| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Cardiac Imaging |
1 From the Departments of Radiological Sciences (J.G.G., S.B.H., M.S.B., A.M.E., D.R.A.), Radiology Physics (M.M.M.G.), and Biomathematics (J.W.S.), UCLA School of Medicine, 10833 Le Conte Ave, CHS B2-200, Los Angeles, CA 90095-1721; Department of Radiology, University of Utah School of Medicine, Salt Lake City (H.C.Y.); and Chicago Medical School, Ill (L.E.G.). Received June 5, 2000; revision requested July 26; final revision received March 30, 2001; accepted April 3. Address correspondence to J.G.G. (e-mail: jgoldin@mednet.ucla.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Seventy subjects asymptomatic for coronary heart disease underwent both electron-beam CT and subsecond spiral CT. In all subjects, two images each were obtained with both scanners. Two experienced readers using three different algorithms scored each of the four scans: one score for the electron-beam CT images and two scores for the spiral CT images.
RESULTS: With a 130-HU threshold for the quantification of calcium, there were no significant differences in interscan and interobserver variation in calcium scores between the electron-beam CT and spiral CT images. There was greater interobserver (P < .001) and interscan (P < .03) variation in scores when a 90-HU threshold was used for spiral CT images. With a 130-HU threshold, when calcium scores were used for clinical risk stratification, there was a significant difference between the results obtained with electron-beam CT and those obtained with spiral CT (P < .05).
CONCLUSION: Spiral CT has not yet proved to be a feasible alternative to electron-beam CT for coronary artery calcium quantification. There are systematic differences between calcium scores obtained with single-detector array subsecond spiral CT and those obtained with electron-beam CT.
Index terms: Calcium • Computed tomography (CT), electron beam, 54.12111, 54.12119 • Coronary vessels, CT, 54.12111, 54.12115, 54.12119 • Coronary vessels, stenosis or obstruction, 54.76, 54.812
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Our present understanding of coronary artery calcium detection, its quantification, and its clinical implications are based almost exclusively on the coronary calcium scoring methods used with electron-beam CT data sets. Two studies (5,6) comparing spiral or conventional CT with electron-beam CT were recently published. Becker et al (5) found that there was a closer correlation between prospectively gated transverse CT and electron-beam CT measurements of coronary calcium than between nongated subsecond spiral CT and electron-beam CT measurements. Carr et al (6) noted a high degree of correlation between coronary artery calcium quantifications performed with a developmental version of retrospectively gated subsecond spiral CT and those performed with electron-beam CT in a group of 36 patients.
The purpose of this study was to determine differences in coronary artery calcium detection, quantification, and reproducibility, as measured by using electron-beam CT and subsecond spiral CT with retrospective ECG gating and a commercially available software package (Smartscore) in an asymptomatic adult population.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Electron-Beam CT Protocol
The standard electron-beam CT protocol was followed; the parameters used were 3-mm beam collimation, 3-mm table incrementation, 130 kVp, 630 mA, and ECG-triggered single-section mode (100-msec acquisition) with acquisitions triggered at 80% of the R-R interval. Thirty to 40 contiguous transverse CT images were obtained in a single breath hold to ensure coverage of the entire heart. Image reconstruction was performed with a 512 x 512-pixel matrix by using the sharp reconstruction algorithm. The reconstruction circle size was adjusted to enable visualization of the entire heart on all transverse images. A display field of view of 26 cm was generally sufficient and yielded a pixel size of approximately 0.5 mm, although occasionally, some subjects required a 30-cm display field of view. Immediately following the first sequence, a second sequence was performed by using identical imaging parameters with no change in subject positioning.
Spiral CT Protocol
The spiral scans (CT/i scanner) through the heart were obtained in one breath hold by using the following parameters: 800-msec revolution, 120 kVp, 200–250 mA, and 3-mm collimation. The pitch was set by using the formula pitch = 1.0 x (BPM/75), where BPM (beats per minute) is the subject’s heart rate during suspended maximal inspiration. An ECG trace was obtained during each sequence and referenced to the image data by using the software package (Smartscore). Imaging was performed to include the entire heart. Immediately following the first sequence, a second sequence was performed by using identical parameters with no change in subject positioning. Retrospective image reconstruction was performed by using the vendor-supplied (GE Medical Systems) segmented reconstruction algorithm with a standard reconstruction filter and a 512 x 512 matrix.
