| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Cardiac Imaging |
1 From the Departments of Cardiology (G.T.L., D.B.B., S.B.F., L.K.) and Radiology (L.J.R., M.C.S., L.A.W.), Concord Repatriation General Hospital, Hospital Rd, 3 West, Concord, NSW 2139, Australia; Vascular Biology Laboratory, ANZAC Research Institute, University of Sydney, Concord, Australia (G.T.L., D.B.B., S.B.F., L.K.); Institute for International Health, University of Sydney, Australia (S.K.L.); and Centre for Thrombosis and Vascular Research, University of New South Wales, Kensington, Australia (L.K.). Received November 12, 2003; revision requested February 5, 2004; final revision received July 4; accepted July 26. Supported by unrestricted grants from the Departments of Cardiology and Radiology, Concord Hospital, and the National Heart Foundation of Australia and a Pfizer Cardiovascular Lipid Research Grant. Address correspondence to L.K. (e-mail: l.kritharides@unsw.edu.au).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSION REFERENCES |
|---|
MATERIALS AND METHODS: The study was approved by the institutional review board Human Research Ethics Committee, and all patients completed written informed consent. Fifty patients (40 men, 10 women) aged 62 years ± 11 (± standard deviation) who were suspected of having coronary artery disease underwent both conventional coronary angiography and multisection coronary CT angiography with CS. Sensitivity and specificity of CS, CT angiography, and both methods combined in demonstrating luminal stenosis greater than or equal to 50% were determined for each arterial segment, coronary vessel, and patient. Receiver operating characteristic (ROC) curves were generated for CS prediction of significant stenosis, and the Mann-Whitney U test was used for comparison of CS between groups.
RESULTS: When used with segment-specific electrocardiographic phase reconstructions, CT angiography demonstrated stenosed segments with 79% sensitivity and 95% specificity. Mean calcium score was greater in segments, vessels, and patients with stenoses than in segments, vessels, and patients without stenoses (P < .001 for all); nine (16%) of 56 stenosed segments, however, had a calcium score of 0. The patient calcium score correlated strongly with the number of stenosed arteries (Spearman 
= 0.75, P < .001). CS was more accurate in demonstrating stenosis in patients than in segments (areas under ROC curve were 0.88 and 0.74, respectively). CT angiography, however, was more accurate than CS in demonstrating stenosis in patients, vessels, and segments. The sensitivity and specificity of CS varied according to the threshold used, but when the calcium score cutoff (ie, >150) matched the specificity of CT angiography (95%), the sensitivity of CS in demonstrating stenosed segments was 29% (compared with 79% for CT angiography). Combining CT angiography with CS (at threshold of 400) improved the sensitivity of CT angiography (from 93% to 100%) in demonstrating significant coronary disease in patients, without a loss of specificity (85%); this finding, however, was not statistically significant.
CONCLUSION: CT angiography is more accurate than CS in demonstrating coronary stenoses. A patient calcium score of greater than or equal to 400, however, can be used to potentially identify patients with significant coronary stenoses not detected at CT angiography.
© RSNA, 2005
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSION REFERENCES |
|---|
Calcification of coronary arteries is characteristic of atherosclerotic disease and can be assessed by using electron-beam and multisection CT (6). Coronary calcification is associated with future cardiac events (7,8), can be modulated by using medical therapy (9), and is associated with coronary luminal stenoses (10,11). At low thresholds, coronary calcium scoring (CS) was reported to be sensitive but not specific in demonstrating luminal stenoses; at high thresholds (eg, patient calcium score thresholds above 310), the reverse was reported (11). It is unknown whether the quantification of arterial calcification can complement multisection CT angiography and improve the sensitivity, specificity, or both of CT angiography in detection of coronary stenoses.
To date, CS and CT angiography have been used almost exclusively to screen patients for risk of coronary artery disease or risk of future cardiac events. It is not clear whether these techniques have the potential to demonstrate new stenotic lesions within individual coronary arteries when patients are followed up prospectively. Such information would extend the use of CS and CT angiography beyond simple screening for risk to include monitoring for the appearance of new coronary lesions during long-term medical therapy. Before such a practical application can be realized, CS and CT angiography must be demonstrably useful in assessing individual coronary segments. Thus, the purpose of the study was to investigate prospectively the relative accuracy of CT angiography, CS, and both methods combined in demonstrating coronary artery stenoses by using conventional angiography as the reference standard.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSION REFERENCES |
|---|
Imaging
Conventional angiography, which was used as the reference standard, was performed with a digital fluorography system (DFP-2000A; Toshiba Medical Systems, Tokyo, Japan) by using a femoral approach with the standard Judkin technique (D.B.B., S.B.F., L.K., each with >10 years of experience). A minimum of four views of the left coronary artery and two views of the right coronary artery were obtained and recorded for analysis. Extra views were obtained if suspected stenoses were inadequately seen on initial views or if there was an overlap of arteries.
