Skip to main content
NCBI home page
Radiology: Cardiothoracic Imaging logoLink to Radiology: Cardiothoracic Imaging
. 2019 Dec 19;1(5):e190069. doi: 10.1148/ryct.2019190069

Improved Evaluation of Lipid-Rich Plaque at Coronary CT Angiography: Head-to-Head Comparison with Intravascular US

Hidenari Matsumoto 1, Satoshi Watanabe 1, Eisho Kyo 1, Takafumi Tsuji 1, Yosuke Ando 1, Evann Eisenberg 1, Yuka Otaki 1, Osamu Manabe 1, Sebastien Cadet 1, Piotr J Slomka 1, Balaji K Tamarappoo 1, Daniel S Berman 1, Damini Dey 1
PMCID: PMC6939749  PMID: 32076671

Abstract

Purpose

To improve the evaluation of low-attenuation plaque (LAP) by using semiautomated software and to assess whether the use of a proposed automated function (LAP editor) that excludes voxels adjacent to the outer vessel wall improves the relationship between LAP and the presence and size of the lipid-rich component (LRC) verified at intravascular US. At coronary CT angiography, quantification of LAP can improve risk stratification. Plaque, defined as the area between the vessel and the lumen wall, is prone to partial volume effects from the surrounding pericoronary adipose tissue.

Materials and Methods

The percentage of LAP (%LAP), defined as the percentage of noncalcified plaque with an attenuation value lower than 30 HU (LAP/total plaque volume) at greater than or equal to 0 mm (%LAP0), greater than or equal to 0.1 mm (%LAP0.1), greater than or equal to 0.3 mm (%LAP0.3), greater than or equal to 0.5 mm (%LAP0.5), and greater than or equal to 0.7 mm (%LAP0.7) inward from the vessel wall boundaries, were quantified in 155 plaques in 90 patients who underwent coronary CT angiography before intravascular US. At intravascular US, the LRC was identified by using echo attenuation, and its size was measured by using the attenuation score (summed score/analysis length) based on the attenuation arc (1 = < 90°, 2 = 90° to < 180°, 3 = 180° to < 270°, 4 = 270°–360°) for every 1 mm.

Results

Use of LAP editing improved the ability for discriminating LRC (areas under receiver operating characteristic curve: 0.667 with %LAP0, 0.713 with %LAP0.1 [P < .001 for comparison with %LAP0]), 0.778 with %LAP0.3 [P < .001], 0.825 with %LAP0.5 [P < .001], 0.802 with %LAP0.7 [P = .002]). %LAP0.5 had the strongest correlation (r = 0.612, P < .001) with LRC size, whereas %LAP0 resulted in the weakest correlation (r = 0.307; P < .001).

Conclusion

Evaluation of LAP at coronary CT angiography can be significantly improved by excluding voxels that are adjacent to the vessel wall boundaries by 0.5 mm.

Supplemental material is available for this article.

© RSNA, 2019

Summary

In low-attenuation plaque measurements, excluding voxels adjacent to the coronary vessel wall boundaries improves assessment of lipid-rich plaque at coronary CT angiography.

Key Point

  • ■ The exclusion of voxels less than 0.5 mm from the standard coronary artery vessel wall boundaries at measurement of low-attenuation plaque yielded the best correlation with the presence and size of the lipid-rich component verified at intravascular US.

Introduction

Plaque rupture resulting in luminal thrombosis is the primary cause of acute coronary syndrome (1). The results of autopsy and intravascular US studies have shown that plaques with a large lipid-rich component (LRC) are the most prone to rupture (1,2). At coronary CT angiography, the LRC is identified as low-attenuation plaque (LAP). Although CT attenuation thresholds for defining LAP have varied among studies (2–7), a threshold of less than 30 HU has been widely accepted (2,3,6,8–16). Quantification of LAP at coronary CT angiography can improve the risk stratification and prediction of subsequent adverse cardiac events (13,14).

Coronary artery plaque is defined as the area between the coronary artery vessel wall and the lumen (17–23). However, the coronary artery is surrounded by low-attenuation adipose tissue, and this factor can lead to overestimation of LAP volume caused by partial volume effects (24). Results of a previous study (24) suggest the use of direct tracing of coronary artery plaque to improve this quantification. However, this method is time consuming and requires substantial operator experience. Semiautomated techniques have been widely used in both research studies and clinical practice to improve the reproducibility and accuracy of plaque analysis at coronary CT angiography (2,17–20).

In this study, an automated function that excludes voxels adjacent to the outer vessel boundaries was used to improve the evaluation of LAP. The aim of this study was to evaluate whether the use of this proposed function, which we named LAP editor, improves the relationship between LAP and the presence and size of LRC verified at intravascular US.

Materials and Methods

All imaging studies were clinically indicated. The local ethics committee approved the study protocol, and written informed consent was obtained from all patients for the use of their data for research. This study was supported by a National Heart, Lung, and Blood Institute grant and partially supported by a grant from the Miriam and Sheldon G. Adelson Medical Research Foundation. The authors of this study had sole control of the data and information submitted for publication.

Study Patients

One hundred sixteen consecutive patients who were known or suspected of having coronary artery disease and underwent coronary CT angiography within 1 month before intravascular US at Kusatsu Heart Center (Kusatsu, Shiga, Japan) were enrolled in this retrospective study. The exclusion criterion was poor image quality at coronary CT angiography or intravascular US, segments with implanted stents, predilatation before intravascular US, and/or presence of extensive acoustic shadowing from calcification (arc ≥ 90° and ≥3 mm in length) at intravascular US, which precludes accurate assessment of the LRC. With these exclusions, a total of 155 atherosclerotic plaques in 90 patients were evaluated (Fig E1 [supplement]).

