Calcium scoring of the coronary artery by electron beam CT: how to apply an individual attenuation threshold.

Abstract

Our aim was to assess the inter- and intraindividual variability of the attenuation threshold used to identify coronary artery calcification on electron beam CT and to illustrate a new threshold method. We measured the soft-tissue attenuation of regions surrounding the coronary arteries at the level of the left main coronary artery ostium (high level) and at the bottom of the heart (low level) in 48 consecutive patients (22 men, 26 women). Mean +/- 2 standard deviations (SD) of soft-tissue attenuation and variance of soft-tissue density and SDs were calculated at each level for every patient. It was assumed that setting an attenuation threshold greater than or equal to 3 SDs above that of soft tissue at each myocardial level would eliminate 99.5% of all scatter artifacts, allowing precise identification of calcific deposits. For the entire patient cohort, the average soft-tissue attenuation was 41 H and 35 H at the high and low levels, respectively (p < 0.01), indicating a large intraindividual variability. The SDs of soft-tissue attenuation measured by the computer software at the high and low levels were not different (26 H at the high level and 28 H at the low level; p = not significant). However, the calculated SD of the individual mean soft-tissue attenuation was 5 H at the high level and 8 H at the low level, again indicating a large intraindividual variability (p < 0.01). The addition of 3 measured SDs above the mean individual soft-tissue attenuation predicted a mean threshold of 120 and 121 H at the high and low levels, respectively, but with a wide interindividual variability (83-193 H at the high level and 79-242 H at the low level). There was a strong correlation between body weight and SD of soft-tissue attenuation at the low level (r = 0.75, p < 0.001) and a weaker but statistically significant correlation between weight and SD of soft-tissue attenuation at the high level (r = 0.51, p < 0.001). For the patients in this study, a threshold of 120 H for the detection of coronary calcification by electron beam CT seemed more appropriate than a threshold of 130 H, which is currently in use. However, given the great inter- and intraindividual variability, a biologic threshold tailored to the individual patient and to each individual imaging level should be used instead of a fixed threshold.