Does quantitative ultrasound imaging enhance precision and discrimination?

The aim of this study was to compare quantitative ultrasound (QUS) measurements obtained using a new calcaneal QUS imaging device with a conventional non-imaging device using fixed transducers. The study group consisted of 340 healthy women with no risk factors associated with osteoporosis (176 premenopausal and 164 postmenopausal) and 83 women with one or more vertebral fractures. All women had QUS measurements performed on the Osteometer DTU-one (imaging) and Walker-Sonix UBA575+ (non-imaging) devices and bone mineral density (BMD) measurements performed at the spine and hip. A subgroup of 81 women had additional dual-energy X-ray absorptiometry (DXA) scans at the calcaneus. Short-term standardized precision (SP = SD/young adult SD) based on duplicate measurements was significantly better on the DTU for broadband ultrasound attenuation (BUA) (SP: DTU 0.15 vs UBA 0.21,p = 0.01) and speed of sound (SOS) (SP: DTU 0.14 vs UBA 0.18, p = 0.01). However, long-term SP of the DTU was comparable to or significantly poorer than the SP of the UBA device. The BUA and SOS measurements obtained on the DTU and UBA were significantly correlated (r = 0.76 and 0.89 for BUA and SOS measurements respectively). The correlations between QUS and BMD measurements were all significant, ranging from 0.53 to 0.72. No significant improvements in the correlation with axial or peripheral BMD were observed using the imaging device. All the QUS measurement parameters showed a significant negative relationship between age and years since menopause in the postmenopausal group. Annual losses were lower for the DTU for BUA (DTU 0.22 dB/MHz per year vs UBA 0.44 dB/MHz per year) but comparable for SOS (DTU 0.29 m/s per-year vs UBA 0.22 m/s per year). However, when these figures were standardized to take into account the clinical range, the annual losses were similar on the DTU and UBA. Age-adjusted odds ratios for each SD decline were similar on the DTU for BUA (DTU 3.2 vs UBA 3.3) and SOS (DTU 3.4 vs UBA 5.1). The corresponding odds ratios for BMD at the lumbar spine, femoral neck and total hip were 2.7, 2.9 and 3.3 respectively. Age-adjusted receiver-operating characteristics analysis yielded values for the area under the curve (AUC) ranging from 0.74 to 0.83. The DTU BUA AUC of 0.83 was significantly greater than the AUC obtained for UBA BUA and BMD measurements at the lumbar spine and femoral neck. Ultrasound imaging at the calcaneus was found to improve the standardized precision of BUA and SOS measurements in the short term but not in the long term. Neither the correlation with BMD nor the discriminatory ability of QUS was improved by utilizing QUS images at the calcaneus. The inconsistencies of the imaging system used for this study demonstrate that further development is required before it will be possible to show improvements in long-term precision.