Assessing Spatial Patterns of Epidermal Nerve Fibers

描述（由申请人提供）： 皮肤水疱活检可以“鸟瞰”整个表皮的表皮神经纤维（ENF）的空间格局。糖尿病和非糖尿病受试者水泡图像的视觉比较显示，随着神经病变的进展，ENF的空间模式变得更加“聚集”。通过空间统计学对这些观察结果进行准确定量将提供一种微创筛查技术，以及评估其他早期神经病变生物标志物的工具。我们提出了一个为期两年的重点研究计划（R21），以评估使用空间统计的准确性和可靠性，以量化观察到的空间模式的ENF水泡图像。我们将：1.计算个体内和个体间的变异以测量空间聚类。我们将使用空间点处理方法来定义在图像内观察到的聚类量，并使用最近广泛收集的水泡图像来评估个体间/个体内的相关变异。2.建立空间聚类措施的标准范围。作为第一个具体目标的副产品，我们将建立按身体部位和糖尿病状态划分的空间聚集度的标准范围，并按年龄和性别分层。3.纳入形态学测量。ENF图像包括关于神经纤维本身的形态结构的信息，包括长度，方向，分支的数量等。我们提出了调查的统计模型，解决这些措施，以隔离由于糖尿病状态ENF的结构变化。每个目标的初步结果是有希望的，但需要在大量图像上复制，以确定其用于筛选的总体效用或作为评估其他生物标志物性能的比较标准。R21程序结构结合最近可用的水泡图像集合提供了一种理想的机制，用于仔细开发和严格评估ENF水泡图像的空间分析，以识别和验证神经病变的进展。

DESCRIPTION (provided by applicant): Skin blister biopsies allows a "bird's eye" view of the spatial pattern of epidermal nerve fibers (ENFs) across the epidermis. Visual comparison of blister images from diabetic and nondiabetic subjects shows the spatial pattern of ENFs becomes more "clustered" as neuropathy advances. Accurate quantification of these observations via spatial statistics would provide both a minimally invasive screening technique, and a tool to assess other proposed biomarkers of early neuropathy. We propose a focused two year research program (R21) to assess the accuracy and reliability of the use of spatial statistics to quantify observed spatial patterns of ENFs in blister images. We will: 1. Calculate intra- and interindividual variation for measures of spatial clustering. We will use spatial point process methods to define the amount of clustering observed within an image and assess the associated between/within individual variation using a recent extensive collection of blister images. 2. Establish normative ranges for measures of spatial clustering. As a by-product of the first specific aim, we will establish normative ranges for the degree of spatial clustering by body site and diabetes status, stratified by age and sex. 3. Incorporate morphologic measurements. ENF images include information regarding the morphologic structure of the nerve fibers themselves including length, direction, numbers of branches, etc. We propose investigation of statistical models addressing such measures to isolate structural changes in ENFs due to diabetes status. Preliminary results for each aim are promising but require replication on a large number of images to determine their overall utility for screening or as a standard of comparison for assessing performance of other biomarkers. The R21 program structure combined with a recently available collection of blister images provide an ideal mechanism for careful development and critical assessment of spatial analyses of ENF blister images for identifying and verifying progression of neuropathy.