An evaluation of dual source computed tomography used with the de Weert classification to detect vulnerable plaque, using IVUS virtual histology as a standard of reference

Background. One of the main risk factors for cerebral ischemic events is atherosclerotic disease of the internal carotid artery (ICA). Nowadays, increasing attention is being paid to the relationship between the morphological features of atherosclerotic plaque and the occurrence of stroke. Several studies have demonstrated that the presence of specific vulnerable plaque types, with a large lipid core and thin fibrous cap, can be used as an independent risk predictor of cerebral ischemic events. Objectives

Ischemic stroke is one of the most common causes of death and long-term disability in industrialized countries.Although stroke rates adjusted for age have continued a downward trend over the past three decades, the aging population could cause the number of strokes to increase in the near future. 1 One of the main risk factors for ischemic stroke is atherosclerotic disease of the internal carotid artery (ICA).For this reason, evaluation of the degree of ICA stenosis has become part of the currently used algorithms for choosing optimal treatment methods.3][4] On the other hand, for over a decade more and more attention has been drawn to the relationship between the morphological features of atherosclerotic plaque and the occurrence of ischemic events, particularly with regard to acute coronary syndrome (ACS). 5The development of novel imaging methods has made it possible to identify in vivo the presence of plaque ruptures and plaque hemorrhages, as well as assess the size of a lipid or necrotic core and the thickness of a fibrous tissue cap.Based on these factors, the concept of "vulnerable plaque" was introduced, referring to lesions with a high tendency to rupture, resulting in embolization or thrombosis. 6everal studies have demonstrated that the presence of a specific vulnerable plaque type with a large lipid core and a thin fibrous cap can be used as an independent risk predictor of ischemic cerebrovascular events. 7,8][11] In 2009, in a study published in Stroke, de Weert et al. proposed an interesting method for indirect assessment of plaque morphology using minimally invasive multidetector computed tomography angiography (MDCTA) and their own classification system. 12The authors correlated the MDCTA results with clinical data regarding the severity of stenosis, cardiovascular risk factors and symptomatology.
The imaging protocol for intravascular ultrasound virtual histology (IVUS-VH) has introduced a new level of quality to the assessment of lesion morphology that makes it possible to put histopathological knowledge to direct use in in vivo intravascular diagnosis. 13The present study is an attempt to develop the method presented by de Weert et al. by correlating the results of Dual Source Computed Tomography (DSCT) with those obtained from IVUS-VH.

Material and methods
From January 2011 to June 2013, a group of 170 patients with ICA stenosis found in ultrasound imaging (US) un-derwent dual source computed tomography (DSCT) to confirm the presence of significant stenosis (above 70%) in the ostium of the ICA.On this basis, the patients were included in the intravascular ultrasound (IVUS) study.
The experimental protocol and informed consent procedure complied with the Helsinki Convention and were approved by a bioethics commission.
Scanning was performed on a 64-slice dual source CT scanner (Somatom Definition, Siemens Medical Solutions, Forchheim, Germany) according to Persson's Dual Energy protocol using 140 kV and 80 kV lamps. 14When an 80 kV lamp is used, both contrast and calcifications are characterized by high density.When a 140 kV lamp is used, calcifications can be differentiated and removed from the image due to their higher density.The DSCT scan range extended from the aortic arch to the circle of Willis.All the patients received 50 mL of contrast material (400 mg/mL Iomeron 400, Bracco Imaging Deutschland GmbH, Konstanz, Germany), followed by 60 mL of saline bolus chaser, both with an injection rate of 6 mL/s.Synchronization between the passage of the contrast ma- terial and data acquisition was achieved by real-time bolus tracking at the level of the ascending aorta.The trigger threshold was set at an increase in attenuation of 110 Hounsfield units above baseline attenuation.
The DSCT images were processed on a Syngo-Via workstation (Siemens Medical Solutions, Forchheim, Germany) using dedicated 3-dimensional analysis software with hard plaque and head bone removal algorithms.Carotid arteries were evaluated with multiplanar reformatting (MPR) software, adjusting oblique planes to assess them in multiple short and long axes.
The surface of a lesion was evaluated and classified according to the de Weert classification into one of the three types: (1) smooth, (2) irregular and (3) ulcerated.Atherosclerotic plaque was classified as ulcerated when the contrast material was observed in the surrounding plaque, beyond the vascular lumen.Plaque classified as irregular showed discreet irregularities within the vessel walls, with an absence of ulcerations.Lesions without any ulcerations or distortions of lumen contour were classified as smooth.
In all the patients undergoing the IVUS procedure, embolic protection devices (Emboshield NAV6 Device, Abbott Vascular, Abbott Park, IL, USA) were used.Every examination was carried out by the same team, with a similar transducer position: 30 mm from the common carotid artery bifurcation.The images were obtained using 20 MHz transducers, a 3.2-Fr IVUS catheter (Eagle Eye Gold, Volcano Corporation, Rancho Cordova, USA) with the options: gray scale, chroma flow and virtual histology (VH) protocol.All the analyses were performed at an independent external core laboratory (Kraków Cardiovascular Research Institute, Krakow, Poland).
Internal carotid artery lesions observed with IVUS were classified into 5 categories 15,16 : z Adaptive intimal thickening (AIT): plaque comprised of mainly fibrous tissue; calcification, fibrofatty and/or necrotic core plaque components < 10%; z Pathological intimal thickening (PIT): fibrofatty plaque component > 10%, and confluent necrotic core or calcium amounting to < 10% of the total plaque cross-sectional area; z Fibroatheroma (FA): a confluent necrotic core comprising > 10% of the total plaque cross-sectional area; z Fibrocalcific (FCa): a confluent area of calcium comprising > 10% of the total plaque cross-sectional area, with necrotic core and fibrofatty plaque each < 10% of the total plaque cross-sectional area; z Thin-cap fibroatheroma (TCFA): high-risk plaque with a necrotic core comprising > 10% of the total plaque cross-sectional area and confluent with the lumen.

