Micro Optical Coherence Tomography for Coronary Imaging

Intravascular optical coherence tomography (IVOCT) that produces images with 10 μm resolution has emerged as a significant technology for evaluating coronary architectural morphology. Yet, many features that are relevant to coronary plaque pathogenesis can only be seen at the cellular level. This issue has motivated the development of a next-generation form of OCT imaging that offers higher resolution. One such technology that we review here is termed micro-OCT (μOCT) that enables the assessment of the cellular and subcellular morphology of human coronary atherosclerotic plaques. This chapter reviews recent advances and ongoing works regarding μOCT in the field of cardiology. This new technology has the potential to provide researchers and clinicians with a tool to better understand the natural history of coronary atherosclerosis, increase plaque progression prediction capabilities, and better assess the vessel healing process after revascularization therapy.

Intravascular optical coherence tomography (IVOCT) that produces images with 10 µm resolution has emerged as a significant technology for evaluating coronary architectural morphology. Yet, many features that are relevant to coronary plaque pathogenesis can only be seen at the cellular level. This issue has motivated the development of a next-generation form of OCT imaging that offers higher resolution. One such technology that we review here is termed micro-OCT (µOCT) that enables the assessment of the cellular and subcellular morphology of human coronary atherosclerotic plaques. This chapter reviews recent advances and ongoing works regarding µOCT in the field of cardiology. This new technology has the potential to provide researchers and clinicians with a tool to better understand the natural history of coronary atherosclerosis, increase plaque progression prediction capabilities, and better assess the vessel healing process after revascularization therapy.
Despite the potential importance of physiological assessment of myocardial ischemia due to significant organic coronary stenosis (20), recent studies have highlighted that the initial interventional strategy with percutaneous coronary intervention or coronary artery bypass does not necessarily result in better clinical outcomes in stable (chronic) coronary artery disease (CAD) patients when compared to optimal medical therapy (21). These results have raised questions regarding what coronary morphological features bestow high risk of a future clinical event, regardless of the severity of the luminal narrowing. To address these questions, OCT with even higher resolution could illuminate the roles of coronary microstructures heretofore unseen, such as individual coronary endothelial cells (22,23), inflammatory cells, cholesterol crystals (24), vascular smooth muscle cells (25), fibroblasts, (micro-)calcifications (26), and components of thrombi such as platelets and fibrin, all thought to play roles in natural history of coronary atherosclerosis and the clinical manifestations of high risk lesions (27).
In 2011, a new mode of OCT termed micro-OCT (µOCT) was demonstrated with a resolution of 1-2 µm (28). The initial µOCT technology was implemented using a bench-top microscope system and has shown broad utility for a variety of in vitro and ex vivo studies and applications (28)(29)(30)(31)(32). Recently, to implement µOCT clinically, a single fiber optic µOCT probe and intracoronary catheter have been created (33, 34)-the technology is now poised to be used in coronaries in vivo (35). In this article, we review the developments in µOCT technology and describe its potential clinical implications for intracoronary imaging. µOCT FOR CORONARY ENDOTHELIAL CELL VISUALIZATION Endothelial cells act as gatekeepers for the passage of lowdensity lipoprotein (LDL) and leukocytes into the intima, and thus endothelial disruption/dysfunction is considered to be an important catalyst of coronary atherogenesis (36,37). Previous ultrastructural studies have demonstrated that endothelial cells cover the intima in a "cobblestone" pattern, also known as "endothelial pavementing" on en-face SEM (22). It has been recognized that coronary plaque erosion characterized by lesions with loss of endothelial cells beneath thrombus is the second most prevalent histopathological cause of acute coronary syndrome (ACS) (27,38,39).
