| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Cardiac Imaging |
1 From Dept of Radiology, State University Groningen, University Hospital Groningen, Hanzeplein 1, 9713 GZ Groningen, the Netherlands (R.V., M.O.); Depts of Epidemiology and Biostatistics (R.V., A.H., J.C.M.W.) and Radiology, Daniel den Hoed Clinic, Erasmus Medical Center (M.O.), Rotterdam, the Netherlands; and Dept of Radiology, First University Hospital, Chengdu, P.R. China (B.S.). Received Oct 10, 2002; revision requested Dec 19; revision received Feb 11, 2003; accepted March 13. Supported in part by the NESTOR program for geriatric research, Netherlands Heart Foundation, Netherlands Organization for Scientific Research, Health Research and Development Council grants 28–2975 and 97–1-364, and Municipality of Rotterdam. Address correspondence to R.V. (e-mail: r.vliegenthart@rad.azg.nl).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Electron-beam CT images were acquired with nonoverlapping 1.5- and 3.0-mm section thickness. Scans were obtained in an anthropomorphic thorax phantom with calcium cylinders of different sizes and densities, as well as in 1,302 study participants. A calcified lesion was defined as a minimum of 2 pixels (area, 0.52 mm2) with a minimum attenuation of 130 HU. The calcified lesions were quantified by means of a volumetric method with isotropic interpolation. From the phantom scans, mean volume scores, SDs, and measurement variations were calculated. From the participant scans, median volume scores and interquartile ranges were calculated. Participants were classified in categories based on cutoff levels for volume score quartiles for the 1.5-mm scans. An intraclass correlation coefficient (
value) was calculated as a measure of correlation between categories.
RESULTS: In the phantom, deviations of calculated volumes from the true cylinder volumes and measurement variations were generally higher for the 3.0-mm protocol than for the 1.5-mm protocol. In the participants, the median volume score was 100 mm3 (interquartile range, 11–409 mm3) for the 3.0-mm protocol and 144 mm3 (interquartile range, 35–513 mm3) for the 1.5-mm protocol. Agreement between classifications of volume scores for the 1.5- and 3.0-mm scans was good ( = 0.62, P < .001). Compared with the quartile classification for the 1.5-mm scan, however, classifications for 370 (28%) participants were put in a different category with the 3.0-mm protocol.
CONCLUSION: In a phantom, electron-beam CT scans with 3.0-mm section thickness yield less accurate estimates of calcified volume than do 1.5-mm scans. Electron-beam CT protocols with thinner sections considerably affect classification of individuals on the basis of the amount of coronary calcification depicted.
© RSNA, 2003
Index terms: Coronary vessels, calcification, 54.81 • Coronary vessels, CT, 54.12118
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Discussion has arisen whether measurement of coronary calcification with electron-beam CT meets these requirements. First, the repeatability of traditional calcium scores obtained on sequential electron-beam CT scans is poor (7–9). Compared with the widely used calcium scoring method according to Agatston, volumetric measurement of coronary calcification yields more reproducible scores (9,10). Second, the commonly used imaging protocol for coronary calcification may cause variability in calcium quantification. Partial volume effects considerably influence results obtained with the standard protocol with 3.0-mm section thickness, which leads to loss of sensitivity for detection of small calcifications with relatively low attenuation, which are also of clinical importance. Some authors suggest that thinner-section protocols may substantially improve the accuracy of calcium scoring as a result of decreased partial volume effects (9,11–13). The purpose of our study was to compare (a) the accuracy of electron-beam CT protocols with 3.0- and 1.5-mm section thickness for calcium quantification in a phantom and (b) the prevalence of coronary calcifications with the two protocols in a study population.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
|
|
Scan Protocol
Imaging was performed with a electron-beam CT scanner (C-150; Imatron, South San Francisco, Calif). Every day that the scanner was used, it was calibrated by using a water phantom. Electron-beam CT was performed with 100-msec acquisition time, 130 kV, and 630 mA. Two scan protocols with different section thicknesses were used: 3.0-mm protocol, 3.0-mm section thickness with 3.0-mm table increment; and 1.5-mm protocol, 1.5-mm section thickness with 1.5-mm table increment. The cardiac CT phantom was positioned with the spine centered on the table and the rear edge of the phantom in the middle of the gantry. We positioned the phantom isocentrically with the horizontal and vertical laser lights. The whole phantom was scanned, and each scan protocol was repeated four times.
