Unravelling the endovascular microenvironment by optical coherence tomography

“The more I see, the less I know for sure.” Barlis et al. have reported the first multicentre experience on the use of optical coherence tomography (OCT) to evaluate the clinical outcomes of a new drug-eluting stent (DES).1 The term substudy in the title and the fact that OCT was not the primary endpoint of the main investigation should not distract the reader from realizing the historical value of this report as it indicates the early signs of a late catch-up phenomenon, not the one commonly associated with late lumen loss. It defines the somewhat late arrival of OCT, a derivative of optical low-coherence reflectometry introduced by James Fujimoto in the early 1990s,2 in cardiology clinical trials providing the corrective lens to allow a clear, close-up look at the vessel wall. In a simplified view, OCT is the optical cousin of IVUS: emitting and detecting backscattered signals. The use of light renders considerable advantages to OCT while creating some challenges—the most noticeable are the inability to see through blood or beyond 2 mm of tissue depth. The high resolution (<20 µm) of OCT enables the clinician to visualize the vascular microenvironment for the first time, a prerogative once only granted to pathologists evaluating tissues under the microscope, with the extra bonus of in vivo imaging. While this technology has been present in European cathlabs since the early days of DES, it has taken several years and a few highly publicized reports challenging DES safety to turn our attention to the value of OCT in device clinical trials.3 Technical, scientific, and economic challenges may explain this somewhat delayed renaissance—or perhaps we were too focused on keeping arteries patent rather than visualizing and understanding the biological processes involved with such an act. Barlis et al. capitalized upon OCT's capabilities to evaluate tissue coverage 9 months after deployment of a biodegradable polymer-based biolimus-eluting stent (BES) compared with a durable polymer-based sirolimus-eluting stent (SES) in a total of 46 patients (SES = 26, BES = 20). After analysing >11 000 struts, the authors found fewer uncovered struts in the BES group compared with the SES group. There are so many ways to interpret these findings that we are left with the feeling that ‘the more we see, the less we know for sure’. Nevertheless, the report of Barlis et al.1 stirs our imagination in the search for a biological explanation for thrombosis post-DES, which was not found to be a clinical problem in the study itself, and poses many difficult questions. Why would a DES strut be covered at 9 months, or why wouldn't it be? What is the meaning of uncovered DES struts? Does an uncovered strut mean delayed re-endothelialization or healing? How does one define healing in vivo? After all, DES were engineered to block cellular proliferation. That cells do not grow freely on their surface should not be surprising—unexpected were the findings that >94% of struts were covered, particularly if one understands that biological factors do not account for all uncovered struts. This metric, however, does not allow the appreciation of the high interindividual and intrasegment variability of the vascular response (well illustrated in Figure 3 of Barlis et al.), and has been described even after balloon angioplasty.4 Clusters of uncovered and malapposed struts overlying an irregular lumen surface will probably behave differently from evenly and sparsely distributed uncovered struts apposed to a smooth surface. However, larger studies with longer follow-up and the utilization of OCT molecular imaging and tissue characterization would be required to address this issue. The study findings were interpreted as ‘implicating the polymer as a potential source of failure to endothelialize stent struts long-term’. This statement is loaded with assumptions and was properly qualified as hypothesis generating by the authors. It nevertheless deserves further discussion. First, one should not forget that it is the drug, not the polymer, which blocks cellular proliferation. Although sirolimus and biolimus affect the same mammalian target of rapamycin (mTOR) pathway, they have different pharmacokinetics and potency, were applied at different amounts and release profiles, and have different vessel wall dwelling and distribution.5 These differences could also have accounted for the study findings. On the other hand ‘bad’ polymers, like the ones tested in the early 1990s,6 can actually induce more strut ‘coverage’ because of inflammatory and proliferative vessel response. Indeed, biodegradable polymers were not selected in the first-generation DES platforms because of concerns with inflammation. The LEADERS data may dispel the early concerns with the use of these polymers, but it is difficult to discriminate the relative contribution of 9 month degradation time, dosage, and the fact that it is deposited only in the abluminal surface to the outcomes observed. The term ‘healing’ after stenting has come to represent a complex array of vascular responses to DES implantation, including re-endothelialization, tissue