摘要目的：:利用视网膜神经纤维层厚度的二维矩阵数据，从空间分布特征的角度探讨正常人视网膜神经纤维层（RNFL）厚度与眼轴长度（AL）及年龄的相关性。方法：:系列病例研究。收集2015年10月至2020年7月于长沙爱尔眼科医院门诊进行体检的正常人群123例（213眼），建立涵盖不同年龄段（18~69岁）以及不同AL（22.07~30.00 mm）范围的正常人眼（包括屈光不正）的平均环视盘RNFL（cpRNFL）厚度数据组，及基于二维矩阵RNFL厚度图横断面的数据库。利用基于Python语言自行开发的计算程序，将年龄、AL、每个像素位置上的RNFL厚度代入基于混合线性模型的计算库进行矩阵运算，生成基于年龄、AL的有效斜率图。采用RNFL厚度/AL变化（μm/mm）表示RNFL厚度随AL增长的变化率。利用混合线性模型及堆叠图分析年龄、AL与RNFL厚度的相关性及其空间分布特征。结果：:cpRNFL厚度与年龄、AL呈负相关（ r=-0.146， P=0.023； r=-1.012， P=0.026）。在空间分布上观察，视网膜颞下区域的RNFL厚度与AL存在正负2种截然不同的相关性。随着AL增长，颞侧上、下象限的RNFL厚度变薄，变化最快的位置靠近视盘的颞下方向（-8.186 μm/mm），最慢位置位于远离视盘的颞下方向（-0.155 μm/mm）。在颞下方靠近颞侧处，存在与AL增加呈正相关区域，变化率最高处为靠近视盘的近颞侧位，变化率为6.292 μm/mm。长眼轴的RNFL厚度堆叠图提示与AL呈正相关的区域与长眼轴病例的RNFL束夹角移位存在重合。 结论：:RNFL厚度与AL、年龄相关，并且相关性及变化率在空间分布上不同。随着AL增长，颞侧RNFL束的夹角移位可能造成相关性改变。

abstractsObjective::To investigate the correlation between the thickness of the retinal nerve fiber layer (RNFL) and axial length (AL), as well as age in normal people from the perspective of spatial distribution characteristics by utilizing the two-dimensional thickness matrix data of RNFL thickness.Methods::This was a serial case research project. The examination results of a total of 123 normal people (213 eyes) who underwent physical examinations in the outpatient clinic of Changsha Aier Eye Hospital from October 2015 to July 2020 were recruited. A data set of mean circumpapillary retinal never fiber layer (cpRNFL) thickness and a cross-sectional database of the RNFL thickness map based on a two-dimensional matrix of normal eyes (including refractive errors) spanning different age groups (18-69 year-old) and different ALs (22.07-30.00 mm) was established. Age, AL and RNFL thickness at each pixel position were substituted into the calculation library based on the mixed linear model for matrix calculation by using a custom-made computing program based on the Python language. The effective slope maps based on age and AL factors were generated. RNFL thickness/AL change (μm/mm) indicated the rate of change of RNFL thickness with the growth of AL. A mixed linear model and stacked maps were used to analyze the correlation between each factor and RNFL thickness and its spatial distribution characteristics.Results::The mean thickness of cpRNFL was negatively correlated with age and AL ( r=-0.146, P=0.023; r=-1.012, P=0.026). However, in terms of spatial distribution, there were positive and negative correlations between the RNFL thickness and AL in the retinal inferotemporal region. As the AL grew, the thickness of the RNFL in the superior and inferior temporal quadrants became thinner. The fastest change was in the location close to the inferior temporal direction of the optic disc (-8.186 μm/mm), and the slowest change was in the location far from the optic disc (-0.155 μm/mm). In the inferior temporal area near the temporal side, there was a positive correlation with an increase in AL. The fastest rate of change was at the proximal temporal position near the optic disc, with a rate of change of 6.292 μm/mm. The RNFL thickness stack map of the longest axis indicated that the area positively correlated with the axis of the eye overlaps with the angle shift of the RNFL bundle in the case of the long axis. Conclusions::The thickness of RNFL is correlated with AL and age. The correlation and rate of change are different in spatial distribution. As the AL grows, the angle shift of the temporal RNFL bundle may cause a change in correlation.