Intravascular US

Image acquisition.—Gray-scale intravascular US examinations were performed in a standard fashion by using commercially available 40-MHz imaging catheters (Boston Scientific; Minneapolis, Minn or Terumo; Tokyo, Japan), as previously described (22). The imaging catheter was advanced beyond the distal portion of the target lesion, and automated pullback was performed at a rate of 0.5 mm/sec.

Data analysis.—Intravascular US analyses were performed by experienced observers with use of dedicated software (echoPlaque; INDEC Medical Systems, Santa Clara, Calif), as previously described (22). Intracoronary atherosclerotic lesions with a maximal plaque thickness greater than 1.0 mm and a plaque burden (vessel area/plaque area) of greater than or equal to 40% were identified at intravascular US (25,26). The proximal and distal reference segments were selected at the points most adjacent to the maximal stenosis, in which there was minimal or no plaque. Vessel (external elastic membrane) and lumen contours were manually delineated every 1 mm, and total plaque volume was calculated as the vessel volume minus the lumen volume.

Intravascular US–verified LRC was defined as hypoechoic plaque with superficial echo attenuation (27–29). Superficial echo attenuation was defined as echo attenuation with an arc of greater than 30° that began closer to the lumen than to the adventitia despite the absence of calcium. Deep attenuation that began closer to the adventitia than to the lumen was not classified as LRC. Figures 1b and 2b show examples of an intravascular US cross-section (top left) with (Fig 1b) and without (Fig 2b) an LRC and corresponding cross-sectional CT angiograms. The arcs of attenuation were measured in degrees, with the protractor centered on the lumen every 1 mm, and graded on the basis of a five-point scale (score of 0 for no attenuation; 1, < 90°; 2, 90° to < 180°; 3, 180° to < 270°; 4, 270°–360°) to obtain an attenuation score (summed score/analysis length) (30).

Figure 1b:

Figure 1b:

Open in a new tab

Low-attenuation plaque (LAP) quantification with and without LAP editing (0.5 mm) in a plaque with an intravascular US (IVUS)–verified lipid-rich component (LRC). (a) Curved planar reformatted coronary CT angiograms show areas of plaque before LAP quantification (left image), and the same areas at LAP quantification without editing (middle image) and with editing (right image). Red areas indicate noncalcified plaque, and tan regions indicate LAP, defined as noncalcified plaque with attenuation of less than 30 HU. (b) Intravascular US cross-section shows an LRC (area of superficial echo attenuation) (arrowheads); corresponding cross-sectional coronary CT angiograms (CCTA) are shown for comparison. Red areas indicate noncalcified plaque and tan areas indicate LAP. (c) Three-dimensional views of a vessel with LAP (orange areas), with and without LAP editing. Purple regions indicate the outer vessel boundaries.

Figure 2b:

Figure 2b:

Open in a new tab

Low-attenuation plaque (LAP) quantification with and without LAP editing (0.5 mm) in a plaque without an intravascular US (IVUS)–verified lipid rich component (LRC). (a) On curved planar reformatted coronary CT angiograms, red areas indicate noncalcified plaque; tan areas, LAP, defined as noncalcified plaque with an attenuation of less than 30 HU; and yellow areas, calcified plaque. (b) Intravascular US cross-section without an LRC. Corresponding cross-sectional coronary CT angiograms (CCTA) are shown for comparison. Red areas indicate noncalcified plaque and tan areas indicate LAP. (c) Three-dimensional drawings depict a vessel assessed with and without LAP editing. Purple regions indicate outer vessel boundaries, and orange-brown regions indicate LAP.

Figure 1c:

Figure 1c:

Open in a new tab

Low-attenuation plaque (LAP) quantification with and without LAP editing (0.5 mm) in a plaque with an intravascular US (IVUS)–verified lipid-rich component (LRC). (a) Curved planar reformatted coronary CT angiograms show areas of plaque before LAP quantification (left image), and the same areas at LAP quantification without editing (middle image) and with editing (right image). Red areas indicate noncalcified plaque, and tan regions indicate LAP, defined as noncalcified plaque with attenuation of less than 30 HU. (b) Intravascular US cross-section shows an LRC (area of superficial echo attenuation) (arrowheads); corresponding cross-sectional coronary CT angiograms (CCTA) are shown for comparison. Red areas indicate noncalcified plaque and tan areas indicate LAP. (c) Three-dimensional views of a vessel with LAP (orange areas), with and without LAP editing. Purple regions indicate the outer vessel boundaries.

Figure 2a:

Figure 2a:

Open in a new tab

Low-attenuation plaque (LAP) quantification with and without LAP editing (0.5 mm) in a plaque without an intravascular US (IVUS)–verified lipid rich component (LRC). (a) On curved planar reformatted coronary CT angiograms, red areas indicate noncalcified plaque; tan areas, LAP, defined as noncalcified plaque with an attenuation of less than 30 HU; and yellow areas, calcified plaque. (b) Intravascular US cross-section without an LRC. Corresponding cross-sectional coronary CT angiograms (CCTA) are shown for comparison. Red areas indicate noncalcified plaque and tan areas indicate LAP. (c) Three-dimensional drawings depict a vessel assessed with and without LAP editing. Purple regions indicate outer vessel boundaries, and orange-brown regions indicate LAP.

