Quantitative automated lesion detection (QALD) in moderate and severe TBI

描述（由申请人提供）： 摘要目前用于临床的神经影像学方案只能检测到中度和重度（m/s）TBI患者的一部分脑异常，部分原因是对 弥漫性轴索损伤（DAI）和轻微弥漫性脑损伤。本研究项目的目标是开发改进的定量自动病变检测（QALD）程序，以全面量化m/s TBI后脑损伤的性质和程度，并将脑损伤的性质和程度与认知结果联系起来。QALD的原理很简单：多模态结构MRI数据在标准坐标系中共同注册，以便TBI患者的局部脑组织特性可以与人口统计学匹配的对照人群进行统计学比较。QALD将利用来自六个高分辨率MR成像序列的数据，这些序列对不同类型的TBI相关损伤具有不同的敏感性：（1）T1图像，以分析组织体积、皮质厚度和皮质模糊（指示灰色/白色交界处的损伤）;（2）T2加权图像，以确保准确的组织分割并使用T1/T2比率评估髓磷脂;（3）液体衰减反转恢复（FLAIR）技术，以帮助量化病变范围;（4）扩散张量成像（DTI）技术，以量化皮质周围纤维和主要白色物质束中的DAI;（5）磁化传递成像（MTI）以量化DAI相关的脱髓鞘，和（6）敏感加权成像（SWI）以检测指示微血管的含铁血黄素残留。在目标1中，我们将使用基于皮质表面的坐标系分析不同皮质区域的体积和组织特性。我们将分析m/s TBI对局部皮质厚度和体积的影响，并分析挫伤、微出血、脱髓鞘和DAI的分布。此外，我们将分析损伤的层状分布，以确定微出血和DAI是否不成比例地发生在灰色/白色边界处。在目标2中，我们将分析主要纤维束的体积和组织特性。纤维束范围将用TRACULA定义，TRACULA使用纤维束图谱，并补充纤维方向数据和受试者特定的解剖学先验。我们还将应用我们实验室开发的程序来自动分割和分析胼胝体的不同区域。在目标3中，我们将分析皮质下灰质结构，包括丘脑、基底神经节、杏仁核、海马和小脑。每个结构将被自动分割，以评价平均组织特性，并扭曲到结构特异性3D模板上，以评价区域异常。最后，在目标4中，我们将使用计算机神经心理学测试来量化神经心理学缺陷，包括记忆力、注意力、执行功能和处理速度。这将使我们能够将脑损伤的位置和程度与m/s TBI患者的认知障碍相关联。我们也将描述不同影像学方法在侦测脑外伤相关病灶的敏感度.因此，QALD将提高m/s TBI的神经影像学研究的敏感性和客观性，阐明m/s TBI的神经病理学，并提高神经影像学研究的预后价值。

DESCRIPTION (provided by applicant): ABSTRACT Current neuroimaging protocols used in clinical settings detect only a subset of brain abnormalities in patients with moderate and severe (m/s) TBI, due in part to insensitivity to diffuse axonal injury (DAI) and subtle diffuse brain damage. The goal of this research program is to develop improved quantitative automated lesion detection (QALD) procedures to comprehensively quantify the nature and extent of brain damage following m/s TBI, and to relate the nature and extent of brain damage to cognitive outcome. The principle of QALD is simple: multimodal structural MRI data is coregistered in a standard coordinate system so that the regional brain tissue properties of TBI patients can be statistically compared with those of demographically-matched control populations. QALD will utilize the data from six high-resolution MR imaging sequences that are differentially sensitive to different types of TBI-related damage: (1) T1 images to analyze tissue volumes, cortical thickness, and cortical blurring (indicative of damage at the gray/white junction); (2) T2 weighted images to assure accurate tissue segmentation and to evaluate myelin using T1/T2 ratios; (3) Fluid Attenuated Inversion Recovery (FLAIR) to assist in quantifying lesion extent; (4) Diffusion Tensor Imaging (DTI) to quantify DAI in pericortical fibers and major white matter tracts; (5) Magnetization Transfer Imaging (MTI) to quantify DAI-associated demyelination, and (6) Susceptibility Weighted Imaging (SWI) to detect hemosiderin residues indicative of micro-hemorrhages. In Aim 1, we will analyze the volume and tissue properties of different cortical regions using a cortical-surface based coordinate system. We will analyze the effects of m/s TBI on regional cortical thickness and volume, and analyze the distribution of contusions, microbleeds, demyelination, and DAI. In addition, we will analyze the laminar distribution of damage to determine if microbleeds and DAI occur disproportionately at the gray/white boundary. In Aim 2, we will analyze the volume and tissue properties of major fiber tracts. Fiber tract extent will be defined with TRACULA, which uses fiber-tract atlases supplemented with fiber-direction data and subject-specific anatomical priors. We will also apply procedures developed in our laboratory to automatically segment and analyze different regions of the corpus callosum. In Aim 3, we will analyze subcortical gray matter structures, including the thalamus, basal ganglia, amygdala, hippocampus, and cerebellum. Each structure will be automatically segmented to evaluate mean tissue properties and warped onto structure-specific 3D templates to evaluate regional abnormalities. Finally, in Aim 4, we will quantify neuropsychological deficits using computerized neuropsychological tests of memory, attention, executive function, and processing speed. This will enable us to correlate the location and extent of brain damage with the cognitive impairments in m/s TBI patients. We will also describe the sensitivity of different imaging modalities in detecting TBI- related lesion. Thus, QALD will improve the sensitivity and objectivity of neuroimaging studies of m/s TBI, clarify m/s TBI neuropathology, and improve the prognostic value of neuroimaging studies.