For CT angiography, the target heart rate was 65 beats per minute (1), and, to avoid excessive bradycardia and hypotension in the research cohort, only those patients with a heart rate of more than 10 beats per minute more than that of the target heart rate were given ß-blocker medication before scanning. Sixteen patients were already taking ß-blocker medication. Nine patients who were not taking ß-blocker medication were given 50–100 mg of metoprolol tartrate (Metohexal; Hexal, Pyrmont, Australia) orally 30 minutes before scanning.
For calcification assessments, unenhanced electrocardiographically (ECG)-gated cardiac CT was performed by using a four–detector row CT scanner (LightSpeed Plus; GE Medical Systems, Milwaukee, Wis) with 2.5-mm section thickness (4 x 2.5-mm collimation), pitch of 1.3–1.5 (depending on heart rate), 0.5-second rotation time, 140 kV, and 250 mA. For contrast material–enhanced CT, the transit time of the contrast material from the intravenous cannula to the ascending aorta was calculated with data from a timing scan obtained by using 20 mL of intravenous nonionic contrast material (iopromide, Ultravist 300; Schering, Berlin, Germany), and 8 seconds was added (as recommended by the CT manufacturer) to give the scanning delay for CT angiography. For assessment of luminal stenoses, single breath-hold ECG-gated CT was performed (4 x 1.25-mm collimation, pitch of 1.3–1.5, 0.5-second rotation time, 140 kV, and 270 mA) with 150 mL of intravenous iopromide administered at 3.5 mL/sec. The mean scanning delay was 26 seconds (range, 18–36 seconds), and the mean breath hold was 32 seconds (range, 26–38 seconds).
Unenhanced CT angiograms were reconstructed from a data acquisition window centered at 70% of the R-R interval at ECG gating. Contrast-enhanced CT angiograms were reconstructed from data acquisition windows centered at 10% intervals from 0% to 90% of the R-R interval. The temporal resolution was 250 msec. Algorithms that used data from different cardiac cycles to reconstruct an image and reduce temporal resolution were not used because the required breath-hold times for scanning would have increased to more than 35 seconds. There were no major adverse reactions to conventional angiography or CT angiography.
Image Analysis
For analysis of arterial segments on conventional angiograms and on CT angiograms, the coronary artery tree was divided into 13 segments. Segments 1–4 corresponded to the proximal, middle, distal right coronary, and posterior descending arteries; segment 5, to the left main artery; segments 6–9, to the proximal, middle, distal left anterior descending, and first diagonal arteries; and segments 11–13 and 15, to the proximal, middle, distal left circumflex, and first obtuse marginal arteries, as defined by the American Heart Association (12). All segments 2.0 mm or larger in diameter, as measured at conventional angiography, were included in the analyses of conventional angiography, CT angiography, and CS. Other segments were not analyzed because of the limitations of CT resolution and because such segments rarely undergo revascularization. Conventional angiograms and CT angiograms were evaluated by independent reviewers (L.K., D.B.B., L.J.R., L.A.W., G.T.L.) who were blinded to each other’s results.
Two cardiologists (L.K. and D.B.B., each with >10 years of experience) independently assessed each segment on the conventional angiogram by using off-line quantitative coronary vessel analysis (DFP-2000A; Toshiba Medical Systems) to quantify the reference diameter and the severity of stenoses in the view that showed the largest reduction in diameter for the segment in question (13). The results of the two reviewers were averaged, except when the results varied by more than 10%. In these cases, the differences were resolved by consensus. Significant luminal stenosis was defined a reduction in lumen diameter greater than or equal to 50% (2–4).
Two radiologists (L.J.R. and L.A.W., with 3 and 2 years of experience, respectively) independently assessed each segment on unenhanced CT angiograms for the amount of calcification. The presence of calcification was determined by using the Agatston method for multi–detector row CT with a 130-HU threshold (14,15). Two radiologists (L.J.R. and G.T.L., each with 3 years of experience) independently assessed each segment on contrast-enhanced CT angiograms for stenosis. All CT angiographic assessments were performed by using a computer workstation (Card IQ, Advantage Workstation 4.0; GE Medical Systems). Differences between the two CT angiographic assessments (presence or absence of 
50% stenosis) were resolved by a joint consensus reading of the CT angiograms. The calcium scores in each segment, vessel, and patient were termed the segment calcium score, the vessel calcium score, and the patient calcium score, respectively. (The vessel calcium score was thus equal to the sum of the segment calcium scores for that vessel, and the patient calcium score was equal to the sum of the segment calcium scores for that patient.) The calcium scores determined by each of the two reviewers were averaged. In two patients, the calcium score of the two reviewers varied by more than 20% (and the absolute difference was >20). The results from these two patients were reviewed, and consensus was reached as to whether calcium was within the coronary tree or external to it. The reviewers then rescored the images of two patients individually.
