This document is protected by international copyright laws. No additional reproduction is authorized. It is permitted for personal use to download and save only one file and print only one copy of this Article. It is not permitted to make additional copies (either sporadically or systematically, either printed or electronic) of the Article for any purpose. It is not permitted to distribute the electronic copy of the article through online internet and/or intranet file sharing systems, electronic mailing or any other means which may allow access to the Article. The use of all or any part of the Article for any Commercial Use is not permitted. The creation of derivative works from the Article is not permitted. The production of reprints for personal or commercial use i not permitted. It is not permitted to remove, cover, overlay, obscure, block, or change any copyright notices or terms of use which the Publisher may post on the Article. It is not permitted to frame or use framing techniques to enclose any trademark, log or other proprietary information of the Publisher.
©
COPYRIGHT 2017 EDIZIONI MINERVA MEDICA
© 2016 EDIZIONI MINERVA MEDICA Online version at http://www.minervamedica.it
The Journal of Cardiovascular Surgery 2017 February;58(1):55-64 DOI: 10.23736/S0021-9509.16.07655-2
ORIGINAL ARTICLE VA S C U L A R S U R G E R Y
Ex vivo characterization of carotid plaques by intravascular ultrasonography and virtual histology: concordance with real plaque pathomorphology Martina FUCHS 1, Peter HEIDER 1, Jaroslav PELISEK 1, Holger POPPERT 2, Henning H. ECKSTEIN
1
1Department
of Vascular and Endovascular Surgery; 2Department of Neurology, Klinikum rechts der Isar der Technischen Universitaet Muenchen, Munich, Germany *Corresponding author:Jaroslav Pelisek, PhD, Department of Vascular and Endovascular Surgery, Klinikum rechts der Isar der Technischen Universitaet Muenchen, Ismaninger Str. 22, 81675 Munich, Germany. E-mail: [email protected]
A B S T RA C T BACKGROUND: The purpose of this study was to evaluate the accuracy of carotid plaque characterisation by virtual histology using intravascular ultrasonography (VH-IVUS) by comparing the results with real morphology. METHODS: Following elective carotid endarterectomy (CEA), atherosclerotic plaques from 36 patients (19 asymptomatic, 17 symptomatic) underwent ex-vivo VH-IVUS examination. Afterwards, tissue specimens were fixed with formalin and embedded in paraffin. Atherosclerotic lesions were characterised following hematoxylin/eosin (HE) and Elastin van Gieson (EvG) staining using AHA classification (stages I to VIII). The plaque composition, cellularity, severity of inflammation, and atheroma-associated macrophages and foam cells were compared with virtual histology. RESULTS: Patients with symptomatic carotid artery stenosis showed most commonly lesion type IV-V (N.=9; 52.9%), followed by type VI (N.=3; 17.6%) and type VII (N.=3, 17.6%), type VIII (N.=1; 5.9%) and type I-III (N.=1; 5.9%). In asymptomatic patients with the main lesion was type VII (N.=8; 42.1%), followed by type I-III (N.=4; 21.1%), type IV-V (N.=3, 15.8%) and type VIII (N.=1; 5.3%). The composition of unstable lesions differed significantly in symptomatic patients compared to asymptomatic subjects (70.1% vs. 31.6%, P=0.03). The concordance between the histological results and the VH-IVUS classification was 86.1% (Cohen`s kappa of 0.72). CONCLUSIONS: In the present study, our findings demonstrated significant correlation between true plaque composition determined by histology and VH-IVUS. Thus, IVUS might be useful as an additional diagnostic method to detect patients with unstable rupture-prone plaques. (Cite this article as: Fuchs M, Heider P, Pelisek J, Poppert H, Eckstein HH. Ex vivo characterization of carotid plaques by intravascular ultrasonography and virtual histology: concordance with real plaque pathomorphology. J Cardiovasc Surg 2017;58:55-64. DOI: 10.23736/S0021-9509.16.07655-2) Key words: Diagnostic imaging - Atherosclerosis - Carotid stenosis.
I
maging is a promising diagnostic tool to analyse atherosclerotic plaque morphology in carotid arteries and to evaluate the potential risk of vulnerability. The resolution of current imaging techniques is so far not sufficient to detect vulnerable plaques postulated by Rotwell and Redgrave, defining unstable carotid lesions with thin-cap fibroatheroma 70%.9 Figure 1.—A) Experimental set-up for ex vivo evaluation of carotid plaques with VH-IVUS; B) carotid plaque specimen with 20-MHz Eagle Eye Gold catheter (Volcano Corporation, Rancho Cordova, CA, USA) for data acquisition.
(CT), ultrasonography, contrast-enhanced ultrasound or intravascular ultrasonography (IVUS) have already been described in carotid plaque imaging.5, 6 Especially IVUS has been used for more than ten years in carotid stenting. Grayscale intravascular ultrasonography (IVUS), a tomographic imaging tool, can visualize atherosclerosis in vivo, elucidating plaque area, plaque distribution, lesion length and degree of stenosis.7 However, quantitative assessment of plaque composition has not been possible with grayscale IVUS. An improvement in plaque imaging was achieved by IVUS-derived virtual histology (VH-IVUS), uses advanced radiofrequency analysis of ultrasound signals to provide a more detailed analysis of plaque morphology.7 In addition, VH-IVUS allows quantitative evaluation of various plaque components, such as fibrous tissue, fibro-fatty tissue, calcification, and lipid/necrotic core.8 However, this technique has still to be further elucidated. The aim of this study was, therefore, to evaluate the accuracy of plaque characterization by VH-IVUS in comparison to the true morphology, determined by histology and immunohistochemistry.
