Author + information
- Received January 28, 2013
- Revision received March 29, 2013
- Accepted April 29, 2013
- Published online August 1, 2013.
- Tomotaka Dohi, MD, PhD∗,†,
- Gary S. Mintz, MD†,
- John A. McPherson, MD‡,
- Bernard de Bruyne, MD, PhD§,
- Naim Z. Farhat, MD⋮,
- Alexandra J. Lansky, MD¶,
- Roxana Mehran, MD†,#,
- Giora Weisz, MD∗,†,
- Ke Xu, PhD†,
- Gregg W. Stone, MD∗,† and
- Akiko Maehara, MD∗,†∗ ()
- ∗Columbia University Medical Center, New York, New York
- †Cardiovascular Research Foundation, New York, New York
- ‡Vanderbilt University Medical Center, Nashville, Tennessee
- §Cardiovascular Center Aalst, OLV Hospital, Aalst, Belgium
- ⋮North Ohio Heart Center/Elyria Memorial Hospital Regional Medical Center, Elyria, Ohio
- ¶Yale University School of Medicine, New Haven, Connecticut
- #Mount Sinai Medical Center, New York, New York
- ↵∗Reprint requests and correspondence:
Dr. Akiko Maehara, Cardiovascular Research Foundation, 111 East 59th Street, 12th Floor, New York, New York 10022.
Objectives The purpose of this study was to determine the clinical impact of non-fibroatheroma lesion phenotype in patients presenting with an acute coronary syndrome (ACS).
Background Although fibroatheromas (FAs) are known to be clinically unstable, the impact of non-FA lesion phenotype on clinical outcomes has not been studied.
Methods In the PROSPECT (Providing Regional Observations to Study Predictors of Events in the Coronary Tree) study, patients presenting with an ACS underwent 3-vessel grayscale and virtual histology intravascular ultrasound (VH-IVUS) after successful percutaneous intervention for all culprit lesions and were followed for 3 years. Patients were divided into those who had only the non-FA phenotype (pathological intimal thickening or fibrotic and/or fibrocalcific lesions) versus those who had at least 1 nonculprit FA.
Results Among 2,880 nonculprit lesions identified by VH-IVUS, 39.8% were non-FAs (1,042 pathological intimal thickening, 72 fibrotic, and 33 fibrocalcific). Nonculprit major adverse cardiac events (MACE) (death, myocardial infarction, or urgent rehospitalization for progressive or unstable angina) were attributed to only 7 non-FA lesions (0.7%) versus 43 FA lesions (2.7%, p < 0.001) during 3 years follow-up. Of 609 patients, 67 (11.0%) patients had only non-FA lesion phenotypes. Patients with only non-FAs tended to be younger and more often female, have fewer nonculprit lesions and less overall plaque burden and necrotic core, and fewer nonculprit lesion MACE compared with patients with at least 1 FA. In the adjusted Cox proportional hazards model, absence of a FA was a significant predictive of a lower 3-year nonculprit MACE rate (hazard ratio: 0.23; 95% confidence interval: 0.06 to 0.95).
Conclusions Non-FA lesions were clinically stable and were rarely associated with clinical events during 3 years of follow-up. The intermediate-term prognosis in patients presenting with ACS in whom all nonculprit lesions are non-FAs is favorable. (PROSPECT: An Imaging Study in Patients With Unstable Atherosclerotic Lesions; NCT00180466)
- acute coronary syndromes
- intravascular ultrasound
- plaque lesion phenotype
- virtual histology
Plaque composition is a key determinant of the propensity for atherosclerotic lesions to become unstable and result in adverse cardiovascular events (1). Fibroatheromas (FA) are recognized as high-risk plaques, especially thin-cap fibroatheroma (TCFA) (2,3). Conversely, non-FA are considered to represent stable plaques, although the relationship between non–lipid rich plaque (NLRP) or non-FA stable lesion phenotype and future events has not been prospectively studied. The PROSPECT (Providing Regional Observations to Study Predictors of Events in the Coronary Tree) study was a prospective, multicenter, natural history study that used 3-vessel multimodality intracoronary imaging to characterize the coronary tree and the relationship between specific atherosclerotic lesions to long-term follow-up events (4). We performed an analysis using the data from PROSPECT to determine the clinical impact of NLRP phenotype on long-term clinical outcomes in patients presenting with an acute coronary syndrome (ACS).
