|Home | About | Journals | Submit | Contact Us | Français|
While our understanding of vulnerable coronary plaque is still at an early stage, the concept that certain types of plaques predispose patients to developing an acute myocardial infarction continues to be at the forefront of cardiology research. Intracoronary optical coherence tomography (OCT) has been developed to both identify and study these lesions due to its distinct resolution advantage over other imaging modalities. We review clinical research conducted at the Massachusetts General Hospital over the past five years to develop, validate, and utilize this technology to improve our understanding of vulnerable plaque. Our results show that intracoronary OCT may be safely conducted in patients and that it provides abundant information regarding plaque microscopic morphology, which is essential to the identification and study of high-risk lesions. Even though many basic biological, clinical, and technological challenges must be addressed prior to widespread use of this technology, the unique capabilities of OCT ensure that it will have a prominent role in shaping the future of cardiology.
Acute myocardial infarction (AMI) is a leading cause of death in the United States and industrialized countries.1,2 Research conducted over the past 15 years has demonstrated that several types of minimally or modestly stenotic atherosclerotic plaques, termed vulnerable plaques, are precursors to coronary thrombosis, myocardial ischemia, and sudden cardiac death. Postmortem studies have identified one type of vulnerable plaque, the thin-cap fibroatheroma (TCFA), as the culprit lesion in approximately 80% of sudden cardiac deaths.3–7 Over 90% of TCFAs are found within the most proximal 5.0 cm segment of each of the main coronary arteries [left anterior descending (LAD); left circumflex (LCx); and right coronary artery (RCA)].3,5 The TCFA is typically a minimally occlusive plaque characterized histologically by the following features: (1) thin fibrous cap (<65 µm), (2) large lipid pool, and (3) activated macrophages near or within the fibrous cap.3,5,7–9 It is hypothesized that these features predispose TCFAs to rupture in response to biomechanical stresses.10,11 Following rupture and the release of procoagulant proteins, such as tissue factor, a substrate for thrombus formation is created, leading to an acute coronary event.12,13 While TCFAs are associated with the majority of AMIs, recent autopsy studies have shown that coronary plaques with erosions or superficial calcified nodules may also precipitate thrombosis and sudden occlusion of a coronary artery.3,5,14,15
Although autopsy studies have been valuable in determining features of culprit plaques, the retrospective nature of these studies limits their ability to quantify the risk of an individual plaque for causing acute coronary thrombosis. For instance, TCFAs are a frequent autopsy finding in asymptomatic or stable patients and are found with equal frequency in culprit and nonculprit arteries in acute coronary syndromes.16 Moreover, disrupted TCFAs have been found in 10% of non-cardiac deaths.16 Recent findings of multiple ruptured plaques17 and increased systemic inflammation in acute patients18–20 has challenged the notion of a single vulnerable plaque as the precursor for AMI.19,21,22 An improved understanding of the natural history and clinical significance of these lesions would accelerate progress in diagnosis, treatment, and prevention of coronary artery disease (CAD).
An attractive approach to studying the evolution of vulnerable plaques is noninvasive or intracoronary imaging of individual lesions at multiple time points. Unfortunately, the microscopic features that characterize vulnerable plaque are not reliably identified by conventional imaging technologies such as intravascular ultrasound (IVUS),23–28 CT,29–32 and MRI.32–35 While experimental intracoronary imaging modalities such as integrated backscatter IVUS,36,37 elastography,32,38 angioscopy,39–43 near-infrared spectroscopy,44 fluorescence spectroscopy,45–47 Raman spectroscopy,48,49 and thermography50,51 have been investigated for the detection of vulnerable plaque, no method to date has been shown to reliably identify all of the characteristic features of these lesions.
