Assessment of bone property using quantitative ultrasound images based on whole calcaneus

Abstract

Osteoporosis is a systemic skeletal disease which leads to bone fragility. It was found that the evaluation of osteoporosis using quantitative ultrasound (QUS) parameters based on the whole calcaneus bone is better than a smaller circular ROI selected from a point with minimum density. Although it was claimed that inhomogeneity of the bone might be the main factor in causing such a difference, other factors, such as phase cancellation might also be the reason. In this paper, a modified contour deformable model (MCDM) was used to automatically detect the calcaneus contour from the QUS images. A three-dimensional calcaneus phantom (CIRS Inc., NV) was first used for the validation of MCDM and the demonstration of the phase cancellation effect. Then, nine subjects were recruited for human testing. From the phantom experiment, it was observed that the correlation coefficient between bone mineral density (BMD) and the parameter (BUA) evaluated based on the ROI selected with the whole calcaneus contour (R = 0.99) is better than other ROIs. The influence of ROI size, including 12, 14, 16 and 18 mm in diameter, and region-selecting methods, such as fixed region (ROIfix), automatic circular region (ROIcir) and calcaneal contour region (ROIanat), were evaluated for subject testing. The precision errors of the measured ultrasonic parameters for ROIanat (1.02 for BUA, 0.07 for SOS, 0.46 for STI) were smaller than ROIfix (3.75–4.23 for BUA, 0.42–0.54 for SOS, 1.79–2.71 for STI) and ROIcir (1.78–2.85 for BUA, 0.19–0.36 for SOS, 1.60–1.91 for STI). In conclusion, calculating QUS parameters using the whole calcaneal region provides a more precise and optimal measurement than a partial circular region (ROIfix or ROIcir). It has clinical implication that it could increase the diagnostic sensitivity of QUS.