56
Data acquisition: VH-IVUS and histology CEA was performed with a great care to remove the plaque en bloc in its entirety. Immediately after removal the plaque was transferred into a basin with a solution of physiological saline (Figure 1A). After positioning of the plaque VH-IVUS examination was performed using a 2.9F, 20-MHz Eagle Eye Gold catheter (Volcano Corporation, Rancho Cordova, CA, USA) over a 0.014-inch wire, and it was pulled back at 0.5mm/s to the carotid bifurcation with an auto-motorized pullback system (Figure 1B). The VH-IVUS data were recorded onto the imaging system hard disk, and analyses were independently performed by experienced analysts unaware of the clinical findings and lesion characteristics. All measurements were drived from Volcano InVision Gold imaging system software. Afterwards plaques were cut into serial cross sections, starting at the site of bifurcation area until the end of the stenosis. Histological analyses were performed on representative sections (2-3 μm) of paraffin-fixed carotid plaque tissue samples. Paraffin sections were first dewaxed with xylene and rehydrated in descending ethanol sequence (100% to 70%). The carotid plaque tissues sections were then routinely stained with hematoxylin and eosin (HE) and Elastin van Gieson (EvG) to assess the tissue structure of all plaque samples, their
The Journal of Cardiovascular Surgery February 2017
IVUS and virtual histologyS FUCHS
Table I.—VH-IVUS and AHA classification according to Cai and Virmani.16-18 VH-IVUS plaque classificastion and lesion assessment
AHA-classification modified
Type I: initial lesion with foam cells Type II: fatty streak with foam cell layers
Type III: preatheroma with extracellular lipid
Type I – II: near normal wall thickness, no calcification
Type 1: adaptive intimal thickening (AIT). Plaque comprised all fibrous tissue (5%), and necrotic core and often calcification (5%) in fibrotic and/or fibrofatty tissue –– with minor calcification (5%)
Type IV: atheroma with a confluent extraxcellular lipid core
Type V: Fibroatheroma (FA)
Type IV – V: plaque with a lipid or necrotic core surrounded by fibrous tissue with possible calcification
Type VI: complex plaque with possible surface defect, hemorrhage or thrombus
Type VI: complex plaque with possible surface defect, hemorrhage or thrombus
Type VII: calcified plaque
Type VII: calcified plaque
Type VIII: fibrotic plaque without lipid core
Type VIII: fibrotic plaque without lipid core and with possible small calcifications
cellular composition, infiltration, plaque rupture, intraplaque hemorrhage, atheroma-associated foam cells and, atheroma-associated macrophages. Histological sections were digitized and analyzed by two independent investigators blinded to the VH-IVUS data acquisition. Fibrous, fibro-fatty, calcium and necrosis were defined accordingly. Densely packed collagen was termed fibrous, significant lipid interspersed in collagen was termed fibro-fatty, calcium deposits without adjacent necrosis were termed calcium and regions comprising cholesterol clefts, foam cells and microcalcifications were identified as calcified necrosis.
Vol. 58 - No. 1
Early lesion
AHA-classification conventional
Type 4a, b: Fibroatheroma (FA), without FA, NC 5-10% –– with minor calcification (5%) Type 5a, b: Thin FC fibroatheroma (ID TCFA), NC >10%, w/o IVUS evidence of FC –– with minor calcification (5%) vulnerability index 2 (VI 2)
Advanced lesion
This document is protected by international copyright laws. No additional reproduction is authorized. It is permitted for personal use to download and save only one file and print only one copy of this Article. It is not permitted to make additional copies (either sporadically or systematically, either printed or electronic) of the Article for any purpose. It is not permitted to distribute the electronic copy of the article through online internet and/or intranet file sharing systems, electronic mailing or any other means which may allow access to the Article. The use of all or any part of the Article for any Commercial Use is not permitted. The creation of derivative works from the Article is not permitted. The production of reprints for personal or commercial use i not permitted. It is not permitted to remove, cover, overlay, obscure, block, or change any copyright notices or terms of use which the Publisher may post on the Article. It is not permitted to frame or use framing techniques to enclose any trademark, log or other proprietary information of the Publisher.
©
COPYRIGHT 2017 EDIZIONI MINERVA MEDICA
Type 5c, d: ID TCFA –– several confluent NCs (one NC without FC suggesting previous rupture(s) with calcification = VI 3 –– high risk ID TCFA; most often found as a cause of the sudden coronary death = VI 4 Type 6: Fibrocalcific plaque –– fibrous plaques with dense calcification (>5%) –– NC0.75 was considered to indicate good concordance, a kappa value between 0.40 and 0.75 to indicate moderate concordance. Receiver-Operator characteristic curves (ROC) were applied for the evaluation of the diagnostic accuracy of VH-IVUS in patients with carotid artery stenosis by the size of the area under ROC curve (AUC). AUC >0.9 indicates the highest diagnostic accuracy, 0.65-0.9 confident diagnostic utility, 0.5-0.65 stands for low diagnostic accuracy. Areas under ROC curve for different tests were compared according to the overlap of their confidence intervals and according u-test. ROC analysis provided also determination of the highest specificity and sensitivity (HSS) of VH-IVUS.
In the present study 36 patients (17 symptomatic, 19 asymptomatic) with high degree carotid artery stenosis were evaluated regarding VH-IVUS and true histology. No differences were found between clinical variables, risk factor profiles, or pharmacological treatment between symptomatic and asymptomatic study group (Table II). The degree of stenosis was identical in each group as well (asymptomatic vs. symptomatic: 87.2±9.9% vs. 84.1±10.6%, P=0.334).
The Journal of Cardiovascular Surgery February 2017
This document is protected by international copyright laws. No additional reproduction is authorized. It is permitted for personal use to download and save only one file and print only one copy of this Article. It is not permitted to make additional copies (either sporadically or systematically, either printed or electronic) of the Article for any purpose. It is not permitted to distribute the electronic copy of the article through online internet and/or intranet file sharing systems, electronic mailing or any other means which may allow access to the Article. The use of all or any part of the Article for any Commercial Use is not permitted. The creation of derivative works from the Article is not permitted. The production of reprints for personal or commercial use i not permitted. It is not permitted to remove, cover, overlay, obscure, block, or change any copyright notices or terms of use which the Publisher may post on the Article. It is not permitted to frame or use framing techniques to enclose any trademark, log or other proprietary information of the Publisher.
©
COPYRIGHT 2017 EDIZIONI MINERVA MEDICA IVUS and virtual histologyS FUCHS
A
B
Figure 3.—Distribution of stable and unstable lesions in patients with symptomatic or asymptomatic carotid artery stenosis according the modified AHA criteria.
C
D
Figure 2.—Expamples of gray-scale (A), virtual (B, C), and corresponding true histology (D) using IVUS image of carotid atherosclerotic plaque. A) Border detection is shown as yellow and red line; B, C) color-coded VH-IVUS data of a thin cap fibroatheroma (TCFA) with greater than 50% narrowing (green=fibrous, yellow-green=fibrolipidic, red=necrotic, white=calcification); D) hematoxylin and eosin staining. Large atheroma (Ath) with a reasonably thick stable fibrous cap (FC).
