Author + information
- Received September 30, 2008
- Revision received November 10, 2008
- Accepted November 16, 2008
- Published online April 1, 2009.
- Nieves Gonzalo, MD,
- Hector M. Garcia-Garcia, MD, MSc,
- Evelyn Regar, MD, PhD,
- Peter Barlis, MBBS, MPH,
- Jolanda Wentzel, PhD,
- Yoshinobu Onuma, MD,
- Jurgen Ligthart, BsC and
- Patrick W. Serruys, MD, PhD⁎ ()
- ↵⁎Reprint requests and correspondence:
Prof. Patrick W. Serruys, Thoraxcenter, Ba583a, Gravendijkwal 230, 3015 CE Rotterdam, the Netherlands
Objectives This study sought to evaluate the in vivo frequency and distribution of high-risk plaques (i.e., necrotic core rich) at bifurcations using a combined plaque assessment with intravascular ultrasound–virtual histology (IVUS-VH) and optical coherence tomography (OCT).
Background Pathological examinations have shown that atherosclerotic plaque rich in necrotic core is prone to develop at bifurcations. High-risk plaque detection could be improved by the combined use of a technique able to detect necrotic core (IVUS-VH) and a high-resolution technique that allows the measurement of the fibrous cap thickness (OCT).
Methods From 30 patients imaged with IVUS-VH and OCT, 103 bifurcations were selected. The main branch was analyzed at the proximal rim of the ostium of the side branch, at the in-bifurcation segment and at the distal rim of the ostium of the side branch. Plaques with more than 10% confluent necrotic core by IVUS-VH were selected and classified as fibroatheroma (FA) or thin-cap fibroatheroma (TCFA) depending on the thickness of the fibrous cap by OCT (>65 or ≤65 μm for FA and TCFA, respectively).
Results Twenty-seven FA (26.2%) and 18 TCFA (17.4%) were found out of the 103 lesions studied. Overall the percentage of necrotic core decreases from proximal to distal rim (16.8% vs. 13.5% respectively, p = 0.01), whereas the cap thickness showed an inverse tendency (130 ± 105 μm vs. 151 ± 68 μm for proximal and distal rim, respectively, p = 0.05). The thin caps were more often located in the proximal rim (15 of 34, 44.1%), followed by the in-bifurcation segment (14 of 34, 41.2%), and were less frequent in the distal rim (5 of 34, 14.7%).
Conclusions The proximal rim of the ostium of the side branch has been identified as a region more likely to contain thin fibrous cap and a greater proportion of necrotic core.
High-risk atherosclerotic plaques (i.e., rich in necrotic core) are prone to develop at bifurcations because of the specific shear stress conditions present in these regions (1,2). Stented bifurcations lesions represent a complex lesion subset at high risk of restenosis and thrombosis (3,4). These phenomena may reflect certain procedural aspects such as incomplete stent apposition, underexpansion, or gap regions (5,6), but may also be associated with specific compositional and morphological plaque features in these regions. Thin-cap fibroatheroma (TCFA) has been described as the plaque with an increased risk of rupture (7). Such lesions are characterized by a large necrotic core with a thin fibrous cap, usually <65 μm in thickness (8). Recently, it has been reported that TCFA detection could be improved by the combined use of intravascular ultrasound–virtual histology (IVUS-VH) and optical coherence tomography (OCT) (9).
IVUS-VH uses spectral analysis of IVUS radiofrequency data to identify 4 tissue types in the atherosclerotic plaque, among them necrotic core (10). Optical coherence tomography is a high-resolution imaging modality that uses reflected near-infrared light, allowing a very precise visualization and measurement of vascular microstructures such as the fibrous cap (11). To our knowledge, in vivo characterization of necrotic core rich plaques at bifurcation regions has not been explored. The objective of the present study was therefore to evaluate in vivo the frequency and distribution of high-risk plaques at bifurcation lesions using a combined plaque assessment with IVUS-VH and OCT.
All of the patients admitted to our hospital between January 2005 and March 2008 in whom IVUS-VH and OCT were performed in the same vessel were investigated for bifurcations adequately visualized by both imaging techniques. The indication for the IVUS-VH and OCT was the assessment of intermediate, nonflow-limiting lesions by angiography or post stent implantation assessment. Only regions located more than 5 mm beyond the stent were included. All patients gave written informed consent.
