| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Medical Radiology, Institute of Diagnostic Radiology, University
Hospital Zurich, Raemistr. 100, 8091 Zurich, Switzerland.
2 Clinic for Cardiovascular Surgery, University Hospital Zurich, Zurich,
Switzerland.
Received March 28, 2008;
accepted after revision June 15, 2008.
Supported by the National Center of Competence in Research, Computer Aided
and Image Guided Medical Interventions, of the Swiss National Science
Foundation.
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. One hundred consecutive patients (42 women and
58 men; mean age, 65.8 ± 6.5 years) with a sinus rhythm and heart rates
< 70 beats per minute were included. Two independent, blinded readers
classified coronary artery segments as being of diagnostic or nondiagnostic
image quality and assessed each segment with diagnostic image quality for the
presence of significant coronary stenoses. Nondiagnostic segments were
excluded from analysis. Sensitivity, specificity, positive predictive value
(PPV), and negative predictive value (NPV) were calculated for all patients
and for the subgroup of patients with a low or high calcium score (group A,
median Agatston score < 316; group B, 
316). Catheter angiography was
used as the reference standard. Effective radiation dose values were
calculated.
RESULTS. In 89 of 100 patients (89%), 1,462 of 1,524 coronary segments (96%) were depicted with diagnostic image quality. The overall sensitivity, specificity, PPV, and NPV were 98%, 99%, 95%, and 100%, respectively. The rate of segments with nondiagnostic image quality was significantly higher (p < 0.001) in group B compared with group A. In group A, sensitivity, specificity, PPV, and NPV were 99%, 99%, 94%, and 100%, respectively, and in group B, 98%, 99%, 94%, and 99%, respectively, with no significant differences between the groups. The average effective radiation dose was 2.6 ± 0.8 mSv (range, 1.2–4.4 mSv).
CONCLUSION. Dual-source CT coronary angiography with use of prospective ECG triggering performs accurately in the assessment of coronary artery disease at low radiation doses. Diagnostic accuracy remains high despite the presence of heavy calcifications but is associated with an increased rate of nondiagnostic segments.
Keywords: calcifications • CT coronary angiography • diagnostic accuracy • low dose • prospective ECG triggering
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Using dual-source CT, Gutstein and colleagues [9] developed and successfully applied selection criteria for predicting patients in whom prospective ECG triggering is likely to be diagnostic. The authors have found the low-dose technique to be feasible in patients with a body mass index (BMI) below 30 kg/m2, a heart rate below 70 beats per minute (bpm), a low heart rate variation, and an Agatston score below 400. To date, however, there are no data available assessing the influence of calcifications on the diagnostic accuracy of coronary CTA with prospective ECG triggering [12].
Thus, the purpose of this study was to investigate the diagnostic accuracy of dual-source coronary CTA with prospective ECG triggering compared with conventional catheter angiography and to determine the influence of vessel wall calcifications.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Reasons for referral were atypical chest pain in 81 patients (74%) and typical chest pain in nine patients (8%). Another 20 patients (18%) were referred for the exclusion of CAD before cardiac valvular surgery. All patients underwent both imaging techniques within a time interval of 10 days (mean, 5.8 ± 2.3 days). For all patients, clinical data were collected including age, sex, body weight, body height, symptoms, and common cardiovascular risk factors. The BMI was calculated from body weight and body height.
General exclusion criteria for contrast-enhanced CT included nephropathy
(serum creatinine level > 150 µmol/L) (n = 2) and known
hypersensitivity to iodine-containing contrast media (n = 0).
Patients with nonsinus rhythm were excluded from the study (n = 4)
because CT using prospective ECG triggering requires a regular heart rhythm
[8]. If a target heart rate of
70 bpm [9] could not be
achieved by the administration of IV β-blockers, the patient was excluded
from the study.
Twenty-five patients showed mean heart rates > 70 bpm during the preparation time of CT. Because none of the patients showed contraindications for the administration of IV β-blockers, metoprolol (Beloc, AstraZeneca) for heart rate reduction was administered in all of these 25 patients (100%). The mean heart rate could be sufficiently lowered in 21 patients (84%), and the remaining four patients (16%) had to be excluded from the study because the heart rate still exceeded 70 bpm.
Thus, 100 patients (62 men and 38 women; mean age, 64.2 ± 6.5 years; age range, 40–82 years) were included in this study. Our local ethics committee approved the study; written informed consent was obtained from all patients.
