Overlapping cross-sections significantly improve the reproducibility of coronary calcium measurements by electron beam tomography: a phantom study.

Purpose: We conducted phantom studies to investigate whether overlapping cross-sections and volumetric scoring would significantly improve interscan reproducibility of electron beam tomography (EBT) for coronary artery calcium quantification.

Method: Fifteen phantoms simulating various amounts of coronary calcification were scanned in five different positions with a slice thickness of 3.0 mm and a table feed of 3.0, 2.5, and 2.0 mm. For the conventional "Agatston score" and a "volume score" (total volume of calcified lesions), interscan variabilities were compared between the three image acquisition protocols.

Results: Agatston score variability was significantly lower for the 2.0 mm table feed than for the 3.0 or 2.5 mm table feed (3.0 mm: 22.9 +/- 10.3%; 2.5 mm: 13.6 +/- 8.2%; 2.0 mm: 8.9 +/- 5.5%). Volume score variability was significantly lower for 2.5 and 2.0 mm table feed than for 3.0 mm table feed (3.0 mm: 21.7 +/- 11.0%; 2.5 mm: 10.9 +/- 5.9%; 2.0 mm: 9.8 +/- 5.9%).

Conclusion: Overlapping cross-sections, especially in combination with volumetric scoring, significantly improved interscan reproducibility of EBT calcium quantification in a phantom study.