The IVUS was performed using the Eagle Eye 20 MHz catheter (Volcano Corp., Rancho Cordova, California) with an automatic continuous pullback at a rate of 0.5 mm/s. Grayscale images and radiofrequency data required for VH analysis were acquired during the same pullback. The VH processing was performed offline with pcVH 2.1 software (Volcano Corp.) that permits semiautomated contour detection and provides the compositional structure of the vessel. The IVUS-VH uses spectral analysis to classify the 4 different components of the atherosclerotic plaque and gives a color-coded map distinguishing between fibrous tissue (green), fibrofatty tissue (light green), necrotic core (red), and dense calcium (white).
The OCT acquisition was performed using a commercially available system for intracoronary imaging and a 0.019-inch ImageWire (LightLab Imaging, Westford, Massachusetts). Red blood cells represent a nontransparent tissue causing multiple light scattering and substantial signal attenuation. Therefore, for adequate OCT image acquisition blood must be temporarily removed from the vessel. In 77% of the cases this was achieved with the occlusion technique in which a proximal, low-pressure (0.4 atm) occlusion balloon (Helios, Goodman Inc., Nagoya, Japan) is inflated with simultaneous distal flush delivery (lactated Ringer solution; flow rate 0.8 ml/s) to remove blood from the vessel lumen. Images were acquired during a pullback rate of 1.0 mm/s. The possibility to increase the pullback speed up to 3 mm/s in the new OCT system permitted 23% of the cases to be acquired exclusively using a nonocclusive technique in which the blood was removed by the continuous injection of contrast (Iodixanol 370, Visipaque, GE Health Care, Cork, Ireland) through the guiding catheter. The nonocclusive technique reduces the procedural time and the incidence of chest pain and electrocardiographic changes during image acquisition without affecting the image quality (12).
Bifurcation selection and analysis
Simultaneous visual assessment of IVUS-VH and OCT pullbacks, in 2 contiguous screens, allowed the selection of all bifurcations that could be identified with both techniques (13). To ensure proper matching between 2 imaging modalities that have different lateral resolutions (20 μm for OCT and 300 μm for IVUS) and depth penetration, a strict selection of the frames was followed. Only the main branch was analyzed. The lesion analysis included: 1) proximal rim of the ostium of the side branch cross-section (first frame proximal to the take-off of the side branch); 2) in-bifurcation cross-section (frame with the larger ostial diameter of the side branch); and 3) distal rim of the ostium of the side branch cross-section (first frame distal to the take off of the side branch) (Fig. 1). In each bifurcation the plaque location in relation to the flow divider was analyzed (Fig. 2).
Plaque type classification
Two experienced observers jointly analyzed the IVUS-VH data and the OCT measurements in the selected frames to characterize the plaque type according to the following hierarchical classification (7,14,15) (Figs. 3 to 5):⇓⇓
1. Adaptive intimal thickening (AIT): intimal thickening of <600 μm for <20% of the circumference.
2. Pathological intimal thickening (PIT): intimal thickening ≥600 μm for more than 20% of the circumference with more than 15% of fibrofatty tissue, and no confluent necrotic core or dense calcium.
3. Fibrotic plaque: plaque constituted predominantly by fibrous tissue without confluent necrotic core (NC) or dense calcium.
4. Fibrocalcific plaque: more than 10% of confluent dense calcium without confluent NC.
5. Fibroatheroma (FA): plaque characterized by the presence of more than 10% confluent NC covered by a fibrous cap thicker than 65 μm.
6. Calcified fibroatheroma (CaFA): fibroatheroma that contains more than 10% of confluent dense calcium.
7. IVUS/OCT-derived thin-capped fibroatheroma (TCFA): defined as the presence of more than 10% confluent NC at the lumen covered by a thin fibrous cap (<65 μm).
8. IVUS/OCT-derived calcified thin-capped fibroatheroma (CaTCFA): TCFA that contains more than 10% of confluent dense-calcium.