CT Data Acquisition and Postprocessing
All CT examinations were performed on a dual-source CT scanner (Somatom
Definition, Siemens Medical Solutions) using prospective ECG triggering. An
initial unenhanced scan was performed for calcium scoring. Then, all patients
received a single 2.5-mg dose of sublingual isosorbide dinitrate (Isoket,
Schwarz Pharma). For coronary CTA, the test bolus technique was used to
determine the circulation time of the contrast media (10 mL of Visipaque 320
[320 mg I/mL], GE Healthcare, and a 50-mL chaser bolus of saline solution) by
the use of a dual-head power injector (Stellant, Medrad). In patients with a
BMI 25 kg/m2, the contrast medium was injected at 1.0 mL/kg of
body weight with an injection rate of 5.5 mL/s. In patients presenting with a
BMI < 25 kg/m2, the contrast medium was injected at 0.8 mL/kg of
body weight at an injection rate of 4.4 mL/s
[13]. All injections were
followed by a chaser bolus diluted by saline solution (20% contrast medium and
80% saline solution) with the same amount as the first phase. Both the
unenhanced and the contrast-enhanced CT scans were obtained in a craniocaudal
direction from the level of the carina to the diaphragm. Data were acquired
using prospective ECG triggering with a detector collimation of 2 x 32
x 0.6 mm, slice collimation of 2 x 64 x 0.6 mm, by means of
a z-flying spot [14],
and gantry rotation time of 0.33 second. Attenuation-based tube current
modulation was used with a reference tube current–time product set at
100 mAs per rotation for calcium scoring and 190 mAs per rotation for coronary
angiography. Prospective ECG triggering with dual-source CT needed a minimum
cycle time of 1.36 seconds for one acquisition and the subsequent table feed.
The data acquisition window of 110 milliseconds was set at 70% of the R-R
interval; the temporal resolution was 83 milliseconds. Patients with a BMI
25 kg/m2 (n = 66) were examined with a tube voltage
of 120 kV; patients with a BMI < 25 kg/m2 (n = 34) were
examined with a tube voltage of 100 kV, as previously indicated
[9,
13].
Unenhanced CT angiograms were reconstructed in a monosegment mode using 3-mm-thick non overlapping slices (reconstruction kernel B35). Contrast-enhanced CT scans were reconstructed in a monosegment mode with a nonoverlapping slice thickness of 0.6 mm, using a medium smooth-tissue convolution kernel (B30f). If the vessel segment was calcified, additional reconstructions were performed using the same slice thickness of 0.6 mm using a sharp-tissue convolution kernel (B45f) to compensate for blooming artifacts. Images were transferred to an external workstation (Multi-Modality Workplace, Siemens Medical Solutions) for further analysis.
CT Data Analysis
Coronary segments were defined according to a scheme proposed by the
American Heart Association
[15]. The intermediate artery
was designated as segment 16, if present, and was considered to belong to the
left anterior descending coronary artery. All diameter measurements were
performed with an electronic caliper tool.
Calcifications were quantified with scoring software (Syngo CaScore, Siemens Medical Solutions). All lesions with a detection threshold of > 130 HU were marked by an experienced observer, and the calcium load in each patient was computed using the method described by Agatston et al. [16].
Coronary CTA data analysis was performed by two independent observers who were blinded to the clinical history and to the results from conventional catheter angiography. All reconstructed images were evaluated using transverse source images and multiplanar reformations.
In a first step, both readers independently rated the image quality of each coronary segment as being diagnostic or nondiagnostic. Reasons for nondiagnostic image quality were assigned to motion artifacts, stairstep artifacts, image noise, or severe vessel wall calcifications. Segments with a diameter of at least 1 mm at their origin were included. Vessel segments distal to occlusions were excluded from analysis. Coronary segments with stents were excluded. If one reader classified a vessel segment as being nondiagnostic, the segment was rated as being nondiagnostic independent of an opposite rating of the other reader.
Then both observers independently assessed all coronary artery segments qualitatively for the presence of significant stenoses, which were defined as a luminal diameter narrowing > 50%. In case of disagreement, both observers appended a consensus reading within a week after the initial, independent CT readout.
Quantitative Coronary Angiography
Conventional catheter angiography was performed according to standard
techniques and at least two views in different planes were obtained for each
coronary artery. One reader, blinded to the CT results, evaluated all
angiograms with regard to the presence or absence of significant stenoses
using computerized quantitative coronary angiography analysis software
(Xcelera, Philips Healthcare). After averaging the results from the two
orthogonal views, narrowing greater than 50% of the luminal diameter in
relation to the reference diameters was defined as a significant stenosis, as
previously shown [17,
18]. Coronary artery segments
were defined according to the same scheme as for CT
[15].
Radiation Dose Estimation
The effective dose values of coronary CTA were calculated as proposed by
the European Working Group for Guidelines on Quality Criteria in CT
[19]. Values were derived from
the dose–length product (DLP) and the conversion coefficient for the
chest (k = 0.017 mSv/mGy x cm) as the examined anatomic region. Scanning
range, mean rate, and DLP were obtained from the protocol that summarized the
relevant individual radiation exposure parameters of each coronary CTA
examination [20].
