve=0.97) image catheter position.
Tissue Characterization at Autopsy
In receiver operating characteristic curve analysis, the corresponding cutoff value for human stented arteries at autopsy was 101.6. Sensitivity and specificity were 89% and 71%, respectively (Figure 4). Of 129 measured areas, 79 (61%) regions were identified as mature and 50 (39%) as immature. By histopathologic assessment, mature tissue revealed significantly higher smooth muscle cell content and lower proteoglycan/collagen, fibrin, and inflammatory cell content (Table 5). Both tissue types had a similar distribution of underlying plaque (Table IV in the online-only Data Supplement).
Prospective Assessment of Tissue Characterization in Patients
GSI of human OCT frames 6 months after DES implantation revealed that 27.7% of the assessed tissue areas above stent struts represented mature neointima when a cutoff value of 91.6 after strut brightness normalization was used. Using a cutoff of 109.7 after normalization to the guide wire level resulted in a slight decrease of mature neointimal tissue areas with 22.9% showing GSI values for mature tissue.
The current study aimed to characterize vascular healing after DES implantation in atherosclerotic rabbits, human autopsy specimens, and patients. As a first step, OCT-acquired morphometric data were validated against histopathology at 2 different time points in an atherosclerotic animal model. In a second step, morphometric OCT data were compared between animals and humans to find corresponding time points of vascular healing after implantation of DES. Next, we attempted to validate a novel surrogate parameter of vascular healing after stent implantation to enable differentiation of mature and immature neointimal tissue using GSI analysis of OCT-acquired data. Finally, after verification and adjustment of this surrogate parameter in human autopsy samples, vascular healing was assessed during invasive imaging surveillance at 6-month follow-up in patients with DES.
In brief, the salient findings of the current study were as follows: (1) histopathologic morphometric parameters and OCT imaging analysis show excellent correlation in an atherosclerotic rabbit model of DES implantation. Although most of the morphometric area measurements showed slightly larger dimensions in OCT compared with histopathology, overall correlation was very high; (2) neointimal growth and vessel healing at 28 days after stent implantation in the animal model are close to what is seen in human stented arteries after 6 months, considering the expected variability of the underlying histopathologic plaque composition observed in patients. Against this, most parameters of healing assessed were further advanced in atherosclerotic rabbits at the time point of 42 days; (3) there is robust evidence that OCT-derived GSI analysis is capable of providing information on tissue characterization after stent implantation, which can reliably distinguish mature from immature neointimal tissue in atherosclerotic rabbits and human autopsy specimens and could be translated to routine assessment of vascular healing in clinical practice.
Validation of OCT Imaging Analysis With Preclinical Animal Models
One of the principal findings in our study was the fact that there is an excellent overall correlation between OCT-acquired morphometric data and their corresponding parameters assessed by histology. Despite the fact that the absolute numeric values were consistently lower for histology, the correlations remained statistically significant among the different parameters studied. Systematic differences in area measurements between OCT and histology have been reported previously and may be explained by tissue shrinkage during histopathologic processing.8,9 Nevertheless, correlations based on a stent strut level analysis of tissue coverage between OCT and histology were very high in our study.
A recent human autopsy study confirmed the excellent overall reliability of OCT in assessing both quality and quantity of tissue growth after stent implantation.10 However, in terms of detecting stent strut coverage with viable tissue, OCT-based analysis resulted in several false-positive counts owing to the limited ability of OCT to differentiate between neointimal tissue, fibrin, and inflammatory cell deposition. At the same time, the investigators found that OCT was unable to distinguish thin cellular layers above stent struts. Indeed, we also showed that the overall number of uncovered stent struts was higher with OCT-based analysis compared with histopathology, which was explained by the inability of OCT to identify small monolayers of nascent neointimal tissue. This clearly demonstrates the limitations of current technology in the range of cellular dimensions.
