In vivo coronary artery plaque assessment with computed tomography angiography: is there an impact of iterative reconstruction on plaque volume and attenuation metrics?

Abstract

Background

Coronary computed tomography angiography (CTA) allows the evaluation of coronary plaque volume and low attenuation (lipid-rich) component, for plaque vulnerability assessment.

Purpose

To determine the effect of iterative reconstruction (IR) on coronary plaque volume and composition.

Material and Methods

Consecutive patients without coronary artery disease were prospectively enrolled for 256-slice CT. Images were reconstructed with both filtered back projection (FBP) and a hybrid IR algorithm (iDose⁴, Philips) levels 1, 3, 5, and 7. Coronary plaques were assessed according to predefined Hounsfield unit (HU) attenuation intervals, for total plaque and HU-interval volumes.

Results

Fifty-three patients (mean age, 53.6 years) were included. Noise was significantly decreased and signal-to-noise ratio (SNR) / contrast-to-noise (CNR) were both significantly improved at all IR levels in comparison to FBP. Plaque characterization was performed in 41 patients for a total of 125 plaques. Total plaque volume ranged from 104.4 ± 120.7 to 107.4 ± 128.9 mm³ and low attenuation plaque component from 40.5 ± 54.7 to 43.5 ± 58.9 mm³, with no statistically significant differences between all IR levels and FBP (P = 0.786 and P ≥ 0.078, respectively).

Conclusion

IR improved image quality. Total and low attenuation plaque volumes were similar using either IR or FBP.

Keywords

Algorithms coronary artery disease atherosclerotic plaque X-ray computed tomography (CT)computer-assisted image processing image reconstruction

Introduction

Coronary computed tomography angiography (CTA) nowadays can have a role in risk stratification and in prediction of coronary events through the identification of coronary artery lumen stenoses (1,2), as well as with the assessment and follow-up of vessel wall characteristics, such as plaque volume and composition. First, total plaque volume as assessed with CTA has shown to be associated with acute coronary syndromes (3,4); second, low-attenuation CTA foci in plaques are histologically related to lipid content (5,6), and evidences suggest that this CT feature can be used as a marker of plaque vulnerability. Histological and intravascular ultrasound studies described thin-cap fibroatheromas lipid-rich content (7) to be a rupture-prone substrate for subsequent major cardiovascular events (8 –11). Intravascular optical coherence tomography has shown good sensitivity/specificity values of CT for the detection of coronary plaques and of non-calcified plaques with a lipid core (12).

Iterative reconstruction (IR) CT techniques improve image quality using various noise reduction algorithms, and allows more aggressive radiation dose reduction strategies (13). IR is progressively replacing traditional filtered back projection (FBP) in clinical and research settings (14). However, there is not enough evidence of the effects of IR on plaque volume and composition measurements and how IR can compare with FBP for vulnerability assessment.

The objective of this study was to assess image quality as well as plaque volume and composition with CTA using a commercially available IR algorithm compared to FBP.

Material and Methods

Study design: This is a cross-sectional study prospectively recruiting patients without coronary artery disease, using a repeated measures design to compare the effect of levels of IR on volume and composition of coronary plaques.

Patient population: Our center is involved in the Canadian HIV and Aging Cohort Study (CHACS), following a prospective cohort of HIV-positive and HIV-negative patients undergoing coronary CTA. The present study is nested in and recruited consecutive patients from the CHACS. To be included, patients had to be asymptomatic, without any history of coronary artery disease and with an intermediate cardiovascular risk level. Exclusion criteria were renal impairment and hypersensitivity to contrast agents. The local Institutional Review Board approved the protocol. All subjects provided written consent.

Fifty-three consecutive patients (50 men, 3 women; mean age, 54.5 ± 6.8 [standard deviation, SD] years; 51 HIV-positive, 2 HIV-negative) of the CHACS were prospectively enrolled in the present study.

CT protocol: Before the scan, patients were given metoprolol 50–75 mg PO if their heart rate was >60 beats per minute (bpm), and 0.4 mg nitroglycerin sublingually, unless contraindicated. The scans were done with a 256-slice CT scanner (Brilliance iCT, Philips Healthcare, Best, The Netherlands). CTA was done using 370 mg/mL iopamidol (Bracco Imaging, Milan, Italy) injected at 5 mL/s after bolus tracking. For heart rates ≤ 70 bpm, prospective ECG-gating was used; otherwise retrospective ECG-gating was used. Gantry rotation time was 270 ms, and slice thickness was 0.625 mm. A non-contrast cardiac scan was also performed for coronary calcium assessment.

