Rationale and objectives: The electron beam tomography coronary calcium score continues to be used without experimental validation. To determine its accuracy, a series of experiments was performed.
Methods: A chest phantom model was constructed with coronary arteries represented by cylindrical holes containing hydroxyapatite granules embedded in a gelatin matrix to simulate coronary arteries. Experiments were performed to determine the relationship between the mass of hydroxyapatite in each of these arteries, the coronary calcium score currently used in coronary screening, and an alternative method of estimating mass from the images. The model was scanned with equal amounts of hydroxyapatite in each artery: 1) when the cylindrical heart was rotated 36 degrees 10 times between scans, and 2) when the particle diameters varied from 0.1 mm to 4 mm. The scores were calculated, and a subtraction algorithm was applied to estimate the exact mass of hydroxyapatite in each artery.
Results: The hydroxyapatite scores varied by 42% with position and by 1.54 x 10(6)% with particle diameter. The estimated masses from the subtraction algorithm were more stable with position and particle size, with maximum percent errors of 10% and 14% for position and particle size, respectively.
Conclusions: These results suggest that the coronary calcium score is invalid, and that more precise and clinically relevant methods, such as the arterial summation method, should be rigorously tested in clinical studies.