Author + information
- Received August 10, 2011
- Revision received December 13, 2011
- Accepted January 17, 2012
- Published online March 1, 2012.
- Joanna J. Wykrzykowska, MD, PhD⁎,†,
- Gary S. Mintz, MD‡,⁎ (, )
- Hector M. Garcia-Garcia, MD, PhD⁎,
- Akiko Maehara, MD‡,
- Martin Fahy, MSc‡,
- Ke Xu, PhD‡,
- Andres Inguez, MD§,
- Jean Fajadet, MD∥,
- Alexandra Lansky, MD¶,
- Barry Templin, MBA#,
- Zhen Zhang, PhD#,
- Bernard de Bruyne, MD⁎⁎,
- Giora Weisz, MD‡,
- Patrick W. Serruys, MD, PhD⁎ and
- Gregg W. Stone, MD‡
- ↵⁎Reprint requests and correspondence:
Dr. Gary S. Mintz, Cardiovascular Research Foundation, 111 East 59th Street, 11th Floor, New York, New York 10022
Objectives In this substudy of the PROSPECT (Providing Regional Observations to Study Predictors of Events in the Coronary Tree) study, we examined the longitudinal distribution of atherosclerotic plaque burden, virtual histology–intravascular ultrasound (VH-IVUS) characterized necrotic core (NC) content and VH–thin-cap fibroatheroma (TCFA) distribution in nonculprit lesions of patients presenting with acute coronary syndromes.
Background Previous analyses suggested that vulnerable plaques and acute myocardial infarction may occur more frequently in the proximal than the distal coronary tree.
Methods A total of 4,234 proximal, mid, and distal 30-mm-long segments of each epicardial coronary artery were compared with each other and to the left main coronary artery (LMCA).
Results Combining IVUS data from all 3 arteries, there was a gradient in plaque burden from the proximal (42.4%) to mid (37.6%) to distal (32.6%) 30-mm-long segments (p < 0.0001). Overall, 67.4% of proximal, 41.0% of mid, and 29.7% of distal 30-mm-long segments contained at least 1 lesion (plaque burden >40%). Proportion of NC, however, was similar in the proximal and mid 30-mm-long segments of all arteries (10.3% [interquartile range (IQR): 4.8% to 16.7%] vs. 10.6% [IQR: 5.0% to 18.1%], p = 0.25), but less in the distal 30-mm-long segment (9.1% [IQR: 3.7% to 17.8%], p = 0.03 compared with the proximal segment and p = 0.003 compared with the mid segment). Overall, 17.3% of proximal, 11.5% of mid, and 9.1% of distal 30-mm-long segments had at least 1 lesion that was classified as VH-TCFA (p < 0.0001). Comparing the LMCA with the combined cohort of proximal left anterior descending, left circumflex, and right coronary artery 30-mm-long segments: 1) plaque burden was less (35.4% [IQR: 28.8% to 43.5%] vs. 40.9% [IQR: 33.3% to 48.0%], p < 0.0001); 2) fewer LMCAs contained at least 1 lesion (17.5%, p < 0.0001); 3) there was less NC (6.5% [IQR: 2.9% to 12.2%] vs. 9.3% [IQR: 4.3% to 15.9%], p < 0.0001); and 4) LMCAs rarely contained a VH-TCFA (1.8%, p < 0.0001).
Conclusions The current analysis appears to confirm that lesions that are responsible for acute coronary events (large, plaque burden–rich in NC) are somewhat more likely to be present in the proximal than the distal coronary tree, except for the LMCA.
The majority of acute coronary syndromes are caused by coronary plaque rupture at the site of a thin-cap fibroatheroma (TCFA) with subsequent local thrombosis (1,2). The PROSPECT (Providing Regional Observations to Study Predictors of Events in the Coronary Tree) study was the first multicenter, natural history study that employed angiography and grayscale and radiofrequency virtual histology-intravascular ultrasound (VH-IVUS) to relate site-specific quantitative and qualitative measures of coronary disease to major adverse cardiac events at 3 years (3). In this substudy of the PROSPECT study, we examined the longitudinal distribution of atherosclerotic plaque burden, nonculprit lesions, VH-IVUS necrotic core (NC) content, and TCFA plaque phenotype in patients presenting with an acute coronary syndrome (ACS).
