Comparison of flex-center, center, and corner methods of corneal endothelial cell analysis

Purpose: The center method of corneal endothelial cell analysis is rapid but excludes the outermost digitized cells of a contiguous group from analysis; the flex-center method (Konan, Inc) is a modification that includes analysis of the outermost cells, which is advantageous in images with few cells. In this study, we examined agreement among the flex-center, center, and corner (standard) methods of endothelial analysis. Methods: Identical cells in endothelial images of 10 normal corneas and 10 corneas after penetrating keratoplasty (PK) were analyzed by each method. Agreement among methods for endothelial cell density (ECD), coefficient of variation of cell area (CV), and the percentage of hexagonal cells (HEX) was assessed by using a Student-Newman-Keuls procedure. Results: In normal corneas, there were small (clinically insignificant) differences among methods for ECD (P < 0.001) (mean ± SD: flex center, 2846 ± 248 cells/mm; center, 2870 ± 253 cells/mm; and corner, 2892 ± 254 cells/mm), CV (P < 0.001) (flex center, 33% ± 3%; center, 30% ± 3%; and corner, 30% ± 3%), and HEX (P = 0.004) (flex center, 60% ± 6%; center, 60% ± 8%; and corner, 58% ± 7%). In PK corneas, the methods agreed for ECD (P = 0.06) (flex center, 908 ± 319 cells/mm; center, 912 ± 307 cells/mm; and corner, 929 ± 333 cells/mm) but disagreed for CV (P = 0.02) (flex center, 35% ± 13%; center, 30% ± 10%; and corner, 35% ± 15%) and HEX (P = 0.02) (flex center, 56% ± 19%; center, 54% ± 17%; and corner, 43% ± 23%). Conclusion: ECD agreed among methods in normal and PK corneas, whereas morphometric data agreed poorly in PK corneas.