To improve temporal resolution, each image was reconstructed by using a 500-msec segment of the raw scan acquisition data. Images were reconstructed with 0.3-mm spacing, which yielded 10 images per 3-mm table travel, and a 25-cm display field of view. The in-plane pixel size was 0.49 mm. The retrospective reconstructions and ECG tracing were transferred to a workstation (AW 3.1; GE Medical Systems) for analysis with the software program (Smartscore). On the basis of the ECG tracing, the software program automatically selected a reduced set of diastolic images from each cardiac cycle. Each reader (S.B.H., J.G.G.) then optimized this set of images by picking the diastolic image from each cardiac cycle with the least amount of motion artifact.
Radiation Dosimetry
With the electron-beam CT scanner, the radiation dose is asymmetric, because it does not complete a full 360° sweep around the patient. The estimated peak radiation dose per scan is 13 mGy (1.3 rad) at the posterior portion of the patient’s thorax; this yields a total dose of 26 mGy (2.6 rad) for two scans. The organ and effective doses for symmetric scanners were estimated by using a software program developed by the National Radiation Laboratory, Christchurch, New Zealand (CTDOSE), which is based on the results of Monte Carlo calculations performed by the National Radiological Protection Board (7). Although this software was not designed for an asymmetric scanner, we calculated an approximate weighted dose for electron-beam CT and determined the time (ie, number of milliampere seconds) that would yield a similar approximate weighted dose for a symmetric scanner that was specified in the software. By using these assumptions and approximations, we determined that the peak organ dose occurs in the lungs (10.0 mGy [1.0 rad]) and esophagus (5.5 mGy [0.55 rad]). The effective dose, or weighted organ dose as specified by guidelines published in International Commission on Radiological Protection publication 60 (8), is estimated to be 3.0 mGy (0.3 rad).
By using the previously described protocol for the spiral CT scanner (CT/i), we estimated the total peak radiation dose to be 27 mGy (2.7 rad) at the skin of the thorax for both scans. According to the National Radiation Laboratory software program (CTDOSE) data, the peak organ doses occur in the breast (18 mGy [1.8 rad]) and lungs (17 mGy [1.7 rad]). The effective dose is estimated to be 4.8 mGy (.48 rad).
Scoring Algorithms
Two readers (S.B.H., J.G.G.) experienced with coronary artery calcium quantification independently scored the four scans obtained with the electron-beam CT and spiral CT scanners by using the electron-beam display console and software (OS version 12.4; Imatron) and the spiral CT workstation and software (Smartscore). Because scoring was performed on separate workstations by using proprietary software, the readers were not blinded to the type of scanner used to acquire the images. Each reader scored both scans obtained in each subject during a single sitting. However, because the two workstations were not close to each other, all four scans could not be scored at one time.
For the scans obtained at electron-beam CT, all high-attenuating foci with two or more contiguous pixels with an attenuation of 130 HU or greater were manually traced by each reader. For the scans obtained at spiral CT, all high-attenuating foci with two or more contiguous pixels with a threshold attenuation of 90 HU rather than 130 HU were manually traced by the same two readers.
The electron-beam CT scans were scored for coronary artery calcium by using the Agatston algorithm (9). With this algorithm, two or more contiguous pixels with an attenuation of 130 HU or greater in the expected location of an epicardial artery are considered a calcified lesion. Lesions are assigned a value between 1 and 4 that is determined by the peak pixel attenuation of the lesion. Lesions with a peak attenuation of 130–200 HU are assigned a value of 1; lesions with a peak attenuation of 201–300 HU, a value of 2; lesions with a peak attenuation of 301–400 HU, a value of 3; and lesions with a peak attenuation value greater than 400 HU, a value of 4. The integer value of each lesion is multiplied by the area of that lesion to yield a lesion-specific calcium score. The sum of all lesion scores yields the total calcium score.