Patient population and IVUS image acquisition
The enrollment criteria and the methodology of the PROSPECT study have been previously described in detail (3). In brief, 697 patients with ACS who underwent successful percutaneous coronary intervention of all lesions responsible for the ACS event and other angiographically significant stenoses (culprit lesions) underwent grayscale and VH-IVUS examination of the proximal 6 to 8 cm of the coronary vessels. Imaging was performed using a synthetic aperture array, 20 MHz, 3.2-F catheter (Eagle Eye, Volcano Corporation, Rancho Cordova, California). During a motorized catheter pullback at 0.5 mm/s, grayscale IVUS was recorded, and radiofrequency data were captured gated to the R-wave (In-Vision Gold, Volcano Corporation). In contrast to conventional grayscale IVUS, VH-IVUS uses spectral analysis in addition to amplitude analysis to classify plaque into 4 components: NC, fibrofatty plaque, fibrotic plaque, and dense calcium. This classification has been correlated to histological samples with high specificity and sensitivity (4).
Qualitative and quantitative coronary angiographic assessment of the entire length of the coronary tree was performed at an independent core laboratory (Cardiovascular Research Foundation, New York, New York) using a proprietary methodology modified from standard Medis CMS software (version 7.0, Leiden, the Netherlands); this analysis included each major epicardial coronary artery and every side branch ≥1.5 mm in diameter. This 3-vessel angiographic analysis served as a roadmap to identify each lesion on the basis of longitudinal axis location (mm). The methodology and classification of the lesions is described in this issue of iJACC by Maehara et al. (5).
All IVUS images were also analyzed at an independent core laboratory (Cardiovascular Research Foundation). Offline grayscale and VH-IVUS analyses were performed using: 1) QCU-CMS (Medis) for contouring; 2) pcVH 2.1 software (Volcano Corporation) for contouring and data output; and 3) proprietary qVH software for segmental qualitative assessment and quantitative data output. External elastic membrane (EEM) and lumen borders were contoured for all recorded frames (median interslice distance = 0.40 mm). Quantitative IVUS measurements included EEM, lumen, plaque and media (P&M, defined as EEM minus lumen) cross-sectional area (CSA), and plaque burden (P&M divided by EEM CSA). The slice with minimal lumen CSA and the slice with maximum NC CSA were identified. VH-IVUS plaque components were color coded as dense calcium (white), NC (red), fibrofatty plaque (light green), and fibrotic plaque (dark green) and reported as CSA and percentages of total plaque CSA (4). Volumetric data were calculated using Simpson's rule and normalized for analysis length. A lesion was defined as a segment with ≥3 consecutive frames with ≥40% plaque burden, and each lesion was classified as: 1) VH-TCFA; 2) thick-cap fibroatheroma; 3) pathological intimal thickening; 4) fibrotic plaque; or 5) fibrocalcific plaque (3,6). Grayscale and VH-IVUS frames were co-registered to the angiographic roadmap using fiduciary branch points to align the imaging modality outputs. Based on motorized transducer pullback, the left anterior descending (LAD), left circumflex (LCX), and right coronary (RCA) arteries were divided into 30-mm-long segments beginning at the ostium of each vessel. Lesions were assigned to proximal, mid, or distal 30-mm-long segments depending on the location of the minimum lumen CSA. When assessing the longitudinal distribution of IVUS-identified lesions, distances were measured from the ostium of the coronary artery to the minimum lumen slice. When assessing longitudinal distribution of VH-TCFAs, distances were measured from the ostium of the coronary artery to the slice with the greatest amount of NC within the VH-TCFA–containing lesion. The left main coronary artery (LMCA) was analyzed separately.
Categorical variables are presented using frequencies and percentages and compared using chi-square statistics, and continuous variables are presented as median (interquartile range [IQR]) and compared using the Mann-Whitney U test. All statistical analyses were performed using SAS version 9.1.3 (SAS Institute, Cary, North Carolina).
