Contemporary invasive imaging modalities that identify and risk-stratify coronary plaques at risk of rupture

Abstract

Atherosclerotic plaque rupture is responsible for the majority of myocardial infarctions, with ruptured plaques exhibiting specific morphological features, including large lipid cores, thinner overlying fibrous caps and micro-calcifications. Contemporary imaging modalities are increasingly able to characterize plaques, potentially leading to the identification of precursor lesions that are at high risk of rupture. Observational studies using invasive imaging consistently find that plaques responsible for an acute coronary event display these high-risk morphological features, and recent prospective imaging studies have now established links between baseline plaque characteristics and future cardiovascular events. Despite these promising advances, subsequent overall event rates remain too low for clinical utility. Novel technologies are now required to refine and improve our ability to identify and risk-stratify lesions at risk of rupture, if plaque-based risk evaluation is ever to become reality.

Keywords:

• angioscopy
• coronary atherosclerosis
• intravascular ultrasound
• near-infrared spectroscopy
• optical coherence tomography
• virtual-histology intravascular ultrasound

Rupture of an atheromatous plaque may be responsible for around two-thirds of all thrombotic coronary events. Autopsy studies have consistently shown that ruptured plaques exhibit different morphological features from non-ruptured plaques, specifically a larger necrotic core, thinner overlying fibrous cap, microcalcification and accumulations of macrophages in the fibrous cap. Such studies also found coronary plaques with similar features that lacked evidence of luminal thrombosis, latterly termed thin-cap fibroatheromata (TCFA) [1]. Although similar to ruptured plaques, TCFAs differ by having smaller necrotic cores, less macrophage infiltration of the cap and less calcification. TCFAs are focal and cluster in the proximal segments of the major coronary arteries [2]. TCFAs are found frequently in patients dying from both cardiovascular and non-cardiovascular causes (40% of hearts studied at post-mortem), but despite this, TCFA only account for 1.6% of the total length of the coronary tree [2]. Although a direct relationship between TCFA and plaque rupture has hitherto not been proven, the concept of a plaque subtype ‘vulnerable’ to rupture presents an attractive target for research. However, it should be noted that the term ‘vulnerable plaque’ is not necessarily synonymous with TCFA, as other plaque subtypes are associated with luminal thrombosis, namely plaque erosions and calcified nodules. Over the past decade, there has been significant interest in diagnostic technologies able to identify TCFA in vivo. However, for reliable identification of TCFA, coronary plaque morphology must be visualized in detail; therefore, any imaging modality must have high axial and lateral resolution. At present, non-invasive imaging modalities, including cardiac computed tomography or MRI, lack the necessary spatial resolution to accurately portray coronary plaque architecture. Research has therefore focused on invasive imaging modalities where image resolution is significantly higher, potentially allowing for reliable and accurate identification of TCFA.

Intravascular ultrasound

The principle of intravascular ultrasound (IVUS) imaging is based on the oscillatory movement of a piezoelectric transducer, producing sound waves when electrically stimulated. Contemporary IVUS catheter transducers are either constructed from an electronic phased array or a single-element design, which are mechanically rotated at high speed (1800 rpm) and typically operate in the 20–40 MHz range. This results in IVUS having an axial resolution of 70–200 μm and a lateral resolution of 200–250 μm. These features have allowed grayscale IVUS to become an established intracoronary diagnostic tool that is routinely used to visualize atherosclerotic plaques. Despite widespread adoption of grayscale IVUS for optimization of percutaneous coronary intervention (PCI) results, the technology has proved inconsistent in its ability to discriminate individual plaque components. Furthermore, the visual interpretation of grayscale IVUS images remains highly subjective and does not allow real-time assessment of plaque composition [3]. In an attempt to overcome these issues, alternative methods for assessing plaque composition were developed, based on differing methods of analyzing the ultrasound backscatter radiofrequency data. The two most prominent methods are virtual histology IVUS (VH-IVUS) and integrated backscatter IVUS. As the majority of clinical data exist for VH-IVUS, this article will predominantly focus on this technology.

