QCT Image Based Measurements For The Structure Parameters Of Animal Bone

Osteoporosis has been one of the most common diseases among the elderly. The analysis of bone diseases based on animal clinical trials has important reference value for the study of the causes, early prevention and diagnosis of osteoporosis. At present, the diagnosis of osteoporosis mainly depends on the measurement of bone mineral density, which only partly reflects the bone strength in terms of bone mineral content, while the bone health status can also be revealed from the perspective of bone's morphological structure. More comprehensive and specific understanding of osteoporosis' pathology can be obtained by the measurement of micro parameters like the trabecular thickness etc., which is helpful for timely diagnosis and treatment of this disease. Although there are a variety of measurements for the mircoscopic parameters of bone, the analysis is still lack of accuracy, efficiency and stability, due to the complicacy and diversity of trabecular bone. Hence further studies are required to improve in these aspects.By means of quantitative CT, trabecular bone's three-dimensional structure can be observed while measuring bone density, so the parameters of the bone structure can also be measured. Trabecular bone structure parameter measurements are majorly involved on its morphological and topological research. In this paper, various methods of bone density and bone structure parameters are studied on the platforms of small animal CT and high resolution CT, and the methods are verified by experiments using simulated data and real data.Animal bone density measurement method is based on the physical principle of linear relationship between CT value and material density. This article selects a more accurate method with standard phantom of bone by a serious of contrastive experiment, and the accuracy verification is completed.Four morphological parameters are researched in this article:1) trabecular thickness; 2) trabecular spacing; 3) trabecular number; 4) bone structure model index (SMI). A three-dimensional algorithm based on distance transform is applied to count thickness and spacing, and trabecular number can also be count by thickness and spacing. Bone volume and bone surface based on Marching-Cubes algorithm are evaluated to calculate SMI with extra precision. Compared with BoneJ's results, the relative errors of trabecular thickness, spacing and number of six groups of boar mandible are less than 3%, and the average relative error of SMI drops down to 3.49%. These methods are proved to have sufficient accuracy and stability.Connectivity and anisotropy are selected as the two parameters for bone's topology research. Pre-processing of 3D denoising is implemented to enhance results accuracy in measurement of connectivity. And both precision and consistency get significantly promoted by iteration of mean intercept length in the measurement of anisotropy. Compared with BoneJ's measurement, the relative error of connectivity is zero, and the average relative error of anisotropy is smaller than 2%, which proves the methods have sufficient accuracy and stability.All the above measurement methods are accomplished on VC++ platform, and finally packaged as a complete bone structure parameters analysis software. By the algorithm optimization, the final system generated from this article performs excellently in succinct operation, intuitive 3D visualization effect, moderate runtime and accurate results, which has vital value on animal bones image parameter measurement.