Intersite comparison of the Hologic QDR-1000 dual energy X-ray bone densitometer

We have compared results from 13 Hologic QDR-1000 bone densitometers: (i) by performing spine and hip scans on two normal volunteers; (ii) by acquiring sets of 10 sequential scans on a Hologic anthropomorphic spine phantom. For each QDR-1000 site visited a set of spine phantom scans was also acquired on a QDR-1000 at Guy's Hospital to serve as a control study. All scans were analysed using the Hologic scan comparison software. Radiographers at each site were asked to perform their own independent analysis of the scans of the two volunteers. The precision of the bone mineral density (BMD), bone mineral content (BMC) and projected area (Area) results for a set of 10 phantom scans was 0.11%, 0.14% and 0.12% respectively. The coefficient of variation (CV) between sites for the sets of phantom scans was 0.58% for BMD, 0.71% for BMC and 0.35% for Area. In comparison, CVs for the phantom scans acquired on a single QDR-1000 were 0.23%, 0.23% and 0.09% respectively. The CV for the BMD results on the two volunteers obtained using the scan comparison software averaged 1.4% for the spine and 2.1% for the femoral neck. The CV for the results obtained by site radiographers averaged 2.2% for the spine and 3.7% for the femoral neck. Significant differences in the technique used for hip analysis were found.

(i) differences in calibration between systems were generally less than 1%; (ii) variations in results resulting from differences in analytical technique were more significant than those resulting from differences in calibration.