To determine the diagnostic accuracy of VH-IVUS imaging, first the correlation between repeated measurements of an identical vascular cross section was determined for one investigator (intra-observer correlation). For all one-dimensional measurement (lumen diameter, vessel diameter, and maximal plaque thickness) Pearson correlation coefficient was r=0.98, for two-dimensional measurement (vessel area, lumen area, and plaque area) the correlation coefficient was between 0.96 and 0.98. The correlation accuracy between two independent investigators (inter-observer correlation) by evaluation of an identical vascular cross section was between 0.90 (maximum plaque thickness) and 0.94 (lumen diameter and vessel diameter), for two-dimensional parameters between 0.83 (plaque area) and 0.95 (vessel area). All these results were comparable to the data of other studies.11, 13, 14 In patients with symptomatic carotid artery stenosis, lesion type IV-V occurred most commonly (N.=9; 52.9%), followed by type VI (N.=3; 17.6%), VII (N.=3;
Vol. 58 - No. 1
Figure 4.—Percentage of concordance of AHA-classification using true histology and VH-IVUS imaging technique.
17.6%), VIII (N.=1; 5.9%), and III (N.=1; 5.9%) (Figure 3). In patients with asymptomatic stenosis, mainly lesion type VII was found (N.=8; 42.1%), followed by type II-III (N.=4; 21.1%), IV-V (N.=3; 15.8%), VI (N.=3; 15.8%), and VIII (N.=1; 5.3%). The distribution of unstable lesion types differed significantly in symptomatic patients as compared to asymptomatic patients (70.1% vs. 31.6%, P=0.03) (Figure 3). Furthermore,
The Journal of Cardiovascular Surgery
59
This document is protected by international copyright laws. No additional reproduction is authorized. It is permitted for personal use to download and save only one file and print only one copy of this Article. It is not permitted to make additional copies (either sporadically or systematically, either printed or electronic) of the Article for any purpose. It is not permitted to distribute the electronic copy of the article through online internet and/or intranet file sharing systems, electronic mailing or any other means which may allow access to the Article. The use of all or any part of the Article for any Commercial Use is not permitted. The creation of derivative works from the Article is not permitted. The production of reprints for personal or commercial use i not permitted. It is not permitted to remove, cover, overlay, obscure, block, or change any copyright notices or terms of use which the Publisher may post on the Article. It is not permitted to frame or use framing techniques to enclose any trademark, log or other proprietary information of the Publisher.
©
COPYRIGHT 2017 EDIZIONI MINERVA MEDICA FUCHS IVUS and virtual histologyS
A
B
C
D
Figure 5.—Content of various plaque compoments in stable and unstable plaques identified by VH-IVUS. A) Fibrous tissue; B) fibro-fatty areas; C) areas of calcification; D) necrotic tissue.
comparing the results of true histology and the results achieved by VH-IVUS classification, the concordance was in average by 86.1% (Cohen’s kappa of 0.72). The individual agreement for the particular types of atherosclerotic lesions between true histology and VH-IVUS are shown in Figure 4. Linear regression analysis showed a strong correlation between histology and VH-IVUS for various plaque features. Fibrous tissue areas were significant larger in stable plaques (r=0.552, P=0.01), whereas unstable plaques showed more dense calcification and lipid/necrotic core (r=0.512, P=0.002 and r=0.385, P=0.022). There was no significant difference for fibrolipidic tissue regarding plaque stability (r=0.030, P=0.87). The mean percentage of fibrous, fibro-fatty, necrotic, and calcification in stable vs. unstable plaques were 63% vs. 53.9% (P=0.01), 18.3% vs. 19.7% (P=0.866), 18.3% vs.
60
6.2% (P=0.002), and 14.6% vs. 20% (P=0.022), respectively (Figure 5). Furthermore, ROC analysis was performed to analyze the performance of plaque composition as well as the stability of plaques (Figure 6). Cut-off values for plaque stability in VH-IVUS analysis were determined by using the Youden index (J=maximum {sensitivity + specificity-1}). The cut-off values of the different detected tissue types were as follows: plaques with more than 62.5% of fibrous tissue or less than 15.5% of necrotic or 4.5% of calcified tissue were classified as stable (Table III). Due to the fact, that carotid plaque thickness is a marker of subclinical atherosclerosis, we also measured total plaque volume, plaque area, and plaque burden in stable and unstable plaques (Figure 7). The mean total plaque volume was 39±11.4 mm³ in stable plaques and 50.5±34.1
The Journal of Cardiovascular Surgery February 2017
This document is protected by international copyright laws. No additional reproduction is authorized. It is permitted for personal use to download and save only one file and print only one copy of this Article. It is not permitted to make additional copies (either sporadically or systematically, either printed or electronic) of the Article for any purpose. It is not permitted to distribute the electronic copy of the article through online internet and/or intranet file sharing systems, electronic mailing or any other means which may allow access to the Article. The use of all or any part of the Article for any Commercial Use is not permitted. The creation of derivative works from the Article is not permitted. The production of reprints for personal or commercial use i not permitted. It is not permitted to remove, cover, overlay, obscure, block, or change any copyright notices or terms of use which the Publisher may post on the Article. It is not permitted to frame or use framing techniques to enclose any trademark, log or other proprietary information of the Publisher.
©
COPYRIGHT 2017 EDIZIONI MINERVA MEDICA IVUS and virtual histologyS FUCHS
A
B
Figure 6.—ROC curve analysis for comparison of the expression of different tissue types in stable and unstable carotid plaques. The curve above the diagonal shows the true-positive against the false-negative rate for determination of possible cut-off values of a prediction algorithm as a tradeoff between sensitivity and specificity. The greater the AUC (area under curve), the more accurate is the model. A) green=dense calcification; blue=necrotic areas; B) green=fibrous tissue, blue=fibrolipidic areas.
Table III.—Combined sensitivity, specificity and positive predictive value (PPV) for optimal cut-off points (plaque stability in VH-IVUS analysis). Cut-off point
Fibrous area Fibro-fatty area Necrotic area Calcified area
62.5 n.s. 15.5 4.5
A
Sensitivity
Specificity
0.87 n.s. 0.70 0.78
0.75 n.s. 0.67 0.75
PPV
72.9 n.s. 63.8 68.7
B
Figure 7.—Mean value of plaque area (A) and plaque burden (B) in percent of stable versus unstable carotid plaques using box-and-whisker-plots with significant differences between the groups (P=0.001 and P=0.002). The boxes demonstrate values between 0.25 and 0.75 quantile with median within the box. The lines (whiskers) show 95% percentiles of the corresponding samples.
