基于图像拼接方法对不同程度干眼症患者角膜上皮基底神经丛的定量分析

目的：利用图像拼接方法对不同程度干眼症患者角膜上皮基底神经丛进行定量分析，探讨其变化规 律。方法：横断面研究。选取2021年4—10月于山西省眼科医院门诊就诊的干眼症患者共91例（91眼）， 其中轻度干眼34例、中度干眼33例、重度干眼24例，同时征集正常健康志愿者作为对照组共27例 （27眼），所有受检者均选择右眼数据分析，对所有受检者分别行常规眼科检查、眼表综合分析仪检 查、角膜共焦显微镜检查后，运用Photoshop CC 2018图像处理软件合成大范围角膜上皮基底细胞层 下神经丛（SNP）结构拼图，截取以涡状结构为中心的、大小为700 μm×700 μm的图像进行分析，结 合Image J图像分析系统计算得出图像中神经纤维总长度（NFL），将所得数据通过图像大小与像素之 间的换算关系计算出单位面积（1 mm2）的神经纤维总长度值即神经纤维密度值（mm/mm2）。采用单 因素方差分析对各组数据进行比较，同时对两两之间进行LSD多重比较；采用Pearson线性相关分析 各组角膜SNP密度与眼表综合分析检查中各项指标间的相关性。结果：4组间性别、年龄比较差异 无统计学意义。对照组SNP密度为（22.71±6.19）mm/mm2，轻度干眼组为（28.58±5.19）mm/mm2， 中度干眼组为（28.17±4.71）mm/mm2，重度干眼组为（5.98±7.32）mm/mm2，4组SNP密度总体差异 有统计学意义（F=88.10，P<0.001）；轻、中、重度干眼SNP密度与对照组差异均有统计学意义（均 P<0.001），轻度干眼组与中度干眼组间差异无统计学意义，重度干眼组与轻、中度干眼组间差异均 有统计学意义（均P<0.001）。干眼症患者角膜SNP密度与泪河高度、睑板腺缺失面积评分（上+下）及 眼表充血评分之间无明显相关性，与非侵入性泪膜破裂时间呈正相关（r=0.51，P<0.001），与脂质层 分级、角膜荧光素钠染色分级呈负相关（r=-0.30，P=0.004；r=-0.68，P<0.001）。结论：角膜SNP密 度及形态改变与干眼症的严重程度分级有关，可作为临床评估干眼程度的有效指标。

Quantitative Analysis of Corneal Epithelial Basal Plexus in Patients with Different Degrees of Dry Eye Based on Image Mosaic Method

Objective: To quantitatively analyze the corneal epithelial basal plexus in patients with different degrees of dry eye by image mosaic method and to explore its changes. Methods: Cross sectional study. A total of 91 eyes of dry eye patients (n=91) with mild dry eye (n=34), moderate dry eye (n=33), and severe dry eye (n=24) were recruited from the outpatient clinic of Shanxi Eye Hospital between April 2021 and October 2021. A total of 27 eyes (n=27) were recruited from normal healthy volunteers as the control group, and all subjects were selected for right eye data analysis. After routine ophthalmic examination, comprehensive ocular surface analyzer examination, and confocal microscopy of the cornea were performed for all subjects, the structural puzzle of the large corneal epithelial basal nerve plexus (SNP) was synthesized using Photoshop CC 2018 and captured as a vortex structure centered at 700 μm×700 μm of the images were analyzed, combined with the calculation of the total length of nerve fibers (NFL) in the images by the image J. The resulting data were converted from the image size to pixels to calculate the total length of nerve fibers per unit area (1 mm2 ), which was the value of nerve fiber density (mm/mm2 ). One way ANOVA was used to statistically compare the data in each group,while pairwise comparisons were made between the means of each group (LSD multiple comparisons), and Pearson's linear correlation was used to analyze the correlation between corneal nerve fiber density in each group and each index in the comprehensive analysis examination of the ocular surface. Result: Comparison of gender and age among the 4 groups showed no significant difference. In the control group, nerve fiber density was (22.71±6.19) mm/mm2 , Mild dry eye group was (28.58±5.19) mm/mm2 , Moderate dry eye group was (28.17±4.71) mm/mm2 , and Severe dry eye group was (5.98±7.32) mm/mm2,corneal SNP density was significantly different between groups (F=88.10, P<0.001). The density of SNP in the control group was significantly different from that in the mild,moderate,and severe dry eye groups (P<0.001), that in the mild dry eye group was not significantly different from that in the moderate dry eye group, and that in the mild,moderate dry eye group was significantly different from that in the severe dry eye group (P<0.001). There were no significant correlations between corneal SNP density and tear meniscus height, meibomian gland missing area score (superior+inferior) or ocular surface hyperemia score, a positive correlation with noninvasive breakup time (r=0.51, P<0.001), lipid layer grade (r=-0.30, P=0.004) and corneal fluorescein sodium staining grade (r=-0.68, P<0.001). Conclusion: The density and morphological changes of the basal nerve plexus in the corneal epithelium are associated with the severity grading of dry eye and can be used as a useful indicator for clinical assessment.