Results
The final study group consisted of 30 patients who underwent both DSCT and IVUS-VH examinations.The DSCT results were classified according to the de Weert classification.Thirteen cases (43%) were characterized as having a smooth wall surface, 10 cases (33%) had discreet wall irregularities, and in 7 cases (23%) incursions of contrast were observed, indicating the presence of ulceration (Table 3).
In the IVUS-VH examinations, 4 out of 30 cases (13%) were identified as AIT, 4 (13%) as PIT, 6 (20%) as FA, 6 (20%) as FCa and 10 (33%) as TCFA (Table 4).The results of the de Weert classification of the DSCT results were compared with the results of plaque morphology classification obtained using IVUS-VH (Fig. 1).It was found that all the patients with confirmed presence of wall ulceration in DSCT were characterized as having high-risk plaque (TCFA) in IVUS-VH.Patients with discreet wall irregularities in DSCT formed a heterogeneous group in IVUS-VH: 6 lesions with a necrotic core (3 FA and 3 TCFA) and 4 lesions without a necrotic core (2 FCA, 1 PIT, 1 AIT).Among the 13 patients with smooth wall surfaces in DSCT, IVUS-VH showed 4 FCA lesions, 3 PIT lesions, 3 AIT lesions and 3 FA lesions (Table 5).

