方法：使用自行搭建的UHR-OCT系统采集18例健康受试者(年龄23.2±0.9岁)左眼的图像， 应用一种基于轴向梯度信息和最短路径的分层算法对得到的SD-OCT图像进行边界探测获取角膜上皮厚度。

结果：自动与手动的上皮分层的位置有很好的相关性(P>0.05，配对t检验)。自动算法获取的6mm范围平均上皮厚度为52.1±2.1μm，范围从50.3±1.8μm到53.2±2.4μm， 与手动分层之间的角膜上皮厚度差异为0.4±0.3μm。 UHR-OCT自动算法重复测量上皮厚度的组内相关系数ICC和可重复性系数CoR分别为0.90±0.05，1.44±0.01μm。

结论：对于超高分辨率OCT，自动分层方法在角膜上皮厚度测量中有着很好的准确性和可重复性。

METHODS: Eighteen left eyes of 18 subjects(23.2±0.9)years; mean age±standard deviation)were imaged twice on non-consecutive days by a custom-built ultra-high resolution OCT(UHR-OCT)system. A hierarchic algorithm based on the axial gradient information and a shortest path search was developed to measure the corneal epithelial thickness profiles from the SD-OCT images.

RESULTS: There was good correlation between theautomatic and manual hierarchy positions of the epithelium( p>0.05, paired t-test). The average epithelial thickness of the 6mm central corneal region determined by the automatic hierarchic algorithm was 52.1±2.1μm, ranging from 50.3±1.8μm to 53.2±2.4μm along the horizontal meridian. The epithelial thickness differences between automatic and manual hierarchy by the UHR-OCT systems were 0.4±0.3μm. The intraclass correlation coefficients(ICC)and coefficients of repeatability(CoR)with repeated UHR-OCT measurements for epithelial thickness were 0.90±0.05 and 1.44±0.01μm respectively.

CONCLUSION:Regarding UHR-OCT system, the custom-developed automatic hierarchy has good repeatability and accuracy for measuring the thickness of corneal epithelium.