Vol. 58 - No. 1
The Journal of Cardiovascular Surgery
61
This document is protected by international copyright laws. No additional reproduction is authorized. It is permitted for personal use to download and save only one file and print only one copy of this Article. It is not permitted to make additional copies (either sporadically or systematically, either printed or electronic) of the Article for any purpose. It is not permitted to distribute the electronic copy of the article through online internet and/or intranet file sharing systems, electronic mailing or any other means which may allow access to the Article. The use of all or any part of the Article for any Commercial Use is not permitted. The creation of derivative works from the Article is not permitted. The production of reprints for personal or commercial use i not permitted. It is not permitted to remove, cover, overlay, obscure, block, or change any copyright notices or terms of use which the Publisher may post on the Article. It is not permitted to frame or use framing techniques to enclose any trademark, log or other proprietary information of the Publisher.
©
COPYRIGHT 2017 EDIZIONI MINERVA MEDICA FUCHS IVUS and virtual histologyS
A
B
Figure 8.—Correlation of plaque area (A) and plaque burden (B) with total vessel area in the total plaque cohort using scattergraph. The figure contains a linear regression line of the best fit between the variables with border lines showing 95% confidence interval (r=0.498 and r=0.542; P=0.001).
mm³ in unstable plaques (P=0.001). Similarly, plaque burden was significantly lower in stable plaques (70.7±7.9%) than in unstable plaques (78.6±5.7%) (P=0.002). Mean plaque area (relatively to vessel area) also showed significantly lower values in stable plaques than in unstable plaques (73.1±7.7% vs. 82.4±5.7%, P=0.001). The mean measured plaque volume of each plaque showed no significant difference in stable and unstable plaques (0.39±0.11 mm³ vs. 0.51±0.34 mm³, P=0.187).���������� The������ relative plaque area of the target lesion correlated with total vessel area (r=0.498, P=0.01) and with the plaque burden (r=0.542, P=0.001). Grayscale and VH-IVUS findings for ruptured plaques, VH-TCFA and non-VH-TCFA plaques showed neither in asymptomatic patients, nor in symptomatic patients any significant differences. Discussion Stroke represents one of the most serious causes of mortality and morbidity in western countries and extra-cranial carotid complications represent approximately one half of all cases. Identification of vulnerable plaques is therefore essential to predict patients at risk. A promising tool to identify such individuals is plaque imaging. However, broad heterogeneity exists in the structure and composition of carotid atherosclerotic plaques. Therefore, it is difficult to recognize the true
62
rupture-prone lesion. Furthermore, many atherosclerotic plaques may cause no symptoms for decades; however, few plaques disrupt and cause thrombosis. Such vulnerable plaques are assumed to be at high short term risk of thrombosis, causing an acute cerebrovascular insult or sudden coronary death.1, 2, 10-12 Up to date, there is no reliable diagnostic method to prospectively identify these vulnerable plaques.11 However, many of the current imaging techniques used to assess carotid artery disease are already able to detect different features of the rupture-prone plaques.13-15 Despite several imaging techniques as well as different serological biomarkers being currently under development, none of them alone provides such all-embracing information to identify individual patients at increased risk of stroke.11 Specific criteria have been recently proposed to assess the relevance of MRI due to high resolution, or IVUS and VH-IVUS that are easy to perform and assess detailed plaque morphology. Nevertheless, these techniques are still under investigation and at present, none of them can reliable identify vulnerable plaques and, most importantly, predict its further development.16 From a clinical point of view, these methods currently assess only some features of atherosclerotic lesions. Thus, combination of several modalities would be of importance in the future to ensure high sensitivity and specificity in detecting vulnerable plaques. In our study we could show high concordance between true histology and VH-IVUS regarding various
The Journal of Cardiovascular Surgery February 2017
This document is protected by international copyright laws. No additional reproduction is authorized. It is permitted for personal use to download and save only one file and print only one copy of this Article. It is not permitted to make additional copies (either sporadically or systematically, either printed or electronic) of the Article for any purpose. It is not permitted to distribute the electronic copy of the article through online internet and/or intranet file sharing systems, electronic mailing or any other means which may allow access to the Article. The use of all or any part of the Article for any Commercial Use is not permitted. The creation of derivative works from the Article is not permitted. The production of reprints for personal or commercial use i not permitted. It is not permitted to remove, cover, overlay, obscure, block, or change any copyright notices or terms of use which the Publisher may post on the Article. It is not permitted to frame or use framing techniques to enclose any trademark, log or other proprietary information of the Publisher.
©
COPYRIGHT 2017 EDIZIONI MINERVA MEDICA IVUS and virtual histologyS FUCHS
atherosclerotic plaque types, symptoms, and plaque instability. Our results demonstrated that composition of plaques from symptomatic patients significantly differs from asymptomatic individuals. The former contain more lipid and cholesterol and less fibrous tissue. Compared to Diethrich et al. we could also demonstrate that fibrous tissue is the most frequent plaque component, followed by fibro-fatty, necrotic, and calcified tissue.17 In our study, the diagnostic accuracy of VH IVUS with true histology in different carotid plaque types was 99.4% in thin-cap fibroatheroma, 96.1% for calcified thin-cap fibroatheroma, 85.9% in fibroatheroma, 85.5% for fibrocalcific, 83.4% in pathological intimal thickening, and 72.4% for calcified fibroatheroma.��������� Furthermore, we detected a higher ratio of dense calcification and necrotic tissue in culprit lesions. Although Granada et al.18 postulated very low specificity and sensitivity of VH-IVUS in coronary lesions, we found high significant results for fibrous, calcified, and also necrotic tissue. Conversely, fibro-fatty tissue was not detected sufficiently, because the current classification cannot differentiate between intramural thrombus and other atherosclerotic components of vessel wall. In a study of Rodriguez-Granillo et al.19 non-culprit and non-obstructive lesions were examined by VH-IVUS in 23 patients with acute coronary syndromes and in 32 patients with stable CAD, demonstrating significantly higher incidence of thin fibrous cap in patients with acute coronary syndromes. Other histological studies have also demonstrated that lesions containing a considerable amount of necrotic core tissue without evidence of an overlying thin fibrous cap lesions may be considered as a precursor of plaque rupture.1, 2, 20, 21 Moreover, we could demonstrate a good correlation between tissue composition and stability of the carotid lesions. For stable and unstable plaques the correlation coefficient was 0.72 for VH-IVUS detection, which indicates a good accuracy for in vivo identification of plaques. Therefore, we can postulate that plaque components determined in carotid plaques reflect the real histological situation. According to the CAPITAL study,17 we showed strong correlation between VH-IVUS classification and stability of carotid atherosclerotic plaques. Nair et al.22 imaged ex vivo 88 plaques from 51 left anterior descending coronary arteries and came to the conclusion that coronary plaque composition can be predicted through the use of IVUS radiofrequency data analysis.