Figure 2c:

Figure 2c:

Open in a new tab

Low-attenuation plaque (LAP) quantification with and without LAP editing (0.5 mm) in a plaque without an intravascular US (IVUS)–verified lipid rich component (LRC). (a) On curved planar reformatted coronary CT angiograms, red areas indicate noncalcified plaque; tan areas, LAP, defined as noncalcified plaque with an attenuation of less than 30 HU; and yellow areas, calcified plaque. (b) Intravascular US cross-section without an LRC. Corresponding cross-sectional coronary CT angiograms (CCTA) are shown for comparison. Red areas indicate noncalcified plaque and tan areas indicate LAP. (c) Three-dimensional drawings depict a vessel assessed with and without LAP editing. Purple regions indicate outer vessel boundaries, and orange-brown regions indicate LAP.

Coronary CT Angiography

Image acquisition.—Coronary CT angiography was performed by using a 64-detector scanner (LightSpeed VCT; GE Healthcare) with use of prospective or retrospective electrocardiographic gating with tube current modulation. All patients were administered nitroglycerin for coronary vasodilatation, and those with a heart rate above 60 beats per minute were given β-blockers unless a contraindication was present. An intravenous bolus of iopamidol (Iopamiron 370; Schering) was continuously injected according to a body weight–adjusted contrast material injection protocol (2.7–5.6 mL/sec) (31). A real-time bolus-tracking technique was used to trigger the initiation of scanning. The scanning parameters were as follows: collimation, 64 × 0.625 mm; rotation time, 350 msec; tube voltage, 100 or 120 kV; and tube current, 450–780 mAs. Transaxial images were reconstructed with a filtered back projection reconstruction algorithm during the cardiac phase, exhibiting minimal cardiac motion. Image reconstruction parameters comprised an individually adapted field of view, a matrix size of 512 × 512 pixels, and a medium soft-tissue convolution kernel.

Data analysis.—An independent observer who was blinded to the intravascular US findings performed quantitative plaque analysis by using dedicated software (Autoplaque, version 2.0; Cedars-Sinai Medical Center, Los Angeles, Calif), as previously described (12,13,20,32–35). Excellent intraobserver reproducibility and interobserver reproducibility have been previously reported (20,34). The proximal and distal reference limits of the plaque were matched to those at intravascular US by using anatomic landmarks such as the distance from the aortocoronary ostium, target lesions, side branches, or calcifications. The registration procedure was performed by a separate investigator who was not involved in the other processes of coronary CT angiography analysis.

Once the proximal and distal limits of the plaque lesions are manually marked, the software computes scan-specific thresholds for epicardial fat, a normal lumen, noncalcified plaque, and calcified plaque from the luminal contrast material attenuation, as previously detailed (20). Then, the total plaque volume and the volume for each plaque component, including the LAP, within the designated segment are automatically quantified. Manual adjustments are made when needed. Differences in lesion length between two modalities (possibly owing to catheter-induced deformation of the coronary artery, cardiac motion, or pullback speed variations) would result in volume measurement differences. Volume parameters at coronary CT angiography were corrected according to the lesion length ([volume parameters at coronary CT angiography times lesion length at intravascular US] divided by lesion length at coronary CT angiography) (36).

LAP at coronary CT angiography was defined as noncalcified plaque with a CT attenuation of less than 30 HU (2,3,6,8–16). Voxels close to the vessel wall can be regarded as LAP because of the partial volume effects from perivascular adipose tissue, which has low CT attenuation. The LAP editor allows the exclusion of voxels with an attenuation of less than 30 HU that are located within a predefined distance from the vessel boundaries, as determined by the software, in the LAP measurement. LAP at greater than or equal to 0 mm (%LAP0), greater than or equal to 0.1 mm (%LAP0.1), greater than or equal to 0.3 mm (%LAP0.3), greater than or equal to 0.5 mm (%LAP0.5), and greater than or equal to 0.7 mm (%LAP0.7) inward from the vessel wall boundaries were quantified. For total plaque volume quantification, no voxel was excluded. The percentage of LAP (%LAP) was derived as (LAP volume/total plaque volume) × 100. Examples of LAP quantification performed by using the LAP editor are shown in Figures 1 (plaque with LRC) and 2 (plaque without LRC).

Figure 1a:

Figure 1a:

Open in a new tab

Low-attenuation plaque (LAP) quantification with and without LAP editing (0.5 mm) in a plaque with an intravascular US (IVUS)–verified lipid-rich component (LRC). (a) Curved planar reformatted coronary CT angiograms show areas of plaque before LAP quantification (left image), and the same areas at LAP quantification without editing (middle image) and with editing (right image). Red areas indicate noncalcified plaque, and tan regions indicate LAP, defined as noncalcified plaque with attenuation of less than 30 HU. (b) Intravascular US cross-section shows an LRC (area of superficial echo attenuation) (arrowheads); corresponding cross-sectional coronary CT angiograms (CCTA) are shown for comparison. Red areas indicate noncalcified plaque and tan areas indicate LAP. (c) Three-dimensional views of a vessel with LAP (orange areas), with and without LAP editing. Purple regions indicate the outer vessel boundaries.

Statistical Analysis

Statistical analysis was performed by using SPSS Statistics 24 (IBM, Armonk, NY). Data were expressed as median with interquartile range (IQR). Categorical variables were expressed as frequency with percentage. Agreement regarding the presence of echo attenuation at intravascular US was assessed by using Cohen κ coefficients. Reproducibility of attenuation scores was examined by Bland-Altman analysis. Quantitative variables between matched pairs were compared by using the paired t test or Wilcoxon signed rank test, where appropriate. The Friedman test was used to examine the difference in %LAP among various degrees of LAP editing. Between-group comparisons for quantitative variables were made by using the unpaired-samples t test or Mann-Whitney U test, where appropriate. The presence of an LAP component between plaques with an LRC and those without this component was compared by using the Fisher exact test. Receiver operating characteristic analysis was applied to determine and compare the diagnostic performance of %LAP for identification of an LRC (37). Correlations between %LAP and attenuation score at intravascular US were assessed by using Spearman rank correlation coefficients. P < .05 was considered to indicate a significant difference.

