PURPOSE

To determine the accuracy and precision of thoracic phantomless bone mineral density (BMD) measurements obtained on coronary artery calcium (CAC) computed tomography (CT) scans by using a variety of commercially available CT scanners.

MATERIALS AND METHODS

The institutional review board approved this Health Insurance Portability and Accountability Act-compliant study. A total of 4126 asymptomatic subjects (2022 [49%] men, 2104 [51%] women; mean age, 63.7 years ± 11.8 [standard deviation]) underwent CAC CT with the use of a quantitative CT calibration phantom for evaluation of subclinical atherosclerosis. Two hundred eighty subjects also underwent CT of the chest, abdomen, and pelvis (C7 through L5). Mean BMD of three consecutive thoracic vertebrae (in the T7-T10 range) was measured in all 4126 subjects. Individual calibration factors for each phantom insert and a general calibration factor for the spine were determined for each CT scanner model. The study population was then divided into three subgroups: All calibration factors were generated from group 1 (n = 1536) and were applied and tested in group 2 (n = 1587), and effects of various image acquisition parameters were assessed in group 3 (n = 1003). Accuracy (bias) and precision of thoracic phantomless BMD measurements across 14 CT scanner models from five manufacturers were determined.

RESULTS

Phantomless BMD values correlated highly with standard phantom-based quantitative CT BMD values (r = 0.987, P < .001). Bias was 3.9% ± 1.4 for phantomless BMD measurements, and the mean coefficient of variation for the general calibration factor was 4.9% ± 2.4.

CONCLUSION

Phantomless BMD can be measured accurately on CAC CT scans acquired with a variety of CT scanners without additional radiation exposure.