Intracoronary optical coherence tomography (OCT) is an invasive microscopic imaging technology that has been developed for the identification of vulnerable plaque.52–55 OCT acquires cross-sectional images of tissue reflectance and, since it may be implemented through an optical fiber probe, it is readily adaptable to coronary catheters56 for insertion into coronary arteries and circumferential imaging of arterial pathology. The first investigation of vascular optical coherence tomography ex vivo demonstrated the potential of this technique to identify arterial microstructure.57 Subsequent development of OCT technology enabled image acquisition at rates sufficient for intracoronary imaging in human patients.58–60 In this manuscript, we review studies conducted with this technology over the past five years at the Massachusetts General Hospital (MGH). Results from these studies show that a wide variety of microscopic features, including those associated with TCFAs, can be identified by OCT imaging both ex vivo and in living human patients. These findings suggest that this technology will play an important role in improving our understanding of coronary artery disease, guiding local therapy, and decreasing the mortality of AMI.
A schematic of the OCT system is shown in Fig. 1.59 Briefly, the system consists of a polarization-diverse fiber-optic non-reciprocal interferometer, which operates in the time domain. The light source is centered at 1300 nm and has a Gaussian spectral full-width-at-half-maximum of 70 nm, providing an axial resolution of approximately 8 µm in tissue. The transverse resolution, determined by the focal spot size produced by the probe, is 25 µm. Group delay scanning at a rate of 2 kHz was conducted by utilizing a phase-control rapidly scanning optical delay (RSOD) in the reference arm.61 Images (500 pixels transverse × 250 pixels axial) were obtained at 4 frames/ s and stored digitally. A custom-built fiber-optic rotary junction was utilized for catheter-based circumferential imaging and a galvanometer mirror was used for the free-space experiments.62 Catheters were constructed by modifying a commercially available 3.0 F (950 µm diameter), rapid-exchange IVUS catheter to incorporate a central fiber, a distal gradient index (GRIN) lens, and a deflecting prism, which were rotated to construct a circumferential image.62
The first steps in validating this imaging modality were to establish and test the accuracy of objective image criteria for discrimination of atherosclerotic plaque types ex vivo. A total of 357 specimens (162 aortas, 105 carotid bulbs, and 90 coronary arteries) were obtained from 90 cadavers (48 male, 42 female, mean age 74.5±13.25). The specimens were imaged fresh at 37°C, less than 72 h postmortem. Registration was accomplished by placing ink marks on the tissue prior to imaging so that both OCT images and histopathology slides contained visibly recognizable reference points. Following imaging, specimens were fixed in Formalin and processed for paraffin embedding, 5-µm sectioning, and staining (H&E, Trichrome, Movat’s Pentachrome, and immunohistochemistry). For the training set, 50 cadaver plaques were imaged by OCT and correlated with histology obtained at the imaging site.55 Fibrous plaques were characterized by homogeneous, signal-rich regions, fibrocalcific plaques by signal-poor regions with sharp borders, and lipid-rich plaques by signal-poor regions with diffuse borders (Fig. 2). Two blinded readers prospectively applied these criteria to images of the remaining 307 plaques (validation set). Using histopathologic diagnosis as the gold standard, the accuracy of OCT for characterizing plaque type was then determined. These criteria yielded a sensitivity and specificity ranging from 71–79% and 97–98%, for fibrous plaques, 95–96% and 97% for fibrocalcific plaques, and 90–94% and 90–92% for lipid-rich plaques, respectively (overall agreement, κ =0.84).55 These results demonstrated that objective OCT criteria are highly sensitive and specific for differentiating lipid-rich plaques from other plaque types.55
Macrophages are central to the etiology of coronary artery disease.12,63–65 Due to the high quantity of intracellular phagolysosomes containing lipid and other cellular debris, we hypothesized that the refractive index contrast provided by the cytoplasm of macrophages would result in a strong optical signal from these cells (Fig. 3). Furthermore, since macrophages are typically heterogeneously distributed in atherosclerotic tissue, the spatial variance of OCT signal in plaques with high macrophage content should also be elevated. In order to test this hypothesis, cap macrophage and smooth muscle densities of 27 necrotic core fibroatheromas were quantified by analyzing indirect horseradish immunoperoxidase staining of paraffin embedded tissue sections incubated with CD68 and smooth muscle actin monoclonal antibodies, respectively.66 Hematoxylin was the counter-stain. Morphometric measurements (single 10× cross-sectional field) of cell density (% area stained) within a 500-µm (lateral) × 125-µm (axial) region of interest were then compared to the normalized standard deviation (NSD) of the OCT signal intensity at corresponding locations.66 We found a high degree of positive correlation between OCT and histologic measurements of fibrous cap macrophage density (r=0.84, p<0.0001) and a negative correlation between OCT and histologic measurements of smooth muscle actin density (r=−0.56, p<0.005). A range of NSD thresholds (6.15–6.35%) yielded greater than 90% sensitivity and specificity for identifying caps containing >10% CD68 staining. This study demonstrated that the high contrast and resolution of OCT enables the quantification of macrophages within fibrous caps of atherosclerotic plaques.
