Osteoporosis has become a growing health threat with rising social and economic consequences. The understanding of the relationship between bone mechanical strength and bone structural quality is important for the diagnosis of osteoporosis and the evaluation of osteoporosis treatments.

A novel cancellous bone segmentation method was developed to separate cancellous bone from cortical bone automatically. The segmentation was based on the three-dimensional images. The basic idea of the method is that a contiguous threedimensional non-bone region that is located inside the whole bone region of interest can be defined as the marrow region. Any bone material that is surrounded by the marrow region is considered cancellous bone while bone that is outside of the marrow region is cortical bone. This method is more objective, more accurate and more precise than a manual method and an established computer method (Helterbrand method).

Twenty-four female Sprague-Dawley rats, 7-8 weeks old, were used in a study of various osteoporosis drugs. After 8 weeks drug treatment, alendronate had more positive effects on cancellous bone volume fraction than 17 (3-Estradiol2, raloxifene and PTH. This striking result may be explained by the age of the rat model because the young animals are still growing rapidly and it is possible that the ovariectomy (OVX) operation results in higher remodeling rate in young animals. When an anti-resorptive drug is given, the individuals with high baseline remodeling rate have more positive response than those with low baseline remodeling rate.

120 female Sprague-Dawley rats, 5.5 months or 6 months old, were used in three studies of experimental drugs. In the Pradama studies, sensitivity analyses and linear regression analyses were conducted. Drug effect analyses were also conducted, but this research did not focus on that.

A new mechanical indentation test was developed. The three-point breaking test and the calculation of volume fraction were also established during the research. By focusing on only Sham group and OVX group, the sensitivity analyses showed that these three new methods were much more sensitive than more traditional methods (Archimedes' density measurements and three-point bending tests).

The linear regression analyses were conducted to explore if the image processing methods can be correlated to the mechanical test methods, such as indentation testing and three-point breaking testing. The analyses showed that the volume fraction calculated from image processing methods explained most of the bone strength, but not all. The R2 of the correlation between the indentation force and the cancellous bone volume fraction (0.542) was a little lower than that between the indentation force and the whole bone volume fraction (0.588). The correlations between the mechanical results from threepoint breaking testing and the volume fractions (cancellous bone and whole bone) were in the range of 0.504 ~ 0.552. So it is better to combine both image analysis and mechanical tests to understand osteoporosis in bone properties and structures.