µOCT has been shown to visualize swine and human coronary endothelial cells ex vivo (40). The capability of µOCT to visualize endothelial cells was validated with the current gold standard scanning electron microscopy (SEM) (22). The histological validation study included a visual comparison of swine coronary endothelial pavementing seen by µOCT, volume-rendered in three-dimension (3D-µOCT), with that seen by en-face SEM (Figure 1) (40). 3D-µOCT images clearly showed the uneven endothelial surface corresponding to coronary endothelial pavementing seen on corresponding SEM. After endothelial stripping (41), the surface roughness disappeared from the 3D-µOCT image, indicating the absence of endothelial cells. Quantitative analysis was performed by calculating surface roughness on a µOCT data-set and the corresponding SEM (40), demonstrating a high degree of correlation between µOCT and the SEM gold standard (R 2 = 0.99, P < 0.01). µOCT has also been used to visualize endothelial cells in human cadaver coronaries ex vivo. Endothelial pavementing was confirmed in early coronary lesions with intimal thickening and fibrous plaque (4) while lesions with superficial nodular calcification and necrotic core (4) lacked endothelial cells. Indeed, the results of surface roughness measurements indicated that endothelial cell distributions diminished over fibroatheromatous and fibrocalcific coronary plaques as compared with intimal-thickening and fibrous lesions (40). 3D-µOCT images of drug-eluting stents (DES) implanted in coronary segments also showed variable presence or absence of the endothelial cell coverage, which standard OCT was unable to identify (Figure 2A).
Studies have suggested that fluid shear stress induces spindleshaped endothelial morphology that is aligned in the direction of flow while those exposed to low endothelial shear stress (ESS) are nonuniformly oriented (42). Furthermore, low and turbulent induce increased vascular permeability (36,42) that may increase the probability of LDL and leukocyte influx. Because, it can be performed on fresh tissue and over large areas in threedimensions, µOCT assessment of endothelial morphology's could increase our understanding of coronary regions altered by shear stress, such as bifurcations and segments at myocardial bridges (43).
Since ruptured coronary plaques account for the majority of ACS (44), an improved understanding of the role of endothelial cells in the progression of atherosclerosis and early identification of plaques at high risk are anticipated to have considerable clinical impact. A number of seminal studies have suggested that certain OCT features of thin cap fibroatheromas (TCFA), such as the thickness of fibrous caps, are critical (13,45). However, OCT cut off values for high risk cap thicknesses are still undetermined (23). Evidence suggests that in TCFA lesions, apoptotic macrophages are not efficiently cleared by efferocytosis and are therefore prone to secondary necrosis, contributing to expansion of the necrotic core and further thinning of the fibrous cap (46). It is thus conceivable that endothelial cell wall border alignment can vary at the weakest point of these caps of TCFA. In this manner, 3D-µOCT visualization and calculation of endothelial surface roughness may augment precision definition of plaque vulnerability in humans.
Since the endothelial monolayer is below the resolution of OCT, the current OCT diagnostic criteria for a coronary erosion is defined as the presence of an intact fibrous cap at the culprit site with overlying thrombus (47). This criterion is a retrospective definition, as thrombus is required to demarcate this entity. In addition, clinically insignificant plaque erosion may occur without increased thrombogenicity resulting in healed plaque (48). Thus, there is a need for a prospective definition of a site that is at high risk of erosion and subsequent thrombus formation. Owing to its capacity to directly visualize the endothelium, µOCT may bridge these gaps in our diagnostic capabilities. Data has shown that µOCT is capable of imaging white thrombus containing fibrin (the type that is common in erosion), small platelets and multiple entrapped cells (28). Whether µOCT can clearly visualize the endothelium beneath thrombus remains an open question.
Compared to conventional OCT, which is incapable of distinctly visualizing endothelial cells, µOCT could make it possible to definitively assess endothelial coverage of stent struts and this information could be potentially used to shorten antiplatelet therapy treatment durations. In the emerging era of biodegradable-polymer DES (49) and bioresorbable scaffolds (50), µOCT should be capable of evaluating standalone stent polymers or polymer-coating overlying metal stents ( Figure 2B). The use of µOCT technology to assess DES strut endothelial coverage may help resolve current questions and controversies regarding novel stent healing responses, potentially leading to a means for determining optimal antiplatelet therapy durations.

