Image Quality in a Low Radiation Exposure Protocol for Retrospectively ECG-Gated Coronary CT Angiography

OBJECTIVE: The purpose of our study was to systematically compare the image quality of dual-source CT coronary angiography using 100 kV instead of 120 kV. SUBJECTS AND METHODS: One hundred patients with a body weight </= 85 kg were included. A dual-source CT scanner was used (330-milliseconds rotation, 0.6-mm collimation, 56 +/- 7 mL of IV contrast agent at 5 mL/s). Each patient was randomized either to scanning protocol group 1 (120 kV and 330 mAs) or protocol group 2 (100 kV and 330 mAs). ECG pulsing was used for all patients. Data sets were assessed by two independent observers for image quality, signal-to-noise ratio, and contrast-to-noise-ratio. Effective dose was determined based on dose-length product. RESULTS: There were no significant differences in body weight or heart rate between the two groups (70 +/- 10 kg and 57 +/- 8 bpm beats per minute vs 70 +/- 9 kg and 59 +/- 8 bpm). Use of 100 kV led to significant reduction of radiation exposure (group 1: 12.7 +/- 1.7 mSv; volume CT dose index CTDI(vol), 47.8 +/- 6.1 mGy and group 2: 7.8 +/- 2.0 mSv; CTDI(vol), 28.6 +/- 6.3 mGy; p < 0.001). Interobserver agreement in assessing image quality (kappa = 0.71) was close. Mean patient-based image quality scores were not significantly different (group 1, 2.7 +/- 0.5 and group 2, 2.6 +/- 0.4; p = 0.75). Also, vessel-based scores showed no significant differences. Beyond the level of significance, group 1 and group 2 showed one and two nonassessable patients and two and three nonassessable vessels, respectively. Mean intraluminal attenuation, contrast enhancement, and image noise were significantly higher for 100 kV, whereas signal-to-noise and contrast-to-noise-ratios were not different between the two scanning protocols. CONCLUSION: The use of lower tube voltage leads to significant reduction in radiation exposure in noninvasive coronary CT angiography. Image quality in nonobese patients is not negatively influenced.