Data on the diagnostic accuracy of multislice computed tomographic coronary angiography (CTA) have been mostly derived from patients with a high pretest likelihood of coronary artery disease. Systematic comparisons with invasive angiography in patients with an intermediate pretest likelihood are scarce. The purpose of the present study was to determine the diagnostic accuracy of CTA in patients without known coronary artery disease with an intermediate pretest likelihood. A total of 61 patients (61% men, average age 57 ± 9 years) who had been referred for invasive coronary angiography underwent additional 64-slice CTA. A total of 920 segments were identified by invasive coronary angiography, of which 885 (96%) were interpretable on CTA. Invasive coronary angiography identified a significant stenosis (≥50% luminal narrowing) in 29 segments, of which 23 were detected on CTA. Thus, the sensitivity, specificity, positive predictive value, and negative predictive value was 79%, 98%, 61%, and 99%, respectively, for CTA. On a patient level, the sensitivity, specificity, positive predictive value, and negative predictive value was 100%, 89%, 76%, and 100%, respectively. CTA correctly ruled out the presence of significant stenosis in 40 (66%) of the 61 patients. In conclusion, the results from the present study have confirmed that CTA has excellent diagnostic accuracy in the target population of patients with an intermediate pretest likelihood. The high negative predictive value allowed us to rule out significant stenosis in a large proportion of patients. CTA can, therefore, be used as a highly effective gatekeeper for invasive coronary angiography.
Recently, noninvasive anatomic imaging has become possible with the introduction of multislice computed tomographic coronary angiography (CTA). Numerous studies have shown that CTA has a high diagnostic accuracy for the evaluation of significant coronary artery disease (CAD) (≥50% luminal narrowing) compared to coronary angiography. Accordingly, the technique has been proposed as a tool to rule out significant CAD and thus serve as a noninvasive gatekeeper for invasive coronary angiography. However, thus far, almost all studies investigating the diagnostic accuracy of CTA have been performed in populations with a high pretest likelihood for CAD. Nevertheless, this population is unlikely to benefit from CTA, because most patients will require invasive coronary angiography. In contrast, patients with an intermediate pretest likelihood for CAD might derive far more benefit from a noninvasive alternative to coronary angiography and might, in fact, represent the target population for this technique. However, only very limited data are available of patients with an intermediate pretest likelihood, and systematic comparisons with invasive coronary angiography are scarce. Therefore, the purpose of the present study was to specifically address the diagnostic accuracy of CTA in patients with an intermediate pretest likelihood for CAD.
Methods
In the present prospective cohort study, 61 patients with an intermediate pretest likelihood for CAD who had been referred for invasive diagnostic coronary angiography underwent additional evaluation with CTA within a 14-day period. An intermediate pretest likelihood was defined according to the Diamond and Forrester criteria as a pretest likelihood of CAD of 13.4% to 87.2%, as previously described. Patients were excluded from the study if they met one of the following exclusion criteria for CTA: cardiac arrhythmias, renal insufficiency (serum creatinine >120 mmol/L), known hypersensitivity to iodine contrast media, and pregnancy. Finally, patients were excluded because of the occurrence of a cardiac event (ie, worsening angina, revascularization, or myocardial infarction) in the period between the 2 examinations. The local medical ethics committee (Medical Center Haaglanden, The Hague, The Netherlands) approved the study, and all patients gave written informed consent.
All examinations were performed using a 64-slice multislice computed tomographic scanner (Lightspeed VR 64, GE Healthcare, Milwaukee, Wisconsin). The patients’ heart rate and blood pressure were monitored before each scan. In the absence of contraindications, patients with a heart rate exceeding the threshold of 65 beats/min were administered β-blocking medication (50 to 100 mg metoprolol orally or 5 to 10 mg metoprolol intravenously).