The participants underwent two consecutive studies within a few minutes of each other: first, the 3.0-mm protocol and then, the 1.5-mm protocol. Before the participants underwent scanning, they exercised adequate breath holding. Between the two protocols, the participants did not move or change position. Images were acquired at 80% of the cardiac cycle, with electrocardiographic triggering. Scans with each protocol were obtained during one breath hold. The scan times varied between 30 and 40 seconds. Images were obtained from the level of the aortic root through the entire heart. The first protocol consisted of acquisition of 38 contiguous 3.0-mm-thick sections; the second protocol consisted of acquisition of 60 contiguous 1.5-mm-thick sections. The effective radiation exposure associated with the two protocols for the participants was approximately 2.1 mSv, namely, 0.8 mSv for the 3.0-mm scan (15) and 1.3 mSv for the 1.5-mm scan. Radiation exposure with the two protocols was considered low because it was the same level as the normal annual background radiation (2–5 mSv). For both the phantom and participants, scan data were reconstructed with a 260-mm field of view by using a 512 x 512 matrix and a sharp reconstruction filter. The pixel area was 0.26 mm2 for both protocols, while the voxel volume was 0.39 mm3 for the 1.5-mm protocol and 0.77 mm3 for the 3.0-mm protocol.
Calcium Scoring Method
Calcifications were quantified (AccuImage; AccuImage Diagnostics, South San Francisco, Calif). A calcification was defined as a minimum of 2 adjacent pixels (area of 0.52 mm2) with attenuation above 130 HU. The scans were reviewed by one of two authors (R.V., B.S.), who both had 3 years of experience in calcium scoring. The readers were blinded to the clinical data for the participants. The phantom scans were reviewed by one reader (R.V.). The reader who scored the 3.0-mm scan also scored the corresponding 1.5-mm scan, in random order. On the phantom scans, a region of interest was placed around each calcium insert depicted. On the participant scans, a region of interest was placed around each high-attenuation lesion in the epicardial coronary arteries. Peak attenuation (in Hounsfield units) and area (in square millimeters) were calculated for the individual calcifications. For this study, the calcifications were quantified by means of a volumetric method with isotropic interpolation, as described by Callister et al (10). These volume scores were calculated because of the reported improved reproducibility for volumetric calcium scoring and because calculation of volumes enabled comparison of results for the phantom directly with the true volumes. For the participants, the scores for individual calcifications were summed, which resulted in a volume score for the entire epicardial coronary system.
Statistical Analysis
For phantom measurements, mean volume scores and SDs were computed. Variation coefficients, defined as the SD divided by the mean, were calculated to assess measurement variation. To compare the mean volume scores for the scans obtained with the two protocols, a paired t test was applied. The calcium cylinder with 1-mm diameter was detected on nine of 24 scans, which resulted in unreliable calcified volume estimates; therefore, measurements for this calcium cylinder were not reported.
The prevalence of calcifications on the two scans of the participants was compared by means of the McNemar test. Because of the skewed distribution of volume scores in the study population, medians and interquartile ranges were computed. Variability of the volume score (VS) between the two protocols was calculated as follows: (VS1.5 - VS3.0)/0.5(VS1.5 + VS3.0). Volume scores were transformed logarithmically by calculating ln(VS + 1) to reduce skewness. A paired t test was applied to compare the mean log volume scores. The systematic error (D) and the limit of agreement between the measurements (D ± 1.96) were determined according to the method described by Bland and Altman (16). Because the magnitude of the differences depended on the mean of the volume scores, we used the logarithmic transformation of the measurements. The limit of agreement was calculated to establish a range of values within which 95% of the differences between the measurements with the two protocols would appear.
Furthermore, to exclude possible breathing artifacts that might have occurred at the end of the 1.5-mm protocol, we scored corresponding sections in a subset of participants (n = 100), consisting of the first 40 1.5-mm sections and the first 20 3.0-mm sections starting from the aortic root. Medians, ranges, variability, and log volume were computed for these sections, and the t test was used to compare mean log volume scores.