Vol. 58 - No. 1
Limitations of the study Although VH-IVUS is a promising imaging modality for plaque characterization, some limitations still remain. High interobserver variability in the prediction of tissue type by visual assessment of VH-IVUS images was observed in several studies. This underlines the need of quantitative methods for the analysis of the ultrasound characteristics of plaque components. Importantly, detection of the thin fibrous cap (
©
COPYRIGHT 2017 EDIZIONI MINERVA MEDICA
© 2016 EDIZIONI MINERVA MEDICA Online version at http://www.minervamedica.it
The Journal of Cardiovascular Surgery 2017 February;58(1):55-64 DOI: 10.23736/S0021-9509.16.07655-2
ORIGINAL ARTICLE VA S C U L A R S U R G E R Y
Ex vivo characterization of carotid plaques by intravascular ultrasonography and virtual histology: concordance with real plaque pathomorphology Martina FUCHS 1, Peter HEIDER 1, Jaroslav PELISEK 1, Holger POPPERT 2, Henning H. ECKSTEIN
1
1Department
of Vascular and Endovascular Surgery; 2Department of Neurology, Klinikum rechts der Isar der Technischen Universitaet Muenchen, Munich, Germany *Corresponding author:Jaroslav Pelisek, PhD, Department of Vascular and Endovascular Surgery, Klinikum rechts der Isar der Technischen Universitaet Muenchen, Ismaninger Str. 22, 81675 Munich, Germany. E-mail: [email protected]
A B S T RA C T BACKGROUND: The purpose of this study was to evaluate the accuracy of carotid plaque characterisation by virtual histology using intravascular ultrasonography (VH-IVUS) by comparing the results with real morphology. METHODS: Following elective carotid endarterectomy (CEA), atherosclerotic plaques from 36 patients (19 asymptomatic, 17 symptomatic) underwent ex-vivo VH-IVUS examination. Afterwards, tissue specimens were fixed with formalin and embedded in paraffin. Atherosclerotic lesions were characterised following hematoxylin/eosin (HE) and Elastin van Gieson (EvG) staining using AHA classification (stages I to VIII). The plaque composition, cellularity, severity of inflammation, and atheroma-associated macrophages and foam cells were compared with virtual histology. RESULTS: Patients with symptomatic carotid artery stenosis showed most commonly lesion type IV-V (N.=9; 52.9%), followed by type VI (N.=3; 17.6%) and type VII (N.=3, 17.6%), type VIII (N.=1; 5.9%) and type I-III (N.=1; 5.9%). In asymptomatic patients with the main lesion was type VII (N.=8; 42.1%), followed by type I-III (N.=4; 21.1%), type IV-V (N.=3, 15.8%) and type VIII (N.=1; 5.3%). The composition of unstable lesions differed significantly in symptomatic patients compared to asymptomatic subjects (70.1% vs. 31.6%, P=0.03). The concordance between the histological results and the VH-IVUS classification was 86.1% (Cohen`s kappa of 0.72). CONCLUSIONS: In the present study, our findings demonstrated significant correlation between true plaque composition determined by histology and VH-IVUS. Thus, IVUS might be useful as an additional diagnostic method to detect patients with unstable rupture-prone plaques. (Cite this article as: Fuchs M, Heider P, Pelisek J, Poppert H, Eckstein HH. Ex vivo characterization of carotid plaques by intravascular ultrasonography and virtual histology: concordance with real plaque pathomorphology. J Cardiovasc Surg 2017;58:55-64. DOI: 10.23736/S0021-9509.16.07655-2) Key words: Diagnostic imaging - Atherosclerosis - Carotid stenosis.
I
maging is a promising diagnostic tool to analyse atherosclerotic plaque morphology in carotid arteries and to evaluate the potential risk of vulnerability. The resolution of current imaging techniques is so far not sufficient to detect vulnerable plaques postulated by Rotwell and Redgrave, defining unstable carotid lesions with thin-cap fibroatheroma 70%.9 Figure 1.—A) Experimental set-up for ex vivo evaluation of carotid plaques with VH-IVUS; B) carotid plaque specimen with 20-MHz Eagle Eye Gold catheter (Volcano Corporation, Rancho Cordova, CA, USA) for data acquisition.
(CT), ultrasonography, contrast-enhanced ultrasound or intravascular ultrasonography (IVUS) have already been described in carotid plaque imaging.5, 6 Especially IVUS has been used for more than ten years in carotid stenting. Grayscale intravascular ultrasonography (IVUS), a tomographic imaging tool, can visualize atherosclerosis in vivo, elucidating plaque area, plaque distribution, lesion length and degree of stenosis.7 However, quantitative assessment of plaque composition has not been possible with grayscale IVUS. An improvement in plaque imaging was achieved by IVUS-derived virtual histology (VH-IVUS), uses advanced radiofrequency analysis of ultrasound signals to provide a more detailed analysis of plaque morphology.7 In addition, VH-IVUS allows quantitative evaluation of various plaque components, such as fibrous tissue, fibro-fatty tissue, calcification, and lipid/necrotic core.8 However, this technique has still to be further elucidated. The aim of this study was, therefore, to evaluate the accuracy of plaque characterization by VH-IVUS in comparison to the true morphology, determined by histology and immunohistochemistry.