FAs and TCFAs are considered high-risk plaques in American Heart Association and Virmani classifications (7,16). The necrotic core was considered confluent when it was forming a major pool. This definition was used to avoid misclassification as high-risk plaques of lesions with isolated islets or individual pixels of necrotic core, which can be artifacts. The presence of confluent NC >10% at the lumen was measured using dedicated in house developed software (MATLAB MathWorks, Natick, Massachusetts) (9). A validation with pathology of a similar algorithm was reported in the CAPITAL (Carotid Artery Plaque Virtual Histology Evaluation) study (14). The diagnostic accuracy of IVUS-VH compared with histology in different carotid plaque types was 99.4% for TCFA, 96.1% for CaTCFA, 85.9% in FA, 85.5% for fibrocalcific, 83.4% in PIT, and 72.4% for CaFA. To overcome the limitations of IVUS-VH in the fibrous cap evaluation, the classification used in our study combines the information about plaque composition provided by IVUS-VH and the measurements of fibrous cap as assessed by OCT (9). The thinnest part of the fibrous cap was measured by OCT in all the plaques that contain more than 10% of confluent NC to distinguish between FA and TCFA. The fibrous cap measurement by OCT was guided by the IVUS-VH to avoid misclassification between lipid pools and calcium. The cap was measured in the area where the NC was closer to the lumen. The reproducibility of fibrous cap measurements has previously been reported (12). If different morphologies were present along the lesion, the highest degree plaque was established as the definite plaque type.
Statistical analyses were performed using SPSS 12.0.1 for Windows (SPSS Inc., Chicago, Illinois). Continuous variables are expressed as mean ± SD. Categorical variables are expressed as percentages. The bifurcation (lesion) was the unit of analysis without corrections for correlated observations in the same subjects. To compare continuous variables between lesions, t test or analysis of variance was used. To compare continuous variables between different segments of the bifurcation, paired samples t test or Wilcoxon signed ranks test for 2 dependent samples was used. Comparison among groups for categorical variables was made with the chi-square method.
Clinical and procedural characteristics
One hundred and three bifurcations were selected in 32 pullbacks performed in 30 patients. The mean age was 60 ± 7 years, and 73% were male. Regarding cardiac risk factors, 48.3%, 6.9%, and 72.4% had hypertension, diabetes mellitus, and hyperlipidemia, respectively, and 27.6% were smokers. Seven percent had prior myocardial infarction, and 43.3% had undergone a prior percutaneous coronary intervention. The clinical presentation was stable angina in 86.7%, unstable angina in 10%, and acute myocardial infarction in 3.3%. The investigated vessel was the left anterior descending artery (LAD) in 34%, left circumflex artery (LCX) in 33% and right coronary artery (RCA) in 33% of the cases. The type of bifurcations studied were LAD/diagonal, LAD/septal branch, LCX/marginal, and RCA/right ventricular branch in 21.4%, 12.6%, 33%, and 33% of the cases, respectively. The indication for IVUS-VH and OCT was assessment of intermediate, non–flow limiting lesions by angiography in 40% of the cases and post-stent implantation assessment in the remaining 60% of cases. All studied lesions were considered nonsignificant by angiographic criteria and had a lumen area >4 mm2 by IVUS. Overall the mean vessel and lumen area and plaque burden were 13.9 ± 3.7 mm2, 7.7 ± 2.2 mm2, and 43.2 ± 13%, respectively. Table 1 shows the number of quadrants containing plaque and the location of the thickest part of the plaque in relation to the side branch.
Frequency of plaque type
The plaque type was analyzed in 103 lesions and 293 cross-sections. Eight distal and 8 proximal rims could not be analyzed because of artifacts. Overall, the frequency of each plaque type per lesion was: AIT 20 (19.4%), PIT 16 (15.5%), fibrocalcific 15 (14.6%), fibrotic 7 (6.8%), FA 8 (7.8%), CaFA 19 (18.4%), TCFA 10 (9.7%), and CaTCFA 8 (7.8%). In the analyzed cross-sections the distribution was as follows: AIT 76 (25.9%), PIT 53 (18.1%), fibrocalcific 42 (14.3%), fibrotic 28 (9.6%), FA 18 (6.1%), CaFA 42 (14.3%), TCFA 21 (7.2%), and CaTCFA 13 (4.4%).
NC and cap thickness distribution
The NC and cap thickness distribution in the proximal rim, in-bifurcation, and distal rim cross-sections are shown in Figure 6. Overall, the mean NC area and mean percentage of NC decreased from proximal to distal, whereas the mean cap thickness showed an inverse tendency.
High-risk plaque distribution and compositional and geometrical analysis
The distribution of the different plaque types in relation to the location is shown in Table 2.Figure 7 shows the number of cross-sections with high-risk plaque morphology in the proximal rim, in-bifurcation, and distal rim. The thin caps were more often located in the proximal rim (15 of 34, 44.1%), followed by the in-bifurcation (14 of 34, 41.2%), and were less frequent in the distal rim (5 of 34, 14.7%). The location of the thin cap in the 18 lesions classified as TCFAs was as follows: In 4 cases the thinning of the cap extended from the proximal rim into the distal rim, in 7 cases the thin cap was located in the proximal rim and in the in-bifurcation; in 3 TCFAs the thinning of the cap was located only in the in-bifurcation and, in 3 cases only in the proximal rim; there was 1 TCFA that presented the thin cap at both rims whereas the cap at the in-bifurcation frame was thicker. There were no cases in which the thin cap was located only in the distal rim. The mean cap thickness in proximal rim, in-bifurcation, and distal rim in TCFAs is shown in Figure 8.
