Purpose: When comparing follow-up endothelial cell (EC) density measurements it is only possible to demonstrate cell loss in large cohorts or in pronounced cases due to the standard deviation of measurements. However, especially in clinical studies or refractive surgery patients, EC stability is an important factor. Thus we developed a computer program to achieve cell-by-cell alignment of conventional specular non-contact EC photographs. This is a pilot study to evaluate whether this new diagnostic technique is applicable in postoperative patients.
Methods: Digitized endothelial cell photos of 30 eyes following implantation of a phakic posterior chamber intraocular lens for correction of high myopia were analyzed. All EC centroids were dotted on each image, and early and late follow-up pictures were automatically aligned on the basis of these centroids. The tracking results were confirmed via alternating image presentation of the corresponding image areas, and were reviewed for loss of individual cells. In addition, conventional EC density measurements were performed.
Results: Mean time interval between first and second postoperative EC image was 1.4 years (range 3 months-2.2 years), with early images taken at a median of 5.5 months after IOL implantation (lower/upper quartile 1.2-12.3 months). Extrapolated central EC density was 2812±500/mm2 at the first time point, and 2797±524/mm2 at the second time point. In 26 out of 30 image pairs, the EC mosaic was successfully retraced; cell loss in this area was excluded via photo flickering. Only in 4 image pairs, the EC mosaic could not be matched.
Conclusions: We demonstrate that the corneal EC mosaic of clinical routine non-contact microscope pictures can be superimposed and compared on single cell level over time with our new computer based program. This new method is valuable to judge on EC stability even in small cohorts since it does not require mean values and standard deviations.