Quartiles of the volume score for the 1.5-mm scan were defined in categories of age and sex. The cutoff levels for volume score quartiles for the 1.5-mm scan were also applied to categorize the volume scores for the 3.0-mm scan. An intraclass correlation coefficient ( value) was calculated as a measure of correlation between categories. Data analysis was performed (SPSS for Windows, version 10.0; SPSS, Chicago, Ill). A statistically significant difference was assumed when the P value was less than .05.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
|
|
|
|
Table 2 shows the classification of age- and sex-adjusted volume score categories for the two protocols. The intraclass correlation coefficient was 0.62 (P < .001), which indicates good agreement of classification between the 1.5- and 3.0-mm scans. Compared with the quartile classification for the 1.5-mm scan, however, the 3.0-mm scans of 28% (n = 370) of all participants would be classified into a different category (333 [90%] of the 370 scans were put in a lower category).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Reproducibility of calcium scores on sequential electron-beam CT scans is poor (7–9). One suggested improvement concerns the scoring algorithm. Drawbacks of the quantification of calcium according to Agatston et al (17) include dependence on the peak voxel attenuation of a calcified lesion, which varies between scans, and an arbitrary scaling factor based on the peak attenuation. Callister et al (10) found reduced variability with a volumetric scoring method with isotropic interpolation in comparison to the standard scoring method (median variability, 9% vs 15%). Findings in the study by Yoon et al (9) confirm improved reproducibility for the volumetric scoring algorithm, as well as for the calculation of calcium mass, on the basis of voxel volume and attenuation. In accordance with these study findings, we quantified calcifications by using a volumetric scoring method.
To increase the accuracy of calcium quantification, not only different scoring algorithms were suggested but also alterations in the commonly used scan protocol (9,10,13). Wang et al (8) found that the variability in 6-mm scans was significantly lower than that in the standard 3.0-mm scans. In a study by Callister et al (11), however, sensitivity for detection of particularly small calcified lesions was significantly less for the 6-mm protocol than that for the 3.0-mm protocol. Findings in several studies suggest that use of section thickness of less than 3 mm may be necessary to decrease partial volume effects and prevent the loss of diagnostic and prognostic information, especially for small calcifications (9,11–13). Ours is the first population study to investigate the accuracy of a 1.5-mm section protocol for electron-beam CT. Results in the current study confirm the inferior detection of calcifications with the standard 3.0-mm protocol compared with the 1.5-mm protocol. Calcified lesions were detected in significantly more participants with 1.5-mm section thickness. Almost half of the small calcifications detectable on 1.5-mm scans were missed on the 3.0-mm scans.
Volume scores calculated for the calcium cylinders in the phantom resulted in overestimation of the true calcified volumes; in general, estimates with the 3.0-mm scan were significantly higher than those with the 1.5-mm scan. The main cause for overestimation of calcifications in the stationary phantom was partial volume effect. The partial volume effect was larger with the 3.0-mm scans, since volume averaging was greater with that protocol. Voxels that contained part of a calcification, but had high density, were considered calcified volumes if the average attenuation of the entire voxel was above 130 HU. This effect was increasingly visible with higher density levels: Volume overestimation increased up to a factor of 3.5.
The partial volume effect can also lead to underestimation of the calcified volume if the average attenuation of a voxel that contains part of a calcification decreases below 130 HU. In the population, the mean volume score for the standard protocol was significantly lower than that for the thinner-section protocol. This can probably be explained by the fact that many calcium deposits in atherosclerotic plaques have a density that is comparable to or lower than that of the calcium cylinder with the lowest density in the phantom. In the population, therefore, partial volume effects may have led to underestimation or even complete disregard of small or relatively low-density calcifications. With 1.5-mm section thickness, the size of the voxel is halved, which reduces the partial volume effect and improves the accuracy. This was shown by the improved volume estimates in the phantom and by the lower variation with the thinner-section protocol. Also in the participants, detection of calcified lesions was better on the 1.5-mm scan than on the 3.0-mm scan, as shown by the increased prevalence of calcifications and the significantly higher mean volume score. The partial volume effect will possibly continue to diminish with section thicknesses less than 1.5 mm.
Some issues concerning the scanning and calcium scoring methods have to be addressed. First, the use of 1.5-mm section thickness implies the acquisition of more sections and thus the necessity for longer breath holds. Before scanning, the participants were trained to hold their breath for the duration of scanning. Especially in the distal part of the 1.5-mm scan, breathing artifacts may have occurred that led to blurred images. We hypothesized that the occurrence of breathing artifacts in the distal scan would not considerably affect the quantification of calcified lesions. Large or high-density calcifications in the distal parts of the coronary arteries are rare. To explore the possible influence of breathing artifacts on 1.5-mm scans, we selected 20 3.0-mm scans and 40 1.5-mm scans in a corresponding region, which started at the aortic root, on scans obtained in 100 subjects.
