Abstract 3269: Head-to-head Comparison Among 64-slice Computed Tomography, Magnetic Resonance Imaging and Conventional Coronary Angiography in the Identification of Significant Coronary Lesions
Purpose: To correlate coronary lesions, ischemia or infarcts visually identified by conventional coronary angiography (CCA), 64-slice computed tomographic angiography (CTA) and cardiac magnetic resonance imaging (CMR), on a head-to-head basis.
Methods: Patients without coronary artery bypass grafting who were referred for workup of chest pain and underwent all CCA, CTA and CMR from 1–27–05 to 1–26–06 were included. Coronary segments with previous intervention and their territories were excluded. CCA was performed as usual. CTA was done with a 64-slice multidetector scanner (Sensation 64, Siemens) and retrospective ECG-gating for image reconstruction. Coronary lesions (≥50% stenosis) were identified for both CCA and CTA. CMR was performed with a 1.5T scanner (Sonata, Siemens). First-pass perfusion scans were done during adenosine stress and rest for perfusion defects. Delayed imaging was performed to identify infarcts. All three imaging modalities were assessed visually by consensus of two observers blinded to other results. CTA and CCA were compared according to the AHA 17-segment model, while CMR was compared to CCA or CTA by coronary territories.
Results: Twenty-eight patients (24 male) were included, median age was 57.0 years (range, 41–87). Twenty-seven coronary segments and 18 corresponding territories were excluded because of prior intervention. There were ten absent segments. Consequently, 439 coronary segments and 66 territories were analyzed. According to CCA, 16.2% (71/439) coronary segments had ≥50% stenosis. The sensitivity, specificity, positive predictive value and negative predictive value of CTA were 93.0%, 95.7%, 80.5% and 98.6%, while those of CMR were 76.9%, 90.0%, 83.3% and 85.7%, respectively. If either positive CTA or CMR was defined as positive, the corresponding values were 100%, 75.0%, 72.2% and 100%. The concordance rate between CTA and CMR was 72.7% (48/66 territories).
Conclusions: This is the first report on a head-to-head basis that correlates 64-slice CTA, CMR and CCA. Sixty-four-slice CTA provides high diagnostic accuracy in assessment of coronary stenosis. The combination of 64-slice CTA and CMR further improves the accuracy with no false negatives, and may help in eliminating unnecessary negative CCA.