Introduction

Approximately one-third of patients admitted with cerebral infarction have significant atherosclerotic stenosis of the ipsilateral carotid artery [1,2,3]. Although carotid plaque with definite stenosis is subject to carotid endarterectomy (CEA) and carotid artery stenting, carotid plaque composition has been proposed as an important risk factor for stroke [4]. It has been reported that thromboembolic phenomena are associated with thinning and subsequent rupture of the fibrous cap on the surface of atherosclerotic plaques, resulting in release of necrotic debris from the plaque substance [5]. Furthermore, there is increasing evidence that the presence of intraplaque hemorrhage (IPH) in patients with carotid artery stenosis is associated with the development of future ischemic cerebrovascular events [6, 7]. Therefore, assessment and understanding of the carotid plaque characteristics are regarded to be essential before starting treatment for patients with carotid artery stenosis.

Although magnetic resonance imaging (MRI) is one of the most widespread and useful diagnostic tools for evaluating carotid plaque component at the pre-procedural assessment, it is not able to directly visualize the morphology of the arterial wall. Intravascular angioscopy, a catheter-based technique that can directly visualize the intraluminal morphology and color grade, is a useful tool for understanding plaque characteristics through the plaque color and surface thrombus formation. Plaque color is normally classified as white or yellow according to the surface color of the plaque as seen during angioscopy. This angioscopic color grade assessment is widely performed to evaluate plaque morphology and to predict the clinical prognosis [8,9,10]. Furthermore, previous studies demonstrated that assessment of coronary tissue characteristics by angioscopy before the procedure could minimize embolic complications during percutaneous coronary intervention [11, 12]. In the present study, we evaluated whether angioscopy is capable of identifying various plaque morphologies, including necrotic core and IPH in the carotid artery by ex vivo histological examinations.

Methods

In total, 9 patients underwent CEA for carotid artery stenosis, and these endarterectomy specimens were analyzed. CEA was indicated for > 70% stenosis in symptomatic, and > 50% stenosis in symptomatic cases, according to the North American Symptomatic Carotid Endarterectomy Trial criteria [13]. In 4 patients (44%), lesions were asymptomatic and in 5 (56%) they were symptomatic. A symptomatic lesion was defined as a carotid artery stenosis that had caused ipsilateral ischemic events, including cerebral infarction, transient ischemic attack, and retinal ischemia, within the previous 6 months. The study protocol was approved by the Institutional Review Board of Hyogo College of Medicine, and the study was performed in accordance with regulatory standards and ethical guidelines for clinical studies recommended by the Declaration of Helsinki. All patients provided written informed consent.

CEA procedure

CEA was performed under general anesthesia. After exposing the carotid artery, an intraluminal shunt was inserted in all patients during removing the atherosclerotic plaques. The plaque was extracted in an en bloq fashion. The specimens were stored in 0.9% saline at a temperature of 37 °C for ex vivo imaging by angioscopy, which was performed within 6 h.

Angioscopic examination

Angioscopic images were obtained using a 2.25F angioscopic catheter (Visible; FiberTech, Tokyo, Japan). Before the insertion of the angioscopy catheter, the white balance was adjusted for color correction. The light power was adjusted to avoid reflection and to obtain images with adequate brightness for determination of the plaque color. The angioscopy catheter was carefully inserted from the proximal edge of the common carotid artery (CCA) and advanced to the distal edge of the internal carotid artery (ICA). Subsequently, angioscopy was manually pulled back from the distal edge of the ICA to the proximal edge of the CCA through bifurcation under continuous flush of normal saline. During the angioscopic observation, when only a focal lumen could be observed with a viewing field, we adjusted the direction bias of the angioscopy manually so that the whole circumference of the lumen could be evaluated. Furthermore, when sufficient evaluation could not be conducted with a single pull back, we attempted repeat observation to obtain an adequate angioscopic image of the arterial wall. The acquired images were stored on the system hard drive and processed by offline analysis.

