Purchase this article with an account.
J.W. McLaren, C.B. Nau, W.M. Bourne; Use of an Automated Method for Estimating Keratocyte Density in Confocal Microscopy After LASIK . Invest. Ophthalmol. Vis. Sci. 2005;46(13):2763.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Purpose: We compared automated and manual methods of assessing keratocyte densities in confocal images of the corneal stroma before and to 60 months after LASIK. Methods: Seventeen eyes of 11 patients were treated with LASIK to correct refractive errors between –2D and –11D. Central corneas were examined by confocal microscopy (Tandem Scanning, Reston, VA) before and at 1 month, 3 months, 6 months, 1 year, 2 years, 3 years, and 5 years after the procedure. Cell density was assessed manually in 6 layers of the stroma (2 video frames per layer) by counting cells in a rectangular area of each selected frame. Cell density in the same frames was also determined by using a custom program that identified and counted bright objects. Detection parameters were set from manually selected cells in 1 or 2 sample scans from each sample time after LASIK and 4 sample scans before LASIK. Cell densities after treatment were compared with cell densities in the same layer before treatment by using Generalized Estimating Equation models (to account for possible correlation between eyes from the same patient) and Bonferroni correction for 7 comparisons. Results: Before LASIK the unweighted mean cell density was 24,222 ± 3,023 cells/mm3 by manual assessment and 24,290 ± 3,095 cells/mm3 by automated assessment. At 1 month, density decreased to 20,947 ± 3,202 cells/mm3 and 21,860 ± 2,617 cells/mm3, and at 60 months density decreased to 17,046 ± 1,285 cells/mm3 and 17,718 ± 1,648 cells/mm3 by manual and automated methods respectively. By both methods mean density decreased significantly (p<0.007) in the posterior half of the flap at 1 month and in the layers anterior to and posterior to the interface (p<0.05) by 1 year, and remained decreased to 5 years (p<0.007). The automated method did not detect differences just posterior to the ablation at 1, 3, or 6 months, although the manual method did. Both methods indicated decreased cell densities in the anterior 90% of the stroma at 5 years (p<0.007). Conclusions: Changes in cell density can be detected by automatic assessment as well as by manual assessment in a longitudinal study. This program will be a consistent, objective, and time–saving tool for assessing cell density in confocal microscopy.
This PDF is available to Subscribers Only