Baseline characteristics of the 697 patients enrolled in the study are shown in Table 1. Briefly, the median age of the patients was 58.1 years, 24% were women, and 17.2% had diabetes.
Plaque burden gradient
Overall, 4,234 coronary artery segments, each 30 mm long, were assessed using grayscale IVUS. In the combined analysis of 4,234 segments, EEM CSA measured 13.09 [IQR: 8.88 to 17.57] mm2, P&M CSA measured 5.08 [IQR: 2.87 to 7.69] mm2, and plaque burden measured 38.7% [IQR: 30.6% to 46.6%].
Combining IVUS data from all 3 arteries, there was a gradient in the median plaque burden from the proximal (42.43%) to the mid (37.58%) to the distal (32.59%) 30-mm-long segments (p < 0.0001) (Table 2). This gradient was also maintained especially in the LAD and LCX (both p < 0.0001). Comparing the LAD versus the LCX versus the RCA, plaque burden in the proximal 30-mm-long segment was greatest in the LAD (vs. the LCX and RCA), whereas plaque burden in the middle and distal 30-mm-long segments was greatest in the RCA (vs. the LAD and LCX) (Table 3).
Overall, by grayscale IVUS, there were 2,645 lesions: 1,463 in proximal 30-mm-long segments, 773 in mid 30-mm-long segments, and 409 in distal 30-mm-long segments. The median distance from the coronary ostium to the minimum lumen area site measured 27.4 [IQR: 11.2 to 47.8] mm. Overall, 67.4% of proximal 30-mm-long segments, 41.0% of mid 30-mm-long segments, and 29.7% of distal 30-mm-long segments contained at least 1 lesion. However, the proximal to distal lesion gradient was least in the RCA compared with the LAD and LCX, with no difference in lesion density present between the mid and distal RCA segments (Table 4).
Overall, 3,466 coronary artery segments, each 30 mm long, were assessed using VH-IVUS. In the combined analysis of all 3,466 segments, the proportion of NC measured 10.3% [IQR: 4.6% to 17.4%], dense calcium measured 3.9% [IQR: 1.5% to 8.6%], fibrofatty plaque measured 20.4% [IQR: 12.6% to 29.9%], and fibrotic plaque measured 60.2% [IQR: 53.8% to 66.0%].
Combining VH-IVUS analysis from all 3 arteries, the proportion of NC was similar in the proximal and mid 30-mm-long segments (10.3% [IQR: 4.8% to 16.7%] vs. 10.6% [IQR: 5.0% to 18.1%], p = 0.25), but was somewhat less in the distal 30-mm-long segment (9.1% [IQR: 3.7% to 17.8%], p = 0.03 compared with the proximal segment and p = 0.003 compared with the mid segment) (Table 2). A similar pattern was seen when the LAD, LCX, and RCA were assessed separately (Table 5). Comparing the LAD versus the LCX versus the RCA, percent NC was greatest in the LAD in each 30-mm-long segment (Table 5).
Overall, there were 609 VH-TCFA–containing lesions: 314 in proximal 30-mm-long segments, 191 in mid 30-mm-long segments, and 104 in distal 30-mm-long segments. In these lesions, the percent NC measured 21.8% [IQR: 15.9% to 27.2%], and the distance from the coronary ostium to the maximum NC site measured 28.8 [IQR: 13.3 to 46.7] mm.
Overall, 18.9% of the 1,492 proximal 30-mm-long segments, 12.5% of the 1,375 mid 30-mm-long segments, and 9.1% of the 899 distal 30-mm-long segments had at least 1 lesion that was classified as a VH-TCFA. However, the proximal to distal VH-TCFA–containing lesion gradient was significant only in the LAD and LCX, but not in the RCA (Table 6). Because lesions classified as a VH-TCFA were equally distributed in the proximal, mid, and distal RCA, VH-TCFA lesions were more common in the distal RCA than in the distal LAD or LCX. Figure 1 illustrates VH-TCFA distribution along the 3 main coronary arteries every 10 mm.
