Predictors of Image Quality in High-Pitch Coronary CT Angiography
AJR:197, October 2011
Paul Stolzmann Robert P. Goetti Pal Maurovich-Horvat Udo Hoffmann Thomas G. Flohr Sebastian Leschka Hatem Alkadhi
OBJECTIVE. High-pitch coronary CT angiography (CTA) shows an alternative coronary CTA protocol to retrospectively ECG-gated imaging that is associated with a significantly lower radiation exposure. Yet, the diagnostic quality of high-pitch coronary CTA cannot be sufficiently predicted on the basis of heart rate (HR) alone. Thus, we sought to prospectively identify the predictors of diagnostic high-pitch coronary CTA.
SUBJECTS AND METHODS. One hundred consecutive patients (16 women, 84 men; mean age, 67 ± 10 years) without restrictions in HR frequency or variability underwent pro-spectively ECG-gated high-pitch coronary artery calcification (CAC) scoring and coronary CTA using dual-source 128-MDCT. High-pitch CAC was graded on the basis of motion artifacts; high-pitch coronary CTA was graded on the basis of image quality by each of two in¬dependent and blinded readers. Predictors of coronary CTA image quality were assessed by multivariate logistic regression, including body mass index, risk of coronary artery disease, Agatston score, HR frequency and variability, and motion artifacts on CAC. If high-pitch coronary CTA was nondiagnostic, the examination was repeated using a retrospectively ECG-gated coronary CTA protocol.
RESULTS. HR frequency (66 ± 20 beats per minute [bpm]) and variability (12 ± 10 bpm) during high-pitch coronary CTA was not significantly different from that during high-pitch CAC. Interobserver agreement for image quality reading was good (k = 0.79) for CAC and excellent (k = 0.88) for coronary CTA. High-pitch CAC showed no motion artifacts in 1304 of 1396 segments (93.4%) in 78 of 100 patients. Diagnostic image quality of high-pitch coronary CTA was found in 1413 of 1457 segments (97%) in 82 of 100 patients. Multivariate logistic regression revealed HR frequency (odds ratio, 1.04; p < 0.05) and motion artifacts on preceding high-pitch CAC (odds ratio, 94.55; p < 0.001) as significant independent predictors of image quality of high-pitch coronary CTA. The mean effective radiation doses of high-pitch CAC and coronary CTA were 0.4 ± 0.1 mSv (0.3-0.5 mSv) and 1.0 ± 0.2 mSv (0.8-1.2 mSv; p < 0.001), respectively.
CONCLUSION. HR frequency and motion artifacts on preceding high-pitch CAC represent significant independent predictors of image quality in high-pitch coronary CTA.