The design, major inclusion and exclusion criteria, endpoints, and definitions from the PROSPECT study have been described in detail (4,5). In brief, ACS patients (ST-segment elevation myocardial infarction [STEMI] beyond 24 h, non-STEMI, or unstable angina) underwent 3-vessel multimodality coronary artery imaging including quantitative coronary angiography (QCA), grayscale intravascular ultrasound (IVUS) imaging, and radiofrequency virtual histology IVUS (VH-IVUS) (Volcano Corporation, San Diego, California) of the proximal 6 to 8 cm of each of the 3 coronary arteries after performance of successful and uncomplicated percutaneous coronary intervention (PCI) of all coronary lesions responsible for the index event and after completion of any other planned interventions. Patients were treated with guideline-based medical therapy, and clinical follow-up was performed for a median of 3.4 years. The study was approved by the institutional review board at each participating center; and all patients signed written informed consent.
In the present study, based on VH-IVUS analysis, subjects were divided into patients who had only nonculprit pathological intimal thickening (PIT) or fibrotic and/or fibrocalcific lesions (non-FA group) versus patients who had at least 1 nonculprit FA (FA group).
All images were analyzed at independent core laboratories (Cardiovascular Research Foundation, New York, New York) blinded to clinical outcomes. QCA of all culprit and angiographically evident nonculprit lesions (>30% visual angiographic diameter stenosis) was performed using Medis CMS software version 7.0 (Leiden, the Netherlands).
All IVUS cross sections and lesions were co-registered to the angiographic road map using fiduciary side branches for alignment. Grayscale and VH-IVUS analyses were performed using: 1) QCU-CMS (Medis) for contouring; 2) pcVH version 2.1 (Volcano) for contouring and data output; and 3) proprietary software (qVH, Cardiovascular Research Foundation) for segmental qualitative and quantitative output (6,7). Quantitative IVUS measurements for each frame (median interslice distance of 0.40 mm) included external elastic membrane (EEM), lumen, and plaque and media (EEM minus lumen) cross-sectional area (CSA), plaque burden (plaque and media divided by EEM), and minimal lumen area (MLA). A nonculprit IVUS lesion had >3 consecutive IVUS cross sections with ≥40% plaque burden. Qualitative grayscale IVUS morphology included plaque rupture (intraplaque cavity communicating with the lumen with an overlying residual fibrous cap fragment) (8,9).
VH-IVUS color-coded plaque components as fibrous (green), fibrofatty (light green), dense calcium (white), and necrotic core (red) and reported as CSA and percentages of total plaque area (10,11). Each lesion was further classified by VH-IVUS phenotype. A FA had >10% confluent necrotic core. If >30° of the necrotic core abutted the lumen in >3 consecutive frames, the FA was classified as VH-TCFA; otherwise, it was categorized as thick-cap fibroatheroma (ThCFA). PIT contained ≥15% fibrofatty tissue with <10% confluent necrotic core and <10% confluent dense calcium. Fibrotic plaque contained mainly fibrous tissue, and fibrocalcific plaque had >10% confluent dense calcium (6).
Endpoints and definitions
The primary endpoint in PROSPECT was nonculprit lesion–related major adverse cardiac events (MACE): the composite of cardiac death, cardiac arrest, myocardial infarction (MI), or rehospitalization due to unstable or progressive angina. On the basis of follow-up angiography, the primary endpoint was adjudicated by a blinded independent clinical events committee to previously treated versus untreated coronary segments (i.e., nonculprit lesions).