LMCA versus proximal segments of the LAD, LCX, and RCA
Comparing the LMCA (n = 550) to the combined cohort of proximal LAD, LCX, and RCA 30-mm-long segments (n = 2,233) imaged using grayscale IVUS analysis, the EEM CSA (23.77 [IQR: 20.54 to 27.58] mm2 vs. 18.08 [IQR: 14.15 to 22.83] mm2, p < 0.0001) and P&M CSA were larger (8.49 [IQR: 6.38 to 10.80] mm2, p < 0.0001); however, plaque burden was less (35.4% [IQR: 28.8% to 43.5%] vs. 40.9% [IQR: 33.3% to 48.0%], p < 0.0001) in the LMCA than in the proximal segments of the other epicardial arteries. This was also true when the LMCA was compared with the proximal 30 mm of the LAD, LCX, and RCA separately (all p < 0.0001). Overall, 17.5% (n = 96) of the 550 LMCA segments contained at least 1 lesion compared with 67.4% of the 1,683 proximal 30-mm-long segments of the LAD, LCX, and RCA (p < 0.0001).
Comparing the LMCA (n = 496) to the combined cohort of proximal LAD, LCX, and RCA 30-mm-long segments (n = 1,974) imaged using VH-IVUS analysis, the LMCA had a smaller proportion of NC (6.5% [IQR: 2.9% to 12.2%] vs. 9.3% [IQR: 4.3% to 15.9%], p < 0.0001) and dense calcium (2.0% [IQR: 0.6% to 4.6%] vs. 3.5% [IQR: 1.3% to 7.7%], p < 0.0001), but more fibrofatty plaque (30.1% [IQR: 21.8% to 40.5%] vs. 24.3% [IQR: 15.8% to 34.6%], p < 0.0001) and fibrotic plaque (56.9% [IQR: 51.2% to 62.9%] vs. 58.3% [IQR: 52.3% to 63.4%], p = 0.004) than the proximal segments of the other 3 epicardial arteries. This was also true when the LMCA was compared with the proximal 30 mm of the LAD, LCX, and RCA separately (all p < 0.0001). Overall, only 1.8% (n = 9) of 496 LMCA segments contained a lesion that was classified as a VH-TCFA compared with 18.9% of the 1,492 proximal 30-mm-long segments of the LAD, LCX, and RCA (p < 0.0001).
The main findings of this PROSPECT substudy are the following: a gradient was observed in the degree of plaque burden, lesion distribution, percent NC, and VH-TCFA frequency from proximal to distal in the 3 major epicardial coronary arteries that was greatest in the LAD and LCX and least in the RCA. The LMCA had a smaller plaque burden, less NC, fewer lesions, and fewer VH-TCFAs compared with the proximal 30 mm of the LAD, LCX, and RCA.
Plaque burden, lesion, NC, and VH-TCFA gradients among the LAD, LCX, and RCA
Previous autopsy, angiographic, IVUS, and optical coherence tomography analyses have reported that atherosclerosis accumulates to a greater degree in the proximal than the distal coronary tree (7–14). Other retrospective studies have suggested that symptomatic coronary artery occlusions also have a proximal coronary predominance. Gibson et al. (11) showed that the first 60 mm of the coronary artery contained 75% of culprit ST-segment elevation myocardial infarction (STEMI) lesions, with the median distance being roughly similar in the LAD (40 mm), LCX (43 mm), and RCA (47 mm). Wang et al. (12) reported that acute coronary occlusions are preferentially located in the proximal third of each of the major epicardial arteries, decreasing distally in all 3 vessels such that the risk diminished by 30% per 10-mm segment in the LAD, 26% per 10-mm segment in the LCX, and 13% per 10-mm segment in the RCA. Conversely, Hong et al. (13), using grayscale 3-vessel IVUS analysis, and Fujii et al. (14), using 3-vessel optical coherence tomography, reported that plaque ruptures and TCFAs were more diffusely and distally distributed in the RCA than the LAD or LCX. This disparity between the clinical and IVUS/ optical coherence tomography studies may be attributed to the paucity of branches in the RCA; the limited number of branches in the RCA may allow retrograde propagation of thrombus from ruptured TCFAs in the distal RCA, thereby angiographically appearing as a mid or proximal RCA occlusion.