Results

The patients’ demographic data and clinical characteristics are listed in Table 1. At intravascular US, an LRC was observed in 85 (55%) of the 155 plaques.

Table 1:

Clinical Characteristics

graphic file with name ryct.2019190069.tbl1.jpg

Open in a new tab

Reproducibility of LRC Assessment at Intravascular US

The reproducibility of echo attenuation assessment at intravascular US was evaluated in 20 randomly selected plaques. κ values for intraobserver and interobserver agreement between CT angiography and intravascular US regarding the presence of echo attenuation were 0.90 and 0.90 (P < .001), respectively. At Bland-Altman analysis, the average differences in attenuation score were 0.005 and 0.004, and the 95% limits of agreement were −0.131 and 0.141, and −0.115 and 0.123 for intraobserver and interobserver measurement, respectively.

Diagnostic Performance in Detecting Intravascular US–Verified LRC at Coronary CT Angiography LAP Evaluation

Coronary CT angiography did not differ significantly from intravascular US in vessel volume measurement, regardless of the presence of an LRC: At CT angiography and intravascular US, the median volume of vessels with an LRC (LRC+) was 321.4 mm3 (IQR, 237.7–461.5) versus 326.4 mm3 (IQR, 239.6–459.1), respectively (P = .518); that of vessels without an LRC (LRC−) was 225.8 mm3 (IQR, 153.2–316.3) versus 227.4 mm3 (IQR, 149.7–317.5), respectively (P = .96). There also were no significant differences in lumen volume measurements (median LRC+ volume: 128.0 mm3 [IQR, 90.4–179.0] vs 131.6 mm3 [IQR, 90.4–183.8], respectively [P = .852]; median LRC− volume: 103.4 mm3 [IQR, 67.3–150.4] vs 101.0 mm3 [IQR, 69.3–160.4], respectively [P = .363]) or plaque volume measurements (median LRC+ volume: 198.5 mm3 [IQR, 140.4–280.4] vs 192.3 mm3 [IQR, 134.8–270.0], respectively [P = .466]; median LRC− volume: 116.6 mm3 [IQR, 82.1–171.7] vs 122.7 mm3 [IQR, 73.3–172.4], respectively [P = .518]). Figure 3 shows Bland-Altman plots for comparisons of vessel, lumen, and plaque volumes between coronary CT angiography and intravascular US.

Figure 3a:

Figure 3a:

Open in a new tab

Bland-Altman plots show comparison of vessel, lumen, and total plaque volumes between coronary CT angiography (CCTA) and intravascular US (IVUS). (a) Plots for plaques with an intravascular US–verified lipid-rich component. Solid lines represent mean biases (vessel: 1.9 mm3; lumen: 0.2 mm3; plaque 1.7 mm3), and dashed lines represent 95% limits of agreement (vessel: −51.7 mm3 and 55.5 mm3; lumen: −36.4 mm3 and 36.8 mm3; plaque: −54.2 mm3 and 57.7 mm3). (b) Plots for plaques without a lipid-rich component. Solid lines represent mean biases (vessel, −1.2 mm3; lumen, −2.4 mm3; plaque, 1.2 mm3), and dashed lines represent 95% limits of agreement (vessel: −51.4 mm3 and 49.1 mm3; lumen: −45.6 mm3 and 40.9 mm3; plaque: −46.6 mm3 and 49.0 mm3).

Figure 3b:

Figure 3b:

Open in a new tab

Bland-Altman plots show comparison of vessel, lumen, and total plaque volumes between coronary CT angiography (CCTA) and intravascular US (IVUS). (a) Plots for plaques with an intravascular US–verified lipid-rich component. Solid lines represent mean biases (vessel: 1.9 mm3; lumen: 0.2 mm3; plaque 1.7 mm3), and dashed lines represent 95% limits of agreement (vessel: −51.7 mm3 and 55.5 mm3; lumen: −36.4 mm3 and 36.8 mm3; plaque: −54.2 mm3 and 57.7 mm3). (b) Plots for plaques without a lipid-rich component. Solid lines represent mean biases (vessel, −1.2 mm3; lumen, −2.4 mm3; plaque, 1.2 mm3), and dashed lines represent 95% limits of agreement (vessel: −51.4 mm3 and 49.1 mm3; lumen: −45.6 mm3 and 40.9 mm3; plaque: −46.6 mm3 and 49.0 mm3).

Data regarding the presence of LAP component are summarized in Table 2. Without LAP editing, the presence of LAP was not significantly different between plaques with an intravascular US–verified LRC and those without one (P = .452). Using LAP editing at less than 0.3 mm, less than 0.5 mm, and less than 0.7 mm from the vessel wall, the presence of LAP was more frequent in plaques with an LRC than in those without one (P < .001). The graphs and table in Figure 4 summarize the comparison of %LAP values between plaques with an LRC and those without one. Use of the LAP editor was associated with a linear reduction in %LAP in both plaques with an LRC and those without one (P < .001 for both). Although the %LAP was greater in plaques with an LRC, regardless of the use of LAP editing (P < .001), use of LAP editing reduced overlaps in the distribution of values.