All statistical analyses were performed using SAS version 9.2 (SAS Institute, Cary, North Carolina). Categorical variables were summarized using percentages and counts, and compared using chi-square tests or Fischer exact tests as appropriate. Continuous variables were presented as medians and interquartile ranges (IQRs); comparisons were conducted by the Kruskal-Wallis test.
For comparison between patients with versus without FA lesions, a model with a generalized estimating equations approach was used to compensate for any potential clustering effect of multiple lesions in the same vessels or in the same patients and was developed using a working correlation structure of compound symmetry and a 2-level nested cluster effect for patient and vessel-within-patient using the PROC GENMOD code in SAS, with results presented as least square means with 95% confidence intervals.
Three-year outcomes are displayed as time-to-event curves, compared by the log-rank test. Multivariable modeling was performed by Cox regression model. The following variables were considered for the multivariable model using stepwise selection: absence of a FA, history of PCI before the current ACS, insulin-dependent diabetes mellitus, plaque burden at MLA site >70%, and smallest MLA among all lesions per patient. Values of p < 0.05 were considered statistically significant.
Of the 697 patients enrolled in PROSPECT, 609 patients with complete 3-vessel VH-IVUS data had at least 1 nonculprit lesion. Overall, 542 (89.0%) patients had at least 1 FA, whereas 67 (11.0%) patients had only non-FAs.
The baseline characteristics of these 2 groups are shown in Table 1. There was a trend for patients in the non-FA group to be younger, more often female, and less often have a family history of coronary artery disease compared with the FA group. Statin therapy at discharge and at 3 years was significantly less common in the non-FA group.
Image data on a per-patient level
The total length of coronary arteries analyzed by QCA and volumetric IVUS did not differ between FA and non-FA groups (data not shown). The number of IVUS nonculprit lesions per patient was significantly less in the non-FA than in the FA group (3 [IQR: 2 to 5] vs. 5 [IQR: 4 to 6]; p < 0.001). Patients in the non-FA group had a smaller overall volumetric plaque burden compared with the non-FA group (47.2% [IQR: 44.4% to 49.6%] vs. 49.3% [IQR: 46.8% to 52.1%]; p < 0.001). The smallest MLA in the non-FA group was larger than in the FA group (Table 2). The frequency of patients with at least 1 lesion with a plaque rupture, an MLA <4.0 mm2, or a plaque burden >70% was significantly less in the non-FA than the FA group (Table 2).
By grayscale and VH-IVUS analysis, there were 216 nonculprit lesions in the non-FA group and 2,664 lesions in the FA group. In the non-FA group, 91.2% (197 of 216) of the lesions were characterized as PIT, and 94% (63 of 67) of patients had at least 1 PIT lesion; the rest of the lesions were either fibrotic (n = 14) or fibrocalcific (n = 5). In the FA group, 65.1% (1,773 of 2,664) of the lesions were characterized as either TCFA or ThCFA; 60.5% (368 of 542) of patients had at least 1 VH-TCFA, 87.5% (474 of 542) patients had at least 1 ThCFA, and the rest of the lesions were either PIT (n = 845), fibrotic (n = 58), or fibrocalcific (n = 28).
The percentage of fibrofatty tissue per patient was significantly larger in the non-FA than in the FA group, whereas the percentage of necrotic core and dense calcium was correspondingly less in the non-FA compared with the FA group.
Overall, among the 2,880 nonculprit lesions, 1,147 (39.8%) were non-FA (1,042 PIT, 72 fibrotic, 33 fibrocalcific), and 1,733 (60.2%) were FA (642 VH-TCFA and 1,091 ThCFA) (Table 3).
Overall, 541 lesions had an MLA <4.0 mm2. Of these, 31.8% (172 of 541) were non-FAs (146 PIT, 18 fibrotic, 8 fibrocalcific), and the others were VH-TCFA (n = 140) or ThCFA (n = 229). Plaque burden at the MLA site was significantly smaller in non-FA versus FA lesions (p < 0.001), although the actual MLA was not different.