While the previously described studies demonstrated accurate characterization of features associated with TCFAs, OCT is capable of identifying additional plaque components that may be associated with acute coronary events. These features are described below.
Calcific nodules have been associated with plaque thrombosis in a minority of cases.3,5 Calcium by OCT appears as a signal-poor region with a sharp delineation between the nodule and the surrounding tissue [Fig. 2(C)]. In our histopathologic study of plaque characterization, we were able to use this criterion to diagnose calcific nodules with 96% sensitivity and 97% specificity.55
Differentiation of thrombus from plaque is critical for accurate plaque characterization. In addition, there is evidence to suggest that thrombus type, platelet-rich versus red-blood-cell-rich, is an indicator of the flow conditions associated with the thrombus formation and could be an important predictor of the efficacy of thrombolytic therapy.67,68 We have conducted a study to correlate histopathologic sections of thrombus with OCT. Our preliminary results suggest that a “red” thrombus (red-blood-cell-rich) rapidly attenuates the signal in a manner similar to whole blood [Fig. 4(A)]. In contrast, a “white” thrombus (platelet-rich) appears to exhibit a homogeneous moderate-strong signal with significantly less attenuation [Fig. 4(B)].60 In our experience, both forms of thrombus are usually easily distinguished from the arterial wall proper.
High-resolution cross-sectional optical imaging affords the unique opportunity to study the location of macrophage accumulations within a given plaque cross-section. Using our macrophage data set,66 we found that 84% of OCT images demonstrated a high signal at the junction between the cap and the lipid pool [Fig. 5(A) and 5(B)] when CD68 staining at this interface was greater than 50%.
Multinucleated macrophages, or giant cells, are an inflammatory response to a foreign body (e.g., cholesterol crystals) within atherosclerotic plaques. Giant cells can be visualized by OCT as large highly reflecting regions within the cap and at the cap-lipid pool interface [Fig. 5(C) and 5(D)].
Research investigating the biomechanical properties of atherosclerotic plaques has shown that the presence of cholesterol crystals increases the stiffness of lipid pools, and as a result, may decrease the likelihood of plaque rupture.69 Images of cholesterol crystals demonstrate oriented, linear, highly reflecting structures within the plaques [Fig. 5(E) and 5(F)].70
Histopathologic validation of qualitative and quantitative image criteria ex vivo provided a foundation for interpreting data obtained from living human patients. Between January 2000 and September 2003, a total of 86 patients undergoing routine percutaneous transluminal coronary intervention (PTCI) were enrolled in a study at the Massachusetts General Hospital (Boston, MA) to investigate the feasibility of intracoronary OCT. Informed consent was obtained for all patients. Imaging was performed in culprit lesions and remote sites with IVUS and OCT pre- and post-coronary intervention.60,71 Clear visualization of the arterial wall was accomplished by use of intermittent saline flushes (8–10 cc) through the guide catheter. All patients tolerated the procedure well without complications. Images demonstrating detailed arterial microstructure were successfully obtained in all patients studied and in all three major coronary arteries.60,71 To compare IVUS and OCT, images were registered by angiographic visualization of the catheter tip, anatomical landmarks, and guide catheter tip location. Summaries of the results from our clinical study are described below:
All of the characteristics of macrophage-rich TCFAs as well as the additional plaque features described in the previous section were observed in vivo; there were no image features in the clinical data that had not been observed ex vivo (Fig. 6).60 These observations suggest that image interpretation criteria and algorithms validated ex vivo can also be applied to images obtained from patients.
