Morphology of Corneal Basal Epithelial Cells by In Vivo Slit-Scanning Confocal Microscopy Devin A. Harrison, M.D., Cara Joos, B.S., and Renato Ambrósio, Jr., M.D. Purpose. To determine corneal basal epithelial cell density and morphology in a group of normal corneas and corneas with various conditions. Methods. The central corneas of a group of 20 normal patients and 27 other patients who either had refractive surgery or had documented corneal pathology were examined with an in vivo slit-scanning confocal microscope. For all these patients in whom a well-defined basal epithelial layer was visualized, the basal cell density, area, and number of sides per cell were determined using customized software. Results. The average basal cell density was 5,274 ± 575 cells/mm 2 , the average cell area was 192 ± 19.6 m 2 , and the average number of cell sides was 5.5 ± 0.1 sides for the group of normal controls. There were no statistically significant differences between basal cell density, cell area, or number of sides between the normal group and any of the other groups in which there were four or more patients. Corneas examined from patients with Fuchs’ endothelial dystrophy or bullous keratopathy had hyperreflective basal epithelial cell borders. The measurement of the number of cell sides is not reliable using software developed for endothelial cell counting. Conclusion. Corneal basal epithelial cell density and area were found to be remarkably consistent in the variety of corneal disorders that we examined. This finding is further evidence that this aspect of corneal epithelial cell differ- entiation and maturation is tightly controlled with little variability between individuals even if a corneal condition is present. Key Words: Confocal microscopy—Corneal epithelium— Epithelial cells—Image analysis—Refractive. Maintenance or reestablishment of the corneal epithelium is essential for a protective barrier against infections and for good optical performance. In 1983, Thoft and Friend 1 proposed the X, Y, Z hypothesis of corneal epithelial maintenance. The hypothesis states that the loss of cells from the corneal epithelial surface (Z) is maintained by the proliferation of basal epithelial cells (X), and the centripetal movement of peripheral epithelial cells (Y). If a steady state of corneal epithelial cells is to be maintained, the loss of cells should be equal to the addition of new cells (X + Y  Z). More recent studies have investigated the maturation and move- ment of corneal epithelial cells. 2–7 The coordination of cell differentiation and division likely involves a complex regulatory mechanism. In vivo corneal confocal microscopy has been shown to be a valuable tool in the evaluation and diagnosis of a variety of corneal disorders including corneal epithelial disorders. 8,9 Other studies have determined basal epithelial cell area and density in a normal population 10 and have shown that basal epithelial cell density is not significantly different by sex or with aging. 11,12 The purpose of the current study was to determine whether there were any differences in basal epithelial cell density or area in a variety of corneal disorders. METHODS Scanning slit confocal microscopy was performed with the Con- foscan 2 (Fortune Technologies, Padova, Italy). 13 Topical anes- thesia was used before examination with the standard 40× objec- tive lens, and the light level set at medium to high. The scans were set to start at the endothelium of the central cornea and progres- sively scan anteriorly, then back to the endothelium for six passes. The Confoscan 2 acquires 350 images of a 300 × 200-m frame at a rate of 25 frames per second. All data were recorded and ana- lyzed with the Confoscan 2 software, a part of NAVIS developed by Fortune Technologies. In approximately 60% of all scans done, a well-defined basal epithelial cell layer was identified. Since the distance between scans should be 10 to 15 m with our settings, it is easy to miss the basal cell layer given the mitigating factor of patient movement. When a well-defined basal cell layer was identified, the image was selected for analysis. The image was first manipulated using the supplied software, so that an “inverted” image was obtained (Figs. 1, 2). The inverted image was then saved and analyzed with the automated cell count feature designed for analyzing endothelial cells. This software can perform similar calculations on inverted basal cell images. However, due to the differences in the layers, manual corrections of the automatic counts are performed to en- sure that each cell is outlined properly. The software is able to identify accurately approximately 50% of the basal cells. The re- maining cells were identified and outlined manually. For images where the basal cell borders were difficult to discern, the borders became more apparent by varying the image contrast and brightness. The cell densities (cells/mm 2 ), cell area (m 2 ), and number of sides of each cell were entered on a computer spreadsheet. Aver- Submitted December 17, 2002. Accepted January 21, 2003. From the Department of Ophthalmology (D.A.H., C.J., R.A.), University of Washington, Seattle, Washington, U.S.A.; and the Department of Oph- thalmology (R.A.), University of São Paulo, São Paulo, Brazil. Supported in part by an unrestricted grant from Research to Prevent Blindness, Inc., New York, NY. The authors have no commercial interest whatsoever with any products or companies mentioned in this report. Address correspondence and reprint requests to Devin A. Harrison, M.D., University of Washington, Ophthalmology, Box 356485, Seattle WA 98195-6485. E-mail: dah@u.washington.edu Cornea 22(3): 246–248, 2003. © 2003 Lippincott Williams & Wilkins, Inc., Philadelphia 246