A new set of calibration standards for estimating the fat and mineral content of vertebrae via dual energy QCT.

A new set of calibration standards has been developed for implementing a dual-energy (DE) quantitative CT technique for estimating the fat and bone content of vertebrae. The QCT technique is based upon a three-component model of bone and utilizes calibration materials that mimic those components in their X-ray attenuation properties. The three components we chose to simulate are bone (mineral plus collagen), fat and a fat-free red marrow. This choice was predicated upon our desire to employ materials that would facilitate later experimental verification of the method. The calibration standards and a set of test samples were manufactured of tissue-simulating epoxy resins. They were employed in studies of the accuracy (consistency) and precision of the technique and in a study of 21 normal postmenopausal women. Estimates of the bone and fat content of the test samples were consistent with the manufacturer's specifications to within 13 mg/ml and 7 vol%, respectively. Long-term reproducibility (coefficient of variation) for both quantities was about 3%. The average bone content of the T12-L3 vertebrae of the human subjects was 262 +/- 32 mg/ml (152 +/- 18 mg/ml calcium hydroxyapatite or mineral) and the average fat content was 63 +/- 8%. Conventional single energy (SE) QCT measurements of these vertebrae were about 23% less than the DE mineral measurements, which is consistent with the differences between SEQCT and ash content that others have determined via chemical analysis. The DE fat content is, in absolute terms, about 15% greater than values reported in the literature and may be due to an error in the assumed composition of red marrow. The true accuracy of the bone and fat estimates is to be determined in a planned human vertebral specimen study.