The vessel area and the plaque burden was significantly greater in the subgroup of lesions considered high risk (FA, CaFA, TCFA, and CaTCFA) than in the rest of lesions (16.5 ± 3 mm2 vs. 14.1 ± 3 mm2, p = 0.002 and 55 ± 9% vs. 42 ± 12%, p < 0.001 for vessel area and plaque burden, respectively), whereas the lumen area was not significantly different (6.8 ± 2 mm2 vs. 7.1 ± 2 mm2, p = 0.39). Table 3 shows the plaque burden, NC percentage, and cap thickness in FA, CaFA, TCFA, and CaTCFA.
Plaque type in relation to the clinical presentation
As exploratory analysis, patients with acute coronary syndromes (ACS) had a significant higher proportion of lesions with high-risk morphology plaques than stable patients (17 of 23 [73.9%] vs. 28 of 80 [35%], p = 0.002). Specifically the number of lesions with TCFA morphology was 6 of 23 (26.1%) vs. 12 of 80 (15%), p = 0.2, for ACS and stable patients respectively, and the number of lesions with FA morphology was 11 of 23 (47.8%) vs. 16 of 80 (20%), p = 0.01, for ACS and stable patients respectively.
To our knowledge this is the first in-vivo study evaluating the frequency and distribution of high-risk plaques at bifurcations in coronary arteries using a combined plaque assessment with IVUS-VH and OCT. The main finding is that the thinning of the fibrous cap occurs more often in the proximal rim of the ostium of the side branch. The NC shows also a differential distribution along the bifurcation being higher at the proximal rim.
Multi-modality plaque assessment
The detection of plaques at potentially high risk of rupture could prevent future occurrence of ACS. At present, multiple techniques are available that evaluate different aspects of the atherosclerotic plaque, such as its structure, composition, or mechanical properties (17). The combined information provided by the different methods is essential for better identification of high-risk coronary lesions. In this study, a combined approach with IVUS-VH and OCT was used. The IVUS-VH allows the detection and quantification of NC, one of the main components of high-risk plaques. However, the limited axial resolution of IVUS (approximately 200 μm) does not permit an accurate evaluation of the fibrous cap. On the contrary, OCT has a very high resolution (10 to 20 μm), allowing a very precise measurement of the fibrous cap. Sawada et al. (9) recently reported that the combined use of IVUS-VH and OCT improved the accuracy for TCFA detection. They studied 126 plaques in 56 patients with angina. Of the 61 plaques diagnosed as TCFA by IVUS-VH criteria, only 28 had a thin fibrous cap, as measured by OCT. In addition, 8 OCT-derived TCFA did not have NC in the VH analysis, mainly because of the misinterpretation in the OCT analysis caused by dense calcium. This source of error in plaque characterization by OCT has previously been described and indicates the difficulty in identifying TCFAs using only OCT (18). To avoid this misclassification in our study, the measurement of the fibrous cap in OCT was guided by IVUS-VH. At the present stage neither modality independently is sufficient for detecting the highest risk plaques; the combined approach seems to be mandatory for the accurate diagnosis of TCFA in vivo.
High-risk plaque frequency, distribution, composition, and geometrical analysis
The in vivo frequency of TCFAs at bifurcations is not known. In our study, in which a highly selected population was included, 18 of 103 bifurcations presented TCFAs (17%). Considering that the number of bifurcations in the complete coronary tree is approximately 15 (19), the frequency of TCFAs per heart would be approximately 2.5. This is in agreement with reported pathological data (1). The bifurcation left main/LAD has been studied with IVUS-VH showing that the plaque burden and the amount of necrotic core are higher in the LAD than in the left main artery. However, plaque morphology in the different segments of other bifurcations has not been previously evaluated in vivo. The present study, which extended the assessment of bifurcation lesions beyond the left main artery, showed that the amount of necrotic core is higher at the proximal rim of the ostium of the side branch with a thin fibrous cap more often identified in the proximal rim. The fibrous cap thickness is determined by the balance between matrix synthesis by the smooth muscle cells and matrix degradation by metalloproteinases (MMP) produced by macrophages (20). It has been shown that the distribution of inflammatory cells in atherosclerotic plaques relates to the direction of the flow with higher concentration of macrophages and MMP activity in the upstream or proximal part (21). One of the mechanisms that have been proposed for the differential distribution of the high-risk plaques along the artery is the influence of endothelial shear stress. Bifurcations are geometrically irregular regions in which disturbed laminar flow occurs, generating abnormal endothelial shear stress patterns that may play a role in plaque destabilization. Previous studies showed that atherosclerosis develops preferentially at low shear stress locations such as the outer wall of bifurcations (2). However, in our data the thickest part of the plaque did not show a preferential location for this region.
