Applegate R A, Nuñez R, Buettner J, Howland H C
Department of Ophthalmology, University of Texas Health Science Center at San Antonio, USA.
Optom Vis Sci. 1995 Nov;72(11):785-92. doi: 10.1097/00006324-199511000-00003.
Surface topography, as opposed to dioptric topography, defines the corneal surface in simple terms without assumptions. Accordingly, it is important to know how well surface topography can be measured with current videokeratometric machines.
The purpose of this paper is to quantify the accuracy with which the TMS-1 Corneal Modeling System can measure the surface topography of calibrated spherical, elliptical, and bicurve surfaces.
The Computed Anatomy TMS-1 videokeratometer was used to measure three spherical, three elliptical, and two bicurve surfaces with known characteristics. Surface characteristics were either back-calculated from the dioptric files or directly obtained from the TMS-1 elevation file for each of 6400 points (256 points in each of 25 rings). The accuracy with which each method determined the true surface was quantified by calculating the root mean squared error (RMSE) of the 6400 measured surface elevations from the known surface elevation at each sampling point.
(1) For spherical and elliptical surfaces, back-calculation of surface elevation from the dioptric file can be made with RMSE of 5 mu or less. (2) For spheres but not elliptical surfaces the TMS-1 elevation file defines the surface with RMSE 5 mu or less. (3) The surface area measured by placido-based videokeratometers varies with surface curvature. (4) RMSE in measured surface elevation increase as the distance from the videokeratometric axis increases. (5) For bicurves, the dioptric maps are smoothed by the TMS-1 over abrupt transitions and for large transitions never recover. Additionally, our back-calculation methods further smooth abrupt transitions, making the RMSE of the bicurve surface that is back-calculated from the dioptric file larger than the RMSE of the surface generated from the TMS-1 elevation file.
Surface elevations can be back-calculated from dioptric files with RMSE of 5 microns or less for spheres and elliptical surfaces as long as there are no areas of abrupt transition. If areas of abrupt transition exist, the TMS-1 elevation file provides more accurate surface profile data.
与屈光地形图不同,表面地形图以简单的方式定义角膜表面,无需假设。因此,了解当前的视频角膜曲率计测量表面地形图的精度很重要。
本文的目的是量化TMS-1角膜建模系统测量校准的球形、椭圆形和双曲面表面地形图的准确性。
使用计算机解剖学TMS-1视频角膜曲率计测量三个具有已知特征的球形、三个椭圆形和两个双曲面表面。表面特征要么从屈光文件中反算得出,要么直接从TMS-1高度文件中获取,每个文件包含6400个点(25个环,每个环256个点)。通过计算每个采样点处6400个测量表面高度与已知表面高度之间的均方根误差(RMSE),量化每种方法确定真实表面的准确性。
(1)对于球形和椭圆形表面,从屈光文件中反算表面高度时,RMSE可低至5μm或更小。(2)对于球形表面而非椭圆形表面,TMS-1高度文件定义表面时RMSE为5μm或更小。(3)基于普拉西多盘的视频角膜曲率计测量的表面积随表面曲率而变化。(4)测量表面高度的RMSE随着与视频角膜曲率计轴距离的增加而增大。(5)对于双曲面,TMS-1会在突然过渡处平滑屈光图,对于大的过渡区域则无法恢复。此外,我们的反算方法会进一步平滑突然过渡,使得从屈光文件反算得到的双曲面表面的RMSE大于从TMS-1高度文件生成的表面的RMSE。
只要没有突然过渡区域,对于球形和椭圆形表面,可以从屈光文件中反算表面高度,RMSE为5微米或更小。如果存在突然过渡区域,TMS-1高度文件可提供更准确的表面轮廓数据。