Department of Automatic Control and Systems Engineering, The University of Sheffield, Sheffield, United Kingdom.
INSIGNEO Institute for in Silico Medicine, The University of Sheffield, Sheffield, United Kingdom.
PLoS One. 2019 Jul 22;14(7):e0219404. doi: 10.1371/journal.pone.0219404. eCollection 2019.
Longitudinal studies of bone adaptation in mice using in vivo micro-computed tomography (μCT) have been commonly used for pre-clinical evaluation of physical and pharmacological interventions. The main advantage of this approach is to use each mouse as its own control, reducing considerably the sample size required by statistical power analysis. To date, multi-scale estimation of bone adaptations become essential since the bone activity that takes place at different scales may be associated with different bone mechanisms. Measures of bone adaptations at different time scales have been attempted in a previous study. This paper extends quantification of bone activity at different spatial scales with a proposition of a novel framework. The method involves applying level-set method (LSM) to track the geometric changes from the longitudinal in vivo μCT scans of mice tibia. Bone low- and high-spatial frequency patterns are then estimated using multi-resolution analysis. The accuracy of the framework is quantified by applying it to two times separated scanned images with synthetically manipulated global and/or local activity. The Root Mean Square Deviation (RMSD) was approximately 1.5 voxels or 0.7 voxels for the global low-spatial frequency or local high-spatial frequency changes, respectively. The framework is further applied to the study of bone changes in longitudinal datasets of wild-type mice tibiae over time and space. The results demonstrate the ability for the spatio-temporal quantification and visualisation of bone activity at different spatial scales in longitudinal studies thus providing further insight into bone adaptation mechanisms.
使用体内 micro-CT（μCT）对小鼠骨骼适应性进行纵向研究已广泛用于物理和药理学干预的临床前评估。这种方法的主要优势在于将每只小鼠作为自身对照，大大减少了统计功效分析所需的样本量。迄今为止，多尺度估计骨骼适应性变得至关重要，因为在不同尺度上发生的骨骼活动可能与不同的骨骼机制有关。之前的研究已经尝试了不同时间尺度的骨骼适应性测量。本文通过提出一种新的框架，扩展了不同空间尺度上的骨骼活性定量。该方法涉及应用水平集方法（LSM）来跟踪小鼠胫骨纵向体内 μCT 扫描中的几何变化。然后使用多分辨率分析估计骨骼的低和高空间频率模式。该框架的准确性通过将其应用于具有全局和/或局部活动的合成处理的两次分开扫描图像来量化。对于全局低空间频率或局部高空间频率变化，均方根偏差（RMSD）分别约为 1.5 体素或 0.7 体素。该框架进一步应用于野生型小鼠胫骨纵向数据集随时间和空间的骨骼变化研究。结果表明，该框架能够在纵向研究中对不同空间尺度的骨骼活性进行时空定量和可视化，从而深入了解骨骼适应机制。