Temporal Differences in Vascular Healing Between Animal Models and Humans
It has been reported in the past that neointimal response after stent implantation is delayed, and the time course of healing is 5 to 6× longer in humans as compared with porcine and rabbit models.5,11,12 It has also been shown that vascular healing is accelerated in healthy porcine coronary arteries compared with healthy and atherosclerotic iliac arteries of the rabbit at similar time points.5 Moreover, in addition to differences in the life span of the species, there are 2 important factors that result in differences in the process of vascular healing after stenting in animals and humans: the underlying atherosclerotic process and technical differences in stent implantation. Advanced coronary atherosclerotic plaques cannot be reliably replicated in contemporary animal models. Indeed, important components of human coronary atherosclerotic lesions, such as calcification and presence of a necrotic core, have not been replicated to date in animal models. For this reason we chose as comparator relatively noncomplex atherosclerotic lesions (AHA type A/B1) in the clinical arm of our study. In addition, coronary stenting in humans is associated with extensive local trauma characterized by plaque splitting and medial disruption, and high-pressure balloon dilatations have been associated with increased neointimal growth.13 Pre- and high-pressure post dilatation are not routinely performed in animal models.
Impact of Tissue Characterization Using OCT Imaging Analysis
Despite its high resolution, one of the unresolved issues in OCT imaging is neointimal tissue characterization. Previous studies focused on the presence of fibrin as a predictor of delayed healing, Templin et al6 compared optical density measurements for stents which were covered with fibrin using morphological information gathered by scanning electron and light microscopy. They found a significant difference in optical density measurements pertaining to peristrut areas among struts covered with fibrin versus neointima.6 However, it should be acknowledged that because of limitations of scanning electron microscopy imaging, fibrin may not be readily distinguishable from other extracellular matrix proteins. To the best of our knowledge, in the current report, we have shown for the first time that smooth muscle cell-rich (mature) tissue can be reliably characterized using GSI measurements acquired by OCT imaging analysis. As the current study was aimed at proof-of-concept for the capability of OCT to distinguish between mature and immature neointimal tissue, further studies applying GSI analysis should be undertaken to validate this approach and to assess its relevance in clinical practice.
GSI Analysis Limitations and Future Perspectives
Because neointimal tissue immaturity has been shown to be an important predictor of late stent thrombosis in human autopsies,1 the ability to differentiate between mature and immature neointimal tissue in our study might have important implications for characterization of vascular healing after stent implantation in humans. As the accuracy of vascular imaging using OCT is limited by its resolution, we aimed to establish OCT-based tissue characterization on a histologically and clinically relevant level. In the current study, ≈approximately 25% of tissue areas examined at 6-month follow-up in patients (depending on normalization methodology) represented mature neointima. These findings underscore that the presence of tissue coverage of stent struts is no indication of completeness of vascular healing. Completeness of vascular healing after DES implantation may be regarded as the presence of mature tissue coverage in the absence of fibrin deposition with complete and physiological endothelialization. An important limitation of OCT-based imaging analysis is that precise cellular composition of the tissue overlying stent struts cannot be distinguished because of spatial resolution constraints. Nevertheless, we believe that the characterization of tissue maturation after DES implantation may be an important step forward in the assessment of vascular healing. However, one of the important limitations in the current investigation of immature tissue areas after DES implantation is the lack of clinical studies examining the clinical impact of this novel surrogate end point at long-term follow-up. As OCT-acquired imaging analysis is now readily available in modern catheterization laboratories, dedicated clinical studies should be undertaken to delineate the causal relationship between immature tissue areas assessed by OCT and adverse clinical outcomes in humans.
Preclinical data were derived from an atherosclerotic animal model. Despite its suitability for the assessment of vascular healing after stent implantation, this animal model remains hampered by the inability to reflect important patient comorbidities and other dependent parameters, such as the presence of diabetes mellitus and the influence of medications that may affect stent healing. As no OCT imaging was performed at the index procedure in patients, the exact morphology of atherosclerotic lesions at the time of stent implantation was not determined. As the nature of the underlying lesion has considerable impact on vascular healing after stent implantation, subsequent studies are warranted to address this question. Furthermore, the extent of positive vessel remodeling could not be assessed in the current study as a result of single-staged OCT analysis. Similarly, we did not attempt to correlate OCT findings in humans at 2 different time points, which may have simulated our animal model more closely.