Discussion
The degree of ICA stenosis is an important predicator of ischemic cerebrovascular events and is universally used in therapeutic decision-making.3][4] However, it has been noted that advanced atherosclerosis may be found in non-stenotic artery segments due to the phenomenon of artery remodeling. 17,181]19 Several studies have demonstrated that the presence of a specific vulnerable plaque type (with a large lipid core and a thin fibrous cap) can be used as an independent risk predictor of ischemic cerebrovascular events. 7,8The clinical consequences of plaque ulceration have been clearly demonstrated in several studies, including NAS-CET. 3,20,21o develop the NASCET and ECST scales for carotid pathology detection, digital subtraction angiography (DSA) was used, which made this method the gold standard in ICA stenosis evaluation.Due to its availability and low cost, ultrasound imaging (US) is currently the most widely used diagnostic method for detecting and assessing the extent of carotid atherosclerosis.However, there is no universally accepted method that provides a satisfactory amount of information in vivo about ICA plaque morphology.In their study, Saba et al. compared the diagnostic efficacy of multidetector computed tomography angiography (MDCTA) and US with echo color Doppler in patients with carotid plaque complicated by ulceration. 22hen the data from MDCTA and US imaging were set side by side with surgical specimens taken during carotid endarterectomy, it was demonstrated that MDCTA has very high sensitivity (93.75%) and specificity (98.59%) in detecting plaque ulcerations of the carotid artery.
To explore the potential of CT in assessing plaque morphology, the authors needed a well-established reference method.The choice fell on IVUS-derived virtual histology (IVUS-VH), which is used for the detection and quantification of non-stenotic atheroma in coronary arteries. 13he IVUS-VH imaging protocol uses advanced radiofrequency analysis of ultrasound signals, enhancing the image with colors in correlation to different amplitudes and frequencies.This makes it possible to recognize vulnerable plaque by assessing their structure and composition, differentiating four primary plaque components: fibrotic tissue, calcifications and lipid or necrotic cores.IVUS-VH findings demonstrate a good correlation with histopathology results. 23,24s already noted, IVUS-VH is used as a reference standard in evaluating other methods of imaging atherosclerotic plaque morphology in coronary arteries. 25However, the authors did not find any publications concerning the usefulness of IVUS-VH in evaluating DSCT results in ICA plaque morphology assessments.Until now, studies have focused on determining a correlation between CT results and histopathological findings gathered during endarterectomies.Fragmentation or damage to the sample -which often occurs during the procedure -makes it impossible to pinpoint the exact location of the histopathological specimen with respect to CT images.For this reason, a correlation can be rendered only with regard to general morphological composition. 26n their study, de Weert et al. noted that they had difficulties with the identification of ulceration in highly stenotic vessels. 12The calcification burden rises with severe stenosis.With MDCTA, differentiating between contrast enhancement and strongly calcified lesions was often impossible, which hampered the assessment of plaque morphology.The current authors did not encounter such difficulties in distinguishing contrast material from calcifications, thanks to the availability of a new dual-source multidetector CT system (DSCT).
Introduced in 2009, DSCT is equipped with 2 X-ray tubes that can be operated at different kV settings (an 80 kV lamp paired with a 120 kV or 140 kV lamp) and different prefiltration.Due to the simultaneous use of two X-ray sources, it is possible to obtain different radiation absorption coefficients (in Hounsfield units) for any given tissue.This allows better tissue differentiation and advanced image post-processing.DSCT permits reliable subtraction of calcified structures such as bones and plaque; therefore, it makes evaluation of the grade of stenosis and plaque morphology possible, even in highly stenotic vessels. 27Dual-energy acquisition and more detector rows make examination times shorter and reduce the radiation absorbed by patients.Shorter scan time also means less contrast (50-60 mL).Therefore, DSCT enhanced with plaque and bone subtraction algorithms (PBS) has the potential to identify patients with vulnerable plaque better than conventional CT.
Work on CT number-based plaque characterization with the use of absolute Hounsfield unit (HU) ranges to separate plaque components began with the advent of new software. 25Unified HU-based atherosclerotic plaque characterization criteria have not yet been created.Metaanalyses indicate that this approach is limited due to its dependence on factors such as lumen contrast opacification, the degree of artery stenosis, image noise, partial volume effects, undershooting artifacts and tube voltage. 28Moreover, existing classifications do not work well for identifying lipid-rich necrotic cores and hemorrhages, because the range of densities associated with these components is similar to those for connective tissue.This overlap severely limits the reliability of individual pixel Hounsfield readings on key components of vulnerable plaque. 29 present, there is no satisfactory method for direct assessment of plaque morphology in CT. 30,31 This is the reason the current authors considered adapting de Weert classification for DSCT data analysis, which uses plaque surface morphology assessment.The indirect approach is, in the current authors' opinion, a feasible and reliable tool for clinical use.Nevertheless, the distinction between atherosclerotic plaque morphology and plaque surface morphology has to be recognized.
In the present study, only a partial correlation was found between the results of DSCT with the de Weert classification and the results of IVUS-VH with plaque morphology classification.Nevertheless, the authors find this outcome clinically important.The data show that all the lesions with surplus contrast (type 3 in the de Weert classification) had the morphological features of TCFA, i.e. the presence of a necrotic component > 10%.According to the plaque morphology classification, TCFA is considered high-risk plaque with a propensity for rupturing.In fact, TCFA bears a strong morphological resemblance to plaque rupture even though it exhibits a smaller necrotic core, a thin fibrous cap infiltrated by a smaller number of macrophages, and a lower degree of calcification. 5 This study has some limitations.Firstly, the study group consisted of only 30 patients.Secondly, IVUS-VH has its own intrinsic limitations as a method of reference.In the case of strong remodeling, the transducer penetration depth (3-6 mm) may be insufficient to evaluate a complete cross-section of the ICA wall.The IVUS transducer itself is more suitable for coronary arteries, and the wider lumen of the carotid artery may cause non-axial settlement that results in an elliptic rather than orthoradial scan of an artery.However, that would more likely cause a misevaluation of the degree of stenosis rather than an incorrect assessment of plaque morphology.
DSCT with the de Weert classification of plaque surface morphology makes the reliable detection of ulcerations possible.The present study demonstrated that surplus the contrast agent in DSCT (indicating the presence of ulceration) correlates with the presence of unstable TCFA plaque in IVUS-VH.The benefits of surgical or endovascular intervention in symptomatic patients who do not meet the criteria for severe stenosis, but in whom the presence of ulcerated plaque has been identified, needs to be demonstrated in larger prospective studies.Nevertheless, the authors are confident that diagnostic algorithms in patients with cerebrovascular ischemic risk should be enriched with an assessment of atherosclerotic plaque morphology.Due to the unavailability of a satisfactory method that allows for direct assessment of plaque morphology in DSCT, an indirect approach using DSCT with de Weert classification could offer a significant new technique to improve current imaging protocols.

Table 1 .
Inclusion and exclusion criteria for the study

Table 2 .
Demographic and epidemiological data of the study group

Table 3 .
Number of cases of atherosclerotic plaque types in DSCT according to the de Weert classification

Table 4 .
Number of cases of each atherosclerotic plaque type according to plaque morphology characteristics (IVUS-VH)