Table 2:

Presence of LAP Component in Plaques with and without an Intravascular US–verified LRC

graphic file with name ryct.2019190069.tbl2.jpg

Open in a new tab

Figure 4a:

Figure 4a:

Open in a new tab

(a) Box-and-whisker plots and (b) corresponding table show comparisons of low-attenuation plaque percentage (%LAP, calculated as LAP volume/total plaque volume) between plaques with (LRC+) and plaques without (LRC−) a lipid-rich component. In a, x-axis (vertical) data represent %LAP values without LAP editing (%LAP0) and %LAP values with LAP editing of 0.1 mm (%LAP0.1), 0.3 mm (%LAP0.3), 0.5 mm (%LAP0.5), and 0.7 mm (%LAP0.7). In b, data in the second and third (%LAP) columns are expressed as median, with the interquartile range (IQR) in parentheses. Use of LAP editing reduced overlap in the distribution of %LAP between plaques with and those without an LRC.

Figure 4b:

Figure 4b:

Open in a new tab

(a) Box-and-whisker plots and (b) corresponding table show comparisons of low-attenuation plaque percentage (%LAP, calculated as LAP volume/total plaque volume) between plaques with (LRC+) and plaques without (LRC−) a lipid-rich component. In a, x-axis (vertical) data represent %LAP values without LAP editing (%LAP0) and %LAP values with LAP editing of 0.1 mm (%LAP0.1), 0.3 mm (%LAP0.3), 0.5 mm (%LAP0.5), and 0.7 mm (%LAP0.7). In b, data in the second and third (%LAP) columns are expressed as median, with the interquartile range (IQR) in parentheses. Use of LAP editing reduced overlap in the distribution of %LAP between plaques with and those without an LRC.

The graph and table in Figure 5 show the diagnostic performance of %LAP in the detection of intravascular US–verified LRC. At receiver operating characteristic analysis, %LAP0 had the lowest capability for discriminating LRC, with an area under the receiver operating characteristic curve (AUC) of 0.667 (95% confidence interval: 0.582, 0.751), and was inferior to the other %LAP values (P < .001 for comparison with %LAP0.1, %LAP0.3, and %LAP0.5; P = .002 for comparison with %LAP0.7). The largest AUC was obtained with use of %LAP0.5 (0.825; 95% confidence interval: 0.762, 0.888).

Figure 5a:

Figure 5a:

Open in a new tab

(a) Graph and (b) table summarize the diagnostic performance of low-attenuation plaque percentage (%LAP) measurements in the identification of a lipid-rich component with use of LAP editing. In b, *P values are those for the comparison of %LAP0 versus the other (based on distance from vessel boundary) %LAP measurements. AUC = area under receiver operating characteristic curve, CI = confidence interval.

Figure 5b:

Figure 5b:

Open in a new tab

(a) Graph and (b) table summarize the diagnostic performance of low-attenuation plaque percentage (%LAP) measurements in the identification of a lipid-rich component with use of LAP editing. In b, *P values are those for the comparison of %LAP0 versus the other (based on distance from vessel boundary) %LAP measurements. AUC = area under receiver operating characteristic curve, CI = confidence interval.

Correlations between LRC Size and %LAP

The LRC sizes determined at intravascular US, expressed as attenuation scores, ranged from 0 to 1.59 (median, 0.12; IQR, 0.0–0.34). Correlations between LRC size and %LAP with various degrees of LAP editing are shown in Figure 6. %LAP0.5 had the strongest correlation (r = 0.612; P < .001) with LRC size, whereas %LAP0 had the weakest correlation (r = 0.307; P < .001).

Figure 6:

Figure 6:

Open in a new tab

Plots show P values and Spearman rank correlation coefficients (r) for correlations between low-attenuation plaque percentage (%LAP) and intravascular US–verified lipid-rich component size (ie, attenuation score).

Discussion

In this study, we found that the exclusion of voxels adjacent to the coronary artery vessel wall improved diagnostic accuracy in the evaluation of LAP at coronary CT angiography and that the best assessment of LRC was achieved by excluding voxels within 0.5 mm inward of the vessel wall. To our knowledge, this is the first study with results that demonstrate the feasibility of using an LAP editor for LRP assessment. The LAP editor can be readily applied to semiautomated plaque quantification software to improve the volumetric measurement of high-risk LAP.

Owing to the limited temporal and spatial resolution of CT scanners, accurate discrimination between the outer vessel boundaries and the pericoronary adipose tissue is still challenging with both manual and semiautomated methods. Voxels adjacent to the vessel boundaries can have attenuation values within the LAP threshold, as these voxels are subject to partial volume effects from neighboring pericoronary adipose tissue with low attenuation or may include pericoronary adipose tissue itself. In addition, the vessel wall comprises fibrotic tissue, and a high-risk lipid-rich necrotic core is usually located close to the lumen and is separated (from the lumen) by a thin fibrous cap (2). In this study, the vessel volumes measured at coronary CT angiography and intravascular US were not significantly different, and almost all plaques had an LAP component, regardless of the presence or absence of an intravascular US–verified LRC. Although the LAP component was eliminated in most plaques without an LRC after the exclusion of voxels adjacent to the vessel boundaries, it remained in plaques with an LRC. Findings suggested that the LAP component adjacent to the vessel boundaries is not associated with a lipid-rich necrotic core. Therefore, the inclusion of voxels with low attenuation adjacent to the vessel boundaries at LAP measurement leads to an overestimation of the true high-risk LRC.