In each section of the ICA, bifurcation, and CCA, which enabled angioscopic observation of the whole circumference, the dominant color tone was evaluated based on the following definitions. White plaque was defined as visually homogeneous white area without color boundaries on the luminal surface (Fig. 1A). Yellow plaque was defined simply as the yellow area on the luminal surface, which have smooth or irregular surface with or without protrusion into the lumen. Red plaques contained red area on the luminal surface of the plaque (Fig. 1B, C). All findings were evaluated by two angioscopic specialists who were blinded to the patient characteristics.

Fig. 1
figure 1

Representative angioscopic images of a carotid artery. A white plaque, B yellow plaque, C red plaque (arrows)

Full size image

Histological preparation and analysis

The carotid specimens were fixed in 10% neutral-buffered formalin immediately after angioscopic examination. Forty-eight hours after fixation, ring-like arterial specimens obtained at the same level as the imaging study were decalcified for 5 h, embedded in paraffin, and cut every 3 to 4 mm into 4-µm transverse sections perpendicular to the longitudinal axis of the artery. All histological sections were stained with hematoxylin–eosin, elastica van Gieson, and Masson’s trichrome. The immunohistochemistry for glycophorin A stain was also performed (1:200 dilution) (Dako, Tokyo, Japan) to identify the IPH. Histological assessment of each cross-section was performed by a single specialist who was blinded to the imaging results. Quantitative analysis was performed on the histological section using measurement software (ImageJ version 1.44; National Institutes of Health, MD, USA). Measurement parameters were minimum lumen diameter, lumen area, and vessel area. Plaque area was defined as the difference between the vessel area and lumen area. The necrotic core was defined as a hypocellular plaque cavity devoid of collagen and containing necrotic debris and cholesterol clefts. The necrotic core area was measured using measurement software. IPH is defined as a large accumulation of erythrocytes lying free in the tissue [14]. The IPH area was measured as the glycophorin A–positive area. As an index of the degree of histological characteristics, the percentage occupied in plaque area was calculated for the necrotic core and IPH. The thickness of the fibrous cap was measured on Masson’s trichrome-stained sections at the site of the thinnest portion.

Statistical analysis

Categorical variables were presented as numbers and percentages. All continuous variables were presented as mean ± standard deviation values. The one-way analysis of variance (ANOVA) was used for comparing continuous variables. Additionally, post-hoc analysis was performed to evaluate variables that were found to be significant by ANOVA. For categorical data, all tests performed were analyzed using chi-square tests or fisher's exact test. Receiver-operating curves (ROCs) were analyzed to assess the best cutoff values of the fibrous cap thickness to determine the presence of red plaques with a maximal accuracy. Difference with P < 0.05 was considered significant. All statistical analysis was performed using JMP software version 13.1.0 (SAS Institute Inc., Cary, North Carolina, USA).

Results

In the present study, 9 patients with carotid artery stenosis participated. The patient characteristics are shown in Table 1. All patients received at least one antiplatelet agent at admission. In total, 27 histological sections were matched to angioscopic images from 9 endarterectomy specimens. Of those, two sections were excluded from analysis due to inadequate histological preparation, and finally 25 histological sections were matched and compared with corresponding angioscopic images.

Table 1 Baseline characteristics of patients
Full size table

Histological findings

Table 2 shows results of quantitative histological analysis in the CCA, bifurcation, and ICA. In the present study, plaque rupture was not detected in any histological sections. Vessel and plaque area tended to be larger in the CCA and bifurcation than in the ICA. The absolute necrotic core and IPH area, and fibrous cap thickness were similar among the 3 locations. Similarly, the percent area occupied by the necrotic core and IPH were similar among the 3 locations.

Table 2 Histological analysis among each carotid segment
Full size table

Angioscopic findings

A total of 7 plaques were graded as white plaques, 10 as yellow, and 8 as red by angioscopy. The frequency distribution of white, yellow, and red plaques in each location is shown in Fig. 2. The prevalence of red plaques tended to be lower in the CCA than in the bifurcation and the ICA (22%, 38% and 38%), while yellow plaques were less frequently observed in the ICA than in the bifurcation and the CCA (25%, 38%, and 56%).