µOCT FOR THE VISUALIZATION OF INFLAMMATORY CELLS
Inflammatory cells, such as leukocytes, monocytes, and macrophages play key roles in developing coronary atherosclerotic lesions (51). Because of its exquisite resolution, µOCT is capable of typing leukocytes based on cellular and intracellular morphology (28,52). Additionally, µOCT has been shown to be quite capable of imaging pseudopods that inform on the activity of these cells (Figure 3A) (28). For example, compared to smaller cells with scant cytoplasm, consistent with lymphocytes, some of large cells seen on the surface had beanshaped nucleus inside, presumably corresponding to monocytes ( Figure 3B) (28). Macrophages are also seen clearly by µOCT as highly scattering, flocculent, round or ellipsoidal cells (28,52), and are frequently observed over and within necrotic core lesions ( Figure 3C) (28,52). Some of these features have been recently demonstrated in the nasal airways in vivo, as µOCT was shown to be able to clearly visualize granulocytes in the mucus and epithelium of patients with cystic fibrosis (32). Inflammatory cells play a pivotal role in all phases of coronary atherosclerosis. The compromised endothelial barrier permits them to invade into the tunica intima and initiate arterial wall thickening. Macrophages contribute to plaque vulnerability by producing proteolytic enzymes that digest extracellular matrix and destroy the integrity of the fibrous cap (46) and through their accumulation and death that form biomechanically unstable lipid deposits. For these and many other reasons, it is important to explore macrophage behavior in atherosclerotic lesions in vivo. Owing to its 3D imaging capabilities, µOCT makes it possible to observe such morphologic phenomena that are rarely seen in 2D cross sections. Due to the cellular resolution capabilities of µOCT, this technology could potentially also identify plaques with neutrophil extracellular trap (NETs) accumulations that induce endothelial cell apoptosis and resultant plaque erosion (53).

µOCT FOR THE VISUALIZATION OF INTIMAL CRYSTALS
In in vitro cell culture experiments, macrophages containing cholesterol crystals demonstrated higher cytoplasmic scattering in µOCT images when compared to those without cholesterol crystals (Figure 4) (52). Of note, there was a discrepancy that cholesterol crystals were detected by the gold-standard polarization microscopy but were not seen on the µOCT image (Figures 4A,B). Using polarization microscopy, the accuracy of cholesterol crystal inclusions in macrophages relied on the size of cholesterol crystals (the size ≥100 mm 2 , 52% vs. <100 mm 2 , 36%, P < 0.05). 3D-µOCT clearly visualized a macrophage cell with the  high scattering cholesterol crystal within its cytoplasm ( Figure 4C).
Cholesterol crystal protrusion toward the lumen has recently been proposed as a possible cause of thrombosis and resultant ACS (54). 3D-µOCT has shown clear delineation of multilayered cholesterol crystal sheets in human cadaver coronary arteries ( Figure 5A) (28) and their protrusions that were similar to what seen by SEM (Figure 5B).
Several recent studies have highlighted that antiinflammatory pharmacotherapeutic strategies (e.g., an interleukin-1β neutralizing human monoclonal antibody, colchicine) (55, 56) may have a high potential to eliminate residual risk of CAD. Intimal crystals have been identified as a possible therapeutic target for cardiovascular disease, due to the potential of these crystals to exacerbate inflammation through inflammasome-mediated cytokine production/activation (57,58). Identification of localized vascular inflammation as intimal crystals surrounded by inflammatory changes using µOCT would be helpful for assessing the effects of these novel therapeutic agents in patients in vivo.

