Tensor Scale-based Methods for Assessment of Trabecular Bone Quality

摘要 骨小梁（TB）是一个复杂的准随机网络互连的支柱和板。结核病不断 重塑以动态地适应它所受到的压力（沃尔夫定律）。在骨质疏松症中， 骨形成和再吸收之间的动态平衡被扰乱，导致骨丢失和结构性骨丢失。 恶化，两者都增加了骨折风险。大多数椎骨骨折发生在结核病多发部位（椎骨， 桡骨，股骨近端）。骨骼的机械能力只能部分地用骨骼的变化来解释 骨密度（BMD），这导致了骨质量的概念，其中主要是建筑， 结核病强度的决定因素。CT和MRI的最新进展现在允许体内TB成像。但 有限的SNR排除了远小于TB厚度（100-150 μ m）的体素尺寸，因此导致图像 是天生模糊的。因此，大多数常规的TB结构的组织形态计量学方法， 评估不适用于体内分辨率方案。该建议引入了一种新的形态测量 通过体内成像和设计测量TB微结构质量的称为张量尺度（t尺度）的指数 实验计划，以评估基于t尺度的TB结构测量的再现性和灵敏度。还有， 我们将检查它们的优势，以发现结核病结构对疾病或治疗的反应变化 进展t尺度的基本原理是将椭球体拟合到局部结构。的独特性质 t尺度的一个重要特征是，椭球体的形状、方向和大小同时决定了拓扑结构（板块与 杆），方向和规模的小梁。我们在这个项目中的主要目标是-（1）开发方法 用于从TB图像计算基于t尺度的结构测量，（2）评估灵敏度， （3）检验t标度测度对实验数据的预测能力; 结核标本的生物力学参数和（4）检查的灵敏度的t尺度措施，以检测 通过体内MRI观察肺结核丢失和抗再吸收治疗的效果。该方法将（1） 避免了二值化的需要，（2）表征拓扑，方向和规模，而不需要 结核分枝杆菌化和（3）检测响应于治疗或疾病进展的早期TB结构变化。 中心假设是，新参数对检测重塑效应更敏感， 比传统的测量方法更容易重现。将对新方法的灵敏度和重现性进行评价 使用合成TB网络，对尸体桡骨远端沿着的TB核心进行微型CT和离体MR成像， 实验生物力学数据，完整标本和人类受试者的MR图像，最后， 来自NIH先前资助的临床研究的患者数据。我们的目标是应用基于t标度的分析 方法纵向和横截面成像研究评估骨质量。

ABSTRACT Trabecular bone (TB) is a complex quasi-random network of interconnected struts and plates. TB constantly remodels to adapt dynamically to the stresses to which it is subjected (Wolff's Law). In osteoporosis, this dynamic equilibrium between bone formation and resorption is perturbed, leading to bone loss and structural deterioration, both increasing fracture risk. Most osteoporotic fractures occur at sites rich in TB (vertebrae, radius, proximal femur). Bone's mechanical competence can only be partly explained by variations in bone mineral density (BMD), which led to the notion of bone quality, chief among which is architecture as a determinant of TB strength. Recent advances in CT and MRI now allow imaging of TB in vivo. However, the limited SNR precludes voxel sizes much smaller than TB thickness (100-150¿m), therefore resulting in images that are inherently fuzzy. Therefore, most conventional histomorphometric approaches to TB structure assessment are not applicable to in vivo resolution regime. This proposal introduces a new morphometric index called tensor scale (t-scale) to measure quality of TB micro architecture via in vivo imaging and designs experimental plans to evaluate reproducibility and sensitivity of t-scale-based TB architectural measures. Also, we will examine their strengths to detect TB architectural changes in response to disease or treatment progression. The fundamental principle of t-scale is to fit an ellipsoid to a local structure. The unique property of t-scale is that the ellipsoid's shape, orientation and size simultaneously determine the topology (plates vs rods), orientation and scale of trabeculae. Our major goals in this project are - (1) to develop the methodology for computing t-scale-based architectural measures from TB images, (2) to evaluate the sensitivity and reproducibility of t-scale-based measures, (3) to examine t-scale measures' ability to predict experimental biomechanical parameters of TB specimens and (4) to examine the sensitivity of t-scale measures to detect the effects of osteoporotic TB loss and antiresorptive treatment via in vivo MRI. The proposed method will (1) obviate the need for binarization, (2) characterize topology, orientation and scale without the need for skeletonization and (3) detect early TB architectural changes in response to treatment or disease progression. The central hypothesis is that the new parameters are more sensitive to detecting remodeling effects and more reproducible than conventional measures. Sensitivity and reproducibility of the new method will be evaluated using synthetic TB networks, micro-CT and ex vivo MR imaging of TB cores from cadaveric distal radii along with experimental biomechanical data, MR images of intact specimens and of human subjects, and finally, patient data from clinical studies previously funded by the NIH. Our objective is to apply t-scale based analysis methods to longitudinal and cross-sectional imaging studies for assessing bone quality.