Two scores were generated for each spiral CT data set by using calcium thresholds of 130 and 90 HU. The first score, that for spiral CT with a 130-HU threshold, was computed by summing the lesions with an attenuation of 130 HU or greater that were multiplied by a weighting factor on the basis of the maximum attenuation of the lesion. The weighting factor was the same as that used in the Agatston method (9). The second score, that for spiral CT with a 90-HU threshold, was calculated by using the same formula but with a threshold of 90 HU, as originally suggested by Broderick et al (10). With this method, a weighting value of 1 is assigned to lesions with an attenuation of 90–199 HU, but for lesions with an attenuation of greater than 200 HU, the weighting factor is the same as that used in the Agatston method.
Statistical Analyses
All statistical analyses were performed by using a statistical software package (SPSS version 9.0; SPSS, Chicago, Ill). A general linear model repeated-measures analysis was performed by using the three scoring methods (ie, electron-beam CT, 130-HU spiral CT, and 90-HU spiral CT), the two CT scan types (ie, electron-beam and spiral) as the within-subjects factors, and the two readers as the random-effects factors. The sphericity assumption was evaluated with the Mauchly W test.
Interobserver agreement was evaluated with the Cohen 
statistic and was measured in three ways: by using the absolute difference in calcium scores, the percentage difference in calcium scores, and the coefficient of variation, which in this study was the ratio of the SD of the two readers’ scores to the mean of their scores. Interscan variation between the two sequences performed with each scanner also was measured by using the absolute difference in calcium scores, the percentage difference in calcium scores, and the coefficient of variation, which in this study was the ratio of the SD of each reader’s two scores to the mean of that reader’s scores. To determine whether the magnitude of the score had an effect on interscan variation, the absolute difference in calcium scores was plotted against the mean calcium score for each of the methods (11). A log10 transformation (score +1) was performed to reduce skewness.
Intermodality variation was assessed by comparing the electron-beam CT score with the 130-HU spiral CT and 90-HU spiral CT scores, as determined by the two readers. Each subject was assigned a clinical cardiovascular risk on the basis of each calcium score by using the clinical management guideline model proposed by Rumberger et al (12). In this model, asymptomatic patients are stratified into one of five categories that are based on their calcium score: very low (calcium score 0), low (calcium score 1–10), moderate (calcium score 11–100), moderately high (calcium score 101–400), and high risk (calcium score >400). In this study, those subjects with low or moderate cardiovascular risk whose calcium score was greater than that seen in 75% of the subjects with matched age and sex were placed in a category one risk higher. Separate risk categories were assigned to each subject for each scanner by using the calcium score of each scan. The levels of predicted cardiovascular risk were then compared between the two imaging modalities.
For statistical analyses, continuous variables were evaluated with the t test. Ordinal variables were evaluated with the Wilcoxon signed rank test. Nominal variables were evaluated with 
2 tests. A statistically significant difference was assumed when the P value was less than .05.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
|
When spiral CT scores were compared, the absolute differences in reader scores, as well as the percentage differences in scores, were significantly lower with use of the 130-HU algorithm than with use of the 90-HU algorithm. As shown in Table 2, the interobserver coefficients of variation with both scoring algorithms were similar for both scans.
Interscan Variation
The differences in calcium scores between the two scans for each scoring algorithm are shown in Figure 2. The slopes of the regression lines describing the interscan variation for each of the three scoring algorithms were not significantly different for either reader; this suggests that the variation in scores between scans was similar across the three algorithms.
|
|
|
When the 130-HU spiral CT calcium scores obtained by reader 1 were compared, four subjects had a score of 0 on one scan but not on the other. The range of values for interscan differences in scores was 5–52, with an average difference in score of 18. Reader 2 identified no subject with a calcium score of 0 on one scan but not on the other.
Reader 1 identified 11 subjects with a 90-HU spiral CT calcium score of 0 on one scan but not on the other. The range of values for interscan differences in scores was 1–23, with an average difference in score of 6. Reader 2 identified six subjects with a 90-HU spiral CT calcium score of 0 on one scan but not on the other. The range of values for interscan differences in scores was 1–20, and the average interscan difference was 5.
Intermodality Comparisons
Reader 1, by using scan 1 calcium scores, identified 33, 27, and 45 subjects with calcium (calcium score >1) detected by using electron-beam CT, 130-HU spiral CT, and 90-HU spiral CT scoring algorithms, respectively. The mean calcium score measured at electron-beam CT was significantly greater than that measured at 130-HU spiral CT by an average difference in score of 13 ± 46 (P = .02). The 90-HU spiral CT score was not significantly greater than the electron-beam CT score: The average difference in score was 16 ± 67 (P = .05).