This study aimed to establish evidence that OCT is capable of differentiating mature from immature neointima. In this regard, we proposed a simple GSI image analysis methodology. After quantification of mean image intensity values, it was possible to classify neointimal tissue as mature or immature with high sensitivity and specificity. However, it is important to recognize that the proposed methodology has some inherent limitations. The gray-scale analysis was performed on log-compressed OCT data, and gray-scale value normalization was performed by calibration based on values of either the brightest strut or the image wire. As a result the reported cutoff values may be applicable only for the datasets used in the current study. However, comparable results were seen when the analysis was undertaken on raw data. Moreover, image intensity variations may not always be related to differences in tissue composition but also to other effects, such as blood clearance, flow dynamics, or, most importantly, to the use of different OCT systems. Complex image analysis methodologies, including optical properties of tissue quantification and textural image analysis, might theoretically overcome these limitations. Accordingly, the evidence created in this study may serve as a background for the development of more sophisticated OCT tissue characterization software. Furthermore, for logistical reasons different OCT systems were used for imaging of preclinical and autopsy samples. Nevertheless, both systems apply similar technical features, and images are of similar resolution.
The current study showed that data derived from histology and OCT imaging show excellent correlation in an atherosclerotic rabbit model of DES implantation. Moreover, neointimal growth and vessel healing at 28 days after stent implantation in the animal model is close to what is seen in human stented arteries after 6 months. Importantly, OCT-derived GSI analysis is capable of providing information on tissue characterization after stent implantation, which can reliably distinguish mature from immature neointimal tissue in atherosclerotic rabbits, human autopsy specimens, and routine clinical practice. These findings provide a basis for the conduct of further studies using OCT imaging and GSI analysis to assess extent and maturity of neointimal coverage after DES implantation in patients.
We thank Lila Adams, CVPath Institute Inc, for performing the immunohistochemistry.
Sources of Funding
Project funding was provided by the European Union under the Seventh Framework Program (FP7): PRESTIGE, grant agreement no.: 260309.
The online-only Data Supplement is available with this article at http://atvb.ahajournals.org/lookup/suppl/doi:10.1161/ATVBAHA.113.301227/-/DC1.
- Received September 29, 2012.
- Accepted March 13, 2013.
- © 2013 American Heart Association, Inc.
- Finn AV,
- Joner M,
- Nakazawa G,
- Kolodgie F,
- Newell J,
- John MC,
- Gold HK,
- Virmani R
- Lowe HC,
- Narula J,
- Fujimoto JG,
- Jang IK
- Cassese S,
- Tada T,
- Kastrati A
- Virmani R,
- Kolodgie FD,
- Farb A,
- Lafont A
- Templin C,
- Meyer M,
- Müller MF,
- et al
- Suzuki Y,
- Ikeno F,
- Koizumi T,
- Tio F,
- Yeung AC,
- Yock PG,
- Fitzgerald PJ,
- Fearon WF
- Murata A,
- Wallace-Bradley D,
- Tellez A,
- Alviar C,
- Aboodi M,
- Sheehy A,
- Coleman L,
- Perkins L,
- Nakazawa G,
- Mintz G,
- Kaluza GL,
- Virmani R,
- Granada JF
- Prati F,
- Zimarino M,
- Stabile E,
- Pizzicannella G,
- Fouad T,
- Rabozzi R,
- Filippini A,
- Pizzicannella J,
- Cera M,
- De Caterina R
- Nakano M,
- Vorpahl M,
- Otsuka F,
- Taniwaki M,
- Yazdani SK,
- Finn AV,
- Ladich ER,
- Kolodgie FD,
- Virmani R
- Farb A,
- Weber DK,
- Kolodgie FD,
- Burke AP,
- Virmani R
- Farb A,
- Sangiorgi G,
- Carter AJ,
- Walley VM,
- Edwards WD,
- Schwartz RS,
- Virmani R
Optical coherence tomography (OCT) is a novel intravascular imaging technology that permits reliable assessment of tissue coverage after coronary stenting but fails to distinguish between tissue types indicating differential degrees of vascular healing. We were able to demonstrate that gray-scale intensity analysis of converted OCT frames facilitates detection of mature versus immature neointimal tissue, thus enabling tissue characterization after stent implantation. Validation of this novel technology was performed in atherosclerotic rabbits, human autopsy specimens, and patients undergoing invasive surveillance after coronary stenting. The results of the study provide basis for new algorithms to differentiate mature from immature neointimal tissue in patients undergoing OCT analysis after stent implantation. This may have important clinical impact for the identification of patients at risk for stent thrombosis and offers a new perspective for future clinical studies assessing long-term safety outcomes after stenting.