Overall, 253 lesions had a plaque burden >70%. Of these, 17.0% (43 of 253) were non-FA lesions (39 PIT, 3 fibrotic, 1 fibrocalcific), and the others were VH-TCFA (n = 75) or ThCFA (n = 135). There were no significant differences in plaque burden at MLA sites comparing non-FAs versus FAs (p = 0.25).
As shown in Table 4, the 3-year cumulative nonculprit lesion MACE rates were significantly lower in the non-FA versus the FA patient group (3.7% vs. 12.9%; p = 0.03). Although there was no occurrence of death, MI, or cardiac arrest in the non-FA group and only 6 MIs in the FA group, the frequency of rehospitalization for unstable or progressive angina was significantly lower in the non-FA versus the FA group (3.7% vs. 11.9%; p = 0.04).
The 3-year cumulative MACE rates are shown in Figure 1. Before adjustment, the absence of any FA was significantly associated with a lower rate of nonculprit lesion MACE during 3-year follow-up (hazard ratio: 0.24; 95% confidence interval: 0.06 to 0.97; p = 0.045). By multivariate analysis the absence of any FA was significantly predictive of a lower non-culprit lesion MACE rate during follow-up (hazard ratio: 0.23; 95% confidence interval: 0.06 to 0.95; p = 0.042) (Table 5).
On a lesion-level analysis, the 3-year nonculprit MACE rate attributed to non-FAs versus FAs was 0.7% versus 2.7%, respectively (p < 0.001). Among untreated lesions with an MLA <4 mm2, the 3-year nonculprit MACE rate attributed to non-FAs versus FAs was 2.4% versus 6.9%, respectively (p = 0.04). Similarly, among untreated lesions with a plaque burden >70%, the 3-year nonculprit MACE rate tended to be lower in non-FA versus FA lesion phenotypes (2.6% versus 10.9%; p = 0.10) (Table 3, Fig. 2).
The present PROSPECT substudy demonstrates that nonculprit coronary lesions without the VH-IVUS phenotype of a FA were associated with a low rate of future cardiovascular events, lower than lesions with a FA phenotype. Furthermore, this association remained significant even after adjustment for clinically important covariates reported in the main PROSPECT analysis. As a result, although recurrent culprit lesion–related and unanticipated nonculprit lesion–related events are common in patients with ACS, those patients with only untreated non-FAs had a favorable long-term prognosis with excellent event-free survival.
The DEFER (Deferral of Percutaneous Coronary Intervention), FAME-I (Fractional Flow Reserve Versus Angiography for Multivessel Evaluation), and the registry arm of FAME-II studies have shown that lesions with a fractional flow reserve (FFR) >0.80 remain stable with medical therapy alone and do not need to be treated with PCI (12–14). However, imaging studies, including the PROSPECT study, have mainly identified high-risk or vulnerable plaques likely to cause future events. Yamagishi et al. (8) used grayscale IVUS to study 114 lesions in 106 patients. At a mean follow-up of 21.6 months, 12 lesions developed an acute occlusion, 10 of which had an echolucent zone (8 superficial and 2 deep). The PROSPECT study demonstrated that the triad of a VH-TCFA, an MLA <4.0 mm2, and a plaque burden >70% predicted lesion-specific events during 3 years of follow-up (4). Similar results were reported by the VIVA (VH-IVUS in Vulnerable Atherosclerosis) investigators (15). Uemura et al. (16) studied 53 patients with optical coherence tomography. At 7 months of follow-up, 13 nonsignificant plaques showed angiographic progression. Optical coherence tomography TCFA and microchannels were associated with angiographic progression. Ohtani et al. (17) used angioscopy to study 552 patients. At a mean follow-up of 57.3 months, 39 patients developed an ACS. Although the maximum color grade of the plaques was not different between ACS versus non-ACS patients, the number of yellow plaques was greater in ACS patients.