Reader 2, by using scan 1 calcium scores, identified 32, 26, and 31 subjects with calcium detected by using electron-beam CT, 130-HU spiral CT, and 90-HU spiral CT scoring algorithms, respectively. The mean calcium score measured at electron-beam CT was significantly greater than that measured at 130-HU spiral CT by an average difference in score of 16 ± 43 (P = .002). The 90-HU spiral CT score was not significantly greater than the electron-beam CT score: The average difference in score was 11 ± 53 (P = .10).
Reader 1, by using scan 2 scores, identified 32, 29, and 44 subjects with calcium detected by using electron-beam CT, 130-HU spiral CT, and 90-HU spiral CT scoring algorithms, respectively. The mean calcium score measured at electron-beam CT was not significantly greater than that measured at 130-HU spiral CT: The average difference in score was 8 ± 45 (P = .15). However, the 90-HU spiral CT score was significantly greater than the electron-beam CT score, with an average difference of 25 ± 81 (P = .01).
Reader 2, by using scan 2 calcium scores, identified 29, 26, and 29 subjects with calcium detected by using electron-beam CT, 130-HU spiral CT, and 90-HU spiral CT scoring algorithms, respectively. The mean calcium score measured at electron-beam CT was significantly greater than that measured at 130-HU spiral CT by an average difference in score of 12 ± 38 (P = .01). The 90-HU spiral CT score also was significantly greater than the electron-beam CT score, with an average difference in score of 15 ± 58 (P = .04).
We compared the number of subjects who had a calcium score of 0 with one scoring algorithm but a nonzero score with another algorithm. Reader 1 identified six and five subjects with calcium on the first and second electron-beam CT scans, respectively, but no calcium on the basis of their respective 130-HU spiral CT scores. Conversely, reader 1 identified one and two subjects with nonzero 130-HU spiral CT scores on the first and second scans, respectively, but no identifiable calcium on the respective electron-beam CT images.
Reader 2 identified seven and five subjects with calcium on the first and second electron-beam CT scans, respectively, but no calcium on the basis of their respective 130-HU spiral CT scores. Reader 2 identified one and two subjects with nonzero 130-HU spiral CT scores on the first and second scans, respectively, but no identifiable calcium on the respective electron-beam CT images.
Comparing electron-beam CT and 90-HU spiral CT calcium scores, reader 1 identified two subjects and one subject with calcium on the first and second electron-beam CT scans, respectively, but no calcium on the basis of the respective 90-HU spiral CT scores. Conversely, reader 1 identified 14 and 13 subjects with nonzero 90-HU spiral CT scores on the first and second scans, respectively, but no calcium on the respective electron-beam CT scans.
Reader 2 identified six and five subjects with calcium on the first and second electron-beam CT scans, respectively, but no calcium on the basis of respective 90-HU spiral CT scores. Reader 2 identified four and five subjects with nonzero 90-HU spiral CT scores on the first and second scans, respectively, but no calcium on the respective electron-beam CT images.
Risk Scores
The results of stratifying subjects according to potential clinical cardiovascular risk, as proposed by Rumberger et al (12), are presented in Tables 4 and 5. There was only moderate agreement for all comparisons of risk stratification based on electron-beam CT and 130-HU spiral CT calcium scores for both readers and both scans. The values ranged from a low of 0.58 for reader 2, based on scan 1 scores at both CT examinations, to a high of 0.64 for reader 2, based on scan 2 scores. There was a significant difference in the clinical risk stratifications determined by using electron-beam CT and 130-HU spiral CT scores (P < .05) for both readers and both scans. In those subjects with different clinical risk assessments based on the calcium scores measured by using the two scanners, the risk stratification based on the 130-HU spiral CT score usually was an underestimation of that determined by using the electron-beam CT score (Table 4).