Fig. 2
figure 2

The frequency distribution of white, yellow, and red plaques in each location (ICA, bifurcation, CCA). There is no statistically significant difference in the distribution of each plaque color among the 3 locations. However, the prevalence of red plaque tended to be lower in the CCA, and the prevalence of yellow plaques tended to be lower in the ICA. CCA common carotid artery, ICA internal carotid artery

Full size image

Comparison of angioscopic findings and histological parameters

Although vessel and plaque area were similar among the 3 locations, lumen area was smallest in red plaques, followed by yellow and white plaques (8 ± 7 mm2, 12 ± 7 mm2, and 20 ± 14 mm2, respectively, P = 0.060). Both the IPH area and the percent area occupied by IPH were significantly larger in red and yellow plaques than in white plaques (10.7 ± 9.3 mm2, 9.4 ± 7.8 mm2, and 2.2 ± 1.7 mm2, respectively, P = 0.074, Fig. 3A; and 25 ± 10%, 19 ± 13%, and 7 ± 5%, respectively, P = 0.008, Fig. 3B). Similarly, the necrotic core area and the percent area occupied by the necrotic core were largest in red plaques, followed by yellow and white plaques (28.0 ± 18.2 mm2, 22.2 ± 13.6 mm2, and 8.5 ± 7.1 mm2, respectively, P = 0.038, Fig. 3C; and 68 ± 15%, 50 ± 26%, and 27 ± 19%, respectively, P = 0.005, Fig. 3D). Figure 3E shows the relationship between angioscopic findings and the histological measurement of the fibrous cap thickness. The thickness of the fibrous cap was significantly thinner in red plaques than in yellow and white plaques (128 ± 34 µm, 328 ± 136 µm, and 285 ± 102 µm, respectively, P = 0.002). The ROC analysis for predicting a presence of red plaque is shown in Fig. 4. The optimal cutoff value of the fibrous cap thickness for predicting a presence of red plaques was 181 mm, with an area under the curve of 0.987 (100% sensitivity, 90% specificity). Figure 5 shows representative images of angioscopic plaque characterization and their corresponding histology.

Fig. 3
figure 3

Comparison of angioscopic findings and histological parameters. A The absolute IPH area was larger in both red and yellow plaques than in white plaques (P = 0.03). B The percent area occupied by IPH was also larger in red and yellow plaques than in white plaques (P < 0.01). C The absolute necrotic core area was larger in red plaques than in yellow and white plaques (P = 0.01). D The percent area occupied by a necrotic core was largest in red plaques, followed by yellow and white plaques (P < 0.01). E The thickness of the fibrous cap was significantly thinner in red plaques than in yellow and white plaques (P < 0.01). An asterisk indicates a significant difference between the plaques. IPH intraplaque hemorrhage

Full size image
Fig. 4
figure 4

The receiver operating curve analysis for predicting a presence of red plaque. The optimal cutoff value of fibrous cap thickness was 181 µm (area under the curve = 0.987, 100% sensitivity, 90% specificity)