The number of scans was selected on the basis of the following considerations. The standard 3.0-mm scan, which consisted of almost 40 sections, required a breath-hold time that can be maintained by most individuals. In addition, by selecting the first 6 cm of the transverse scan, the largest part of the coronary arteries was able to be evaluated. Scanning of the first 6 cm of the heart was previously applied by Wang et al (8). When the selected scan section was evaluated, the mean difference between the volume scores obtained with the two protocols and the variability were comparable to those for the entire scan. Therefore, it is unlikely that possible breathing artifacts on distal 1.5-mm sections caused the difference in volume scores between with the two protocols. Second, in the study population, images were acquired at 80% of the cardiac cycle. Findings in a recent study (18) show that electrocardiographic triggering tailored to the heart rate of the individual results in improved reproducibility and less motion artifact compared with those with triggering at 80%. However, optimization of individual electrocardiographic triggering will be difficult for screening purposes. Therefore, we think that a standardized electrocardiographic triggering protocol must be established in the future. Third, the calcium mass is also a reproducible measure of calcified plaque burden (19). At the time of our study, calcium scoring software did not allow calculation of calcium mass. Therefore, it was not possible to assess the possible effect of thinner sections on the estimates of calcium mass.
Electron-beam CT of the heart is a sensitive noninvasive method to detect calcified lesions in coronary arteries. Quantification of coronary calcification may provide not only a means to categorize asymptomatic individuals according to their risk of coronary heart disease (1–4) but also a means to evaluate progression or regression of atherosclerosis (5,6). Accurate measurement of coronary calcification, especially for the latter purpose, is obligatory. With cutoff values of volume scores based on age- and sex-adjusted quartiles for the 1.5-mm scan, a considerable percentage of individuals was classified into a different category with the 3.0-mm scan. Additional studies are necessary to investigate the reproducibility of the 1.5-mm protocol and to determine whether this level of accuracy in calcium scoring is necessary for accurate stratification of cardiovascular risk in patients seen in clinical practice.
In conclusion, results of this study show that electron-beam CT scans with section thickness of 3.0 mm yield less accurate calcified volume estimates than do 1.5-mm scans. Smaller calcifications are neglected with 3.0-mm section thickness. The use of thinner sections considerably affects the classification of individuals on the basis of the amount of coronary calcification.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. Rutten, I. Isgum, and M. Prokop Coronary Calcification: Effect of Small Variation of Scan Starting Position on Agatston, Volume, and Mass Scores Radiology, December 1, 2007; 246(1): 90 - 98. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Schlosser, P. Hunold, T. Voigtlander, A. Schmermund, and J. Barkhausen Coronary Artery Calcium Scoring: Influence of Reconstruction Interval and Reconstruction Increment Using 64-MDCT Am. J. Roentgenol., April 1, 2007; 188(4): 1063 - 1068. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Horiguchi, H. Yamamoto, N. Hirai, Y. Akiyama, C. Fujioka, K. Marukawa, H. Fukuda, and K. Ito Variability of repeated coronary artery calcium measurements on low-dose ECG-gated 16-MDCT. Am. J. Roentgenol., July 1, 2006; 187(1): W1 - W6. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. E. Clouse, J. Chen, H. M. Krumholz, M. E. Clouse, J. Chen, and H. M. Krumholz Noninvasive Screening for Coronary Artery Disease With Computed Tomography Is Useful Circulation, January 3, 2006; 113(1): 125 - 146. [Full Text] [PDF]
|
|
|
|
|
|
A. L. Daniell, N. D. Wong, J. D. Friedman, N. Ben-Yosef, R. Miranda-Peats, S. W. Hayes, X. Kang, M. G. Sciammarella, L. de Yang, G. Germano, et al. Concordance of Coronary Artery Calcium Estimates Between MDCT and Electron Beam Tomography Am. J. Roentgenol., December 1, 2005; 185(6): 1542 - 1545. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Horiguchi, Y. Shen, Y. Akiyama, N. Hirai, K. Sasaki, M. Ishifuro, T. Nakanishi, and K. Ito Electron Beam CT Versus 16-MDCT on the Variability of Repeated Coronary Artery Calcium Measurements in a Variable Heart Rate Phantom Am. J. Roentgenol., October 1, 2005; 185(4): 995 - 1000. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. L. Higgins, S. A. Marvel, and J. D. Morrisett Quantification of Calcification in Atherosclerotic Lesions Arterioscler. Thromb. Vasc. Biol., August 1, 2005; 25(8): 1567 - 1576. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Horiguchi, H. Yamamoto, Y. Akiyama, N. Hirai, K. Marukawa, H. Fukuda, and K. Ito Variability of Repeated Coronary Artery Calcium Measurements by 16-MDCT with Retrospective Reconstruction Am. J. Roentgenol., June 1, 2005; 184(6): 1917 - 1923. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