Investigators in an ex vivo study (24) compared two plaque delineation methods for the assessment of LAP: indirect delineation of plaque, defined as the area between the vessel and the lumen boundaries, versus direct delineation of plaque by means of manual tracing. Although direct delineation yielded better discriminative capability for histologically verified lipid-rich necrotic core plaque compared with indirect delineation, it led to significantly underestimated total plaque volumes (24). Indirect delineation is used in all semiautomated plaque quantification procedures at coronary CT angiography and intravascular imaging (17,20,22). Results of this study indicated excellent agreement in total plaque volume between intravascular US and coronary CT angiography with good image quality achieved by using the semiautomated algorithm based on scan-specific thresholds. When all voxels with an attenuation of less than 30 HU were included in the LAP measurement, the capability of LAP for discrimination of LRC in this study (AUC, 0.667) was similar to that assessed by using indirect delineation in previous studies (5,24) and was improved with use of LAP editing. It was found that 0.5 mm–based LAP editing yielded the largest AUC (0.825), which was similar or superior to the discriminative ability for histologically verified LRC assessed with direct delineation (7,24). Excessive exclusion of voxels can result in an underestimated or missed true lipid-rich necrotic core and thus reduced sensitivity. The highest sensitivity was achieved by using 0.5-mm LAP editing, and sensitivity was reduced with 0.7-mm LAP editing.

In addition to facilitating the detection of LAP, accurate quantification yields incremental clinical information regarding patient risk stratification and the prediction of adverse cardiac events. Results of prior studies (12–14) have shown that the absolute LAP volume can be used to predict adverse coronary artery disease events and ischemia-causing coronary lesions. Investigators in studies to examine the effect of lipid-lowering therapy have used the LAP volume on serial CT angiographic scans as a surrogate marker (16,35). In our study, the %LAP was greater in plaques with an LRC than in those without one, regardless of the use of the LAP editor. However, there was substantial overlap in the distribution of %LAP values between plaques with and those without an LRC. This overlap became less significant after LAP editing, resulting in superior diagnostic performance in LRC evaluation. In keeping with the results of binary LRC detection, 0.5-mm LAP editing yielded the best correlation with LRC size at intravascular US. This is reasonable from an imaging standpoint, as this distance is similar to the detector row width of the scanner. These findings suggest that use of LAP editing can lead to improved accuracy of LAP measurements, which may be important in future clinical research studies.

Clinical Implications

Our study results support the assertion of improved LRP evaluation with use of the LAP editor in combination with scan-specific threshold-based plaque measurement performed by using semiautomated plaque quantification software. Use of the proposed method could lead to improved risk stratification and evaluation of therapy in longitudinal studies, without additional time and labor.

Limitations

There were several limitations to this study. First, it was a single-center retrospective study. Second, a fixed attenuation threshold (<30 HU) was used to define LAP. The 30-HU threshold is recommended in the Coronary Artery Disease Reporting and Data System expert consensus document (11) and has been most widely used in clinical practice and research studies (2,3,6,8–16). Finally, gray-scale intravascular US was used as the standard of reference. However, superficial echo attenuation at gray-scale intravascular US, which was used to define an LRC in this study, has been demonstrated to be associated with a histologically verified lipid-rich necrotic core (27). Use of radiofrequency signal–based analysis, near-infrared spectroscopy intravascular US, or optical coherence tomography imaging would partly overcome this limitation.

Conclusion

In a head-to-head comparison with intravascular US, LAP evaluation at coronary CT angiography was found to be significantly improved with the exclusion of voxels that were adjacent to the vessel wall boundaries by 0.5 mm.

SUPPLEMENTAL FIGURES

Figure E1:
ryct190069suppf1.jpg (92.1KB, jpg)

Disclosures of Conflicts of Interest: H.M. disclosed no relevant relationships. S.W. disclosed no relevant relationships. E.K. disclosed no relevant relationships. T.T. disclosed no relevant relationships. Y.A. disclosed no relevant relationships. E.E. disclosed no relevant relationships. Y.O. disclosed no relevant relationships. O.M. disclosed no relevant relationships. S.C. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: received Autoplaque software royalties from Cedars-Sinai Medical Center; has a patent US8885905B2, “Method and system for plaque characterization”, related to noncalcified and calcified plaque characterization from coronary CT angiography. Other relationships: disclosed no relevant relationships. P.J.S. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: grants or grants pending from Siemens Medical Systems; receives royalties from Cedars-Sinai Medical Center; has a patent US8885905B2, “Method and system for plaque characterization”, related to noncalcified and calcified plaque characterization from coronary CT angiography. Other relationships: disclosed no relevant relationships. B.K.T. disclosed no relevant relationships. D.S.B. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: received software royalties from Cedars-Sinai Medical Center, has a patent US8885905B2, “Method and system for plaque characterization”, related to noncalcified and calcified plaque characterization from coronary CT angiography. Other relationships: disclosed no relevant relationships. D.D. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: is an employee of and receives software-licensing royalties from Cedars-Sinai Medical Center, has a patent US8885905B2, “Method and system for plaque characterization”, related to noncalcified and calcified plaque characterization from coronary CT angiography. Other relationships: disclosed no relevant relationships.

Study supported by a National Heart, Lung, and Blood Institute grant. Supported partially by a grant from the Miriam and Sheldon G. Adelson Medical Research Foundation.