The current PROSPECT substudy suggests that imaging techniques may also be useful in identifying stable lesions and patients. Of 2,880 nonculprit lesions identified by VH-IVUS, 1,147 were non-FA (1,042 PIT, 72 fibrotic, 33 fibrocalcific). The 3-year MACE rate associated with these lesions was only 0.7%, comparable to that expected from FFR (14). Thus, the current study suggests that deferral of non-FAs may be safe without FFR measurement. The converse, however, is not necessarily true; and future studies are required to determine the event rates in FAs (TCFAs and ThCFAs) that are FFR positive versus negative.
In the present study, the majority of stable non-FAs consisted of the PIT phenotype (91%). The histopathological definition of PIT is intimal thickening with extracellular lipid pools in the absence of necrotic core formation (18). Although precise mechanisms underlying the sequential formation of PIT are unknown, PIT may be the morphological and chemical bridge to more advanced plaques (19). The temporal stability of these lesions has not been thoroughly characterized. A serial VH-IVUS study by Kubo et al. (20) demonstrated that 71% of PITs remained unchanged during 12-month follow-up in patients with stable ischemic heart disease. Thus, maximal medical treatment of lesions with a stable phenotype might be the key to prevent development of more unstable phenotypes. In support of this, the present study showed that patients with all non-FAs remained clinically stable during 3 years of follow-up, even when initially presenting with an ACS.
VH-IVUS phenotype classification depends critically on subjective assessments of plaque morphology that may vary between observers. Wu et al. (21) demonstrated that interobserver and intraobserver variability were sufficiently low for the PIT phenotype (k = 0.90 and k = 0.92, respectively). In addition, the variability for other stable plaque phenotypes—fibrous and fibrocalcific lesions—were also acceptable (k = 0.85 and k = 0.86; k = 0.81 and k = 0.83, respectively). The image analysis in PROSPECT was performed by investigators at a core laboratory who were experienced in grayscale and VH-IVUS interpretation. Therefore, to generalize, the reproducibility of VH-IVUS measurements may be challenging. However, and importantly, the implication of the current study is that <10% necrotic core, which is an automated output of the VH-IVUS algorithm and does not require interpretation, indicates the absence of a FA and, therefore, plaque stability.
Several prior single-center studies also demonstrated sufficient reproducibility between observers, catheters, and repeated pullbacks (22–24). The between-center reproducibility was recently evaluated in 4 European IVUS centers (25). VH-IVUS measurements of plaque and vessel dimensions and plaque composition were highly correlated for the different comparisons. Overall intraclass correlations for fibrous, fibro-lipidic, necrotic core, and calcified volumes were 0.96, 0.94, 0.98, and 0.99, respectively. However, significant differences for both geometrical and compositional measurement were seen. Of the plaque components, fibrous tissue and necrotic core showed, on average, the highest measurement reproducibility.
Our study cohort included only nonculprit lesions in ACS patients and not culprit plaques. The PROSPECT study enrolled a relatively low-risk cohort of ACS patients in whom PCI was successful and uncomplicated. Because low-risk plaques tend to be dispersed throughout the coronary tree (i.e., more distally than the proximal 6- to 8-cm length that was imaged), the actual prevalence of non-FA lesions might be higher than recorded. The ability of VH-IVUS to correctly recognize plaque composition has been debated. Although several validation studies showed high predictive accuracy for evaluating tissue composition versus ex vivo human coronary arteries, the accuracy of VH-IVUS analysis, especially behind calcium, continues to be questioned (6,26). In this regard, it is unclear how often there is a definable signal compared with mostly noise behind calcium from images obtained in vivo. In cases with grayscale shadowing, VH-IVUS often shows the color red, which is an artifact. We did not exclude data behind grayscale calcification. In addition, accuracy may depend on the thickness, density, and amount of calcium. In general, VH-IVUS overestimates, rather than underestimates, the amount of necrotic core; as such, its specificity for stable lesions (i.e., non-necrotic core–containing lesions) should be high. There is also currently no specific VH-IVUS algorithm to identify thrombus (6,11,27). Thrombus appears green or light green (fibrotic or fibrofatty plaque) depending on its age. In the VH-IVUS analysis in the PROSPECT study, no attempt was made to identify thrombus and to exclude it from the overall plaque area. Therefore, thrombus would cause a TCFA to be misclassified as a ThCFA, and a FA to be classified as a non-FA.
