Influence of vertebral fat content on quantitative CT density

Abstract

Single- and dual-energy (85 and 130 kVp) computed tomographic (CT) measurements of bone density were made in 74 lumbar (L-3) vertebral specimens. Single-energy CT densities at 130 kVp consistently underestimated actual ash density by about 25 mg/cm3 in men and 40 mg/cm3 in women. CT densities overestimated age changes by 73% in women and 26% in men. These errors of the single-energy CT approach seemed due to increased marrow fat caused by age. At 130 kVp, there was a large decrease in apparent bone density (13 mg/cm3) for each increase of 100 mg/cm3 in fat content, but the decrease was lower at 85 kVp (11 mg/cm3), suggesting the use of lower energies for quantitative CT. In the vertebrae from the younger subjects (up to age 69), the relative error was 20%-31%, but in the oldest group it amounted to 31%-45%. The 95% confidence interval for an individual determination was +/- 38 mg/cm3. Dual-energy CT greatly reduced the above errors. Ash density was predicted with only a small accuracy error (7 mg/cm3) using a postprocessing dual-energy calculation, but the prediction error was 19 mg/cm3 for determinations at 130 kVp. The large uncertainty in the fat content of marrow (+/- 110 mg/cm3) and the variability in distribution of fat make dual-energy approaches necessary for accurate determinations of vertebral mineral density.