To define optimal ECG phases for each segment, images were reconstructed from all the ECG phases by using transverse images, three-dimensional multiplanar reconstructions (maximal intensity projection), and automated vessel-tracking curvilinear reconstructions. The optimal ECG phase for image quality was then selected and recorded (L.J.R.). Because diastolic phases are more commonly used in the literature (2–5,16), if a segment was seen equally well in multiple phases, preference was given to mid-diastolic phases when nominating the optimal phase. For CT angiography, each segment was classified as significantly stenosed (50% stenosis), not significantly stenosed, or not assessable.
Combining CS with CT Angiography
We investigated two approaches for combining CS with CT angiography. First, we determined whether a high calcium score could be used to detect stenoses missed at CT angiography, and, second, we determined whether a low calcium score (that is, a score of 0) could reliably exclude stenoses erroneously detected at CT angiography (17). A high calcium score cutoff of 400 per patient (specificity of 97%) was chosen because this score was compatible with previously published thresholds that were highly predictive of coronary disease (18,19) and excluded coronary stenoses in patients with an arbitrary a priori desired specificity of more than 95%. The same level of specificity (97%) was then applied to CS evaluations of vessels and segments to obtain high calcium score cutoffs for CS and CT angiography combined. At evaluation of combined high CS information with results from CT angiography, a segment, vessel, or patient with a calcium score that exceeded the upper calcium score threshold indicated the presence of a stenosis, even if results from CT angiography had indicated absence of disease. When a calcium score of 0 was used to exclude stenoses, a segment, vessel, or patient with a calcium score of 0 indicated absence of stenosis, regardless of the results at CT angiography. The CT angiographic result was used for the remainder of the segments.
Statistical Analysis
SPSS software (version 10.0; SPSS, Chicago, Ill) was used for all statistical analyses and graphs, and PASS 2002 software (Number Cruncher Statistical Systems, Kaysville, Utah) was used for power calculations. Sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of CT angiography and of conventional angiography were calculated for each segment, vessel, and patient. As in previous studies (1–4,20), for comparison of CT angiography with conventional angiography, segments that were not assessable at CT angiography were deemed to be free of stenoses because this assumption affects sensitivity, specificity, positive predictive value, and accuracy less than if nonassessable segments are assumed to have stenoses. Classification of such segments as free of stenoses will, however, cause underestimation of sensitivity. (For completeness, calculations were repeated with the assumption that stenoses were present in segments that were not assessable at CT angiography and after exclusion of segments not assessable at CT angiography [5,21,22].)
Receiver operating characteristic curves were generated for CS prediction of significant stenosis in segments, vessels, and patients. Student t test was used to test the differences between age and sex, and the Mann-Whitney U test was used for comparing results of CS between groups. A two-tailed P value of <.05 indicated a statistically significant difference. Intraclass correlation coefficient (2,1) was used to assess interrater variability. Spearman was used to assess the association between patient calcium score and the number of stenosed vessels in a patient. The McNemar test was used to assess the improvement in sensitivity of CS and CT angiography combined compared with that of CT angiography alone, and a sample size of 54 patients was used to identify a 14% improvement with 80% power at a one-sided significance level of .05, assuming an estimated CT angiography sensitivity of 85% (1).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSION REFERENCES |
|---|
CT Angiography
Overall, 407 (85%) of 479 coronary segments were deemed assessable at CT angiography. Of the 15% that were not assessable, 41 (9%) of 479 segments were affected by motion artifact, 17 (4%) were obscured by heavy calcification, and 14 (3%) were inadequately opacified by contrast material. The most frequently nonassessable segment was the middle right coronary artery (n = 11). The average heart rate during scanning was 62 beats per minute ± 9 (range, 46–80 beats per minute), and only two patients had a heart rate of more than 75 beats per minute.