Although IVUS is unable to resolve microstructural features associated with vulnerable plaque, it can identify nonatherosclerotic (normal) vessels, large thrombi, calcific deposits, and pronounced arterial disruptions. While unconfirmed, indirect evidence suggests that IVUS may detect large lipid deposits.24 A study was conducted, comparing 17 OCT-IVUS image pairs obtained from 10 patients.60 In all cases where IVUS identified these characteristics, blinded OCT observations were consistent (Table 1). OCT detected additional cases of intimal hyperplasia, thrombus, intimal disruption, and lipid pool not identified by IVUS.60
In this analysis, TCFA was defined as a plaque with lipid area ≥2 quadrants and cap thickness <65 µm. A summary of the OCT findings, obtained from 57 patients with varying clinical presentations of coronary artery disease, is depicted in Table 2.72 From these 57 patients, 20 presented with AMI, 20 with an acute coronary syndrome (ACS), and 17 with stable angina pectoris (SAP).73 TCFAs identified by OCT were found in 13 AMI patients (65%), 9 with ACS (45%), and 3 with SAP (18%).72 TCFAs were more prevalent in acute presentations (AMI and ACS) of CAD (55% versus 18%, p=0.012).72 Plaque disruption was found more frequently in acute CAD (20% versus 12%, p=0.053) and calcifications were more frequent in stable disease (41% versus 12%, p=0.049).72 These findings represent the first observation of presentation-dependent plaque morphology in living human patients and confirm our current knowledge of the relationship between morphology and patient outcome that has been obtained in previous autopsy studies.3–7
For this study, we created NSD (macrophage density) images (Fig. 7) and analyzed the macrophage distributions in both culprit and remote plaques (n=225) of 49 patients with different clinical syndromes. Macrophage density was found to be significantly higher in the AMI (5.54±1.48%) and ACS (5.86±2.01%) groups compared to the SAP group (4.14±1.81%) (p<0.003).74,75
There is great interest in understanding the role of focal versus multi-focal plaque features in the pathogenesis of acute coronary thrombosis. The unique ability of OCT to quantify macrophage content and observe the spatial distribution of plaque macrophages provides a valuable tool for investigating this critical question. Supporting the concept of focal risk, sites of plaque rupture demonstrated a greater macrophage density than nonrupture sites (6.95±0.48%, 5.71±0.37%; p=0.01).75 In addition, macrophage density was significantly higher at the surface (first 50 µm) of culprit plaques in comparison to remote lesions, indicating that superficial macrophages confer a higher risk for developing an acute thrombus.75 This finding represents a new understanding of the role of macrophages in the pathogenesis of acute coronary disease and may provide an additional parameter to assess individual plaque risk.
We also found evidence in support of the multi-focal hypothesis. Macrophage densities at remote sites were correlated with measurements at culprit sites within the same patient (r=0.66; p=0.01).75 Fibrous plaques, which are not considered to be high-risk lesions, had a higher macrophage content in patients with acute disease, compared with stable patients (p=0.025).75 Taken together, these results suggest that both focal and multi-focal macrophage distributions play important roles governing the severity of CAD.
Studies conducted with OCT clearly demonstrate the potential of this technology for impacting the management of CAD. However, limitations of current technology may preclude widespread use. Most importantly, due to its scattering and absorption, removal of blood interposed between the catheter and the arterial wall is required in order to obtain a clear, unobstructed view. Saline purging adequately removes blood from the field, but at current imaging frame rates (4–8 frames/ s), intracoronary OCT with saline flushing reduces to a single cross-sectional measurement. As a result, large area screening of vessel pathology required for widespread adoption is untenable with this paradigm. Balloon occlusion with saline purging, such as is commonly employed during angioscopy, remains a viable option for certain interventional communities.42,76 However, this technique is not favored in the United States. Strategies such as reducing the scattering of blood by administering an index matching fluid or administration of a transparent oxygen-carrying blood substitute have been proposed to reduce the blood-attenuation problem.77,78 More research needs to be conducted in these areas to determine the clinical viability of these methods. Another line of attack for overcoming the blood-attenuation problem is to increase the frame rate significantly, which can commensurately increase the information yield per purge. While conventional time-domain OCT systems have a frame rate of approximately 4–8 frames/ s, new frequency-domain OCT technology, spectral-domain OCT (SD-OCT)79–83 and optical frequency domain imaging (OFDI), enable more than an order of magnitude higher frame rates.84–86 Due to their high speeds, these second-generation OCT systems further open up the possibility of conducting three-dimensional comprehensive volumetric microscopy of long arterial segments.