In an ex vivo model, Nair et al. were able to validate VH-IVUS by identifying four different plaque tissue components, namely fibrous tissue, fibrofatty tissue, calcified plaque and necrotic core with diagnostic accuracies in the range 79.7–92.8%. The accuracy of the technology was confirmed in further ex vivo analysis, using automated advanced analysis algorithms, with predictive accuracies exceeding 93.5% for all tissue components [4]. More recently, Obaid and colleagues conducted an independent validation autopsy study that examined both plaque composition and plaque type; this study showed that VH-IVUS could identify both fibroatheromata and TCFA with diagnostic accuracies of 79 and 76%, respectively [5].

Based on such encouraging validation studies, prospective studies were designed to determine whether VH-IVUS-defined plaque classification could be used to predict future adverse cardiovascular events. The VH-IVUS in Vulnerable Atherosclerosis (VIVA) study recruited 170 patients with stable angina or troponin-positive acute coronary syndrome (ACS) undergoing PCI, to three-vessel VH-IVUS [6]. Patients were followed for a median of 625 days from enrolment, the primary endpoint being non-restenotic major adverse cardiac events (MACE), a composite of death, myocardial infarction and unplanned revascularization. In total, 1096 plaques were classified with 19 lesions, resulting in MACE events. VH-TCFA was the only plaque subtype associated with MACE events (hazard ratio [HR]: 7.53 [95% CI: 1.12–50.55]; p = 0.038), along with plaque burden (PB) ≥70% (HR: 8.13 [95% CI: 1.63–40.56]; p = 0.01).

The PROSPECT study (A Prospective Natural-History Study of Coronary Atherosclerosis), reporting in the same year, reported similar results [7]. Investigators recruited 697 patients presenting with an ACS to undergo three-vessel VH-IVUS imaging. The primary endpoint of MACE was composed of cardiac death, cardiac arrest, myocardial infarction or rehospitalization due to either unstable/progressive angina. Patients were followed for a median of 3.4 years. Again VH-TCFA and PB ≥70% were associated with future MACE, but PROSPECT also reported an association with plaque minimal luminal area ≤4 mm². Combinations of these measures act to increase the risk further (VH-TCFA+PB≥70%+ minimal luminal area ≤4 mm²; HR: 11.05 [95% CI: 4.39–27.82]; p < 0.001). Finally, the ATHEROREMO-IVUS study recruited 581 patients to VH-IVUS imaging of a non-culprit artery, following a planned PCI procedure [8]. Independent plaque predictors of MACE, defined as mortality, ACS or unplanned revascularization, were VH-TCFA (HR 1.98 [95% CI: 1.09–3.60], p = 0.026) and PB≥70% (HR: 2.90 [95% CI: 1.60–5.25]; p < 0.001).

Although these studies are encouraging and confirm the biological importance of VH-TCFA, the actual event rate per ‘high-risk’ lesion was low at <10% over 3 years. Accordingly, there has been recent studies to assess whether using biomechanical modeling can improve the ability of VH-IVUS to predict clinical events. In a hypothesis generating study, Teng et al. performed finite element analysis simulations on coronary plaques from 53 patients undergoing VH-IVUS [9]. Plaque structural stress was found to be elevated in patients presenting with ACS, specifically in those plaque regions associated with future MACE (e.g., PB ≥ 70%). Calculation of plaque structural stress was also found to have an incremental value to plaque imaging alone, with the authors concluding that this novel technique may refine our ability to predict plaque rupture.

