C. Kang et R. Speller, COMPARISON OF ULTRASOUND AND DUAL-ENERGY X-RAY ABSORPTIOMETRY MEASUREMENTS IN THE CALCANEUS, British journal of radiology, 71(848), 1998, pp. 861-867
Ultrasound measurements have been proposed as a means of providing str
uctural information on bone but have also been shown to correlate with
density. High correlation coefficients have been obtained in vitro, b
ut are lower for in vivo work. The aim of this study was to investigat
e the relationship of broadband ultrasound attenuation (BUA), speed of
sound (SOS) and stiffness, to bone mineral density (BMD) measured in
the calcaneus at a closely matched region of interest (ROI). Measureme
nts were made in 55 post-menopausal and 20 young, normal women. Calcan
eus BMD measurements were made using an ROI (fixed by reference to ext
ernal axes) to represent the area and location of the ultrasound trans
ducers and an ROI based on anatomical markers, positioned in the poste
rior part of the calcaneus. BUA was significantly correlated with calc
aneus BMD in the fixed ROI (r = 0.77, p < 0.0001) and the anatomical R
OI (r = 0.78, p < 0.0001). Correlation of BUA with axial BMD was lower
at r = 0.63, p < 0.0001 for the spine (L2-L4) and r = 0.59, p < 0.000
1 for the femoral neck. Results for SOS and stiffness were very simila
r. From the calcaneus dual energy X-ray absorptiometry (DXA) scan imag
es the region representing the ultrasound measurement area was found t
o contain cortical bone edges in 60% of cases. In 16% of scans a small
part (4.5 +/- 4.0%) of the ROI measurement area fell partially outsid
e the heel. However, there was no obvious discrepancy in the ultrasoun
d results in the individual cases where this occurred. The correlation
between calcaneus BMD and BUA was improved from r = 0.72 to r = 0.86
when scans with a cortical edge in the measurement ROI were excluded.
The values for SOS were similar. In summary, BMD of the calcaneus, whe
n closely matched to the site of ultrasound measurement shows signific
ant correlation with ultrasound measurements. By excluding scans in wh
ich the ultrasound measurement appeared to include bone edges, correla
tions of approximately r = 0.86 were obtained. However, the BMD result
still does not explain all of the variance in the ultrasound results.