One common misconception regarding OCT is that it can replace intravascular ultrasound. In fact, each technology excels at different tasks. IVUS obtains images through blood with approximately 100-µm resolution and a depth of penetration of approximately 1.0 cm. OCT on the other hand obtains images with much higher resolution, but cannot penetrate as deeply (~2 mm). As a result, OCT is particularly well suited to investigating microscopic features at the surface of the arterial wall, which are characteristic of high-risk vulnerable plaques. However, the infrared light utilized by OCT does not reach the back wall of thick atherosclerotic lesions and therefore this technology is probably not appropriate for evaluating pathology such as remodeling that is manifested by alterations of the elastic laminae.
Separate from the depth of penetration issue, conventional time-domain technology also has a limited depth range over which signals can be obtained. This limited ranging depth presents difficulties visualizing the entire arterial circumference, notably when the catheter is eccentrically placed within a large diameter vessel. Auto-ranging, which adaptively adjusts the RSOD galvanometer to follow the lumen of the arterial wall, is one viable solution to this problem.87 Additionally, second generation OFDI systems84 inherently provide up to 7.0 mm of ranging depth, which is sufficient to visualize the entire artery even at the coronary ostia.
To date, optical coherence tomography has had a tangible impact on the quest to understand and identify the vulnerable plaque. It is the only method demonstrated to be capable of measuring all of the microscopic features associated with TCFAs. Our knowledge of the morphology of plaques associated with AMI, previously predicated on retrospective autopsy studies, has now been confirmed in living human patients with this technology. New information regarding macrophage distributions in patients with severe CAD, and insight into the focal and diffuse nature of the inflammatory atherosclerotic process has been uncovered.
The promise of intracoronary OCT is great, yet important challenges still remain. Much needs to be learned about the disease in order to ascertain the eventual role of this or any other intravascular modality in clinical practice. The prevalence, incidence, and natural history of high-risk lesions are presently poorly understood. Studies are currently underway using OCT and other imaging modalities to investigate the evolution of individual coronary lesions. Tests capable of screening a large population to determine patients that require more detailed intracoronary evaluation must be developed. Local treatment strategies have been proposed, including lesion stabilization with drug-eluting stents88,89 and photodynamic therapy,90,91 but the long-term clinical viability of these approaches is currently unknown. While there is still much to be done, we anticipate that the unique capabilities of OCT as an investigational tool for high-risk lesions will serve the cardiology community well as it advances to understand, identify, and ultimately treat the vulnerable plaque.
Studies by the authors described in this manuscript were funded in part by the Center for Integration of Medicine and Innovative Technology (development of the imaging platform), Guidant Corporation, and the National Institutes of Health (grants R01-HL70039 and R01-HL76398). We gratefully acknowledge Masamichi Takano for his assistance with OCT imaging of thrombus.
Guillermo J. Tearney, Wellman Center for Photomedicine, Department of Pathology, Massachusetts General Hospital, 50 Blossom Street, BAR703, Boston, Massachusetts 02114.
Ik-Kyung Jang, Massachusetts General Hospital, Cardiology Division, 55 Fruit Street, BAR703, Boston, Massachusetts 02114.
Brett E. Bouma, Wellman Center for Photomedicine, Massachusetts General Hospital, 50 Blossom Street, BAR703, Boston, Massachusetts 02114.