Visualization techniques for multislice CT datasets of coronary arteries: Correlation of axial, multiplanar, three-dimensional, and virtual endoscopic imaging with coronary angiography

Objective: To evaluate the performance of 4 different visualization techniques for multislice CT datasets from the coronary arteries in comparison to coronary angiography. Material and Methods: 42 patients were examined prospectively using both multislice CT and angiocardiography, All CT scans were performed at collimation 4x1mm, pitch 1.5 and a rotation time of 500 msec. Using retrospective ecg-gating, image reconstruction parameters were 1.25 mm slice thickness, 0.5 mm increment and kernel B30. Each patient's CT dataset subsequently was explored using axial, multiplanar, three-dimensional, and virtual endoscopic Visualization modes. Results: Axial scans showed the highest sensitivity in the evaluation of stenoses (66.7%), followed by virtual endoscopy (55.9%), multiplanar (48.6%), and three-dimensional reformations (33.3%). With regard to the detection of atherosclerotic plaques, axial scans (71.2%), three-dimensional reformations (70.1%), and virtual endoscopy (69.1%) displayed comparable sensitivities, whereas multiplanar reformations showed distinctly lower results (55.6%). On combining the techniques a sensitivity of 74.2% for the detection of atherosclerotic plaques, of 72.0% for the identification of high-grade stenoses, respectively, was obtained. The specificity of all four visualization modes amounted to 97.9% and above. Best results have been obtained within the RIVA (sensitivity 85.2%), Conclusion: For the detection of atheromatous changes CT angiocardiography does not achieve a sensitivity as high as angiocardiography and is restricted to the three major branches. At heart rates below 60 bpm axial scans showed the best correlation in the quantification and classification of atherosclerotic plaques.