Statistical Analysis
All statistical analyses were performed by using commercially available
SPSS software, version 15.0.0. Continuous variables were expressed as mean
± SD, and categoric variables were expressed as frequencies or
percentages. Sensitivity, specificity, positive predictive value (PPV), and
negative predictive value (NPV) for the detection of significant coronary
artery stenosis were calculated from chi-square tests of contingency, and the
95% CI was calculated from binomial expression. Quantitative coronary
angiography (QCA) was considered as the standard of reference. Per-segment,
per-vessel, and per-patient accuracy were calculated for all segments with
diagnostic image quality. Nondiagnostic segments were excluded from
analysis.
The median Agatston score was used as a cutoff point to subdivide the patients into two Agatston score groups. Comparisons of the two groups regarding Agatston score, age, heart rate, scanning time, scanning range, and contrast volume were performed using the Mann-Whitney U test. Differences of the two subgroups regarding the rate of patients with a BMI < 25 kg/m2, nondiagnostic segments, and significant coronary artery stenosis were tested for significance by using the chisquare test or the Fisher's exact test, if appropriate. Interobserver agreements concerning the classifications of coronary segments as having significant coronary artery stenoses were evaluated using kappa statistics. Fisher's exact test was used to test for significant differences in the rate of required consensus reading between the groups. Comparisons of the two groups with regard to the diagnostic accuracy were performed using the chi-square test.
Comparisons of the protocols using different kilovolt settings and comparisons of dose estimates were performed by using the Mann-Whitney U test. Differences regarding the rate of segments with nondiagnostic image quality and the diagnostic accuracy were tested for significance using the chi-square test or the Fisher's exact test, if appropriate. A p value of < 0.05 was considered statistically significant.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
316 are
summarized in Table 1.
|
Quantitative Coronary Angiography
QCA identified 206 coronary segments with significant stenosis. These were
present in 55 patients (55%). In group A, the rate of coronary segments
(p < 0.001), the rate of coronary vessels (p < 0.001),
and the rate of patients (p < 0.05) with significant stenosis were
significantly lower compared with group B
(Table 1).
Image Quality
A total of 1,524 coronary artery segments were evaluated with CT.
Seventy-six segments were missing because of anatomic variants (n =
25), were not included due to a diameter of < 1 mm at their origin
(n = 36), were excluded because of the presence of stents (n
= 8), or were distal to an occluded segment (n = 7).
A total of 1,462 of the 1,524 segments (96%) were depicted with diagnostic image quality. Sixty-two segments (4%) in 11 patients (11%) were considered to be of nondiagnostic image quality. Reasons for nondiagnostic image quality were motion artifacts in 24% (15/62) of the segments in five patients, stairstep artifacts in 31% (19/62) of the segments in seven patients, and vessel wall calcifications in 45% (28/62) of the segments in seven patients. Image noise was not found as the cause of nondiagnostic image quality in any patient. The rate of segments with nondiagnostic image quality was significantly higher (p < 0.001) in group B compared with group A (Table 1).
Diagnostic Performance
Interobserver agreement for the detection of significant coronary stenosis
was good in all patients, excellent in group A, and good in group B (
=
0.77, all patients; = 0.81, group A; and = 0.74, group B).
Consensus reading was required in 52 segments (n = 22 in group A,
n = 30 in group B), with no significant difference between groups A
and B (p = 0.26).
The diagnostic accuracy on a per-segment basis for dual-source coronary CTA using prospective ECG triggering was 99% (95% CI, 96–100%). Per-segment, per-vessel, and per-patient sensitivity, specificity, PPV, and NPV are shown in Table 2.
|
Per-segment, per-vessel, and per-patient analyses with regard to the
diagnostic accuracy of dual-source CT coronary angiography in group A
(Agatston score < 316) and group B (Agatston score 316) are shown in
Table 3 (Figs.
1A,
1B,
2A,
2B,
3A, and
3B).
|
|
|
|
|
|
|
The differences in diagnostic performance on a per-segment basis between groups A and B were not significantly different (sensitivity, p = 0.92; specificity, p = 0.97; PPV, p = 0.92; and NPV, p = 0.99). Similarly, no level of significance was reached on a per-vessel analysis (sensitivity, p = 0.96; specificity, p = 0.73; PPV, p = 0.85; and NPV, p = 0.95) as well as on a per-patient analysis (sensitivity, p = 0.84; specificity, p = 0.94; PPV, p = 0.96; and NPV, p = 0.82) when comparing group A with group B.