Abbreviations:

AUC
area under the receiver operating characteristic curve
IQR
interquartile range
LAP
low-attenuation plaque
LRC
lipid-rich component
%LAP
percentage of LAP

References

  • 1.Falk E, Nakano M, Bentzon JF, Finn AV, Virmani R. Update on acute coronary syndromes: the pathologists’ view. Eur Heart J 2013;34(10):719–728. [DOI] [PubMed] [Google Scholar]
  • 2.Maurovich-Horvat P, Ferencik M, Voros S, Merkely B, Hoffmann U. Comprehensive plaque assessment by coronary CT angiography. Nat Rev Cardiol 2014;11(7):390–402. [DOI] [PubMed] [Google Scholar]
  • 3.Motoyama S, Kondo T, Anno H, et al. Atherosclerotic plaque characterization by 0.5-mm-slice multislice computed tomographic imaging. Circ J 2007;71(3):363–366. [DOI] [PubMed] [Google Scholar]
  • 4.Obaid DR, Calvert PA, Gopalan D, et al. Atherosclerotic plaque composition and classification identified by coronary computed tomography: assessment of computed tomography-generated plaque maps compared with virtual histology intravascular ultrasound and histology. Circ Cardiovasc Imaging 2013;6(5):655–664. [DOI] [PubMed] [Google Scholar]
  • 5.Han D, Torii S, Yahagi K, et al. Quantitative measurement of lipid rich plaque by coronary computed tomography angiography: a correlation of histology in sudden cardiac death. Atherosclerosis 2018;275:426–433. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Marwan M, Taher MA, El Meniawy K, et al. In vivo CT detection of lipid-rich coronary artery atherosclerotic plaques using quantitative histogram analysis: a head to head comparison with IVUS. Atherosclerosis 2011;215(1):110–115. [DOI] [PubMed] [Google Scholar]
  • 7.Schlett CL, Maurovich-Horvat P, Ferencik M, et al. Histogram analysis of lipid-core plaques in coronary computed tomographic angiography: ex vivo validation against histology. Invest Radiol 2013;48(9):646–653. [DOI] [PubMed] [Google Scholar]
  • 8.Motoyama S, Ito H, Sarai M, et al. Plaque characterization by coronary computed tomography angiography and the likelihood of acute coronary events in mid-term follow-up. J Am Coll Cardiol 2015;66(4):337–346. [DOI] [PubMed] [Google Scholar]
  • 9.Williams MC, Moss AJ, Dweck M, et al. Coronary artery plaque characteristics associated with adverse outcomes in the SCOT-HEART study. J Am Coll Cardiol 2019;73(3):291–301. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Nerlekar N, Ha FJ, Cheshire C, et al. Computed tomographic coronary angiography-derived plaque characteristics predict major adverse cardiovascular events: a systematic review and meta-analysis. Circ Cardiovasc Imaging 2018;11(1):e006973. [DOI] [PubMed] [Google Scholar]
  • 11.Cury RC, Abbara S, Achenbach S, et al. CAD-RADS coronary artery disease: reporting and data system—an expert consensus document of the Society of Cardiovascular Computed Tomography (SCCT), the American College of Radiology (ACR) and the North American Society for Cardiovascular Imaging (NASCI). Endorsed by the American College of Cardiology. J Cardiovasc Comput Tomogr 2016;10(4):269–281. [DOI] [PubMed] [Google Scholar]
  • 12.Gaur S, Øvrehus KA, Dey D, et al. Coronary plaque quantification and fractional flow reserve by coronary computed tomography angiography identify ischaemia-causing lesions. Eur Heart J 2016;37(15):1220–1227. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Hell MM, Motwani M, Otaki Y, et al. Quantitative global plaque characteristics from coronary computed tomography angiography for the prediction of future cardiac mortality during long-term follow-up. Eur Heart J Cardiovasc Imaging 2017;18(12):1331–1339. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Chang HJ, Lin FY, Lee SE, et al. Coronary atherosclerotic precursors of acute coronary syndromes. J Am Coll Cardiol 2018;71(22):2511–2522. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Motoyama S, Kondo T, Sarai M, et al. Multislice computed tomographic characteristics of coronary lesions in acute coronary syndromes. J Am Coll Cardiol 2007;50(4):319–326. [DOI] [PubMed] [Google Scholar]
  • 16.Inoue K, Motoyama S, Sarai M, et al. Serial coronary CT angiography-verified changes in plaque characteristics as an end point: evaluation of effect of statin intervention. JACC Cardiovasc Imaging 2010;3(7):691–698. [DOI] [PubMed] [Google Scholar]
  • 17.Boogers MJ, Broersen A, van Velzen JE, et al. Automated quantification of coronary plaque with computed tomography: comparison with intravascular ultrasound using a dedicated registration algorithm for fusion-based quantification. Eur Heart J 2012;33(8):1007–1016. [DOI] [PubMed] [Google Scholar]
  • 18.Rinehart S, Vazquez G, Qian Z, Murrieta L, Christian K, Voros S. Quantitative measurements of coronary arterial stenosis, plaque geometry, and composition are highly reproducible with a standardized coronary arterial computed tomographic approach in high-quality CT datasets. J Cardiovasc Comput Tomogr 2011;5(1):35–43. [DOI] [PubMed] [Google Scholar]
  • 19.Komatsu S, Hirayama A, Omori Y, et al. Detection of coronary plaque by computed tomography with a novel plaque analysis system, ‘Plaque Map’, and comparison with intravascular ultrasound and angioscopy. Circ J 2005;69(1):72–77. [DOI] [PubMed] [Google Scholar]