Effective radiation dose was calculated by multiplying dose-length product (DLP) with a conversion coefficient for the chest (k = 0.014 mSv · cm/mGy).

Image reconstruction: Raw CT data were reconstructed using a commercial IR algorithm (Philips iDose⁴, Philips Healthcare). iDose⁴ is a hybrid statistical iterative reconstruction algorithm that iterates both in the raw data domain and in the image data domain (15,16). It first works in the raw data domain by applying an iterative edge-preserving denoising algorithm directly on the projection raw data (17). The noisy data are penalized and edges are preserved (15). The algorithm then works in the image domain. In our study, CTA image reconstructions were done at five different IR noise reduction levels: 0 (equivalent to standard FBP), 1, 3, 5, and 7. Levels 1, 3, 5, and 7 are associated with noise reduction of 10.6, 22.5, 36.8, and 55.5%, respectively (15). Images from the non-contrast scan were reconstructed only at IR level 3.

Quantitative image quality parameters: Measurements were taken in a picture archiving and communication system (IMPAX 6.3.1, AGFA HealthCare N.V., Mortsel, Belgium). Image quality was assessed in three locations: left main coronary artery (LMA), middle right coronary artery (RCA), and descending thoracic aorta. Image quality was evaluated for noise (SD of mean attenuation), signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR). Data were acquired using regions of interest (ROIs) in the LMA, RCA, aorta, and in the epicardial fat. Between IR levels, the ROIs were identical in size and location.

SNR and CNR were calculated using these formulas (18):

SNR = \frac{RO I_{organ}}{S D_{organ}}

(1)

CNR = \frac{RO I_{organ} - RO I_{background}}{\sqrt{\frac{1}{2} \cdot (S {D^{2}}_{organ} + S {D^{2}}_{background})}};

(2)

where ROI_organ and SD_organ: mean attenuation (HU) and noise (HU) in the target tissue;

ROI_background and SD_background: mean attenuation and noise in the surrounding epicardial adipose tissue, respectively.

Plaque analysis: Among the 53 patients, coronary plaque analysis was performed in the first 41 consecutive patients, for a total of 125 plaques. Coronary plaques were identified by a board-certified radiologist (14 years of experience in cardiac CT), then postprocessing was performed by a research assistant trained in plaque volumetric analysis. Analysis was done in multiplanar reformat (MPR), using a semi-automated software (Aquarius iNtuition 4.4.6, TeraRecon Inc, Foster City, CA, USA).

Proximal and distal plaque boundaries were traced by manual segmentation. For consistency, plaque length and position along the artery had to be similar between all segmentations across IR and FBP levels. Total plaque volume was automatically obtained from plaque segmentation. Plaque composition was then assessed using attenuation-stratified measurements in the plaque volume: <51 HU, 51–100 HU, 101–150 HU, 151–350 HU, and >350 HU (Fig. 1). Relative plaque composition was calculated by dividing the plaque volume of each attenuation interval by the total plaque volume and expressed as percentage of total plaque.

Fig. 1.

Plaque composition analysis. 256-slice CTA with prospective ECG-gating and postprocessing showing a coronary plaque of the middle segment of the left anterior descending artery. Multiplanar reformat in long axis (a, b) and short axis (c, d) with color-coding according to predefined plaque attenuation HU strata (a, b, d): yellow <51 HU, red 51–100 HU, green 101–150 HU, blue 151–350 HU, white >350 HU. (e) Histogram of the respective volume of the attenuation-defined components of the plaque.

Inter- and intra-rater reliability: iDose⁴ is used at strength level 3 in our clinical practice setting. For inter-rater reliability analysis, a second observer performed plaque measurements at IR level 3, independently and blinded to the results of the first. For intra-observer reliability measurement, the second observer performed a second plaque assessment at IR level 3, ≥1 month after the first session.

Statistical analysis: Continuous variables are expressed as mean ± SD, and discrete variables expressed as count (percentage of total). Data were tested for normality, then repeated measures analyses of variance (ANOVA) were performed, followed by post hoc Student's paired t-tests (with Bonferroni adjustment) when needed. If data were not normally distributed, data transformation was attempted using inverse hyperbolic sine. If non-normality remained, non-parametric Friedman tests were used, followed by post hoc Wilcoxon rank-sum tests (with Bonferroni adjustment).