Full size image
Fig. 5
figure 5

Representative images on carotid angioscopy with the corresponding histology. Representative angioscopic image of red plaque (A), yellow plaque (B), and white plaque (C). Corresponding histology of the red plaque has a large necrotic core (a, Hematoxylin-eosin stain, upper image, scale bar = 1 mm; bottom image, scale bar = 200 µm) and a large IPH (b, glycophorin A stain; upper image, scale bar = 1 mm; bottom image, scale bar = 200 µm). The necrotic core area and percent necrotic core area measured 18.4 mm2 and 82%, respectively. IPH area and percent IPH area were 7.1 mm2 and 39%, respectively. The fibrous cap thickness calculated 156 µm (c, Masson’s trichrome stein; upper image, scale bar = 1 mm; bottom image, scale bar = 100 µm). The corresponding histological image of yellow plaque shows a large necrotic core (d, Hematoxylin-eosin stain, upper image, scale bar = 1 mm; bottom image, scale bar = 100 µm) and IPH (e, glycophorin A stain; upper image, scale bar = 1 mm; bottom image, scale bar = 100 µm). The necrotic core area and percent necrotic core area calculated 29.5 mm2 and 68% respectively. IPH area and percent IPH area were 16.6 mm2 and 50%, respectively. The fibrous cap thickness measured 570 µm (f, Masson’s trichrome stain; upper image, scale bar = 1 mm; bottom image, scale bar = 500 µm). The corresponding histological image of white plaque showed a small necrotic core (g, Hematoxylin-eosin stain, upper image, scale bar = 1 mm; bottom image, scale bar = 100 µm) and IPH (h, glycophorin A stain; upper image, scale bar = 1 mm; bottom image, scale bar = 100 µm). The necrotic core area and percent necrotic core area calculated as 16.0 mm2 and 42%, respectively. IPH area and percent IPH area measured 5.4 mm2 and 29%, respectively. The fibrous cap thickness was 368 µm (i, Masson’s trichrome stain; upper image, scale bar = 1 mm; bottom image, scale bar = 200 µm). IPH intra plaque hemorrhage, HE Hematoxylin-eosin, MTc Masson’s trichrome

Full size image

Discussion

The main findings of the present study were as follows: (1) plaques in the carotid artery could be classified as either white, yellow, or red according to the surface color of plaque on angioscopy, (2) IPH area and the necrotic core area was the largest in red plaques, and (3) the thickness of the fibrous cap was thinnest in red plaques. At this time, we believe that this is the first study to demonstrate the capability of angioscopy to identify plaque containing IPH/necrotic core with thin fibrous cap in the carotid artery by ex vivo histological examinations.

In recent years, the operation volume of stent implantation had exceeded CEA due to the improvement of surgical technique. Carotid artery stenting is one of the first choices for the treatment of carotid stenosis together with CEA. However, there are still some concerns regarding the potential higher risk of embolic complication by carotid artery stenting [15]. Therefore, the identification of high-risk plaque for distal embolization and subsequent neurological complications is a major challenge in the field of carotid artery disease. MRI has emerged as a non-invasive imaging modality that enables accurate identification of the components of the atherosclerotic plaque, evaluated by histology [16]. However, an MRI may still be limited in precise atherosclerotic plaque imaging due to its lower resolution. In addition, it cannot be performed in the catheterization laboratory during diagnostic carotid angiography. Intravascular ultrasound is a catheter-based imaging technique that provides invasive, real time, high-resolution imaging of the carotid plaque. Although an intravascular ultrasound allows for precise and highly reproducible measurements of artery dimensions, it does not have the capability of identifying vulnerable plaques due to its limited axial resolution [17]. Intravascular angioscopy is also a catheter-based imaging modality, which allows for visualization of the morphology of the luminal surface of atherosclerotic plaques, and is widely used. Previous clinical studies reported that angioscopy can identify high risk plaques that may be prone to embolization during stent implantation in the coronary artery [8,9,10]. The current study showed that carotid plaques could be classified as either white, yellow, or red.

There are two types of angioscopy system available in the current era, non-obstructive angioscopy and occlusion-type angioscopy. Occlusion-type angioscopy needs vessel occlusion by balloon catheter for the removal of blood to clarify the visual field. This is a critical limitation for the clinical use in the coronary artery, because this may cause myocardial ischemia. Non-obstructive angioscopy obtains a visual field by injecting low-molecular-weight dextran into the space between the probing catheter and the fiber. Therefore, non-obstructive angioscopy is relatively safe, because adequate blood flow is maintained throughout the process of image acquisition. Initially this method was used in the coronary artery, and it has since been applied to other larger arteries such as the renal artery and pulmonary artery [8,9,10, 18, 19]. Likewise, a previous study reported the safety and feasibility of angioscopic observation during carotid angioplasty with stent placement [20]. We believe this imaging technique can be applied to the carotid artery to evaluate plaque morphology. However, it should be noted that there might be some limitations in conducting angioscopic examination in the carotid artery. In some case, angioscopy could not acquire the images of whole vessel wall due to vessel tortuosity and might have missed some plaques. Furthermore, it might be difficult to advance the 2.25Fr angioscopic fiber into a tortuous siphon-like ICA in some case. It should be careful to advance the proving catheter and angioscopic fiber to avoid traumatic dissection, especially in the tortuous artery.