56
Data acquisition: VH-IVUS and histology CEA was performed with a great care to remove the plaque en bloc in its entirety. Immediately after removal the plaque was transferred into a basin with a solution of physiological saline (Figure 1A). After positioning of the plaque VH-IVUS examination was performed using a 2.9F, 20-MHz Eagle Eye Gold catheter (Volcano Corporation, Rancho Cordova, CA, USA) over a 0.014-inch wire, and it was pulled back at 0.5mm/s to the carotid bifurcation with an auto-motorized pullback system (Figure 1B). The VH-IVUS data were recorded onto the imaging system hard disk, and analyses were independently performed by experienced analysts unaware of the clinical findings and lesion characteristics. All measurements were drived from Volcano InVision Gold imaging system software. Afterwards plaques were cut into serial cross sections, starting at the site of bifurcation area until the end of the stenosis. Histological analyses were performed on representative sections (2-3 μm) of paraffin-fixed carotid plaque tissue samples. Paraffin sections were first dewaxed with xylene and rehydrated in descending ethanol sequence (100% to 70%). The carotid plaque tissues sections were then routinely stained with hematoxylin and eosin (HE) and Elastin van Gieson (EvG) to assess the tissue structure of all plaque samples, their
The Journal of Cardiovascular Surgery February 2017
IVUS and virtual histologyS FUCHS
Table I.—VH-IVUS and AHA classification according to Cai and Virmani.16-18 VH-IVUS plaque classificastion and lesion assessment
AHA-classification modified
Type I: initial lesion with foam cells Type II: fatty streak with foam cell layers
Type III: preatheroma with extracellular lipid
Type I – II: near normal wall thickness, no calcification
Type 1: adaptive intimal thickening (AIT). Plaque comprised all fibrous tissue (5%), and necrotic core and often calcification (5%) in fibrotic and/or fibrofatty tissue –– with minor calcification (5%)
Type IV: atheroma with a confluent extraxcellular lipid core
Type V: Fibroatheroma (FA)
Type IV – V: plaque with a lipid or necrotic core surrounded by fibrous tissue with possible calcification
Type VI: complex plaque with possible surface defect, hemorrhage or thrombus
Type VI: complex plaque with possible surface defect, hemorrhage or thrombus
Type VII: calcified plaque
Type VII: calcified plaque
Type VIII: fibrotic plaque without lipid core
Type VIII: fibrotic plaque without lipid core and with possible small calcifications
cellular composition, infiltration, plaque rupture, intraplaque hemorrhage, atheroma-associated foam cells and, atheroma-associated macrophages. Histological sections were digitized and analyzed by two independent investigators blinded to the VH-IVUS data acquisition. Fibrous, fibro-fatty, calcium and necrosis were defined accordingly. Densely packed collagen was termed fibrous, significant lipid interspersed in collagen was termed fibro-fatty, calcium deposits without adjacent necrosis were termed calcium and regions comprising cholesterol clefts, foam cells and microcalcifications were identified as calcified necrosis.
Vol. 58 - No. 1
Early lesion
AHA-classification conventional
Type 4a, b: Fibroatheroma (FA), without FA, NC 5-10% –– with minor calcification (5%) Type 5a, b: Thin FC fibroatheroma (ID TCFA), NC >10%, w/o IVUS evidence of FC –– with minor calcification (5%) vulnerability index 2 (VI 2)
Advanced lesion
This document is protected by international copyright laws. No additional reproduction is authorized. It is permitted for personal use to download and save only one file and print only one copy of this Article. It is not permitted to make additional copies (either sporadically or systematically, either printed or electronic) of the Article for any purpose. It is not permitted to distribute the electronic copy of the article through online internet and/or intranet file sharing systems, electronic mailing or any other means which may allow access to the Article. The use of all or any part of the Article for any Commercial Use is not permitted. The creation of derivative works from the Article is not permitted. The production of reprints for personal or commercial use i not permitted. It is not permitted to remove, cover, overlay, obscure, block, or change any copyright notices or terms of use which the Publisher may post on the Article. It is not permitted to frame or use framing techniques to enclose any trademark, log or other proprietary information of the Publisher.
©
COPYRIGHT 2017 EDIZIONI MINERVA MEDICA
Type 5c, d: ID TCFA –– several confluent NCs (one NC without FC suggesting previous rupture(s) with calcification = VI 3 –– high risk ID TCFA; most often found as a cause of the sudden coronary death = VI 4 Type 6: Fibrocalcific plaque –– fibrous plaques with dense calcification (>5%) –– NC0.75 was considered to indicate good concordance, a kappa value between 0.40 and 0.75 to indicate moderate concordance. Receiver-Operator characteristic curves (ROC) were applied for the evaluation of the diagnostic accuracy of VH-IVUS in patients with carotid artery stenosis by the size of the area under ROC curve (AUC). AUC >0.9 indicates the highest diagnostic accuracy, 0.65-0.9 confident diagnostic utility, 0.5-0.65 stands for low diagnostic accuracy. Areas under ROC curve for different tests were compared according to the overlap of their confidence intervals and according u-test. ROC analysis provided also determination of the highest specificity and sensitivity (HSS) of VH-IVUS.
In the present study 36 patients (17 symptomatic, 19 asymptomatic) with high degree carotid artery stenosis were evaluated regarding VH-IVUS and true histology. No differences were found between clinical variables, risk factor profiles, or pharmacological treatment between symptomatic and asymptomatic study group (Table II). The degree of stenosis was identical in each group as well (asymptomatic vs. symptomatic: 87.2±9.9% vs. 84.1±10.6%, P=0.334).
The Journal of Cardiovascular Surgery February 2017
This document is protected by international copyright laws. No additional reproduction is authorized. It is permitted for personal use to download and save only one file and print only one copy of this Article. It is not permitted to make additional copies (either sporadically or systematically, either printed or electronic) of the Article for any purpose. It is not permitted to distribute the electronic copy of the article through online internet and/or intranet file sharing systems, electronic mailing or any other means which may allow access to the Article. The use of all or any part of the Article for any Commercial Use is not permitted. The creation of derivative works from the Article is not permitted. The production of reprints for personal or commercial use i not permitted. It is not permitted to remove, cover, overlay, obscure, block, or change any copyright notices or terms of use which the Publisher may post on the Article. It is not permitted to frame or use framing techniques to enclose any trademark, log or other proprietary information of the Publisher.
©
COPYRIGHT 2017 EDIZIONI MINERVA MEDICA IVUS and virtual histologyS FUCHS
A
B
Figure 3.—Distribution of stable and unstable lesions in patients with symptomatic or asymptomatic carotid artery stenosis according the modified AHA criteria.
C
D
Figure 2.—Expamples of gray-scale (A), virtual (B, C), and corresponding true histology (D) using IVUS image of carotid atherosclerotic plaque. A) Border detection is shown as yellow and red line; B, C) color-coded VH-IVUS data of a thin cap fibroatheroma (TCFA) with greater than 50% narrowing (green=fibrous, yellow-green=fibrolipidic, red=necrotic, white=calcification); D) hematoxylin and eosin staining. Large atheroma (Ath) with a reasonably thick stable fibrous cap (FC).
To determine the diagnostic accuracy of VH-IVUS imaging, first the correlation between repeated measurements of an identical vascular cross section was determined for one investigator (intra-observer correlation). For all one-dimensional measurement (lumen diameter, vessel diameter, and maximal plaque thickness) Pearson correlation coefficient was r=0.98, for two-dimensional measurement (vessel area, lumen area, and plaque area) the correlation coefficient was between 0.96 and 0.98. The correlation accuracy between two independent investigators (inter-observer correlation) by evaluation of an identical vascular cross section was between 0.90 (maximum plaque thickness) and 0.94 (lumen diameter and vessel diameter), for two-dimensional parameters between 0.83 (plaque area) and 0.95 (vessel area). All these results were comparable to the data of other studies.11, 13, 14 In patients with symptomatic carotid artery stenosis, lesion type IV-V occurred most commonly (N.=9; 52.9%), followed by type VI (N.=3; 17.6%), VII (N.=3;
Vol. 58 - No. 1
Figure 4.—Percentage of concordance of AHA-classification using true histology and VH-IVUS imaging technique.