Inter-rater and intra-rater reliability were determined with the intraclass correlation coefficient (ICC) using a two-way random model for absolute agreement. An ICC of 0.60–0.74 implied good agreement; 0.75–1.00, excellent agreement (19).

For power analysis, a sample size of 116 plaques was calculated to give a power of 80% to detect a difference of 3.3 mm³ between IR level 3 and FBP (paired t-test; P ≤ 0.05) (20,21). Calculation is based on previous data from Fuchs et al. (22), reporting a mean plaque volume difference of 3.3 mm³ between FBP and adaptive statistical iterative reconstruction (blending level 40%), and data from a local pilot phase which showed a SD of the distribution of volume differences of 12.6 mm³. To account for potential incomplete data, sample size for volume analysis was set at 125 plaques.

A two-tailed P value < 0.05 was considered statistically significant. Statistical analyses were performed using SPSS (SPSS version 21, IBM Corp., Armonk, NY, USA).

Results

Patient population: The 53 patients had a total of 160 coronary plaques: 83 (52%) on the left anterior descending artery (LAD), 47 (32%) on the RCA, and 30 (16%) on the left circumflex artery (LCX). Patient demographics and scan data are summarized in Tables 1 and 2.

Table 1.

Patient demographics. Data is expressed as $\bar{x}$ ± SD or count (percentage of total).

Patients (n)	53
Female (n)	3 (5.7%)
Age (years)	54.5 ± 6.8
HIV-infected (n)	51 (96%)
BMI (kg/m²)	26.0 ± 4.7
Waist circumference (cm)	96.6 ± 13.3
Framingham score	10.6 ± 5.5
Smoker (n)	22 (42%)
Alcoholic (n)	2 (3.8%)
History of substance abuse (n)	25 (47%)
Diabetes (n)	8 (15%)
Hypertension (n)	15 (28%)
Positive family history of CAD (n)	29 (55%)
Dyslipidemia (n)	11 (21%)
Calcium score (Agatston)	179.6 ± 314.2

Table 2.

Scan data. Data are expressed as $\bar{x}$ ± SD or count (percentage of total).

CCTA scans (n)	53
Coverage (mm)	153.5 ± 17.9
Voltage 100 kV (n)	11 (20.8%)
Voltage 120 kV (n)	42 (79.2%)
Current (mA)	692.7 ± 157.0
Heart rate during CCTA (min⁻¹)	62.7 ± 16.8
Prospective ECG-gating (n)	31 (82%)
Retrospective ECG-gating (n)	7 (18%)
DLP of CCTA (mGy·cm)	397.2 ± 217.9
Effective radiation dose of CCTA (mSv)	5.56 ± 3.05
Total DLP (mGy·cm)	527.0 ± 220.3
Total effective radiation dose (mSv)	7.38 ± 3.08

Quantitative image quality parameters: Attenuation in the aorta and coronary arteries was marginally changed (≤1.6%) with IR compared to FBP. The change was not statistically significant, except for LMA and RCA at high IR levels (Table 3). Noise was significantly decreased (9.3% to 53%) at all IR levels in comparison to FBP. SNR and CNR were both significantly improved (11% to 109% and 11% to 97%, respectively) at all IR levels in comparison to FBP.

Table 3.

Quantitative image quality parameters and mean attenuation in descending aorta, LMA, and RCA across iterative reconstruction levels, and % change compared to IR level 0 (FBP). Differences in noise, SNR and CNR were statistically significant for all three arteries (P < 0.0005) and for all levels of IR.