  • 20.Dey D, Cheng VY, Slomka PJ, et al. Automated 3-dimensional quantification of noncalcified and calcified coronary plaque from coronary CT angiography. J Cardiovasc Comput Tomogr 2009;3(6):372–382. [DOI] [PubMed] [Google Scholar]
  • 21.Nakazato R, Shalev A, Doh JH, et al. Quantification and characterisation of coronary artery plaque volume and adverse plaque features by coronary computed tomographic angiography: a direct comparison to intravascular ultrasound. Eur Radiol 2013;23(8):2109–2117. [DOI] [PubMed] [Google Scholar]
  • 22.Mintz GS, Nissen SE, Anderson WD, et al. American College of Cardiology Clinical Expert Consensus Document on Standards for Acquisition, Measurement and Reporting of Intravascular Ultrasound Studies (IVUS). A report of the American College of Cardiology Task Force on Clinical Expert Consensus Documents. J Am Coll Cardiol 2001;37(5):1478–1492. [DOI] [PubMed] [Google Scholar]
  • 23.Tearney GJ, Regar E, Akasaka T, et al. Consensus standards for acquisition, measurement, and reporting of intravascular optical coherence tomography studies: a report from the International Working Group for Intravascular Optical Coherence Tomography Standardization and Validation. J Am Coll Cardiol 2012;59(12):1058–1072. [DOI] [PubMed] [Google Scholar]
  • 24.Schlett CL, Ferencik M, Celeng C, et al. How to assess non-calcified plaque in CT angiography: delineation methods affect diagnostic accuracy of low-attenuation plaque by CT for lipid-core plaque in histology. Eur Heart J Cardiovasc Imaging 2013;14(11):1099–1105. [DOI] [PubMed] [Google Scholar]
  • 25.Stone GW, Maehara A, Lansky AJ, et al. A prospective natural-history study of coronary atherosclerosis. N Engl J Med 2011;364(3):226–235. [DOI] [PubMed] [Google Scholar]
  • 26.van der Giessen AG, Toepker MH, Donelly PM, et al. Reproducibility, accuracy, and predictors of accuracy for the detection of coronary atherosclerotic plaque composition by computed tomography: an ex vivo comparison to intravascular ultrasound. Invest Radiol 2010;45(11):693–701. [DOI] [PubMed] [Google Scholar]
  • 27.Pu J, Mintz GS, Biro S, et al. Insights into echo-attenuated plaques, echolucent plaques, and plaques with spotty calcification: novel findings from comparisons among intravascular ultrasound, near-infrared spectroscopy, and pathological histology in 2,294 human coronary artery segments. J Am Coll Cardiol 2014;63(21):2220–2233. [DOI] [PubMed] [Google Scholar]
  • 28.Kang SJ, Mintz GS, Pu J, et al. Combined IVUS and NIRS detection of fibroatheromas: histopathological validation in human coronary arteries. JACC Cardiovasc Imaging 2015;8(2):184–194. [DOI] [PubMed] [Google Scholar]
  • 29.Pu J, Mintz GS, Brilakis ES, et al. In vivo characterization of coronary plaques: novel findings from comparing greyscale and virtual histology intravascular ultrasound and near-infrared spectroscopy. Eur Heart J 2012;33(3):372–383. [DOI] [PubMed] [Google Scholar]
  • 30.Wu X, Mintz GS, Xu K, et al. The relationship between attenuated plaque identified by intravascular ultrasound and no-reflow after stenting in acute myocardial infarction: the HORIZONS-AMI (Harmonizing Outcomes With Revascularization and Stents in Acute Myocardial Infarction) trial. JACC Cardiovasc Interv 2011;4(5):495–502. [DOI] [PubMed] [Google Scholar]
  • 31.Matsumoto H, Watanabe S, Kyo E, et al. Effect of tube potential and luminal contrast attenuation on atherosclerotic plaque attenuation by coronary CT angiography: in vivo comparison with intravascular ultrasound. J Cardiovasc Comput Tomogr 2019 13(4):219–225. [DOI] [PubMed] [Google Scholar]
  • 32.Dey D, Schepis T, Marwan M, Slomka PJ, Berman DS, Achenbach S. Automated three-dimensional quantification of noncalcified coronary plaque from coronary CT angiography: comparison with intravascular US. Radiology 2010;257(2):516–522. [DOI] [PubMed] [Google Scholar]
  • 33.Dey D, Achenbach S, Schuhbaeck A, et al. Comparison of quantitative atherosclerotic plaque burden from coronary CT angiography in patients with first acute coronary syndrome and stable coronary artery disease. J Cardiovasc Comput Tomogr 2014;8(5):368–374. [DOI] [PubMed] [Google Scholar]
  • 34.Cheng VY, Nakazato R, Dey D, et al. Reproducibility of coronary artery plaque volume and composition quantification by 64-detector row coronary computed tomographic angiography: an intraobserver, interobserver, and interscan variability study. J Cardiovasc Comput Tomogr 2009;3(5):312–320. [DOI] [PubMed] [Google Scholar]
  • 35.Tamarappoo B, Otaki Y, Doris M, et al. Improvement in LDL is associated with decrease in non-calcified plaque volume on coronary CTA as measured by automated quantitative software. J Cardiovasc Comput Tomogr 2018;12(5):385–390. [DOI] [PubMed] [Google Scholar]
  • 36.Park HB, Lee BK, Shin S, et al. Clinical feasibility of 3D automated coronary atherosclerotic plaque quantification algorithm on coronary computed tomography angiography: comparison with intravascular ultrasound. Eur Radiol 2015;25(10):3073–3083. [DOI] [PubMed] [Google Scholar]
  • 37.DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 1988;44(3):837–845. [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Figure E1:
ryct190069suppf1.jpg (92.1KB, jpg)

Articles from Radiology: Cardiothoracic Imaging are provided here courtesy of Radiological Society of North America

RESOURCES