|
|
values ranged from a low of 0.55 for reader 1, based on scan 1 calcium scores at both CT examinations, to a high of 0.67 for reader 2, based on scan 2 scores. For reader 1, there was a significant difference between the clinical risk stratifications determined by using electron-beam CT scores and those determined by using 90-HU spiral CT scores. For reader 2, there was no significant difference in clinical risk stratification between the two modalities for either scan 1 (P = .61) or scan 2 (P = .45). In those subjects with different clinical risk assessments based on the calcium scores measured by using the two scanners, the risk stratification based on the 90-HU spiral CT score usually was an overestimation of that based on the electron-beam CT score. |
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Important considerations in interpreting the results of the current study include the relatively small study size and the relative paucity of calcium measurements across a broad clinical spectrum. Nevertheless, we believe this study is unique because it is, to our knowledge, the largest CT calcium measurement comparison study published to date and the only investigation in which repeated measurements were performed in the same subjects. Furthermore, because all scans were scored independently by two experienced readers, the degree of interobserver variation was measurable.
It is also important to note that the spiral CT data sets were obtained by using a single detector array and 0.8-second revolution speed, retrospective image segmentation, and ECG tagging. The differences incurred from this configuration are not generalizable to other spiral technologies that involve different numbers of detector arrays, different methods of image segmentation and ECG registration (eg, prospective or retrospective), and different acquisition times (0.5–1.0 second) and must be evaluated separately. Finally, the Agatston calcium scoring algorithm (9) was used in this study, because it is currently the most commonly used algorithm in clinical practice. The effect of using alternative potentially more useful algorithms, such as calcium volumes (15), was not addressed.
The effect of scanner variables in this study was minimized to the extent that was possible, because both scanners use 3-mm collimation, 512 x 512 image matrices, and approximately the same field of view; these parameters result in nearly identical pixel sizes. The other differences, such as those in kilovolt peak, milliampere value, and scan acquisition time, were suggested by the scanner manufacturers for coronary artery calcium imaging and thus were not altered.
Interobserver Variation
Interobserver variation is an important source of differences in calcium score. In the case of electron-beam CT, intraobserver and interobserver reliability for calcium scoring has been shown to be excellent, with coefficients greater than 0.99 for both intra- and interobserver reliability (16). We are not aware of any published studies on intra- or interobserver variability in calcium scoring at spiral CT. In this study, the interobserver variation between the two experienced readers was comparable between the calcium scores derived from electron-beam CT and 130-HU spiral CT algorithms. There was a trend toward less variation between readers at electron-beam CT image scoring, but a significant difference was not demonstrated. This trend may have been related to less cardiac motion and the lack of choice in selecting images for scoring when electron-beam CT data sets were used. The 90-HU spiral CT algorithm was associated with significantly greater interobserver variation (P < .01) compared with that seen with the electron-beam CT and 130-HU spiral CT algorithms. This may have been due to the effect of the higher calcium scores obtained when the 90-HU gray-scale threshold was used.
Interscan Variation
The reproducibility, or precision, of calcium measurement is an important consideration in the clinical utility of coronary calcium scoring. This variable is particularly important with respect to assessing for interval changes in measurable calcium following interventions (17). Studies (18–20) have addressed interscan variability in calcium scores by using electron-beam CT, and values have ranged from 24% to 49%. We are not aware of any published studies involving the evaluation of interscan variation in calcium scoring at spiral CT. There are many potential subject-based sources of measurement error that contribute to interscan variation, including positioning, cardiac and respiratory motion, the anatomy of the coronary vessels, and poor ECG gating.
Scanner sources of measurement error, such as tube heating, scanning parameters, reconstruction algorithms, detector alignment, and calibration, also are possible, but they are usually easier to control. With the method involving spiral CT software (Smartscore), a further contributor to interscan variation is the requirement to select an optimal image from each anatomic level. The results from this study show that there was no significant difference in the interscan variation in scores for either reader, regardless of whether the data were derived from the electron-beam CT scanner or the spiral scanner with a 130-HU threshold. For reader 1, interscan variation was smaller for calcium scores measured at electron-beam CT compared with those measured at spiral CT with the 90-HU threshold, but this was not true for reader 2. Therefore, our study results suggest that interscan variation is probably comparable between scanners, although the factors that cause the variation may be different for the two units.