	Aorta	Change	LMA	Change	RCA	Change
Attenuation
IR 0	438.7 ± 75.5		446.6 ± 83.4		453.7 ± 104.9
IR 1	437.8 ± 75.4	−0.2%	446.2 ± 82.8	−0.1%	453.1 ± 106.5	−0.1%
IR 3	438.5 ± 75.5	−0.0%	445.2 ± 81.8	−0.3%	452.8 ± 107.8	−0.2%
IR 5	438.1 ± 74.9	−0.1%	444.7 ± 82.3	−0.4%	448.2 ± 108.0	−1.2%*
IR 7	439.6 ± 78.1	0.2%	442.9 ± 82.2	−0.8%^†	446.3 ± 110.5	−1.6%^‡
Noise
IR 0	40.2 ± 14.8		37.9 ± 14.0		36.7 ± 20.1
IR 1	35.3 ± 11.5	−12%^§	33.9 ± 12.5	−11%^§	33.3 ± 19.0	−9.3%^§
IR 3	30.7 ± 9.7	−24%^§	31.1 ± 12.3	−18%^§	30.3 ± 18.7	−17%^§
IR 5	25.3 ± 7.9	−37%^§	26.8 ± 10.6	−29%^§	27.9 ± 17.9	−24%^§
IR 7	18.9 ± 5.9	−53%^§	22.6 ± 10.3	−40%^§	24.5 ± 17.6	−33%^§
SNR
IR 0	12.1 ± 4.2		13.2 ± 5.2		15.8 ± 8.6
IR 1	13.5 ± 4.6	12%^§	14.8 ± 5.7	12%^§	17.5 ± 9.9	11%^§
IR 3	15.5 ± 5.2	28%^§	16.3 ± 6.4	23%^§	20.0 ± 12.5	27%^§
IR 5	18.8 ± 6.2	55%^§	18.9 ± 7.5	43%^§	21.9 ± 14.3	39%^§
IR 7	25.3 ± 8.6	109%^§	23.6 ± 10.9	79%^§	26.6 ± 20.5	68%^§
CNR
IR 0	14.9 4.8		15.5 ± 5.4		16.0 ± 6.1
IR 1	16.6 ± 5.2	11%^§	17.2 ± 5.9	11%^§	17.7 ± 6.6	11%^§
IR 3	19.0 ± 6.0	28%^§	19.3 ± 6.7	25%^§	20.0 ± 7.8	25%^§
IR 5	22.7 ± 7.3	52%^§	22.5 ± 7.9	45%^§	22.7 ± 9.4	42%^§
IR 7	29.4 ± 9.4	97%^§	27.7 ± 10.8	79%^§	27.4 ± 12.1	71%^§

P = 0.017.

†

P = 0.039.

‡

P = 0.007.

P < 0.0005.

Influence of IR on total plaque volume assessment: The first 41 consecutive patients in which plaques were assessed had a total of 125 plaques, all of which were analyzed. Forty plaques (32%) were in the RCA, 65 (52%) in the LAD, and 20 (16%) in the LCX.

Total volumes per plaque for IR levels 0 (FBP), 1, 3, 5 and 7 were similar: 104.4 ± 120.7 mm³, 107.4 ± 128.9 mm³, 105.1 ± 124.5 mm³, 106.2 ± 126.6 mm³, and 105.7 ± 124.4 mm³, respectively (P = 0.786) (Fig. 2 and Table 4).

Fig. 2.

Total plaque volume as well as plaque composition (expressed as absolute volume values per specific HU-defined interval) across all IR levels. Total plaque volume was similar between all IR levels (P = 0.786). Differences between IR levels were statistically significant for HU intervals 101–150 (P = 0.009) and 151–350 (P < 0.005). Height of each bar, mean; brackets, standard error.

Table 4.

Absolute coronary plaque volume across IR levels. Values are in mm³.

	Total	<51	51–100	101–150*	151–350^†	>350^‡
IR 0	104.4 ± 120.7	40.5 ± 54.7	14.4 ± 17.1	11.6 ± 12.8	24.7 ± 25.6	12.4 ± 20.1
IR 1	107.4 ± 128.9	42.6 ± 60.9	14.9 ± 17.7	11.6 ± 13.0	24.8 ± 26.2	12.5 ± 20.9
IR 3	105.1 ± 124.5	40.9 ± 57.9	14.6 ± 17.8	11.6 ± 13.0	24.5 ± 25.6	12.2 ± 20.0
IR 5	106.2 ± 126.6	42.7 ± 57.8	14.9 ± 19.0	11.4 ± 13.0	24.1 ± 26.1	12.2 ± 20.7
IR 7	105.7 ± 124.4	43.5 ± 58.9	14.4 ± 17.1	11.0 ± 12.2	23.7 ± 24.7	12.0 ± 20.5

P = 0.009 for ANOVA across all IR levels.

†

P < 0.0005 for ANOVA across all IR levels.

‡

P = 0.033 for ANOVA across all IR levels.

Influence of IR on absolute plaque composition assessment: The comparison of HU-interval plaque volumes as assessed using IR levels 0 (FBP), 1, 3, 5, and 7 showed a small number of slight but statistically significant differences. These were slight plaque volume measurement decreases with high IR levels (levels 5 and 7) compared to FBP or IR level 1, in the 101–150 and 151–350 HU intervals. More specifically, plaque volume within the 101–150 HU interval was decreased by 5.2% with IR level 7 compared to FBP (P = 0.037). Plaque volume within the 151–350 HU interval was decreased by 2.4% with IR level 5 compared to FBP (P = 0.01), by 4.0% with IR level 7 compared to FBP (P = 0.03), and by 4.4% with IR level 7 compared to IR levels 1 (P = 0.01) (Fig. 2, Tables 4 and 5).