The present prospective study confirms and extends the results of these prior studies, verifying that the absolute and relative amount of plaque burden, number of lesions, proportion of necrotic core, and VH-TCFA plaque phenotypes is greatest in the proximal coronary tree, intermediate in the mid coronary tree, and least in the distal coronary tree, except in the RCA, where the gradient of each high-risk plaque measure is more evenly distributed. The proximal to distal vulnerable plaque gradient is not just an imaging curiosity. In 2 myocardial infarction studies, proximal lesions were associated with a higher incidence of in-hospital mortality compared with mid or distal lesions, presumably because of the greater amount of myocardium affected (15,16).
Previous intravascular imaging studies have compared the LMCA with the proximal LAD, showing a reduced percent NC and frequency of VH-TCFA lesion phenotype in the LMCA (17,18). The current study extends this analysis to a comparison of the LMCA to the proximal 30-mm-long segment of each of the 3 epicardial arteries, not just to the LAD. In patients with established atherosclerosis, but who do not have symptoms referable to the LMCA, the plaque burden, percent NC, and number of lesions and VH-TCFAs are less in the LMCA than in the proximal 30-mm-long segments of the other 3 coronary arteries. The relative paucity of high-risk plaque is consistent with: 1) the CASS (Coronary Artery Surgery Study), which found asymptomatic LMCA lesions in a minority (1,477 of 20,137) of patients (19); 2) the observation of Wang et al. (12) who reported only 1 LMCA acute occlusion out of 208 patients presenting with an acute STEMI; and 3) recent studies of Pappalardo et al. (20), who reported that only 48 of 5,261 STEMI/non-STEMI patients presented with an LMCA culprit lesion, and Pedrazzini et al. (21), who reported that only 348 of 9,075 STEMI patients presented with an LMCA culprit lesion.
Determinants of atherosclerotic plaque distribution
The current descriptive study does not explain the mechanisms underlying the development, distribution, and morphology of atherosclerotic plaque. In this regard, several conflicting observations are worth noting. First, atherogenesis and progression cannot just be a function of longitudinal vessel location. The fact that the extent of atherosclerosis and NC is greater proximally than distally only applies consistently to the LAD and LCX, and not to the RCA. Furthermore, it does not explain the findings that less atherosclerosis and NC develop in the LMCA than the proximal LAD, LCX, and RCA. Second, vessel size cannot explain all the observations in the present study. The LAD and LCX taper more than the RCA, as does the degree of plaque burden and number of discrete lesions and VH-TCFAs; in the current analysis, more disease and unstable plaque morphology were observed in larger segments (larger mean EEM CSA) than in smaller segments (Table 3). On one hand, plaques in the distal 30-mm-long segment of the RCA are less severe compared with the proximal LAD and LCX and more severe compared with the mid LAD and LCX; similarly, mean EEM CSA in the distal 30-mm-long segment of the RCA is midway between the proximal and mid segments of the LAD. Plaque burden, lesion frequency, and VH-TCFA frequency were thus greater in the distal RCA than in the distal LAD or LCX. On the other hand, again, the findings in the LMCA (the coronary artery segment with the largest EEM CSA) do not support this concept.
Thus, it is not clear whether the differences in atherosclerosis distribution between the RCA versus the LAD or LCX or between proximal, mid, and distal coronary artery segments are due to the greater tapering of the LAD or LCX, the lack of side branches in the RCA, different patterns in shear stress among the 3 vessels (which may also be influenced by branch points), or the larger size of the distal RCA compared with the distal LAD or LCX (22–24). Nevertheless, pathology studies have shown that lesions and TCFAs develop and rupture repeatedly at locations within the coronary artery tree to cause symptomatic coronary thrombosis or, if silent, to frequently result in disease progression (25).
