1998 Volume 54 Issue 2 Pages 241-248
Although dual-energy x-ray absorptiometry(DXA)instruments operate on the same basic principles, bone mineral density(BMD)values measured with different DXA systems are not identical. Therefore, a fundamental study of cross-calibration of lumbar vertebral BMD values measured with different DXA systems was performed. The QDR-1000, DPX-L and XR-26 were used as DXA systems, and 30 patients, a rod phantom, a COMAC phantom, excised bone, a Hologic standard phantom, and a human bone phantom were used as subjects for measurement. The BMDs per unit volume in the three systems were 0.371 g/cm^3 for the QDR-1000, 0.411 g/cm^3 for the DPX-L, and 0.361 g/cm^3 for the XR-26. The relation between bone area and BMD values was nearly the same for the QDR-1000 and DPX-L;however, these results were markedly different from that obtained with the XR-26. In vivo conversion formulae for different DXA systems were calculated from the data of the 30 patients. By analyzing the functional relation, the conversion formulae QDR=0.826DPX+0.042 and QDR=0.921XR+0.097 were derived. Comparison of the in vivo conversion formula with the in vitro phantom data indicated that the external shape of the phantom was responsible for the errors between the two. In particular, the difference was remarkable between different bone edge detection methods. In conclusion, it was considered that the external shape of the phantom used plays an important role in vitro cross-calibration for lumbar vertebral BMD values measured between DXA systems.
