The objective of this research is to develop new ultrasonic techniques to better estimate bone’s physical properties and provide diagnostic means for the early prediction of osteoporosis. The work consists of four parts. First, frequency scanning method was proposed as an simple, precise means for the measurement of broadband ultrasound attenuation (BUA). This study compared the frequency scanning method with the widely used broadband pulse method and demonstrated that the BUA derived from the new technique was more reliable and precise. The new method generated higher correlation of BUA to bone properties than the broadband pulse technique. Most significantly, the frequency scanning method has turned some once insignificant correlations between BUA and bone properties into significant ones, such as the increase of the absolute revalue o f the correlation between BUA in the longitudinal direction and BMD (from 0.25 to 0.62) and the correlation between medial-lateral BUA and the medial-lateral stiffness (from 0.31 to 0.55).

Second, a stratified model was proposed to investigate the ultrasonic wave propagation in the trabecular structure. The results have demonstrated that ultrasound velocity in trabecular bone was highly correlated with the bone apparent density (r=0.97). Moreover, a significantly consistent pattern between ultrasound attenuation coefficient and BMD has been observed between simulation using this model and experimental data. BUA derived from the simulation results revealed that BUA was nonlinear with respect to trabecular porosity and BMD. The relationship between BUA and BMD was parabolic in shape and agreed with the published experimental data. The peak magnitude o f BUA was observed at approximately 60% of bone porosity. This result has demonstrated that the stratified model reveals that reflection and transmission are important factors in the ultrasonic wave propagation in the trabecular bone.

Third, statistic study was performed on the relationship between ultrasound properties and bone properties. Data showed that ultrasound velocity was mainly correlated to bone density (r>0.78) and bone stiffness (r>0.68). Furthermore, the linear prediction analysis demonstrated that the correlation between ultrasound velocity and the bone stiffness was not solely dependent to the bone density. It was found that ultrasound velocity had up to 6.3% o f additional stiffness information independent to bone density. Ultrasound attenuation was greatly biased by the anisotropy of the trabecular structure. The r-values had great variations to bone density (from 0.62 to 0.90) and bone stiffness (from 0.55 to 0.78) when the direction of the ultrasound measurement was changed. Its capacity o f new stiffness information independent to bone density was also affected by the anisotropy. Except for the BUA in the antero-posterial direction, which contributed 8.4% more of the stiffness information, BUA in other directions did not contain any new stiffness information other than bone density. BUA was also indicative of the trabecular space and trabecular thickness. Up to 14% of new trabecular space information and 12% o f new trabecular thickness information, both independent to bone density, was found in BUA in longitudinal direction. However, no new information of trabecular space and trabecular thickness independent to the bone density was present in ultrasound velocity was found.

The study also showed that ultrasound velocity was superior to the ultrasound attenuation and BUA in the estimation of the trabecular stiffness. The average percent of stiffness information predicted by ultrasound velocity was 54% for the pulse velocity and 48% for the tone burst velocity, which were much greater than the 10% for the ultrasound attenuation coefficient and 26% for the BUA. When the combine ultrasound velocity and attenuation or BUA was used, the ability of ultrasound to predict the trabecular stiffness has improved. The amount of improvement relied on the amount of the stiffness information embedded in the ultrasound attenuation coefficient or BUA, which was independent to the ultrasound velocity. It was also shown that the stiffness information in the ultrasound attenuation coefficient was independent to the ultrasound velocity, while the BUA, which exhibited more stiffness information than the ultrasound attenuation, had the overlapped stiffness information with the ultrasound velocity.

Finally, a novel technique of acoustic scanning was introduced. The partial demineralization of bone specimens demonstrated that the acoustic scanning technique had the ability to detect the local ultrasound parameters and thus the local properties of bone. While the ultrasound velocity did not vary in the non-demineralized area throughout the process, the ultrasound velocity in the demineralized area showed steady decrease from 2500m/s to 2050m/s. The acoustic scanning technique can also make ultrasound attenuation coefficient more accountable to the bone properties than its nonscanned counterpart. Results showed that ultrasound the scanned ultrasound attenuation coefficient was equivalent to the non-scanned BUA, which had higher correlation than the non-scanned ultrasound attenuation coefficient.

Preliminary clinic test on human calcaneus demonstrated the images of ultrasound attenuation coefficient and BUA were able to show the position and region of interest of calcaneus. The consistency of the acoustic scanning was satisfactory. Based on the multiple tests on the same human subject in a week period, the data showed that the coefficient of variation was 4.38% for BUA, 15.4% for the attenuation coefficient and 1.87% for the ultrasound velocity. Therefore, the images of ultrasound attenuation coefficient and BUA can be used to identify the region of interest on the bone and the ultrasound velocity in the regional of interest can be used as a good indicator of the bone’s physical properties.

In summary, ultrasound is a promising physical modality in the estimation of bone’s properties. The acoustic scanning technique has enabled the ultrasound to detect the changes of bone’s local properties and therefore provided the potential means of early prediction of osteoporosis.