17.6%), VIII (N.=1; 5.9%), and III (N.=1; 5.9%) (Figure 3). In patients with asymptomatic stenosis, mainly lesion type VII was found (N.=8; 42.1%), followed by type II-III (N.=4; 21.1%), IV-V (N.=3; 15.8%), VI (N.=3; 15.8%), and VIII (N.=1; 5.3%). The distribution of unstable lesion types differed significantly in symptomatic patients as compared to asymptomatic patients (70.1% vs. 31.6%, P=0.03) (Figure 3). Furthermore,
The Journal of Cardiovascular Surgery
59
This document is protected by international copyright laws. No additional reproduction is authorized. It is permitted for personal use to download and save only one file and print only one copy of this Article. It is not permitted to make additional copies (either sporadically or systematically, either printed or electronic) of the Article for any purpose. It is not permitted to distribute the electronic copy of the article through online internet and/or intranet file sharing systems, electronic mailing or any other means which may allow access to the Article. The use of all or any part of the Article for any Commercial Use is not permitted. The creation of derivative works from the Article is not permitted. The production of reprints for personal or commercial use i not permitted. It is not permitted to remove, cover, overlay, obscure, block, or change any copyright notices or terms of use which the Publisher may post on the Article. It is not permitted to frame or use framing techniques to enclose any trademark, log or other proprietary information of the Publisher.
©
COPYRIGHT 2017 EDIZIONI MINERVA MEDICA FUCHS IVUS and virtual histologyS
A
B
C
D
Figure 5.—Content of various plaque compoments in stable and unstable plaques identified by VH-IVUS. A) Fibrous tissue; B) fibro-fatty areas; C) areas of calcification; D) necrotic tissue.
comparing the results of true histology and the results achieved by VH-IVUS classification, the concordance was in average by 86.1% (Cohen’s kappa of 0.72). The individual agreement for the particular types of atherosclerotic lesions between true histology and VH-IVUS are shown in Figure 4. Linear regression analysis showed a strong correlation between histology and VH-IVUS for various plaque features. Fibrous tissue areas were significant larger in stable plaques (r=0.552, P=0.01), whereas unstable plaques showed more dense calcification and lipid/necrotic core (r=0.512, P=0.002 and r=0.385, P=0.022). There was no significant difference for fibrolipidic tissue regarding plaque stability (r=0.030, P=0.87). The mean percentage of fibrous, fibro-fatty, necrotic, and calcification in stable vs. unstable plaques were 63% vs. 53.9% (P=0.01), 18.3% vs. 19.7% (P=0.866), 18.3% vs.
60
6.2% (P=0.002), and 14.6% vs. 20% (P=0.022), respectively (Figure 5). Furthermore, ROC analysis was performed to analyze the performance of plaque composition as well as the stability of plaques (Figure 6). Cut-off values for plaque stability in VH-IVUS analysis were determined by using the Youden index (J=maximum {sensitivity + specificity-1}). The cut-off values of the different detected tissue types were as follows: plaques with more than 62.5% of fibrous tissue or less than 15.5% of necrotic or 4.5% of calcified tissue were classified as stable (Table III). Due to the fact, that carotid plaque thickness is a marker of subclinical atherosclerosis, we also measured total plaque volume, plaque area, and plaque burden in stable and unstable plaques (Figure 7). The mean total plaque volume was 39±11.4 mm³ in stable plaques and 50.5±34.1
The Journal of Cardiovascular Surgery February 2017
This document is protected by international copyright laws. No additional reproduction is authorized. It is permitted for personal use to download and save only one file and print only one copy of this Article. It is not permitted to make additional copies (either sporadically or systematically, either printed or electronic) of the Article for any purpose. It is not permitted to distribute the electronic copy of the article through online internet and/or intranet file sharing systems, electronic mailing or any other means which may allow access to the Article. The use of all or any part of the Article for any Commercial Use is not permitted. The creation of derivative works from the Article is not permitted. The production of reprints for personal or commercial use i not permitted. It is not permitted to remove, cover, overlay, obscure, block, or change any copyright notices or terms of use which the Publisher may post on the Article. It is not permitted to frame or use framing techniques to enclose any trademark, log or other proprietary information of the Publisher.
©
COPYRIGHT 2017 EDIZIONI MINERVA MEDICA IVUS and virtual histologyS FUCHS
A
B
Figure 6.—ROC curve analysis for comparison of the expression of different tissue types in stable and unstable carotid plaques. The curve above the diagonal shows the true-positive against the false-negative rate for determination of possible cut-off values of a prediction algorithm as a tradeoff between sensitivity and specificity. The greater the AUC (area under curve), the more accurate is the model. A) green=dense calcification; blue=necrotic areas; B) green=fibrous tissue, blue=fibrolipidic areas.
Table III.—Combined sensitivity, specificity and positive predictive value (PPV) for optimal cut-off points (plaque stability in VH-IVUS analysis). Cut-off point
Fibrous area Fibro-fatty area Necrotic area Calcified area
62.5 n.s. 15.5 4.5
A
Sensitivity
Specificity
0.87 n.s. 0.70 0.78
0.75 n.s. 0.67 0.75
PPV
72.9 n.s. 63.8 68.7
B
Figure 7.—Mean value of plaque area (A) and plaque burden (B) in percent of stable versus unstable carotid plaques using box-and-whisker-plots with significant differences between the groups (P=0.001 and P=0.002). The boxes demonstrate values between 0.25 and 0.75 quantile with median within the box. The lines (whiskers) show 95% percentiles of the corresponding samples.
Vol. 58 - No. 1
The Journal of Cardiovascular Surgery
61
This document is protected by international copyright laws. No additional reproduction is authorized. It is permitted for personal use to download and save only one file and print only one copy of this Article. It is not permitted to make additional copies (either sporadically or systematically, either printed or electronic) of the Article for any purpose. It is not permitted to distribute the electronic copy of the article through online internet and/or intranet file sharing systems, electronic mailing or any other means which may allow access to the Article. The use of all or any part of the Article for any Commercial Use is not permitted. The creation of derivative works from the Article is not permitted. The production of reprints for personal or commercial use i not permitted. It is not permitted to remove, cover, overlay, obscure, block, or change any copyright notices or terms of use which the Publisher may post on the Article. It is not permitted to frame or use framing techniques to enclose any trademark, log or other proprietary information of the Publisher.