Before the helical scan, a nonenhanced electrocardiographically gated scan, prospectively triggered at 75% of the RR interval, was performed to measure the coronary calcium score and to determine the start and end positions of the helical scan. After the calcium scan, a retrospectively electrocardiographically gated helical scan was performed using the following scan parameters: collimation 64 × 0.625 mm, rotation time 0.35 seconds, tube voltage 120 kV, and tube current 600 mA (with tube modulation to reduce the radiation dose). A bolus of 80 ml iomeprol (Iomeron 400, Bracco, Milan, Italy) was injected at 5 ml/s followed by a 40-ml saline flush. The helical scan was automatically triggered using a bolus tracking technique (SmartPrep, Waukesha, Wisconsin), when the attenuation level in the region of interest reached the predefined threshold (baseline attenuation +100 Hounsfield units). Data sets were reconstructed from the retrospectively gated raw data with an effective slice thickness of 0.625 mm. The coronary arteries were evaluated using the reconstructed data set with the fewest motion artifacts, typically an end-diastolic phase.
Postprocessing of the multislice computed tomographic calcium scans and coronary angiograms was performed on a dedicated workstation (Advantage, GE Healthcare, Waukesha, Wisconsin). The total calcium score was calculated from the nonenhanced calcium scan using the Agatston method. Subsequently, the coronary anatomy was evaluated using the contrast-enhanced helical examinations. The coronary arteries were divided into 17 segments according to a modified American Heart Association classification. All studies were interpreted by 2 experienced observers who were unaware of the coronary angiographic results. First, the image quality was assessed and scored as good, average (reduced image quality but of diagnostic quality), and poor (low diagnostic image quality). Next, the presence of significant stenosis (≥50% luminal narrowing) was evaluated using axial slices, curved multiplanar reconstructions, and maximum intensity projections.
Invasive diagnostic coronary angiography was performed according to standard techniques. The coronary angiograms were evaluated by an observer who was unaware of the results from CTA using off-line quantitative software (QCA-CMS, version 6.0, Medis, Leiden, The Netherlands) for quantitative coronary angiography. The coronary arteries were divided into 17 segments according to a modified American Heart Association classification, and quantitative coronary angiography was performed in lesions with >30% luminal narrowing on visual assessment. Each segment was evaluated for the presence of ≥50% luminal narrowing on the quantitative coronary angiogram. Obstructive CAD was defined as luminal narrowing of ≥50%. Accordingly, the sensitivity, specificity, positive and negative predictive values (including 95% confidence intervals), and positive and negative likelihood ratios for the detection of stenoses with ≥50% luminal narrowing on the quantitative coronary angiogram were calculated on the segmental, vessel, and patient level.
Results
All 61 patients had been clinically referred for invasive diagnostic coronary angiography because of chest pain suspect for CAD and an intermediate pretest likelihood according to the Diamond and Forrester criteria. The characteristics of the study population are listed in Table 1 . In brief, the average patient age was 57 ± 9 years, and 61% were men. Most patients presented with atypical angina (82%), with nonanginal chest pain observed in 13% and typical angina in 5%.
| Characteristic | Value |
|---|---|
| Gender | |
| Men | 37 |
| Women | 24 |
| Age (years) | |
| Mean ± SD | 57 ± 9 |
| Range | 35–75 |
| Heart rate (beats/min) | |
| Mean ± SD | 58 ± 8 |
| Range | 41–78 |
| Average calcium score (Agatston) | |
| Mean ± SD | 198 ± 323 |
| Range | 0–1,505 |
| β-Blocking medication | 37 (61%) |
| Diabetes mellitus | 15 (25%) |
| Hypertension | 38 (62%) |
| Hypercholesterolemia ⁎ | 38 (62%) |
| Current smoker | 20 (33%) |
| Body mass index ≥30 kg/m 2 | 14 (23%) |
| Nonanginal chest pain | 8 (13%) |
| Atypical angina pectoris | 50 (82%) |
| Typical angina pectoris | 3 (5%) |
| Coronary arteries narrowed on invasive coronary angiography (n) | |
| None | 45 (74%) |
| 1 | 8 (13%) |
| >1 | 8 (13%) |
⁎ Defined as total serum cholesterol ≥230 mg/dl and/or serum triglycerides ≥200 mg/dl or the use of a lipid-lowering agent.