Table 5.

Results of statistical analyses on differences in absolute plaque volume across IR levels.

HU interval	Omnibus test sig.	Post-hoc tests
Total	P = 0.786	–
<51	P = 0.472	–
51–100	P = 0.432	–
101–150	P = 0.009	−5.2% from IR 0 to 7 (P = 0.037)
151–350	P < 0.0005	−2.4% from IR 0 to 5 (P = 0.01)
		−4.0% from IR 0 to 7 (P = 0.03)
		−4.4% from IR 1 to 7 (P = 0.01)
>350	P = 0.033	P ≥ 0.114 for all tests

Within the > 350 HU interval, a marginally statistically significant difference was found when testing absolute volumes across all IR levels (P = 0.033). However, no significant difference persisted after post hoc pairwise IR level comparisons (P ≥ 0.114) (Table 5).

Plaque volume within the <51 and 51–100 HU intervals showed no significant differences between IR levels (P = 0.472 and 0.432, respectively).

Influence of IR on relative plaque composition assessment: The comparison of HU-interval relative plaque volumes as assessed using IR levels 0 (FBP), 1, 3, 5, and 7 also showed a small number of slight but statistically significant differences. These were slight plaque volume measurement decreases with IR level 7 compared to FBP and IR levels 1 and 3, in the 101–150 HU interval. More specifically, relative plaque volume within the 101–150 HU interval was decreased by 5.0% with IR level 7 compared to FBP (P = 0.02), by 4.2% compared to level 1 (P = 0.01), and by 4.2% compared to level 3 (P = 0.02) (Fig. 3, Tables 6 and 7).

Fig. 3.

Relative plaque composition as relative values (expressed as relative volume values (%) per specific HU-defined interval) across IR levels. Differences between IR levels were statistically significant for HU interval 101–150 (P < 0.005). Height of each bar, mean; brackets, standard error.

Table 6.

Relative plaque volume across IR levels. Values are in percentages.

	<51*	51–100	101–150^†	151–350	>350^‡
IR 0	37.7 ± 14.9	14.9 ± 15.7	12.1 ± 5.1	26.0 ± 8.2	10.8 ± 8.9
IR 1	37.2 ± 14.5	15.1 ± 6.0	12.0 ± 4.6	26.3 ± 8.7	10.7 ± 8.8
IR 3	37.0 ± 14.2	14.7 ± 5.8	12.0 ± 5.0	26.1 ± 8.2	10.9 ± 8.7
IR 5	38.0 ± 14.4	15.9 ± 6.3	11.9 ± 5.0	25.7 ± 8.5	10.6 ± 8.4
IR 7	39.0 ± 15.0	14.8 ± 6.4	11.5 ± 5.0	25.5 ± 8.5	10.1 ± 8.4

P = 0.041.

†

P < 0.0005.

‡

P = 0.034.

Table 7.

Results of statistical analyses on differences in relative plaque volume across IR levels.

HU interval	Omnibus test sig.	Post-hoc tests
<51	P = 0.041	P ≥ 0.078 for all tests:
		4.8% from IR 1 to 7 (P = 0.078)
		5.4% from IR 3 to 7 (P = 0.078)
51–100	p = 0.373	–
101–150	p < 0.0005	−5.0% from IR 0 to 7 (P = 0.02)
		−4.2% from IR 1 to 7 (P = 0.01)
		−4.2% from IR 3 to 7 (P = 0.02)
151–350	P = 0.300	–
>350	P = 0.034	P ≥ 0.161 for all tests

Within the <51 and >350 HU intervals, a marginally statistically significant difference was found when testing relative plaque volumes across all IR (P = 0.041 and P = 0.034, respectively). However, no statistically significant difference persisted after post hoc pairwise IR level comparisons (P ≥ 0.078 and P ≥ 0.161, respectively) (Table 7).

Relative plaque volume within the 51–100 and 151–350 HU intervals showed no significant difference among all different IR levels and FBP (P = 0.373, P = 0.300, respectively).

Inter-rater agreement: Inter-rater agreement for total plaque volume assessment was excellent, with an ICC of 0.951. Agreement for plaque composition was also excellent, with ICCs of 0.860, 0.967, 0.962, 0.959, and 0.933, for the < 51 HU, 51–100 HU, 101–150 HU, 151–350 HU, and >350 HU intervals, respectively.