CATHETER-BASED INTRAVASCULAR-µOCT
Recently, the optical imaging elements required to conduct intravascular µOCT were demonstrated (33,34) and integrated into a catheter that had a size that was suitable for human coronary imaging (35). The imaging capability of the intravascular µOCT catheter was shown in human cadaver coronary arteries ex vivo and atherosclerotic rabbit aortae in vivo (35). µOCT circumferential views displayed cellular and subcellular coronary structures that were not readily identified by the standard OCT. For instance, small or large cholesterol crystal sheets were consistently noted in human lipid-rich plaques ex vivo (Figure 6) that were sometimes difficult to interpret in corresponding convention IVOCT images. As with the ex vivo bench top studies, smooth muscle cells could be clearly visualized as low-intensity, slit-like structures within the intima. Likewise, cross-sectional µOCT showed macrophage diapedesis in human coronaries ex vivo. 3D-rendering of µOCT images exhibited that individual macrophages residing on the surface of fibroatheromatous plaques that appeared to be transmigrating  through the endothelium toward a deposit of intimal cholesterol crystals (Figures 7A,B) or with their pseudopods opposing each other (Figures 7C,D). Thrombus could be noted with cells that were consistent with leucocytes embedded in the intraluminal mass. The findings of this study (35) indicate that we are on the threshold of conducting intracoronary µOCT in vivo and await the development of clinical versions of these devices for the first-in-human studies.

LIMITATIONS OF µOCT AND TECHNOLOGICAL BARRIERS FOR CLINICAL APPLICATIONS
To achieve clinical intracoronary µOCT in the cardiac catheterization lab, several limitations and barriers still need to be resolved. First, to attain such high resolution images, µOCT is currently conducted at a shorter light wavelength (centered at 800 nm) that is lower than that of standard OCT (1,300 nm). The use of this shorter wavelength decreases the penetration depth of light in tissue, potentially further compromising its ability to assess intimal thickness and thus plaque remodeling. Second, imaging with µOCT collects ∼3 orders of magnitude more data than imaging with standard OCT, and so the image sizes of a µOCT pullback will be 1,000 times greater than those of a standard OCT pullback. This 1,000-fold increase in data puts a great strain on data acquisition sensitivity and electronics, so currently the frame rate of acquiring µOCT images is significantly slower than that of standard OCT. With today's technology, imaging the entire length of a coronary artery with isotropic 1-2 µm resolution would require many pullbacks, each needing a radiocontrast flush for blood clearance. Interpreting the immense amount of information provided by µOCT may be difficult for interventional cardiologist; it is likely that artificial intelligence will be needed to aid image analysis. 3D visualization of µOCT will also likely facilitate image understanding in real time. The development of machine learning algorithms and rapid 3D rendering are ongoing topics of investigation in the µOCT field. Future technological developments will be focused on addressing these limitations to enable practical application of µOCT in the cath lab.

POTENTIAL SCIENTIFIC AND CLINICAL IMPLICATIONS OF µOCT
Clinical applications of µOCT for coronary imaging has the potential to be extensive. In the very beginning of coronary plaque development, altered shear-stress affects the endothelial cell alignment and effectuates endothelial dysfunction (42). Although such a relationship between shear-stress and endothelial cell orientation has been known for a long while, the finding has not been demonstrated in living patients in vivo. Furthermore, visualization of ongoing inflammatory cell adhesion, plaque disruption and blood coagulation remain elusive in humans in vivo. For interventional cardiology, µOCT will provide precise information for acute thrombotic formation around stent struts (23) at a microscopic level. A longitudinal view of stent architecture by 3D-µOCT has been demonstrated in rabbit aorta in vivo (Figure 8) (35). The detailed information regarding stent malapposition (59) or stent fracture (60) could help in the early detection of a precursor of procedure-related stent thrombosis. Visualization of DES polymer cracking (61) and the tissue response to anti-proliferative agents delivered by drug-coated balloons (62) could help clinicians to predict the arterial healing process.

CONCLUSION
µOCT is a next-generation form of OCT that provides an order of magnitude increase in axial and lateral resolution. Our group has demonstrated that µOCT enables the visualization of structures relevant to coronary atherosclerosis pathogenesis and stent healing at cellular/subcellular levels. The recently developed µOCT coronary catheter brings this technology close to clinial use. Clinical studies will be conducted with intracoronary µOCT in the near future. Results will potentially change the landscape of coronary imaging and our understanding of coronary disease and its treatment.