On invasive coronary angiography 920 coronary segments were identified, of which 35 (3.8%) were uninterpretable using CTA, leaving 885 segments for additional analysis. The image quality was good in 753 segments (85%), with average image quality observed in 100 segments (11%) and poor image quality in the remaining 32 segments (4%). Using invasive coronary angiography, a significant stenosis was identified in 29 segments (3.3%). CTA correctly ruled out the presence of significant stenosis in 841 of 856 segments, and significant stenosis was correctly identified in 23 of 29 segments. CTA overestimated the disease severity in 15 segments without significant disease found using invasive coronary angiography, and 6 segments were incorrectly scored as nonsignificant using CTA. Thus, the sensitivity and specificity was 79% and 98%, respectively, on a segment level. The positive and negative predictive value was 61% and 98%, respectively. The positive likelihood ratio was 39.5, and the negative likelihood ratio was 0.21.
A total of 183 vessels were identified. Significant stenosis was observed in 26 vessels using invasive coronary angiography. CTA correctly identified a significant stenosis in 22 of 26 vessels. Significant stenosis was ruled out in 148 of 157 vessels. CTA overestimated the disease severity in 9 vessels without significant stenosis using invasive coronary angiography, and the disease severity was underestimated in 4 vessels. The sensitivity, specificity, and positive and negative predictive value on a vessel level was 85%, 94%, 71%, and 93%, respectively. The positive and negative likelihood ratio was 14.2 and 0.16, respectively.
Of the 61 patients, significant CAD was observed in 16 (26%) by invasive coronary angiography. CTA correctly identified significant CAD in all 16 of these patients. In contrast, CTA correctly ruled out significant disease in 40 of 45 patients without significant stenosis using invasive coronary angiography, representing 66% of the total population. The disease severity was overestimated in 5 patients without significant stenosis on invasive coronary angiography. Importantly, CTA did not underestimate the disease severity in any patient. This resulted in a sensitivity and specificity of 100% and 89%, respectively. The positive and negative predictive value was 76% and 100%, respectively. The positive likelihood ratio was 9.1, and the negative likelihood ratio was 0.00. The results of all analyses, including the positive and negative predictive values, with the 95% confidence intervals, are listed in Table 2 .
| Variable | Segmental Analysis | Vessel Analysis | Patient Analysis |
|---|---|---|---|
| Excluded | 35/920 (3.8%) | 0% | 0% |
| Sensitivity | 79% | 85% | 100% |
| Patients (n) | 23/29 | 22/26 | 16/16 |
| 95% Confidence interval | 64–94% | 71–99% | 100–100% |
| Specificity | 98% | 94% | 89% |
| Patients (n) | 841/856 | 148/157 | 40/45 |
| 95% Confidence interval | 97–99% | 90–98% | 80–98% |
| Positive predictive value | 61% | 71% | 76% |
| Patients (n) | 23/38 | 22/31 | 16/21 |
| 95% Confidence interval | 46–77% | 55–87% | 58–94% |
| Negative predictive value | 99% | 97% | 100% |
| Patients (n) | 841/847 | 148/152 | 40/40 |
| 95% Confidence interval | 99–100% | 94–100% | 100–100% |
| Diagnostic accuracy | 98% | 93% | 92% |
| Patients (n) | 864/885 | 170/183 | 56/61 |
| 95% Confidence interval | 97–99% | 89–97% | 85–99% |
| Positive likelihood ratio | 39.5 | 14.2 | 9.1 |
| Negative likelihood ratio | 0.21 | 0.16 | 0.00 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