Evaluation of individual coronary segments requires optimal assessment of each segment, and, to maximize the number of segments that were assessable, variable phase analysis was used. Mid-diastolic ECG phases (70% and 80% of the R-R interval) were optimal for most segments (Table 1). For 29% of segments, however, the optimal ECG phase was outside this range. In 12% of segments, the optimal ECG phase was the isometric relaxation phase of early diastole (40% phase). Ten (18%) of 56 segments with stenosis were best seen at this early diastolic phase, and a number of artifacts apparent on middiastolic images could be resolved by using the 40% phase image (Fig 1). Individual segments within the same vessel had different optimal ECG phases in 60 (40%) of 149 arteries, excluding the left main coronary artery.
|
|
Results from CT angiography varied slightly according to whether analysis was performed according to arterial segment, vessel, or patient (Table 2), but for all analyses the sensitivity of CT angiography was greater than or equal to 79% and specificity was greater than or equal to 85%. CT angiographic results correctly indicated 17 (85%) of 20 patients without significant disease, 11 (79%) of 14 patients with single-vessel disease, and 15 (94%) of 16 patients with multivessel disease.
|
|
|
|
= 0.75) (Fig 4). CT angiography was more accurate than CS in demonstrating stenoses, as is shown by the position of CT angiography to the left of the receiver operating characteristic curves for CS in Figure 2.
|
|
|
|
When the results of CS and CT angiography were combined, a calcium score of 0 did not reliably exclude stenosis. Nine (16%) of 56 arterial segments with a calcium score of 0 contained angiographic stenoses, and seven of these were correctly identified at CT angiography. Similarly, CT angiography was used to correctly identify one patient who had stenosis, despite a calcium score of 0. Combining a calcium score of 0 with results from CT angiography decreased the sensitivity of the combined test relative to that of CT angiography alone (79% vs 70% for segments, 80% vs 78% for vessels, and 93% vs 90% for patients) without any improvement in specificity.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSION REFERENCES |
|---|
To the best of our knowledge, this is the first study to show that the quantification of coronary artery calcium at CS has the potential to enhance diagnosis of coronary disease in patients at CT angiography. A patient calcium score greater than or equal to 400 increases the sensitivity of CT angiography without a loss of specificity in demonstrating significant coronary disease in patients. This may improve the identification of patients requiring further investigation with conventional coronary angiography. Moreover, this result represents a practical improvement in CT diagnosis because it uses a potential confounder of CT angiography—excessive arterial calcification—to improve the accuracy of CT angiography. We suggest that CS be quantified from unenhanced CT angiograms whenever a patient is referred for CT angiography. If CT angiography does not show stenosis but the patient’s calcium score is greater than or equal to 400, then the patient should be considered likely to have unrecognized stenotic coronary artery disease.
Study Limitations
Because the sensitivity of CT angiography in demonstrating stenotic coronary artery disease in patients was higher than expected (93% versus 85%), an improvement from 93% to 100% was not statistically significant. Additionally, all patients in the study had clinical indications for conventional coronary angiography. Thus, a larger study will be needed both to validate the proposed CS and CT angiography algorithm in this group and to screen patients in an asymptomatic population. It is also possible that the use of CT scanners and image processing techniques with better temporal and spatial resolutions may further reduce the benefit of the combined CS and CT angiography approach.
Optimizing CT Angiography
Coronary artery visualization varies with phases of the cardiac cycle. In most multisection CT angiographic studies of the coronary arteries (1–5,20–22,24), investigators have routinely analyzed images for stenosis from a single ECG phase, usually after selecting the ECG phase with least motion artifact for each vessel or patient. Improved vessel imaging could be achieved in 40% of cases if the best ECG phase for each segment was used rather than the single best phase for the whole vessel, although this may be impractical on some computer workstations. In addition, a number of segments were best evaluated during isometric relaxation in early diastole (40% phase). Ventricular volume is most constant during this time and during mid-diastole (70% and 80% phases) (Fig 5). If images from mid-diastolic to late diastolic phases (approximately 50%–80% of the R-R interval) had been used exclusively as described (1), 92 (19%) of 479 segments would have been suboptimally assessed. The use of segment-specific selection of ECG phases from systole and diastole with our four–detector row CT scanner resulted in a sensitivity (79%) and specificity (95%) comparable to those obtained in other studies with a 16–detector row CT scanner (73%–95% and 86%–94%, respectively) (Table 3), despite the superior spatial (0.75-mm vs 1.0-mm section thickness) and temporal (
210 msec vs 250 msec) resolution of 16–detector row technology.
|
|
Radiologic techniques are less sensitive than histologic analysis in demonstrating calcification (25); thus, small amounts of calcium that are present in some lesions will not be detected at CT. Moreover, lipid-rich plaques, which are prone to rupture, often contain less fibrous tissue and less calcium than some stable plaques (26). Segmental disease, therefore, should not be excluded on the basis of CS alone. This is supported by findings from previous studies (11,27) that showed that a low calcium score or a score of 0 did not reliably exclude patients from having substantial coronary disease.
