Non-invasive assessment of bone strength

这项研究的长期目标是开发非侵入性评估骨和关节疾病患者的骨质量和强度的方法。该方案的核心是使用新建立的人类微型计算机断层扫描(µCT)和临床CT来生成特定于患者的有限元(FE)模型，这些模型可以确定骨强度。该研究计划有四个具体目标，旨在改善对骨质疏松症患者的护理，这些目标是基于十年的工作。首先，Work将推进我们的定制FE代码，该代码旨在使用标准桌面工作站以快速、高效的方式求解超大型模型(10‘S，数百万个元素)。这与以前使用“超级计算机”的必要性有很大不同，并使这项技术更容易获得。重要的建模进展将包括复杂的非线性材料模型，以改进骨强度的估计。其次，将开发分析复杂的3D CT图像中包含的人类股骨和/或脊椎的能力，并形成自动化过程的核心，以便能够从CT估计骨强度。在第三个目标中，用潜在的材料结构信息丰富CT数据将基于开发一种新的骨图谱，该图谱包含详细的微结构模式，以增强骨强度的有限元预测。与目前使用CT密度值假设材料属性的简单方法相比，这种方法将是一个重大进步。最后，将开发机器学习工具，以利用基于图像的有限元模型预测骨强度，从而可以计算出特定对象的骨强度估计。这一研究重点是重要的，因为它将建立一种新的方法，通过结合最先进的成像和专门的有限元建模来早期检测和监测骨强度。这种方法将为更好地理解骨质量和骨折负荷之间的关系提供基础，并为更好地管理骨质疏松等破坏性疾病提供亟需的工程贡献。

The long-term goal of this research is to develop methods for non-invasive assessment of bone quality and strength in patients suffering from bone and joint diseases. The core of this proposal uses newly established human micro-computed tomography (µCT) and clinical CT to generate patient-specific finite element (FE) models that can determine bone strength. There are four specific aims of the research program working towards the goal of improving care for patients with osteoporosis, and these are based on a decade of work. First, work will advance our custom FE code designed to solve extremely large models (10's of millions of elements) in a fast, efficient manner using standard desktop workstations. This is a major departure from the previous necessity of using 'super computers' and makes the technology significantly more accessible. Important modeling advances will include sophisticated non-linear material models for an improved estimate of bone strength. Second, the ability to analyse a human femur and/or vertebrae that is contained in complex 3D CT images will be developed, and form the core of an automated process to enable bone strength estimation from CT. In the third aim, enriching CT data with underlying material fabric information will be based on the development of a novel bone atlas that contains detailed microarchitectural patterns to enhance the FE prediction of bone strength. This approach will be a major advancement over the currently simplistic approach of using CT density values to assume material properties. Finally, machine-learning tools will be developed to utilize the image-based FE model predictions of bone strength so that a subject-specific estimate of bone strength can be calculated. This research focus is important, as it will establish a new approach for early detection and monitoring of bone strength by combining state-of-the-art imaging with specialized FE modeling. This approach will provide a basis for better understanding of the relationship between bone quality and fracture load, and provide a much-needed engineering contribution to better management of devastating diseases such as osteoporosis.