An evaluation of image quality and accuracy of eye bank measurement of donor cornea endothelial cell density in the Specular Microscopy Ancillary Study
Jonathan H Lass,Robin L Gal,Katrina J Ruedy,Beth Ann Benetz,Roy W Beck,Keith H Baratz,Edward J Holland,Andrea Kalajian,Craig Kollman,Francis J Manning,Mark J Mannis,Kristen McCoy,Monty Montoya,Doyle Stulting,Dongyuan Xing,Cornea Donor Study Group
DOI: https://doi.org/10.1016/j.ophtha.2004.10.045
IF: 14.277
Ophthalmology
Abstract:Purpose: The Specular Microscopy Ancillary Study was designed to examine donor corneal endothelial specular image quality, compare the central endothelial cell density determined by eye banks with the endothelial cell density determined by a central specular microscopy reading center, and evaluate donor factors that may have an impact on specular image quality and endothelial cell density accuracy. Design: Nonrandomized comparative trial. Participants: Endothelial specular images of donor corneas assigned in the Cornea Donor Study. Methods: Certified readers assessed donor image quality (analyzable from fair to excellent vs. unanalyzable) and determined the central endothelial cell density. Independent adjudication was performed if there was a difference in the quality of grading or if the endothelial cell density varied by > or =5.0% between readers. Average reading center-determined endothelial cell density was compared with the endothelial cell density determined by each eye bank. Main outcome measures: Evaluation of image quality and accuracy of endothelial cell density. Results: Of 688 donor endothelial images submitted by 23 eye banks, 663 (96%) were analyzable (excellent, 40 [6%]; good, 302 [44%]; fair, 321 [47%]), and 25 (4%) were unanalyzable by reading center standards. In situ retrieval and greater epithelial exposure correlated with a higher image quality grading. The eye bank-determined endothelial cell density of 434 of the 663 (65%) analyzable images were within 10% of the endothelial cell density determined by the reading center, whereas 185 (28%) were more than 10% higher and 44 (7%) were more than 10% lower. Greater variation in endothelial cell density between the eye banks and the reading center was observed with shorter time of death to preservation, presence of an epithelial defect, folds in Descemet's membrane, lower image quality, and the use of fixed-frame or center method endothelial cell density analysis. Conclusions: Overall, donor endothelial specular image quality and accuracy of endothelial cell density determination were good. However, the data suggest that factors that may affect image quality and contribute to variation in interpretation of the endothelial cell density should be addressed, because the donor endothelial cell density is an important parameter for assessing long-term corneal graft survival.