Untreated non-FA lesions in human coronary arteries in ACS patients enrolled in the PROSPECT study remained stable during 3 years of follow-up, rarely resulting in adverse cardiovascular events. In particular, after successful PCI in ACS, the intermediate-term prognosis of patients without any untreated FAs is favorable.
Dr. Dohi has received grant support from the Banyu Life Science Foundation International. Dr. Mintz has received grant support from and is a consultant to Boston Scientific and Volcano Corporation. Dr. McPherson is a consultant to CardioDx. Dr. de Bruyne has received grant support from Abbott Vascular, Medtronic, and St. Jude Medical. Dr. Mehran has received grant support from The Medicines Company, Bristol-Myers Squibb/Sanofi, and Eli Lilly and Company/Daiichi Sankyo; and is a consultant to Abbott Vascular, AstraZeneca, Boston Scientific, Covidien, Janssen Pharmaceuticals, Regado Biosciences, Maya Medical, and Merck & Co. Dr. Weisz is a consultant to InfraReDx. Dr. Stone is a consultant to Boston Scientific, InfraReDx, and Volcano Corporation. Dr. Maehara has received grant support from and is a consultant to Boston Scientific. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose. Sherif Nagueh, MD, served as Guest Editor for this paper.
- Abbreviations and Acronyms
- acute coronary syndromes
- cross-sectional area
- external elastic membrane
- fractional flow reserve
- interquartile range
- intravascular ultrasound
- major adverse cardiac event(s)
- myocardial infarction
- minimal lumen area
- percutaneous coronary intervention
- pathological intimal thickening
- quantitative coronary angiography
- ST-segment elevation myocardial infarction
- thin-cap fibroatheroma
- thick-cap fibroatheroma
- virtual histology
- Received January 28, 2013.
- Revision received March 29, 2013.
- Accepted April 29, 2013.
- American College of Cardiology Foundation
- Stary H.C.
- Virmani R.,
- Burke A.P.,
- Farb A.,
- Kolodgie F.D.
- Cheruvu P.K.,
- Finn A.V.,
- Gardner C.,
- et al.
- Maehara A.,
- Cristea E.,
- Mintz G.S.,
- et al.
- Diethrich E.B.,
- Pauliina Margolis M.,
- Reid D.B.,
- et al.
- Yamagishi M.,
- Terashima M.,
- Awano K.,
- et al.
- Nair A.,
- Kuban B.D.,
- Tuzcu E.M.,
- Schoenhagen P.,
- Nissen S.E.,
- Vince D.G.
- Pijls N.H.,
- van Schaardenburgh P.,
- Manoharan G.,
- et al.
- Pijls N.H.,
- Fearon W.F.,
- Tonino P.A.,
- et al.
- Uemura S.,
- Ishigami K.,
- Soeda T.,
- et al.
- Ohtani T.,
- Ueda Y.,
- Mizote I.,
- et al.
- Kolodgie F.D.,
- Burke A.P.,
- Nakazawa G.,
- Virmani R.
- Stary H.C.,
- Blankenhorn D.H.,
- Chandler A.B.,
- et al.
- Kubo T.,
- Maehara A.,
- Mintz G.S.,
- et al.
- Nasu K.,
- Tsuchikane E.,
- Katoh O.,
- et al.