Advancing and calibrating anisotropic diffusion MR imaging brain connectome with Taxon brain network diffusion phantoms

各向异性扩散磁共振成像（AdMRI）的可靠性存在一个关键的差距。这一差距可能是 通过使用地面实况测量能力进行填充，该测量能力允许对填充的水进行必要的参数控制 在微米尺度上的管的几何形状，可以产生代表数百万个轴突的路径， 脑束轨迹扩散张量成像（DTI）出版物报告了具有临床意义的系统误差， 混淆了跨仪器和时间的准确定量评估。提供精确误差的参考幻象 这些指标将推进MRI生物物理学科学和临床定量准确性。使用参考数据的校正算法 可以减少系统测量误差，实现精确的可重复测量，并提供交叉扫描仪规范 用于AdMRI病理学。该项目将使用"Taxons ™"（纺织品）提供第一个可行的AdMRI体模"地面实况" 轴突形状的纳米管），并应用先进的双组分聚合物纳米级生产方法 来创造出与人体组织学相匹配的结构。在这样做的过程中，我们将在800纳米处提供轴突规模的分类单位。 直径，具有每平方毫米100万个分类单位的堆积密度，与实际的人类胼胝体轴突相匹配 测量.在第一阶段，我们提出并交付了具有12微米内径管的分类单元，其包装密度为 1241/mm2，可以用水填充并在人体组织范围内进行FA测量。我们实际上"超过- 递送"，超过每mm2 1，000，000个的包装密度，覆盖人轴突组织范围。我们现在可以 精确地参数化控制直径、堆积密度、受限/受阻和各向同性水分数以进行测试， 改进了扩散的主房室模型。我们创造了一个纤维束路由机器，可以在可行的成本， 与人体组织相匹配的人类尺度纤维束路径，例如2000万个路由分类群的视觉系统眼睛到LGN。 1比1比例分类单元网络模型量化了具有100微米路径的每个分类单元路径的dMRI测量精度 沿着轨迹的精确度。我们在十个地点扫描了第一阶段的幻影。我们在实证研究中发现， 是实质性的系统性、跨仪器和测量误差（例如，5x TBI效应大小），误差稳定，以及 可以修正（消除94%的系统误差）。该项目的第二阶段将：1）提供第一个AdMRI体模 用于地面实况测量，以量化dMRI生物物理学、空间均匀性和路由精度; 2）提供全面的 自动量化测量的准确性和可重复性; 3）评估20多个部位的AdMRI精密度， 在1.5、3、7、9.4和14T场强下的测量误差; 4）开发一套路由模型（Eye> LGN> V1，Spinal 脊髓和皮质束）。这些体模和/或子组件将用非MRI方法（共聚焦& 电子显微镜）使用NIST可追踪测量。参与第一阶段扫描的研究人员和中心主任 结果的审查非常积极，30多个研究中心提供免费扫描时间来使用体模， 产生的质量保证报告。放射学已经取得了基于体模的关键性成功（即，CT亨氏体模 1990年代）。该项目将提供定量AdMRI体模，使MRI指标在以下范围内变得准确： 供应商和时间实施定量质量保证（QQA）。

There is a critical gap in the reliability of anisotropic diffusion magnetic resonance imaging (AdMRI). This gap can be filled by using a ground truth measurement capability that allows for the necessary parametric control of water filled geometries of tubes at the micron scale that can produce paths representative of the millions of axons across centimeters in brain tract trajectories. Diffusion Tensor Imaging (DTI) publications report clinically significant systematic error that confounds accurate quantitative assessments across instruments and time. Reference phantoms that provide exact error metrics will advance MRI biophysics science and clinical quantitative accuracy. Correction algorithms using reference data can reduce systematic measurement error, enabling accurate reproducible measurement and provide cross scanner norms for AdMRI pathology. This project will deliver the first viable AdMRI phantom “ground truth” using ‘Taxons™’ (textile axon shaped nanotubes), invented by this team, and apply advanced bi-component polymer nanoscale production methods to create structures matched to human tissue histology. In doing this we will deliver axon scale taxons at 800 nanometer diameter, with a packing density of one million taxons per mm2, matched to actual human corpus callosum axon measurements. In Phase I we proposed and delivered taxons with 12 micron inner diameter tubes with a packing density of 1241 per mm2 that could be filled with water and produce FA measurement in the human tissue range. We actually “over- delivered”, exceeding a packing density of 1,000,000 per mm2 covering the human axonal tissue range. We can now precisely parametrically control the diameters, packing density, restricted/hindered, and isotropic water fractions to test and improve leading compartmental models of diffusion. We created a fasciculus routing machine that can, at viable cost, create human scale fasciculus routes matched to human tissue, such as the optic system eye to LGN, of 20 million routed taxons. The 1 to 1 scale taxonal network phantoms quantify dMRI measurement accuracy for each taxon path with 100 micron path precision along the trajectory. We scanned the phase I phantoms at ten sites. We established in empirical studies that there is substantial systematic, cross instrument and measurement error (e.g., 5x the TBI effect size), that the error is stable, and can be corrected for (removed 94% of systematic error). Phase II of this project will: 1) provide the first AdMRI phantom for ground truth measurement to quantify dMRI biophysics, spatial homogeneity, and routing precision; 2) provide fully automated quantification of accuracy and repeatability of measurement; 3) assess AdMRI precision of 20+ sites, quantifying measurement error at 1.5, 3, 7, 9.4 and 14T field strength; and4) develop a set of routing phantoms (Eye>LGN> V1, spinal cord and cortical tracts). These phantoms and/or subcomponents will be measured with non-MRI methods (confocal & electron microscope) using NIST traceable measurements. Researchers and center directors involved in Phase I scanning and reviewing of the results were very positive, with 30+ sites offering free scanning time to use the phantom, and to utilize the resulting quality assurance reports. Radiology has had phantom based pivotal successes (i.e., CT Hounsfield phantoms in the 1990s). This project will deliver a quantitative AdMRI phantom, enabling MRI metrics to become accurate across vendors and time implementing quantitative quality assurance (QQA).