Hip fracture is the most serious consequence of osteoporosis. In 2002, the cost of treating hip fractures consumed 63% of the national direct expenditure for OP, or more than $10 billion. By 2050, due to a rapidly aging population, the annual cost of treating hip fractures is projected to increase 6 fold to more than $60 billion. The ability to accurately estimate an individual's risk of osteoporotic hip fracture is key in preventing the expected increase of hip fractures because it would identify those individuals needing the most potent available fracture reduction therapy. Bone Mineral Density (BMD) measurement is currently used for diagnosis of osteoporosis. However, it has only limited utility in predicting hip fracture. This is because Individual fracture risk is also strongly influenced by local bone architecture and structure, bone turnover, cortical thickness, and, fall biomechanics. We recently developed (SBIR Phase 1) non-invasive automated imaging technology using ordinary radiographs that can measure cortical and trabecular parameters that are comparable to those measured by 3D uCT. In general it involves x-ray digitization, identification of regions of interest, trabecular extraction, background subtraction to obtain an image of trabecular structures and binerization plus skeletonization of those structures. Parameters of trabecular structure are measured using algorithms. 2D measurements of trabecular structure in ordinary radiographs of cores of proximal cadaveric femora correlated well with 3D mu/CT measurements. Those 2D measurements also correlated with biomechanical failure loads applied to those cores as well to whole cadaver proximal femora(J Bone Min Res 2003 ;18 Suppl 2, Abst # 1218, J Bone Min Res; 18 Suppl 2 Abst # 107 2003.). Our hypotheses in this proposal are: targeted, focal assessment of bone structure in hip x-ravs yields vastly improved ability to predict osteoporotic fracture as compared with BMD measurements. To assess the validity of this hypothesis, we will determine if the technology developed in Phase I of this project has the ability to discriminate between healthy postmenopausal women and postmenopausal women who have retrospectively experienced osteoporotic fractures. We will also test the short term reproducibility in vivo of our x-ray based measurements of bone microstructure and bone geometry by measuring them twice in those same groups no more than two weeks apart as the root mean square of the individual reproducibility errors.
