Kotyk Taras, Varkey Thomas C, Demydchuk Anastasiia, Shamalo Svitlana, Tokaruk Nadiya, Bedei Viktoriia, Yurakh Omelian, Popadynets Oksana
Department of Human Anatomy, Ivano-Frankivsk National Medical University, Halytska Street, 2, Ivano-Frankivsk, Ukraine.
Department of Neurology, The University of AZ College of Medicine - Phoenix, Phoenix, AZ, USA.
Anat Sci Int. 2025 Mar;100(2):191-197. doi: 10.1007/s12565-024-00801-6. Epub 2024 Sep 10.
Despite advancements in automatic approaches for histomorphometry analysis of peripheral nerves, manual and semi-automated methods are widely utilized. Standard software functions are often unsuitable for analysis due to their irregular shapes, especially in pathological conditions. This study aims to assess the reproducibility of nerves morphometric analysis and compare results obtained using both default and new alternative algorithms. Sciatic nerves from Wistar rats (untreated and after administration of intraperitoneal hydrargyrum chloride), previously embedded in resin, were used. Morphometric analysis (diameters, myelin thickness, g-ratio, and circularity) was conducted using ImageJ on semithin sections, with axon and myelin boundaries manually outlined. Default diameters were calculated as the mean of Feret diameters, with subsequent calculations for myelin thickness and g-ratio. The alternative approach estimated diameters based on the geometric center of axons, iterating through selected coordinates; myelin thickness was obtained using line equations. In the control group, inter-rater agreement was higher or within expected reliability (0.8 ± 0.05). However, in the experimental group, myelin thickness, g-ratio, and axon circularity showed lower agreement (0.66, 0.58, and 0.68, respectively) without visible patterns on Bland-Altman plots. The alternative approach did not reveal significant differences between approaches, except for g-ratio in the control group and fiber diameter in the experimental group (p < 0.05), with effect sizes of 0.29-0.30 and 0.19-0.20, respectively. This study highlights reduced agreement among investigators analyzing nerve fibers under pathological conditions, raising concerns about the current standard measurement methods. The proposed approach, based on a single geometric center, provides more natural estimations for irregular fibers, and can be implemented in automated nerve fibers acquisition systems.
尽管外周神经组织形态计量分析的自动方法取得了进展，但手动和半自动方法仍被广泛使用。由于外周神经形状不规则，标准软件功能通常不适用于分析，尤其是在病理情况下。本研究旨在评估神经形态计量分析的可重复性，并比较使用默认算法和新替代算法获得的结果。使用先前包埋在树脂中的Wistar大鼠（未处理和腹腔注射氯化汞后）的坐骨神经。使用ImageJ在半薄切片上进行形态计量分析（直径、髓鞘厚度、g比值和圆度），手动勾勒轴突和髓鞘边界。默认直径计算为Feret直径的平均值，随后计算髓鞘厚度和g比值。替代方法基于轴突的几何中心估计直径，通过选定坐标进行迭代；使用线性方程获得髓鞘厚度。在对照组中，评分者间一致性较高或在预期可靠性范围内（0.8±0.05）。然而，在实验组中，髓鞘厚度、g比值和轴突圆度的一致性较低（分别为0.66、0.58和0.68），Bland-Altman图上没有明显模式。除对照组的g比值和实验组的纤维直径外（p<0.05），替代方法未显示出两种方法之间的显著差异，效应大小分别为0.29-0.30和0.19-0.20。本研究强调了在病理条件下分析神经纤维的研究人员之间一致性降低，引发了对当前标准测量方法的担忧。所提出的基于单个几何中心的方法为不规则纤维提供了更自然的估计，并且可以在自动神经纤维采集系统中实现。