Intermodality Variation
Differences in calcium scores between different CT technologies are an area of considerable interest and increasing debate. Becker et al (5) compared electron-beam CT, gated nonspiral CT, and spiral CT with variable pitch for the quantification of coronary artery calcium. These investigators observed a high correlation between scores across all modalities but better agreement between electron-beam CT and gated nonspiral CT data. Similar to us in the current study, Carr et al (6) compared electron-beam CT with retrospectively gated subsecond spiral CT and found good correlation between the two imaging modalities. However, correlation analysis may not be the best approach, given the effect of the clustering of scores around a few values, especially in subjects with no identifiable calcium, and the relatively poor precision of calcium scoring methods. This issue is further complicated by the paucity of clinical outcomes data to determine the degree of precision and accuracy needed for clinical relevance.
The presence or absence of calcium is one well-defined cut point that has been suggested to have clinical utility (eg, in middle-aged patients who present with atypical chest pain). In our study, both readers identified more subjects who had a 130-HU spiral CT score of zero and a nonzero electron-beam CT score than the converse. Thus far, all of the cut points used to stratify subjects according to event risk and management have been based on electron-beam CT data (13,21–24). If these values are to be applied to non–electron-beam CT–derived scores, correlation alone may be insufficient.
When the existing commercially available scoring algorithms for the described two scanners are used, there are significant differences between the scores generated with spiral CT and those generated with electron-beam CT. However, the threshold value used with spiral CT algorithms determines whether the calcium score is likely to be higher or lower than that expected with electron-beam CT. When the lower 90-HU threshold is applied, calcium scores are consistently higher than those derived from electron-beam CT data; the opposite is true when the 130-HU threshold is applied to spiral CT data.
The clinical relevance of these differences is not known, since there currently are no available clinical outcomes data associated with calcium scores determined by using spiral CT. Although mathematical correction algorithms are possible, a better approach may be to investigate whether a new threshold between 90 and 130 HU is preferable or to develop reference values based on alternative scoring algorithms, such as calcium volume, that would be specific for subjects with scores derived by using the spiral CT software (Smartscore). However, given the large clinical database of scores derived from electron-beam CT scans, a method to correlate spiral CT scores with electron-beam CT scores would be clinically useful.
Clinical Risk Stratification
To assess the clinical effect of the different calcium scores, we used the guidelines for clinical stratification suggested by Rumberger et al (12). Our study data suggest that coronary artery calcium scores obtained by using gated retrospective reconstruction algorithms from a spiral CT scanner are not directly comparable to those obtained by using electron-beam CT; this discrepancy results in differences in clinical risk stratification. When a threshold of 130 HU is used for calcium scoring of spiral CT data, a larger number of subjects are given a lower cardiovascular risk assessment and more subjects have no detectable calcium, as compared with the number of subjects with calcium depicted by electron-beam CT. Conversely, when a threshold of 90 HU is used for calcium scoring of spiral CT images, a larger number of subjects are given a higher risk assessment and more are shown to have apparent calcific lesions, compared with the number of subjects with calcium depicted by electron-beam CT.
Radiation Dose
The spiral CT protocol currently used results in a higher radiation dose to the subject than that delivered with the electron-beam CT protocol. The estimated dose for a single coronary calcium scan at electron-beam CT is 1.5 mGy (.15 rad), whereas that with the described spiral CT scanner is 2.4 mGy (.24 rad). These values can be compared to background radiation that is from natural origin, which is 3.0 mSv per year (25). Hence, electron-beam CT scanning results in an effective dose that is equivalent to about 6 months of background radiation, whereas scanning performed with the described spiral CT scanner results in a dose equivalent that is to nearly 10 months of background radiation.
In conclusion, spiral CT has not yet proved to be a feasible alternative to electron-beam CT for coronary artery calcium quantification. The calcium scores are not equivalent. There are systematic differences between calcium scores obtained by using single–detector array, subsecond spiral CT and those obtained by using electron-beam CT. Combining coronary artery calcium scores obtained from different CT technologies may confound attempts to develop reference values for different clinical management recommendations. Using electron-beam CT databases to risk stratify patients who undergo spiral CT may lead to incorrect clinical assessments. Further research is warranted to determine which detector technology, imaging protocol, and scoring algorithm permit the greatest agreement between data obtained by using the two CT scanners.