Intra-rater agreement: Intra-rater agreement for total plaque volume assessment was excellent, with an ICC of 0.933. Agreement for per-HU plaque composition was also excellent, with ICCs of 0.845, 0.926, 0.941, 0.940, and 0.912, for the < 51 HU, 51–100 HU, 101–150 HU, 151–350 HU, and >350 HU intervals, respectively. Figs 4 and 5 show Bland–Altman plots for plaque analysis between the two observers and for both readings of the second observer, respectively.

Fig. 4.

Bland–Altman plots for inter-rater reliability of absolute plaque volume (mm³) at IR 3. Plots show the agreement between observer 1 and observer 2 for (a) total plaque volume, and for plaque volume of the (b) < 51 HU, (c) 51–100 HU, (d) 101–150 HU, (e) 151–350 HU, and (f) > 350 HU intervals, respectively.

Fig. 5.

Bland–Altman plots for intra-rater reliability of absolute plaque volume (mm³) at IR 3. Plots show the agreement for both measurements taken by observer 2 for (a) total plaque volume, and for plaque volume of the (b) < 51 HU, (c) 51–100 HU, (d) 101–150 HU, (e) 151–350 HU, and (f) > 350 HU intervals, respectively.

Discussion

This is a prospective study investigating the effect of a hybrid statistical iterative reconstruction algorithm on plaque volume and composition in 125 coronary artery plaques. Image quality was improved with the iDose⁴ IR algorithm compared to FBP, at all IR strength levels. The total coronary plaque volume as measured with IR was similar to FBP. Per-HU composition analysis revealed slight differences between IR and FBP at high IR strength levels in the intermediate HU intervals. No difference was shown in the lower HU intervals. Total plaque volume measurement and plaque characterization were associated with a high inter-observer agreement.

Previous studies have demonstrated that coronary plaque burden and composition as assessed with CTA correlate with histology (5,6) and invasive modalities (12), and are predictive of cardiovascular events (4). In many previous studies, FBP was used for CT image reconstruction. Results from the CONFIRM (COronary CT Angiography EvaluatioN For Clinical Outcomes: An InteRnational Multicenter) registry showed a predictive value for the presence of calcified or mixed plaque (23). However, non-calcified plaques had no significant correlation with mortality, even if known robust arguments suggest that non-calcified plaques are associated with acute coronary syndromes (3) and more vulnerable to rupture and future cardiac events (7,24). As stated by the authors, this discordant result may be due to the heterogeneity of the equipment and reconstruction algorithms used in the CONFIRM multicenter study (23). Recently, Puchner et al. (25), in an ex vivo study in human hearts, assessed the CT detection of lipid-core plaque (plaque area < 60 HU) against histology using an area-based methodology and observed an improved accuracy with iterative reconstruction algorithms.

Two retrospective studies assessed the in vivo CT measurement of plaque burden and composition with IR, using a volume-based methodology. The studies of Takx et al. (26) and Fuchs et al. (22) involved the assessment of 50 and 55 coronary plaques with advanced statistical iterative reconstruction (ASIR) and iDose⁴, respectively. Both studies showed that total plaque volume was similar using IR or FBP (22,26). Takx et al. assessed plaque composition on a scale of three separate components: lipid (10–69 HU), fibrous (70–129 HU), and calcified (>400 HU) plaque, and found no difference in plaque characterization between iDose⁴ and FBP (26). Fuchs et al. (22) used a similar three-component scale, as well as a scale of eight HU-mapping intervals. They showed a slight but significant decrease in the amount of plaque component of 401 to 500 HU with ASIR. No other impact on plaque composition measurement was reported.

Our study was prospective, based on whole plaque analysis and involved the assessment of 125 coronary artery plaques with the iDose⁴ IR algorithm. Total plaque and low-attenuation plaque component volumes were similar using either IR at any strength or FBP. Plaque composition analysis with HU stratification revealed small although significant volume measurement differences of 2.4% to 5.2% within two strata in the summed range of 101–350 HU, detected at high IR levels (levels 5 and 7) in comparison to FBP. This modest effect is probably not clinically significant. High levels of IR generate an unfamiliar “plastic” appearance (27) and are not used in general. Instead, levels 2 or 3 of iDose⁴ are usually used (28,29). Attenuation-stratified plaque volume measurements were similar across all other HU intervals, among which the low-attenuation plaque component (<50 HU). Previous studies have shown the prognostic value of vulnerability markers such as total and low-attenuation plaque component volumes (3,4). Our prospective study on 125 coronary plaques provides data supporting that these two CT markers can be assessed safely with either the iDose⁴ or the FBP algorithms.