Optical coherence tomography

Another invasive imaging tool that may improve clinicians’ ability to identify rupture-prone atherosclerotic plaques is optical coherence tomography (OCT), which uses near-infrared light to characterize plaque composition. The catheter is composed of a rotating lens, connected to an external light source via a single fibreoptic wire, producing a resolution far superior to IVUS (axial 15–20 μm and lateral 20–90 μm). Unlike IVUS, the vessel must be cleared of blood for imaging as the circulating red cells reflect light, causing marked signal attenuation. Validation studies highlight the different appearances of coronary plaques using OCT, with fibrous plaque appearing as homogenous, signal-rich regions, whereas fibrocalcific plaques had signal-poor regions with sharply delineated borders [10]. Lipid-rich plaques (LRP) also had signal-poor regions, but with diffuse borders and a sudden drop-off on OCT signal. Based on these descriptions, OCT has sensitivities of around 75% for fibrous, 95% for fibrocalcific and 92% for LRP. In addition, OCT has also been assessed for its ability to identify specific plaque cellular features. Tearney et al. assessed the reliability of OCT to identify macrophages within the fibrous cap, correlating the raw signal with histological sections stained with CD68 [11]. They found a strong positive correlation between raw-OCT data and histological measures of macrophage density (r = 0.84; p < 0.001), with raw-OCT demonstrating 100% sensitivity and specificity for identifying fibrous caps containing >10% CD68 staining. However, the dynamic range of raw-OCT data is too high to be displayed on a standard computer monitor. As a result, the data is processed and compressed by taking the base 10 logarithm. Analysis of the compressed data revealed that OCT and histological correlations were moderate at best (r = 0.47; p < 0.05), with a sensitivity of 70% and specificity of 75%.

Following the introduction of OCT into clinical practice, the technology has been used in a variety of clinical settings to describe the appearance of lesions responsible for an ACS presentation. Jang et al. first assessed culprit lesion morphology in 57 patients and found that LRPs were more commonly responsible for ACS presentation [12] (61.1 vs 20.0%; p = 0.018) and that the median minimum fibrous cap thickness was lower in plaques responsible for ACS (p = 0.034). In a study of 89 consecutive patients presenting with both ST-segment elevation myocardial infarction and non-ST-segment ACS, Ino et al. found that plaque rupture, the presence of OCT-defined TCFA and intraluminal thrombus were significantly higher in the ST-segment elevation myocardial infarction group [13]. Mizukoshi et al. performed OCT imaging on 115 patients presenting with unstable angina and found that the prevalence of plaque rupture and OCT-TCFA were significantly higher with increasingly unstable clinical symptoms (p < 0.001) [14]. Spotty calcification, consisting of calcified plaque of <90° arc by OCT was also more common, with fibrous cap thickness reduced.

Increasingly, OCT is being used to visualize and describe the pattern of coronary atherosclerosis in non-culprit lesions. In a three-vessel OCT study of 55 patients (165 coronary arteries) presenting with either stable angina or an ACS, 94 OCT-TCFAs were identified, distributed across all three coronary territories [15]. Although left anterior descending artery OCT-TCFAs were located proximally (within the first 30 mm), other lesions were evenly distributed throughout the right coronary artery and left circumflex artery. These findings are similar to that of the histological series describing the spatial distribution of TCFA [2]. More recently, Tian et al. systematically assessed the prevalence of OCT-TCFA in a study of 255 patients, aiming to explore the relationship between TCFA and coronary artery stenosis [16]. Overall, 643 plaques were identified after three-vessel IVUS and OCT imaging, with 111 being classified as an OCT-TCFA. Consistent with histological studies, the absolute number of OCT-TCFA was greatest in angiographically mild lesions compared with severe lesions (58 vs 33 OCT-TCFA); however, the prevalence of OCT-TCFA was double in severe lesions (36 vs 18%; p = 0.002). Interestingly, angiographically severe (>70% diameter stenosis) OCT-TCFA had increased markers of plaque risk, including a higher prevalence of microvessels and cholesterol crystals (p < 0.001 and p = 0.002, respectively), coupled with thinner overlying fibrous caps (49.0 ± 9.2 vs 57.0 ± 6.6 μm; p < 0.001).

At present, there is only one published prospective study associating atherosclerotic OCT plaque features at baseline to subsequent clinical events. In a small study of 53 patients attending for PCI, Uemura et al. imaged 69 angiographically moderate (<50% diameter stenosis) plaques [17]. Repeat angiography was performed between 6 and 9 months to assess plaque progression; plaques with progression were found to have a significantly higher proportion of microvessels and macrophages, and were more likely to be classified as an OCT-TCFA (76.9 vs 14.3%, p < 0.01) at baseline. However, there are no small or large prospective observational studies assessing whether OCT-TCFA or other OCT-defined plaque features are associated with future MACE. It therefore remains to be seen whether OCT-defined coronary plaque assessment will provide a more powerful tool for the identification of plaques vulnerable to rupture.