|
FOOTNOTES |
|---|
Author contributions: Guarantors of integrity of entire study, J.G.G., H.C.Y.; study concepts, J.G.G., H.C.Y., D.R.A.; study design, J.G.G., H.C.Y., S.B.H., L.E.G., M.M.M.G.; literature research, J.G.G., H.C.Y., S.B.H.; clinical studies, J.G.G., S.B.H., A.M.E.; data acquisition, J.J.G., M.M.M.G., M.S.B.; data analysis/interpretation, J.G.G., J.W.S., A.M.E., H.C.Y., L.E.G.; statistical analysis, J.W.S., A.M.E., J.G.G., H.C.Y.; manuscript preparation, J.G.G., H.C.Y., A.M.E.; manuscript definition of intellectual content, J.G.G., H.C.Y.; manuscript editing, H.C.Y., J.G.G., A.M.E., D.R.A.; manuscript revision/review, H.C.Y., J.G.G., A.M.E.; manuscript final version approval, H.C.Y., J.G.G.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  E S Houslay, S J Cowell, R J Prescott, J Reid, J Burton, D B Northridge, N A Boon, D E Newby, and on behalf of the Scottish Aortic Stenosis and Lipi Progressive coronary calcification despite intensive lipid-lowering treatment: a randomised controlled trial Heart, September 1, 2006; 92(9): 1207 - 1212. [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]
|
|
|
|
 |
|
 R. C. Detrano, M. Anderson, J. Nelson, N. D. Wong, J. J. Carr, M. McNitt-Gray, and D. E. Bild Coronary Calcium Measurements: Effect of CT Scanner Type and Calcium Measure on Rescan Reproducibility--MESA Study Radiology, August 1, 2005; 236(2): 477 - 484. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
U. J. Schoepf, C. R. Becker, B. M. Ohnesorge, and E. K. Yucel CT of Coronary Artery Disease Radiology, July 1, 2004; 232(1): 18 - 37. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Horiguchi, H. Yamamoto, Y. Akiyama, K. Marukawa, N. Hirai, and K. Ito Coronary Artery Calcium Scoring Using 16-MDCT and a Retrospective ECG-Gating Reconstruction Algorithm Am. J. Roentgenol., July 1, 2004; 183(1): 103 - 108. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. R. Van Hoe, K. G. De Meerleer, P. Ph. Leyman, and P. K. Vanhoenacker Coronary Artery Calcium Scoring Using ECG-Gated Multidetector CT: Effect of Individually Optimized Image-Reconstruction Windows on Image Quality and Measurement Reproducibility Am. J. Roentgenol., October 1, 2003; 181(4): 1093 - 1100. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. A. Rumberger and L. Kaufman A Rosetta Stone for Coronary Calcium Risk Stratification: Agatston, Volume, and Mass Scores in 11,490 Individuals Am. J. Roentgenol., September 1, 2003; 181(3): 743 - 748. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Takahashi and K. T. Bae Quantification of Coronary Artery Calcium with Multi-Detector row CT: Assessing Interscan Variability with Different Tube Currents—Pilot Study Radiology, July 1, 2003; 228(1): 101 - 106. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Hecht, J. A. Rumberger, M. J. Budoff, R. Park, R. Detrano, P. Fu, Y. Ibrahim, M. Xiang, L. LaBree, and S. Azen C-Reactive Protein and Electron Beam Tomography * Reply Circulation, May 13, 2003; 107 (18): e123 - e124. [Full Text] [PDF]
|
|
|
|
|
|
L. Berlin Liability of Performing CT Screening for Coronary Artery Disease and Lung Cancer Am. J. Roentgenol., October 1, 2002; 179(4): 837 - 842. [Full Text] [PDF]
|
|
|
|
|
|
B. Lu, N. Zhuang, S.-S. Mao, J. Child, S. Carson, H. Bakhsheshi, and M. J. Budoff EKG-triggered CT Data Acquisition to Reduce Variability in Coronary Arterial Calcium Score Radiology, September 1, 2002; 224(3): 838 - 844. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. J. Budoff, G. A. Diamond, P. Raggi, Y. Arad, A. D. Guerci, T. Q. Callister, and D. Berman Continuous Probabilistic Prediction of Angiographically Significant Coronary Artery Disease Using Electron Beam Tomography Circulation, April 16, 2002; 105(15): 1791 - 1796. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||