Literature shows that the utilization of IR significantly reduces coronary calcification scores (such as Agatston (30) or volume scores) as compared to FBP (31 –35). The effect of IR on calcium scoring was not assessed in our study. However, within the >350 HU interval of plaques, we found no statistically significant difference between FBP and IR level, as also reported by others (26).

There are some limitations to this study. First, there was no correlation to histology or IVUS. Second, other markers of plaque vulnerability were not assessed, such as positive remodeling (4) or napkin ring sign (36). Finally, our results were obtained with one statistical IR algorithm, and may have differed if other statistical or more recent model-based IR algorithms had been used.

In conclusion, this prospective study compared CT image reconstruction with a statistical IR algorithm and conventional FBP on 125 coronary plaques. An improved image quality was shown with IR. Quantitative evaluation of plaque volume and attenuation-stratified lipid-rich plaque component volume was similar when using IR or FBP. No significant impact on plaque vulnerability assessment should be expected when using IR versus FBP.

Footnotes

Declaration of Conflicting Interests

The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Bracco (research collaboration, CCL, GS), Terarecon (research collaboration, CCL), Philips (equipment support, CHUM, CCL), Bayer (equipment support, CHUM, CCL).

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: RBIQ (9th grant cycle) and SIDA-MI Networks of the Fonds de recherche du Québec – Santé (FRQS), Canadian Institutes of Health Research (Team Grant, no. 201206), COPSE program (grant summers 2014–2015) of the University of Montreal.

References

Thomas

Divarkan

Villines

. Management of coronary artery calcium and coronary CTA findings. Curr Cardiovasc Imaging Rep 2015; 8: 18–32.

Cheezum

Subramaniyam

Bittencourt

. Prognostic value of coronary CTA vs. exercise treadmill testing: results from the Partners registry. Eur Heart J Cardiovasc Imaging 2015; 16: 1338–1346.

Hoffmann

Moselewski

Nieman

. Noninvasive assessment of plaque morphology and composition in culprit and stable lesions in acute coronary syndrome and stable lesions in stable angina by multidetector computed tomography. J Am Coll Cardiol 2006; 47: 1655–1662.

Motoyama

Sarai

Harigaya

. Computed tomographic angiography characteristics of atherosclerotic plaques subsequently resulting in acute coronary syndrome. J Am Coll Cardiol 2009; 54: 49–57.

Becker

Nikolaou

Muders

. Ex vivo coronary atherosclerotic plaque characterization with multi-detector-row CT. Eur Radiol 2003; 13: 2094–2098.

Schroeder

Kuettner

Leitritz

. Reliability of differentiating human coronary plaque morphology using contrast-enhanced multislice spiral computed tomography: a comparison with histology. J Comput Assist Tomogr 2004; 4: 449–454.

Virmani

Burke

Farb

. Pathology of the vulnerable plaque. J Am Coll Cardiol 2006; 47: C13–18.

Burke

Farb

Malcom

. Coronary risk factors and plaque morphology in men with coronary disease who died suddenly. N Engl J Med 1997; 336: 1276–1282.

Libby

Theroux

. Pathophysiology of coronary artery disease. Circulation 2005; 111: 3481–3488.

10.

Stone

Maehara

Lansky

. A prospective natural-history study of coronary atherosclerosis. N Engl J Med 2011; 364: 226–235.

11.

Calvert

Obaid

O'Sullivan

. Association between IVUS findings and adverse outcomes in patients with coronary artery disease: the VIVA (VH-IVUS in Vulnerable Atherosclerosis) study. J Am Coll Cardiol Imag 2011; 4: 894–901.

12.

Soeda

Uemura

Morikawa

. Diagnostic accuracy of dual-source computed tomography in the characterization of coronary atherosclerotic plaques: Comparison with intravascular optical coherence tomography. Int J Cardiol 2011; 148: 313–318.

13.

Leipsic

Heilbron

Hague

. Iterative reconstruction for coronary CT angiography: finding its way. Int J Cardiovasc Imaging 2012; 28: 613–620.

14.

Boiselle

Bardo

Rubin

. Expert opinion: is there still a role for filtered-back projection reconstruction in cardiothoracic CT? J Thorac Imaging 2015; 30: 219.

15.

