Imaging ultrasonometry of the calcaneus: Dependence on calcaneal area

The purpose of this study was (1) to investigate the dependence of broadban d ultrasound attenuation (BUA) and speed of sound (SOS) measured in a circu lar region of interest (ROI) having a fixed size on calcaneal area and (2) to examine whether the normalization of ultrasonic variables for the area o f the calcaneus provides better differentiation of diseased subjects from h ealthy individuals. Ultrasound variables were estimated in 169 healthy post menopausal women (mean age 66.5 years, range 42-87 years) and 39 women with vertebral fractures (mean age 72.9 years, range 51-86 years). A minimum at tenuation ROI, 15 mm in diameter, with a commercial imaging ultrasonometer was used. Significant relationships were found between both ultrasonic vari ables and calcaneal area (r(2) = 0.06, P < 0.001 for BUA, r(2) = 0.12, P < 0.0001 for SOS). Normalization of ultrasound variables (BUAn and SOSn) was based on the regression equations of the relationships among BUA, SOS, and calcaneal area. In a precision study, nine women were examined five times e ach to determine the errors arising from both the repositioning of the foot and selection of the calcaneal area. The reproducibility errors of BUA. SO S, BUAn, SOSn, and area were 0.87%, 0.20%, 1.07%, 0.27%, and 3.72%, respect ively. Significant differences were found between the areas under the ROC c urve for BUAn and BUA (area under the curve = 0.93 for BUAn versus 0.90 for BUA, P = 0.003) as well as for SOSn and SOS (area under the curve = 0.85 f or SOSn versus 0.79 for SOS, P = 0.003). Normalization of ultrasound variab les for calcaneal area improves the discrimination of clinical studies.