Dual X-ray absorptiometry: clinical evaluation of a new cone-beam system

The DMS Lexxos is the first cone-beam dual X-ray absorptiometry (DXA) system capable of performing bone mineral density (BMD) measurements of the spine and hip. By using a two-dimensional (2-D) detector array rather than the linear array used with conventional fan-beam DXA systems, image acquisition time on Lexxos is only 1.5 s. However, the need to correct for the large signal from scattered radiation reaching the detector is a potential source of error in cone-beam DXA. The aim of this clinical evaluation of the Lexxos bone densitometer was to investigate the relative accuracy of cone-beam BMD measurements compared with conventional DXA by performing an in vivo cross-calibration study with an established fan-beam system, the Hologic QDR-4500. Spine (L1-L4) and hip BMD measurements were performed in 135 patients (111 women, 24 men) referred for a bone densitometry examination. Duplicate Lexxos measurements were performed in 27 female patients to evaluate precision. On average, Lexxos spine and femoral neck BMD measurements were 2% lower than those on the QDR-4500, whereas total hip BMD was 5% higher. Larger differences were found for the trochanter and Ward's triangle regions. For all sites, Lexxos BMD measurements showed a strong linear relationship with those measured on the QDR-4500 with correlation coefficients in the range r = 0.95 to 0.97 for the clinically important spine, femoral neck, and total hip regions. The root mean standard error (RMSE) between Lexxos and QDR-4500 BMDs ranged from 0.037 g/cm(2) for the femoral neck to 0.060 g/cm(2) for the spine, whereas Lexxos precision was 1.3% for total hip, 2.0% for femoral neck, and 2.3% for spine BMD. Although for the hip BMD sites the RMSE and precision of Lexxos measurements were similar to studies of pencil-beam and fan-beam DXA systems, the results for the spine were poorer than expected. The findings of this study suggest that Lexxos corrects accurately for the effects of scattered radiation at the detector, but that the precision of spine BMD measurements may be limited by involuntary patient movement between the high and low energy X-ray exposures.