Radiation Dose Estimations
No significant differences were found for the scanning length (p =
0.46) and the heart rate (p = 0.91) between the two protocols with
100 and 120 kV. The 32 CT examinations with a tube voltage of 100 kV had a
mean effective dose of 1.7 ± 0.4 mSv (range, 1.2–2.2 mSv),
whereas the 68 examinations with a tube voltage of 120 kV had a mean effective
dose of 3.1 ± 0.5 mSv (range, 2.3–4.4 mSv). The 100-kV protocol
resulted in a significantly lower effective dose (p < 0.001)
compared with the 120-kV protocol. The mean effective dose for all 100
examinations was 2.6 ± 0.8 mSv (range, 1.2–4.4 mSv). The rate of
segments with nondiagnostic image quality was not significantly different
between the two protocols (p = 0.73).
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Our study results corroborate recent publications showing the feasibility of the prospective ECG triggering algorithm [8, 9]. To depict all coronary segments with diagnostic quality in a single reconstruction phase, scanning at relatively low and regular heart rates is required [21–23]. With scanning and reconstruction of only a single phase in middiastole, we were able to depict 96% of all coronary vessel segments with diagnostic image quality. Our study shows that the diagnostic performance of the prospective ECG triggering algorithm is at a similarly high level when compared with that of previous studies using retrospective ECG gating [1–7], with an overall diagnostic accuracy on a per-segment basis of 99%.
It is known that arterial wall calcifications may adulterate the results of coronary CTA [1, 2, 7, 9, 24]. Using retrospectively ECG-gated 64-MDCT, Ong et al. [24] showed that the number of segments with nondiagnostic image quality increased from 3% to 13% in patients with higher calcium scores. Using dual-source CT with prospective ECG triggering, we found a 6.6% rate of nondiagnostic segments in patients with a higher calcium load—significantly higher than the rate of nondiagnostic segments in patients with a lower calcium score.
Alkadhi and colleagues [7] showed that in states of calcified coronary atherosclerosis, coronary CTA with retrospective ECG gating becomes less reliable because of an increase in false-positive ratings through stenosis overestimation caused by blooming artifacts. However, the associated lower specificity and PPV did not reach a level of significance [7]. Similar to their study using retrospective ECG gating, we found an increase in false-positive ratings that also did not lower the diagnostic accuracy in patients with higher calcium loads.
With dual-source CT, an increase in dose relative to single-source CT has been mitigated through the use of cardiac beam-shaping filters, 3D adaptive noise reduction, increased pitch for higher heart rates, ECG-based tube current modulation with narrow temporal windows, and fast tube current transitions [25]. Radiation doses of dual-source CT with retrospective ECG gating are in the range of 7–9 mSv [20]. In contrast to retrospective ECG gating, the algorithm of prospective ECG triggering acquires data only at selected time cardiac phases that are associated with a low radiation dose [8]. In our study, the effective doses were on average 1.7 mSv for patients who were examined at 100 kV and 3.1 mSv at 120 kV.
Our study has some limitations. The BMI, a factor known to influence the image quality in coronary CTA, has not been thoroughly investigated in this study. We have accounted for different BMIs by varying the tube voltage and by using the technique of automated tube current modulation, as previously suggested [13, 26]. With this approach, we observed no differences in the rate of nondiagnostic coronary segments among normal-weight or overweight patients. General limitations of the prospective ECG triggering algorithm include its selected feasibility only in patients with regular heart rates below 70 bpm when using dual-source CT. In addition, no information on the global and regional ventricular function or valvular function can be obtained when acquiring data only in a single cardiac phase in middiastole.
In conclusion, dual-source coronary CTA with prospective ECG triggering performs accurately in the assessment of CAD at low radiation doses. Diagnostic accuracy remains similarly high despite the presence of heavy calcifications but is associated with an increased rate of nondiagnostic segments.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  R. J. Gibbons, P. A. Araoz, and E. E. Williamson The Year in Cardiac Imaging J. Am. Coll. Cardiol., February 2, 2010; 55(5): 483 - 495. [Full Text] [PDF]
|
|
|
|
 |
|
 G. M. Schuetz, N. M. Zacharopoulou, P. Schlattmann, and M. Dewey Meta-analysis: Noninvasive Coronary Angiography Using Computed Tomography Versus Magnetic Resonance Imaging Ann Intern Med, February 2, 2010; 152(3): 167 - 177. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Pontone, D. Andreini, A. L. Bartorelli, S. Cortinovis, S. Mushtaq, E. Bertella, A. Annoni, A. Formenti, E. Nobili, D. Trabattoni, et al. Diagnostic accuracy of coronary computed tomography angiography: a comparison between prospective and retrospective electrocardiogram triggering. J. Am. Coll. Cardiol., July 21, 2009; 54(4): 346 - 355. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