IPH is considered to be caused by rupture of the fragile neovessels formed within the plaque [21]. IPH is thought to be an important atherogenic stimulus by contributing cholesterol to the necrotic core of the plaque and thinning the fibrous cap, thus weakening it and making the plaque more vulnerable [22]. Therefore, the identification of a massive IPH, large necrotic core, and thin fibrous cap may be important to predict plaque rupture and complications during carotid artery stenting. Although an MRI is used to detect IPH with good sensitivity in vivo [23], it is not able to distinguish between IPH and a necrotic core. Moreover, it is still limited in precise fibrous cap imaging due to its lower resolution. Recent studies reported that optical coherence tomography (OCT) has the capability to identify thin fibrous cap because of its high axial resolution [24, 25]. Moreover, OCT can differentiate lipid tissue from fibrous tissue. Because OCT uses near-infrared light and cross-sectional images are generated by measuring the time delay and intensity of light signal that is reflected or backscattered from the arterial wall, OCT has the potential for characterizing tissue morphology by measuring backscattered infrared light. Kume et al has reported that sensitivity and specificity for detecting lipid-rich plaque by OCT was a > 90% in 166 sections from 108 coronary arterial segments of 40 consecutive human cadavers [26]. Lipidic tissue is appeared as low signal intensity regions with diffuse borders because of strong multiple scattering of the OCT light signal at the lipid surface. Therefore, OCT cannot distinguish large necrotic core and small necrotic core. Angioscopy may have an advantage in this respect.

In the present study, we revealed the capability of angioscopy to detect a massive IPH, large necrotic core, and thin fibrous cap by identifying red plaque. The average thickness of the fibrous cap was 128 ± 34 µm for red plaques, which is undetectable by intravascular ultrasound due to its low spatial resolution. The cutoff value of fibrous cap thickness to predict the presence of red plaque on angioscopy was 181 µm with high sensitivity and specificity. A previous histological evaluation of 526 patients undergoing endarterectomy showed that the median cap thickness was 150 µm in ruptured plaques, and the optimum cut-offs for distinguishing between ruptured and non-ruptured plaques was a cap thickness < 200 µm [27]. Moreover, our group has recently demonstrated that 95% of ruptured plaques in the carotid artery had a fibrous cap thickness that was thinner than 130 µm [24]. Therefore, the identification of red plaques by angioscopy in the carotid artery may have benefits in terms of a reduction in not only periprocedural complication, but also future neurological events.

Limitations

This study has several limitations. First, this study was conducted at a single center with a small sample size. Therefore, further investigations must be performed with larger samples to confirm the conclusions. Second, angioscopy in vivo is normally performed under the removal of blood flow by sustained flushing of low molecular dextran. In this study, imaging was conducted under a sustained flush of saline in an ex vivo environment. Third, we have not evaluated quantitative calculation in the longitudinal assessment because pathological sections were not evaluated by longitudinal serial sections. Fourth, the color grades of the plaques were determined visually and unaided, and therefore could be subjective. Finally, only one third of all strokes are caused by thromboembolism from unstable carotid atherosclerotic plaques [28]. It is difficult to predict the other cause of strokes on angioscopy.

Conclusions

The prevalence of red plaques on intravascular angioscopy may represent the existence of plaques containing relatively larger necrotic core and IPH with a thin fibrous cap.