Different pathological and IVUS studies have confirmed that plaque rupture occurs usually at sites of significant plaque accumulation associated with positive remodeling (2,22). This is in agreement with our data showing that the vessel area and the plaque burden were higher in high-risk plaques compared with stable lesions. In line with pathological series, TCFAs in our study were located in areas with nonsignificant lumen compromise, with a mean luminal area of 6.8 ± 2 mm2 (8). Similarly, NC in TCFAs in this population is concordant with previously published pathological findings (8).
High-risk plaques in relation to clinical presentation
Although the comparison between ACS and stable patients in this study was exploratory, patients with unstable clinical presentation showed a high-risk profile of plaque types at bifurcations with a higher proportion of TCFAs and FA. This is in agreement with previous data regarding TCFA detection with IVUS-VH showing that ACS patients present a significantly higher prevalence of IVUS-derived TCFA than stable patients (10).
Currently VH-derived necrotic core rich plaques can only be considered as allegedly high-risk lesions because it has not been shown whether they are associated with a higher incidence of clinical events at follow-up. In the present study, OCT and VH were restricted to 1 or 2 vessels; therefore, and unlike pathological studies, this does not allow us to draw conclusions on the incidence of TCFA in bifurcations within the complete coronary tree. No comparison with nonbifurcation lesions was performed. Still the detailed plaque assessment combining 2 imaging modalities has given the first insight into the distribution of these allegedly high-risk lesions at bifurcations.
This study has given unique, in vivo data on the localization of plaque occurring at bifurcation lesions. Further, the proximal rim of the ostium of the side branch has been identified as a region more likely to contain thin fibrous cap and a greater proportion of necrotic core.
- Abbreviations and Acronyms
- acute coronary syndrome
- adaptative intimal thickening
- calcified fibroatheroma
- calcified thin-cap fibroatheroma
- intravascular ultrasound–virtual histology
- left anterior descending artery
- left circumflex artery
- matrix metalloproteinase
- necrotic core
- optical coherence tomography
- pathological intimal thickening
- right coronary artery
- thin-cap fibroatheroma
- Received September 30, 2008.
- Revision received November 10, 2008.
- Accepted November 16, 2008.
- American College of Cardiology Foundation
- Virmani R.,
- Burke A.,
- Farb A.,
- Kolodgie F.D.,
- Finn A.V.,
- Gold H.K.
- Al Suwaidi J.,
- Berger P.B.,
- Rihal C.S.,
- et al.
- Costa R.A.,
- Mintz G.S.,
- Carlier S.G.,
- et al.
- Alfonso F.,
- Suarez A.,
- Perez-Vizcayno M.J.,
- et al.
- Virmani R.,
- Kolodgie F.D.,
- Burke A.P.,
- Farb A.,
- Schwartz S.M.
- Virmani R.,
- Burke A.P.,
- Farb A.,
- Kolodgie F.D.
- Sawada T.,
- Shite J.,
- Garcia-Garcia H.M.,
- et al.
- Rodriguez-Granillo G.A.,
- Garcia-Garcia H.M.,
- Mc Fadden E.P.,
- et al.
- Kume T.,
- Akasaka T.,
- Kawamoto T.,
- et al.
- Diethrich E.B.,
- Margolis P.,
- Reid D.B.,
- et al.
- Garcia-Garcia H.M.,
- Gonzalo N.,
- Barlis P.,
- Serruys P.W.
- Stary H.C.,
- Chandler A.B.,
- Dinsmore R.E.,
- et al.
- Scanlon P.J.,
- Faxon D.P.,
- Audet A.M.,
- et al.
- Dirksen M.T.,
- van der Wal A.C.,
- van den Berg F.M.,
- van der Loos C.M.,
- Becker A.E.