Jonescu-Cuypers C P, Chung H S, Kagemann L, Ishii Y, Zarfati D, Harris A
Department of Ophthalmology, University of Cologne, Cologne, Germany.
Br J Ophthalmol. 2001 Mar;85(3):304-9. doi: 10.1136/bjo.85.3.304.
Accurate Heidelberg retina flowmeter (HRF) measurements require correct manual setting of the HRF photodetector sensitivity. The neuroretinal rim produces a weak signal relative to the peripapillary retina. A newly developed HRF alignment and sensitivity protocol, capable of accurate rim measurement, was investigated.
18 eyes of nine healthy volunteers were examined by HRF. Three images of each eye were taken using three different imaging methods. Method 1: a conventional image (optic nerve head centred image with photodetector sensitivity optimised for the strong signal from the peripapillary retina); method 2: the setting of method 1 with photodetector sensitivity optimised for the weak signal from the rim; and method 3: the setting of method 2 with the temporal rim margin tangent to the lateral image border to remove the overpowering signal from the temporal peripapillary retina. The neuroretinal rim was defined by the Heidelberg retina tomograph (HRT). Blood flow and reflectivity values (DC component) in the rim area were compared for the three methods using pointwise analysis. Coefficients of variation of repeated measurements in 12 subjects have been calculated for method 3.
The neuroretinal rim area measured by method 1 had a significantly lower brightness compared with method 2 and 3 (p=0.0002 and p=0.0002, respectively). Method 2 provided proper sensitivity for the weak signals of the rim area based on rim tissue DC values; however, this sensitivity setting was too high for the strong signal from the peripapillary retina. Method 3 avoided the strong peripapillary signal with the proper signal from the rim and provided significantly higher flow values of the rim area at 75 and 90 percentile pixels (p=0.0065 and p=0.0038 respectively) compared with method 2. Interobserver repeatability ranged from 16.85% to 21.96% for the different parameters.
Method 3 provides an accurate and reproducible flow measurement of the neuroretinal rim area through proper sensitivity for the weak rim signal, alignment, and removal of the strong temporal signal from the image. This new method is recommended to improve accuracy of blood flow measurement in the neuroretinal rim.
准确的海德堡视网膜血流仪（HRF）测量需要正确手动设置HRF光电探测器的灵敏度。相对于视乳头周围视网膜，神经视网膜边缘产生的信号较弱。对一种新开发的能够准确测量边缘的HRF校准和灵敏度方案进行了研究。
对9名健康志愿者的18只眼睛进行HRF检查。使用三种不同的成像方法为每只眼睛拍摄三张图像。方法1：传统图像（以视神经乳头为中心的图像，光电探测器灵敏度针对视乳头周围视网膜的强信号进行优化）；方法2：方法1的设置，光电探测器灵敏度针对边缘的弱信号进行优化；方法3：方法2的设置，颞侧边缘与外侧图像边界相切，以去除来自颞侧视乳头周围视网膜的过强信号。通过海德堡视网膜断层扫描（HRT）定义神经视网膜边缘。使用逐点分析比较三种方法在边缘区域的血流和反射率值（直流分量）。计算了方法3中12名受试者重复测量的变异系数。
方法1测量的神经视网膜边缘区域的亮度明显低于方法2和方法3（分别为p = 0.0002和p = 0.0002）。方法2根据边缘组织直流值为边缘区域的弱信号提供了适当的灵敏度；然而，这种灵敏度设置对视乳头周围视网膜的强信号来说过高。方法3通过边缘的适当信号避免了视乳头周围的强信号，与方法2相比，在第75和第90百分位数像素处边缘区域的血流值明显更高（分别为p = 0.0065和p = 0.0038）。不同参数的观察者间重复性在16.85%至21.96%之间。
方法3通过对边缘弱信号的适当灵敏度、校准以及从图像中去除强颞侧信号，提供了准确且可重复的神经视网膜边缘区域血流测量。推荐使用这种新方法来提高神经视网膜边缘血流测量的准确性。
