Noninvasive assessment of coronary artery disease by multislice spiral computed tomography using a new retrospectively ECG-gated image reconstruction technique.

The present study was designed to investigate the accuracy of multislice spiral computed tomography (MSCT) in detecting coronary artery disease, compared with coronary angiography (CAG), using a new retrospectively ECG-gated reconstruction method that reduced cardiac motion artifact. The study group comprised 54 consecutive patients undergoing MSCT and CAG. MSCT was performed using a SOMATOM Volume Zoom (4-detector-row, Siemens, Germany) with slice thickness 1.0 mm, pitch 1.5 (table feed: 1.5 mm per rotation) and gantry rotation time 500 ms. Metoprolol (20-60 mg) was administered orally prior to MSCT imaging. ECG-gated image reconstruction was performed with the reconstruction window (250 ms) positioned immediately before atrial contraction in order to reduce the cardiac motion artifact caused by the abrupt diastolic ventricular movement occurring during the rapid filling and atrial contraction periods. Following inspection of the volume rendering images, multiplanar reconstruction images and axial images of the left main coronary artery (LMCA), left anterior descending artery (LAD), left circumflex artery (LCx) and right coronary artery (RCA) were obtained and evaluated for luminal narrowing. The results were compared with those obtained by CAG. Of 216 coronary arteries, 206 (95.4%) were assessable; 10 arteries were excluded from the analysis because of severe calcification (n=4), stents (n=3) or insufficient contrast enhancement (n=3). The sensitivity to detect coronary stenoses >or=50% was 93.5% and the specificity to define luminal narrowing <50% was 97.2%. The positive predictive value and the negative predictive value were 93.5% and 97.2%, respectively. The sensitivity was still satisfactory (80.6%) even when non-assessable arteries were included in the analysis. The new retrospectively ECG-gated reconstruction method for MSCT has excellent diagnostic accuracy in detecting significant coronary artery stenoses.

Noninvasive coronary artery imaging by multislice spiral computed tomography.

Although the excellent spatial resolution of multislice spiral computed tomography (MSCT) enables the coronary arteries to be visualized, its limited temporal resolution results in poor image reproducibility because of cardiac motion artifact (CMA) and hence limits its widespread clinical use. A novel retrospectively ECG-gated reconstruction method has been developed to minimize CMA. In 88 consecutive patients, the scan data were reconstructed using 2 retrospectively ECG-gated reconstruction methods. Method 1: the end of the reconstruction window (250 ms) was positioned at the peak of the P wave on ECG, which corresponded to the end of the slow filling phase during diastole immediately before atrial contraction. Method 2 (conventional method): relative retrospective gating with 50% referred to the R-R interval was performed so that the beginning of the reconstruction window (250 ms) was positioned at the halfway point between the R-R intervals of the heart cycle. The quality of the coronary artery images was evaluated according to the presence or absence of CMA. The assessment was applied to the left main coronary artery (LMCA), the left anterior descending artery (LAD, segments #6, #7 and #8), the left circumflex artery (LCx, segments #11 and #13) and the right coronary artery (RCA, segments #1, #2 and #3). The first diagonal artery (#9-1), the obtuse marginal artery (#12-1), the posterior descending artery (#4-PD), the atrioventricular node branch (#4-AV) and the first right ventricular branch (RV) were also evaluated. Of the 88 patients, 85 were eligible for image evaluation. Method 1 allowed visualization of the major coronary arteries without CMA in the majority of patients. The LCA system (segments #5-7, #11 and #13) and the proximal portion of the RCA were visualized in more than 94% of patients. Artifact-free visualization of the distal portion of the LAD (segment #8) and RCA (#4PD and #4AV), and side branches (#9-1, #12-1 and RV) was also achieved in more than 80% of patients. On the other hand, CMA occurred frequently on images obtained by Method 2. The LCx and RCA systems were the most affected by CMA, revealing only 41% artifact-free visualization of the segment #13, 39% of #1, 15% of #2 and 32% of #3. Thus, Method 1, which avoids the ventricular motion occurring during the rapid filling and atrial contraction phases, gives superior image quality over the conventional ECG-gated reconstruction method.