Near-infrared spectroscopy

Near-infrared spectroscopy (NIRS) relies on the principle that differing organic molecules absorb and scatter near infrared light to varying degrees and wavelengths. Analysis of the received signal generates a chemogram, which provides a probability that a lipid core is present within 1 mm of the plaque surface. The ability of NIRS to detect lipid core plaques was validated in a large autopsy study of 212 coronary segments from 84 hearts, where the technology identified lipid core plaques with a receiver–operator characteristic area of 0.80 (95% CI: 0.76–0.85) [18]. Given that rupture of LRPs is the most common cause of ACS, it has been suggested that NIRS may play an important role in the identification of high-risk plaques in the future. The NIRS substudy of the European Collaborative Project on Inflammation and Vascular Wall remodeling in Atherosclerosis, where 203 patients underwent NIRS in a non-culprit segment, demonstrated that a lipid PB index ≥43.0 was associated with a fourfold increased risk of MACE at 1-year follow-up [19]. The PROSPECT II and LRP studies, which are currently recruiting, aim to utilize a hybrid NIRS–IVUS catheter to evaluate whether suspected high-risk plaques are associated with future cardiovascular events. If the technology is shown to improve upon the predictive ability of IVUS alone, subsequent studies can examine whether pre-emptive treatment of large-burden, LRPs has a favorable impact on long-term clinical outcomes.

Coronary angioscopy & intravascular thermography

Coronary angioscopy uses white light emitted through catheter-incorporated glass fibers to directly visualize the color of the arterial wall, allowing identification of thrombus, yellow/white plaques and subtle luminal irregularities. Histological studies have identified white-colored plaques as predominantly fibrous and yellow plaques as fibroatheromata or as degenerated fibrous plaques with areas of necrosis. In the 552-patient study by Ohtani et al., the presence of yellow plaques was associated with an increased risk of subsequent ACS, suggesting that coronary angioscopy may have a role in plaque-based risk stratification [20]. However, the requirement for a saline injection to clear blood from the area of interest and plaque misclassification (superficial calcium may also be associated with the presence of yellow plaques) limits the widespread use of coronary angioscopy in clinical practice. Similarly, intravascular thermography, measuring plaque temperature as a surrogate of inflammatory activity rather than plaque components has also been assessed as a tool to detect plaque vulnerability. Although initial studies appeared promising, further investigations have suggested that temperature differences may not be as great as previously suggested, limiting utility in this area. Intravascular thermography is also affected by blood flow and arterial pressure, making results challenging to interpret.

Expert commentary

Recent advances in intracoronary imaging have resulted in significant increase in our ability to visualize and describe coronary atherosclerosis. Validation studies confirm that these techniques can reliably characterize plaque composition, but each individual technology has limitations that should be recognized when interpreting the results of clinical trials. Observational studies consistently demonstrate that plaques responsible for an acute event display features consistent with the TCFA hypothesis. Encouragingly, prospective studies using VH-IVUS have now established links between specific plaque features at baseline and future cardiovascular events, but overall event rates remain far too low for clinical utility. Well-designed natural history studies, using new or combined/‘hybrid’ imaging technologies, are now required to assess whether plaque-based prognostication can ever become reality.

Future perspectives

Invasive imaging modalities and technologies to identify and risk-stratify coronary plaques at highest risk of rupture are likely to increase. Recent publications illustrate the significant efforts being made to refine and augment the current imaging modalities, through incorporation of novel biomarkers, including mechanical stress. Others are currently aiming to combine the optimal features of each individual modality, creating hybrid devices that are able to perform dual functions, such as OCT-spectroscopy. Meanwhile, novel techniques like intravascular high-resolution IVUS may allow more accurate visualization of plaque architecture, potentially allowing more accurate identification of histological TCFA. Whether these technologies will ultimately improve our ability to accurately risk-stratify plaques, allowing targeted therapy to prevent future clinical events, is still a subject of great debate.

Financial & competing interests disclosure

AJ Brown receives funded from the British Heart Foundation (FS/13/33/30168); NEJ West has acted as a consultant to Abbott Vascular and MR Bennett is funded by the British Heart Foundation. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

No writing assistance was utilized in the production of this manuscript.