Tomita G, Kitazawa Y
Department of Ophthalmology, Gifu University School of Medicine, Japan.
Nippon Ganka Gakkai Zasshi. 1991 Apr;95(4):348-53.
The reproducibility of topographic measurements of the optic nerve head was studied using a prototype Humphrey Retinal Analyzer system (Allergan Humphrey SKB, San Leandro, CA). Digital simultaneous stereographic images were obtained using programs in the Analyzer. The analyzer expressed the results in terms of microns of elevation at approximately 300 locations on the optic nerve head surface. At least 3 separate stereoscopic digital images were made of each of 8 eyes of 5 subjects. One randomly selected image of each disc was analyzed twice in order to determine intra-image reproducibility, and topographic analyses of two randomly selected images of each disc were compared to estimate the inter-image variability. We calculated the coefficients of variation (CV) of measurements at the center point of the optic disc. Further, we computed the absolute value of the difference in the degree of elevation among the measured locations on the optic disc and calculated the median and 95 percentile value of that group of approximately 300 differences. The CV, medians, and the 95 percentile values thus obtained for the 8 discs ranged from 0 to 6.1%, from 6 to 20 microns, and from 20 to 69 microns, respectively for intra-image reproducibility. Those for inter-image reproducibility ranged from 1.5 to 59.1%, from 6 to 33 microns, and from 36 to 92 microns, respectively.
使用原型汉弗莱视网膜分析仪系统（爱尔康汉弗莱SKB，加利福尼亚州圣莱安德罗）研究了视神经乳头地形图测量的可重复性。使用分析仪中的程序获取数字同步立体图像。分析仪以视神经乳头表面约300个位置处的高度微米数来表示结果。对5名受试者的8只眼睛中的每只眼睛至少拍摄3张独立的立体数字图像。对每个视盘随机选择的一张图像进行两次分析，以确定图像内的可重复性，并比较每个视盘随机选择的两张图像的地形图分析，以估计图像间的变异性。我们计算了视盘中心点测量值的变异系数（CV）。此外，我们计算了视盘上测量位置之间高度差的绝对值，并计算了该组约300个差值的中位数和第95百分位数。对于图像内可重复性，8个视盘得到的CV、中位数和第95百分位数分别为0至6.1%、6至20微米和20至69微米。对于图像间可重复性，它们分别为1.5至59.1%、6至33微米和36至92微米。