Noël

Fringerle

Renger

. Initial performance characterization of a clinical noise-suppressing reconstruction algorithm for MDCT. Am J Roentgenol 2011; 197: 1404–1409.

16.

Willemink

de Jong

Leiner

. Iterative reconstruction techniques for computed tomography Part 1: technical principles. Eur Radiol 2013; 23: 1623–1631.

17.

Beister

Kolditz

Kalender

. Iterative reconstruction methods in X-ray CT. Phys Med 2012; 28: 94–108.

18.

Willemink

Leiner

de Jong

. Iterative reconstruction techniques for computed tomography part 2: initial results in dose reduction and image quality. Eur Radiol 2013; 23: 1632–1642.

19.

Cicchetti

. Guidelines, criteria, and rules of thumb for evaluating normed and standardized assessment instruments in psychology. Psychol Assess 1994; 6: 284–290.

20.

Dupont

Plummer

. Controlled clinical trials. Power and sample size calculations: a review and computer program. Control Clin Trials 1990; 11: 116–128.

21.

PS Power and Sample Size Calculations Version 3.0. Available at: http://biostat.mc.vanderbilt.edu/wiki/Main/PowerSampleSize.

22.

Fuchs

Stehli

Fiechter

. First in vivo head-to-head comparison of high-definition versus standard-definition stent imaging with 64-slice computed tomography. Int J Cardiovasc Imaging 2013; 29: 1409–1416.

23.

Hadamitzky

Achenbach

Al-Mallah

. Optimized prognostic score for coronary computed tomographic angiography: results from the CONFIRM registry (COronary CT An- giography EvaluatioN For Clinical Outcomes: An InteRnational Multicenter Registry). J Am Coll Cardiol 2013; 62: 468–476.

24.

Naghavi

Libby

Falk

. From vulnerable plaque to vulnerable patient: a call for new definitions and risk assessment strategies: Part II. Circulation 2003; 108: 1772–1778.

25.

Puchner

Ferencik

Maurovich-Horvat

. Iterative image reconstruction algorithms in coronary CT angiography improve the detection of lipid-core plaque – a comparison with histology. Eur Radiol 2015; 1: 15–23.

26.

Takx

Willemink

Nathoe

. The effect of iterative reconstruction on quantitative computed tomography assessment of coronary plaque composition. Int J Cardiovasc Imaging 2014; 30: 155–163.

27.

Leipsic

Labounty

Heilbron

. Estimated radiation dose reduction using adaptive statistical iterative reconstruction in coronary CT angiography: the ERASIR study. Am J Roentgenol 2010; 195: 655–660.

28.

Nepveu

Stevens

Chartrand-Lefebvre

. Iterative reconstruction in 256-slice computed tomography angiography of coronary artery bypass grafts - assessment of radiation dose reduction and image quality. Am J Roentgenol 2014; 202: W599.

29.

Deseive

Chen

Korosoglou

. Prospective Randomized Trial on Radiation Dose Estimates of CT Angiography Applying Iterative Image Reconstruction: The PROTECTION V Study. JACC Cardiovasc Imaging 2015; 8: 888–896.

30.

Agatston

Janowitz

Hildner

. Quantification of coronary artery calcium using ultrafast computed tomography. J Am Coll Cardiol 1990; 15: 827–832. Mar 15.

31.

Gebhard

Fiechter

Fuchs

. Coronary artery calcium scoring: influence of adaptive statistical iterative reconstruction using 64-MDCT. Int J Cardiol 2013; 167: 2932–2937.

32.

Kurata

Dharampal

Dedic

. Impact of iterative reconstruction on CT coronary calcium quantification. Eur Radiol 2013; 23: 3246–3252.

33.

Van Osch

Mouden

van Dalen

. Influence of iterative image reconstruction on CT-based calcium score measurements. Int J Cardiovasc Imaging 2014; 30: 961–967.

34.

Takahashi

Kimura

Umezawa

. Comparison of adaptive statistical iterative and filtered back projection reconstruction techniques in quantifying coronary calcium. J Cardiovasc Comput Tomogr 2016; 10: 61–68.

35.

Szilveszter

Elzomor

Károlyi

. The effect of iterative model reconstruction on coronary artery calcium quantification. Int J Cardiovasc Imaging 2016; 32: 153–160.

36.

Otsuka

Fukuda

Tanaka

. Napkin-ring sign on coronary CT angiography for the prediction of acute coronary syndrome. JACC Cardiovasc Imaging 2013; 6: 448–457.