Study limitations and analytical considerations
The current analysis used plaque burden, rather than the absolute P&M CSA, as a measure of atherosclerotic plaque burden; P&M CSA correlates with vessel size (EEM CSA), whereas plaque burden normalizes P&M CSA for the vessel size. Similarly, the current study emphasized percent NC rather than absolute NC CSA, which correlates with P&M CSA and introduces bias for vessel size (26). Lesions were assigned to specific segments according to the location of the minimum lumen slice. Because lesion length measured 11.5 [IQR: 5.7 to 21.6] mm, lesions could extend from 1 segment to another; however, lesions tend to have a unique minimum lumen slice (27). Patients were enrolled after successful percutaneous coronary intervention treatment of the culprit lesion(s); culprit lesions were not assessed using grayscale or VH-IVUS and were not included in the current analysis. Thus, the present study findings apply to plaque and NC distribution likely to contribute to future events, as demonstrated in the PROSPECT study (3). There were more proximal and mid 30-mm-long segments than distal 30-mm-long segments assessed, especially in the LCX. Side branches were not studied, and thus no implications can be drawn regarding the distribution and morphology of atherosclerosis at bifurcation lesions.
The current analysis from the prospective PROSPECT study has demonstrated the proximal predominance of large plaques with large N and VH-TCFAs in the LAD and LCX, the diffuse nature of such plaques in the RCA, and the relative paucity of high-risk plaque in the LMCA. As such, lesions pathologically and clinically (3) have been shown to represent vulnerable plaque, the lesions that are responsible for ACS including ST-segment elevation and non–ST-segment elevation ACS, these findings reconcile previous conflicting angiographic and pathological reports regarding the location of acute myocardial infarction and coronary thrombosis, and underscore the coronary regions at greatest risk for future coronary events.
This work was funded by Abbott Vascular Corporation. Dr. Mintz has received research grant support from and is a consultant for Volcano Corporation; and has received grant support from Boston Scientific. Dr. Maehara has received research/grant support from Boston Scientific Corporation and speaker honoraria from Volcano Corporation. Dr. Inguez is a consultant for Abbott. Mr. Templin and Dr. Zhang are employees of Abbott Vascular Corporation. Dr. Stone is a member of the scientific advisory boards for and has received honoraria from Boston Scientific and Abbott Vascular; has received research support from Volcano and InfraReDx; and is a consultant to Medtronic, Abbott Vascular, Boston Scientific, and Volcano. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- acute coronary syndrome(s)
- cross-sectional area
- external elastic membrane
- intravascular ultrasound
- left anterior descending coronary artery
- left circumflex coronary artery
- left main coronary artery
- necrotic core
- plaque and media
- right coronary artery
- ST-segment elevation myocardial infarction
- thin-cap fibroatheroma
- virtual histology
- Received August 10, 2011.
- Revision received December 13, 2011.
- Accepted January 17, 2012.
- American College of Cardiology Foundation
- Burke A.P.,
- Kolodgie F.D.,
- Farb A.,
- Weber D.,
- Virmani R.
- Shah P.K.
- Maehara A.,
- Cristea E.,
- Mintz G.S.,
- et al.
- Strong J.P.,
- Eggen D.A.,
- Oalmann M.C.
- El Fawal M.A.,
- Berg G.A.,
- Wheatley D.J.,
- et al.
- Jost S.,
- Deckers J.W.,
- Nikutta P.,
- et al.
- Wang J.C.,
- Normand S.T.,
- Mauri L.,
- et al.
- Hong M.K.,
- Mintz G.S.,
- Lee C.W.,
- et al.
- Fujii K.,
- Kawasaki D.,
- Masutani M.,
- et al.
- Valgimigli M.,
- Rodriguez-Granillo G.A.,
- Garcia-Garcia H.M.,
- et al.
- Taylor H.A.,
- Deumite N.J.,
- Chaitman B.R.,
- Davis K.B.,
- Killip T.,
- Rogers W.J.
- Pappalardo A.,
- Mamas M.A.,
- Imola F.,
- et al.
- Pedrazzini G.B.,
- Radovanovic D.,
- Vassalli G.,
- et al.
- Murray C.D.
- Burke A.P.,
- Kolodgie F.D.,
- Farb A.,
- et al.
- Mintz G.S.,
- Nissen S.E.,
- Anderson W.D.,
- et al.