coverage, thrombus and fibrin deposition, and strut apposition.7 Regardless of its precise definition, one could classify ‘healing’ into suboptimal, optimal, and exaggerated. The latter is classically represented by restenosis, our long-term nemesis. Optimal healing fits somewhere between suboptimal and exaggerated, and would require DES to have endothelium-enhancing strategies in addition to potent smooth muscle cell inhibition. Such a requirement essentially places all current DES platforms in the suboptimal healing category, although millions of patients worldwide have benefited from these humanly imperfect, yet highly effective devices to treat coronary artery disease. Similarly puzzling is the concept of re-endothelialization, one of the many components of healing. We have learned that disruption of the endothelium is the initiating injury leading to atherosclerosis. As such, devices are deployed onto dysfunctional or absent endothelium. We should also consider the fact that a dysfunctional endothelium may acquire a prothrombotic and proinflammatory phenotype, and, while detecting endothelial cell markers [e.g. endothelial nitric oxide synthase (e-NOS) expression] suggests their presence, it does not establish their functional status. Studies have also shown that complete recovery of endothelial function is a slow process even after balloon angioplasty,8 a therapy imposing no biological or permanent mechanical impact to the vessel wall. It should be no surprise that the endothelium, if present, remains dysfunctional long after DES or bare metal stent (BMS) deployment. One may speculate that the endothelium recovers linearly over time, and that DES downshifts the slope of the recovery timeline. However, this assumption does not match clinical observations, as rates of stent thrombosis do not seem to decrease over time. Indeed, 6–18 months appears to be the safest period after DES deployment. Superimposing the various possible scenarios of re-endothelialization and observed rates of stent thrombosis may help reconcile these two entities (Figure 2). A bi-modal profile of re-endothelialization with a theoretical late ‘de-endothelialization’ could fit our clinical observations, but this remains largely speculative. Graphic illustration of the evolution of interventional cardiology technologies as it relates to imaging resolution requirements of new clinical endpoints. The introduction of new devices (dates shown in the horizontal axis) has increased the requirements for image resolution, as we evolve from the simplistic effectiveness binary restenosis endpoints to micron-scale tissue coverage safety measures. Current imaging technologies still lack resolution to evaluate endothelialization, the ultimate measure of optimal healing. Bars depict the maximal image resolution of coronary cineangiography (left), intravascular ultrasound (middle), and optical coherence tomography (right). The arrows show the dates these imaging technologies were introduced. Illustration of non-cumulative DES thrombosis rates at individual time frames over a 3 year follow-up. Overlays show scenarios of temporal development of endothelialization after drug-eluting stent deployment. Markers represent the actual incidence of stent thrombosis as reported by Daemen et al.:10 0–30 days = 90/7041; 30 days–1 year = 26/5549; 1–2 years = 25/2852; 2–3 years = 11/989. Solid grey lines represent our current understanding of the endothelialization process after stenting [thin lines = bare metal stent (BMS); thick lines = drug-eluting stent (DES)]. Corresponding grey dashed lines illustrate hypothetical scenarios of endothelialization (BMS = 1, 2; DES = 1–4). Note that DES hypothesis 4 best fits the actual incidence of clinical thrombosis. It is hard to imagine how stent technology would have evolved had OCT been available when we implanted the first DES in humans.9 If one recalls, a plaque with large lipid content proximal to the stent edge was identified at 12 months by IVUS and dominated our mechanistic interpretation of the actual first case of DES late thrombosis 2 months later.9 While this remains a very plausible explanation, detection of ‘vulnerable’ struts by OCT in the same case would probably have changed the fate of what we now call the DES era. Our excitement with OCT imaging, however, should not be tempered by the understanding of its current limitations: we still cannot differentiate between small amounts of neointimal hyperplasia and thrombus, and do not have enough resolution to visualize the endothelium. Nevertheless, LEADERS together with two other contemporary studies, the first randomized trial using OCT as a primary endpoint, ODESSA (OCT for DES Safety, G. Guagliumi, personal communication, TCT meeting, Washington DC, 2008), and the first large randomized OCT study in patients with ST-elevation myocardial infarction, HORIZON (G. Guagliumi, personal communication, AHA meeting, New Orleans, 2008), an OCT substudy, mark the beginning of a new era of high-resolution imaging in interventional clinical trials which will probably redefine the way we develop and test our future vascular technologies. Conflict of interest: none declared.