A novel model to test accuracy and reproducibility of MDCT scan protocols for coronary calcium in vivo

Abstract

Objectives: We compared the accuracy and reliability of prospectively triggered, retrospectively ECG gated, and non-gated CT image reconstruction for measurements of coronary artery calcification (CAC) in vivo using a novel animal model.

Materials and methods: In six Yorkshire farm pigs, prefabricated chains of cortical bone fragments were sutured over the epicardial bed of the major coronary arteries. Using a 4-slice MDCT scanner, each animal was imaged with two different protocols: sequential acquisition with prospective ECG triggering, and spiral acquisition with retrospectively ECG gated image reconstruction- non-gated reconstructions were also generated from these latter scans. Two independent observers measured the 'Agatston score' (AS), the calcified volume (CV), and mineral mass (MM). To calculate accuracy of MM measurements the ash weight of the burned bone fragments was compared to MDCT derived MM.

Results: Six pigs successfully underwent surgery and CT imaging (mean heart rate: 86+/-12 bpm). MM measurements from prospectively ECG triggered CT sequential scans were more accurate (p<0.02) and reproducible (p=0.05) than sequential CT scans without ECG triggering or spiral acquisition using retrospective ECG gating.

Conclusions: At high heart rates prospective ECG triggered image reconstruction is more accurate and reproducible for CAC scoring than retrospective ECG gated reconstruction and non-gated reconstruction.