Monitoring of fractures with internal fixators using weight-bearing quantitative cone beam CT

项目总结/摘要 骨折的医疗负担对于患有骨不连和延迟愈合的患者来说是加重的。 骨不连的预测和刺激骨生长的新疗法的开发受到缺乏 定量、非侵入性测试，同时评估骨愈合的两个主要方面：（i）矿物质 骨痂和骨折间隙的密度;和（ii）负重下的机械稳定性。为了解决这个 挑战，我们提出使用一种新型的四肢锥束CT系统（CBCT），提供了一个独特的能力， 高空间分辨率的承重三维成像。这将允许测量骨骼的运动 通过估计碎片在负重和非负重扫描之间的位移， 上肢的此外，与传统CT非常相似，CBCT可以进行骨密度（BMD）测量 骨折处为了能够在四肢CBCT上定量评估骨折修复的负重， 必须减轻由于金属固定硬件引起的伪影和图像不均匀性。科学的前提是 这项工作是，金属硬件的影响可以最小化的组合，新的双能（DE） 适用于四肢CBCT和先进的基于模型的图像重建（MBIR）的技术， 对手术器械的先验知识。DE成像将提供衰减值的稳健校正 由于射束硬化而导致的不准确性。DE CBCT的高效、单次扫描实施将使用 四肢CBCT扫描仪的创新多源配置。已知分量重构 算法（KCR）将用于解决金属诱导的光子饥饿和非线性部分体积效应， 利用金属部件的形状和姿态的先验知识。这种方法的本质是 部件配准步骤，将提供植入物在负重下变形的精确估计， 从而产生一种评估骨折稳定性的新方法。将追求以下具体目标：1）使 通过一种新的DE MBIR算法和优化的DE从多源CBCT数据进行双能量CBCT 成像方案，以检测体模中约5%的骨矿物质密度相对变化; 2）整合 通过利用精确（~0.5 mm目标），在MBIR DE重建中对手术硬件的先验知识 配准误差）骨折固定硬件的可变形3D-2D配准，以估计组件姿态， 变形; 3）在植入的尸体中执行已知分量DE算法的临床翻译 四肢在控制负荷和试点患者研究。骨折患者将在2、4、8和12时进行成像 骨折后10周，以证明骨折修复过程中骨痂矿化的变化。这 研究将建立一种创新的定量成像方法，用于同时进行非侵入性评估 骨折修复的两个主要生物标志物：骨痂和骨折间隙的矿化和机械稳定性 骨-植入物结构在载荷下的应力。

PROJECT SUMMARY / ABSTRACT The healthcare burden of fractures is exacerbated for patients who suffer from non-unions and delayed unions. Prediction of non-unions and development of new therapies stimulating bone growth is challenged by a lack of quantitative, non-invasive tests to simultaneously assess the two primary aspects of bone healing: (i) mineral density of the callus and fracture gap; and (ii) mechanical stability under weight-bearing. To address this challenge, we proposed to use a novel extremity cone-beam CT system (CBCT) that provides a unique capability of weight-bearing 3D imaging at high spatial resolution. This will allow measurement of the motion of bone fragments by estimating their displacement between weight-bearing and non-weight-bearing scans of the extremity. In addition, much like conventional CT, CBCT can perform bone mineral density (BMD) measurements of the fracture. To enable quantitative weight-bearing assessment of fracture repair on extremities CBCT, artifacts and image nonuniformity due to metal fixation hardware must be mitigated. The scientific premise of this work is that the effects of metal hardware can be minimized by a combination of novel Dual Energy (DE) techniques suitable for extremities CBCT and advanced model-based image reconstruction (MBIR) incorporating prior knowledge of the surgical hardware. DE imaging will provide a robust correction of the attenuation value inaccuracy due to beam hardening. Efficient, single-scan implementation of DE CBCT will be achieved using the innovative multi-source configuration on the extremities CBCT scanner. The Known-Component Reconstruction algorithm (KCR) will be used to address metal-induced photon starvation and nonlinear partial volume effects by exploiting prior knowledge of the shape and pose of the metal component. Inherent in this approach is a component registration step that will provide a precise estimate of implant deformation under weight-bearing, resulting in a novel approach to asses fracture stability. The following specific aims will be pursued: 1) Enable Dual Energy CBCT from multi-source CBCT data by means of an novel DE MBIR algorithm and optimized DE imaging protocols to yield detection of ~5% relative change in bone mineral density in phantoms; 2) Integrate prior knowledge of surgical hardware in MBIR DE reconstruction by exploiting accurate (~0.5 mm Target Registration Error) deformable 3D-2D registration of fracture fixation hardware to estimate component pose and deformation; 3) Perform clinical translation of the Known-Component DE algorithms in implanted cadaveric extremities under controlled load and in pilot patient study. Fracture patients will be imaged at 2, 4, 8 and 12 weeks post-fracture to demonstrate detection of changes in callus mineralization during fracture repair. This research will establish an innovative quantitative imaging approach for simultaneous, non-invasive assessment of two primary biomarkers of fracture repair: mineralization of the callus and fracture gap, and mechanical stability of the bone-implant construct under load.