©
COPYRIGHT 2017 EDIZIONI MINERVA MEDICA FUCHS IVUS and virtual histologyS
A
B
Figure 8.—Correlation of plaque area (A) and plaque burden (B) with total vessel area in the total plaque cohort using scattergraph. The figure contains a linear regression line of the best fit between the variables with border lines showing 95% confidence interval (r=0.498 and r=0.542; P=0.001).
mm³ in unstable plaques (P=0.001). Similarly, plaque burden was significantly lower in stable plaques (70.7±7.9%) than in unstable plaques (78.6±5.7%) (P=0.002). Mean plaque area (relatively to vessel area) also showed significantly lower values in stable plaques than in unstable plaques (73.1±7.7% vs. 82.4±5.7%, P=0.001). The mean measured plaque volume of each plaque showed no significant difference in stable and unstable plaques (0.39±0.11 mm³ vs. 0.51±0.34 mm³, P=0.187).���������� The������ relative plaque area of the target lesion correlated with total vessel area (r=0.498, P=0.01) and with the plaque burden (r=0.542, P=0.001). Grayscale and VH-IVUS findings for ruptured plaques, VH-TCFA and non-VH-TCFA plaques showed neither in asymptomatic patients, nor in symptomatic patients any significant differences. Discussion Stroke represents one of the most serious causes of mortality and morbidity in western countries and extra-cranial carotid complications represent approximately one half of all cases. Identification of vulnerable plaques is therefore essential to predict patients at risk. A promising tool to identify such individuals is plaque imaging. However, broad heterogeneity exists in the structure and composition of carotid atherosclerotic plaques. Therefore, it is difficult to recognize the true
62
rupture-prone lesion. Furthermore, many atherosclerotic plaques may cause no symptoms for decades; however, few plaques disrupt and cause thrombosis. Such vulnerable plaques are assumed to be at high short term risk of thrombosis, causing an acute cerebrovascular insult or sudden coronary death.1, 2, 10-12 Up to date, there is no reliable diagnostic method to prospectively identify these vulnerable plaques.11 However, many of the current imaging techniques used to assess carotid artery disease are already able to detect different features of the rupture-prone plaques.13-15 Despite several imaging techniques as well as different serological biomarkers being currently under development, none of them alone provides such all-embracing information to identify individual patients at increased risk of stroke.11 Specific criteria have been recently proposed to assess the relevance of MRI due to high resolution, or IVUS and VH-IVUS that are easy to perform and assess detailed plaque morphology. Nevertheless, these techniques are still under investigation and at present, none of them can reliable identify vulnerable plaques and, most importantly, predict its further development.16 From a clinical point of view, these methods currently assess only some features of atherosclerotic lesions. Thus, combination of several modalities would be of importance in the future to ensure high sensitivity and specificity in detecting vulnerable plaques. In our study we could show high concordance between true histology and VH-IVUS regarding various
The Journal of Cardiovascular Surgery February 2017
This document is protected by international copyright laws. No additional reproduction is authorized. It is permitted for personal use to download and save only one file and print only one copy of this Article. It is not permitted to make additional copies (either sporadically or systematically, either printed or electronic) of the Article for any purpose. It is not permitted to distribute the electronic copy of the article through online internet and/or intranet file sharing systems, electronic mailing or any other means which may allow access to the Article. The use of all or any part of the Article for any Commercial Use is not permitted. The creation of derivative works from the Article is not permitted. The production of reprints for personal or commercial use i not permitted. It is not permitted to remove, cover, overlay, obscure, block, or change any copyright notices or terms of use which the Publisher may post on the Article. It is not permitted to frame or use framing techniques to enclose any trademark, log or other proprietary information of the Publisher.
©
COPYRIGHT 2017 EDIZIONI MINERVA MEDICA IVUS and virtual histologyS FUCHS
atherosclerotic plaque types, symptoms, and plaque instability. Our results demonstrated that composition of plaques from symptomatic patients significantly differs from asymptomatic individuals. The former contain more lipid and cholesterol and less fibrous tissue. Compared to Diethrich et al. we could also demonstrate that fibrous tissue is the most frequent plaque component, followed by fibro-fatty, necrotic, and calcified tissue.17 In our study, the diagnostic accuracy of VH IVUS with true histology in different carotid plaque types was 99.4% in thin-cap fibroatheroma, 96.1% for calcified thin-cap fibroatheroma, 85.9% in fibroatheroma, 85.5% for fibrocalcific, 83.4% in pathological intimal thickening, and 72.4% for calcified fibroatheroma.��������� Furthermore, we detected a higher ratio of dense calcification and necrotic tissue in culprit lesions. Although Granada et al.18 postulated very low specificity and sensitivity of VH-IVUS in coronary lesions, we found high significant results for fibrous, calcified, and also necrotic tissue. Conversely, fibro-fatty tissue was not detected sufficiently, because the current classification cannot differentiate between intramural thrombus and other atherosclerotic components of vessel wall. In a study of Rodriguez-Granillo et al.19 non-culprit and non-obstructive lesions were examined by VH-IVUS in 23 patients with acute coronary syndromes and in 32 patients with stable CAD, demonstrating significantly higher incidence of thin fibrous cap in patients with acute coronary syndromes. Other histological studies have also demonstrated that lesions containing a considerable amount of necrotic core tissue without evidence of an overlying thin fibrous cap lesions may be considered as a precursor of plaque rupture.1, 2, 20, 21 Moreover, we could demonstrate a good correlation between tissue composition and stability of the carotid lesions. For stable and unstable plaques the correlation coefficient was 0.72 for VH-IVUS detection, which indicates a good accuracy for in vivo identification of plaques. Therefore, we can postulate that plaque components determined in carotid plaques reflect the real histological situation. According to the CAPITAL study,17 we showed strong correlation between VH-IVUS classification and stability of carotid atherosclerotic plaques. Nair et al.22 imaged ex vivo 88 plaques from 51 left anterior descending coronary arteries and came to the conclusion that coronary plaque composition can be predicted through the use of IVUS radiofrequency data analysis.
Vol. 58 - No. 1
Limitations of the study Although VH-IVUS is a promising imaging modality for plaque characterization, some limitations still remain. High interobserver variability in the prediction of tissue type by visual assessment of VH-IVUS images was observed in several studies. This underlines the need of quantitative methods for the analysis of the ultrasound characteristics of plaque components. Importantly, detection of the thin fibrous cap (