Although a calcium score of 0 did not reliably exclude disease in individual segments, extensive calcification was used to identify patients and arteries with significant coronary disease and to aid in the detection of coronary disease at CT angiography. Because calcification is a source of inaccuracy at CT angiography, the ability to usefully incorporate calcification data during CT angiography may improve patient categorization in future studies.
In conclusion, although CT angiography is more accurate than CS in demonstrating stenotic coronary artery disease, a calcium score greater than or equal to 400 may help diagnose disease that is not detected at CT angiography, and CT angiography and CS combined have the potential to improve the detection of coronary stenoses relative to CT angiography alone. Segment-specific ECG phase reconstructions may improve the accuracy of CT angiography in demonstrating coronary artery stenoses. CS is of greater use in demonstrating coronary stenoses in patients than in localizing stenoses to individual coronary vessels or segments. The absence of detectable calcium does not reliably exclude substantial coronary disease.
|
FOOTNOTES |
|---|
Authors stated no financial relationship to disclose.
Author contributions: Guarantor of integrity of entire study, G.T.L.; study concepts and design, G.T.L., L.K.; literature research, G.T.L., L.K., L.J.R.; clinical studies, G.T.L., L.J.R., M.C.S., L.K., D.B.B., S.B.F.; data acquisition, G.T.L., L.J.R., M.C.S.; data analysis/interpretation, all authors; statistical analysis, G.T.L., S.K.L.; manuscript preparation, definition of intellectual content and editing, G.T.L., L.K.; manuscript revision/review and final version approval, all authors
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSION REFERENCES |
|---|
50% in diameter in each of the three major arteries by amounts of calcium detected by electron beam computed tomographic scanning in patients with chest pain. Am J Cardiol 1998; 81:778-780.[CrossRef][Medline]
This article has been cited by other articles:
|
|
 |
|
  A. Sarwar, L. J. Shaw, M. D. Shapiro, R. Blankstein, U. Hoffman, R. C. Cury, S. Abbara, T. J. Brady, M. J. Budoff, R. S. Blumenthal, et al. Diagnostic and Prognostic Value of Absence of Coronary Artery Calcification J. Am. Coll. Cardiol. Img., June 1, 2009; 2(6): 675 - 688. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Sundaram, S. Patel, N. Bogot, and E. A. Kazerooni Anatomy and Terminology for the Interpretation and Reporting of Cardiac MDCT: Part 1, Structured Report, Coronary Calcium Screening, and Coronary Artery Anatomy Am. J. Roentgenol., March 1, 2009; 192(3): 574 - 583. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. O'Brien, M. O. Baerlocher, M. Asch, and A. Myers Role of radiology in geriatric care: A primer for family physicians Can Fam Physician, January 1, 2009; 55(1): 32 - 37. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S Leschka, H Scheffel, L Desbiolles, A Plass, O Gaemperli, P Stolzmann, M Genoni, T Luescher, B Marincek, P Kaufmann, et al. Combining dual-source computed tomography coronary angiography and calcium scoring: added value for the assessment of coronary artery disease Heart, September 1, 2008; 94(9): 1154 - 1161. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. K. Vanhoenacker, M. H. Heijenbrok-Kal, R. Van Heste, I. Decramer, L. R. Van Hoe, W. Wijns, and M. G. M. Hunink Diagnostic Performance of Multidetector CT Angiography for Assessment of Coronary Artery Disease: Meta-analysis Radiology, August 1, 2007; 244(2): 419 - 428. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. T. Lau, L. J. Ridley, P. G. Bannon, L. A. Wong, J. Trieu, D. B. Brieger, H. C. Lowe, B. S. Freedman, and L. Kritharides Lumen Loss in the First Year in Saphenous Vein Grafts Is Predominantly a Result of Negative Remodeling of the Whole Vessel Rather Than a Result of Changes in Wall Thickness Circulation, July 4, 2006; 114(1_suppl): I-435 - I-440. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Schwaiger, S. Ziegler, and S. G. Nekolla PET/CT: Challenge for Nuclear Cardiology J. Nucl. Med., October 1, 2005; 46(10): 1664 - 1678. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
) and CS and CT angiography combined, which was determined by using high-specificity threshold for stenoses (
), and the point on the receiver operating characteristic curve with calcium score threshold (CS) of equivalent sensitivity to that of CT angiography alone (arrow) are noted.
