Document resource
Purpose To develop and characterize a novel Hybrid (HY) software-assisted method to quantitively determine corneal endothelial cell density (ECD) and morphological parameters CV% (coefficient of variation) and HEX% (Hexagonality index) using light microscopy (LM) images. To evaluate agreement of the HY method with the evaluation techniques routinely performed at the eye banks, i.e. LM and specular microscopy (SM).Methods A total of 283 images from endothelia of 31 corneas unsuitable for transplantation in hypothermic storage were obtained and analysed with standard LM and SM techniques. For HY method analysis, the same LM images were processed with a SM-dedicated software to semi-automatically determine ECD, CV% and HEX%. Agreement between the three methods and interoperator bias was statistically investigated with Bland-Altman analysis. The success rates of the LM and SM techniques—i.e., how often corneal endothelia could be evaluated using LM and SM—were assessed in 311 analyses performed on 70 corneas stored at 2–8°C between 2023 and 2024.Results HY showed agreement with SM in ECD determination, while LM statistically differed both from HY (mean difference: 134 cells/mm²) and SM (mean difference: 115 cells/mm²). HY showed agreement with SM in HEX% determination, while CV% was statistically different between the two methods (mean difference: 3.4%). Analysis of inter-operator variability showed significant differences in both ECD and endothelial cell morphology evaluations performed with LM. Differently, ECD and HEX% did not differ significantly between operators with HY method, whereas CV% displayed a statistically significant bias (mean difference: 3.1%). The differences in ECD, CV% and HEX%, while statistically significant, were not clinically relevant according to the criteria for tissue suitability for transplantation. The LM technique produced analysable images in 96.8% of evaluations, compared to 72.2% with the SM technique, with a statistically significant difference in success rates. The HY method, which relies on LM images for analysis, matched the LM success rate (96.8%) and significantly increased the frequency of successful evaluations. Additionally, it enabled the collection of quantitative morphological data.Conclusions The HY method utilizes images obtained from LM and results in simple and quick quantitative investigation of ECD as well as of endothelial morphological data CV% and HEX%, which had previously been limited only to SM. The HY method showed statistical agreement with SM in determining ECD and HEX%, while a statistically significant but clinically acceptable bias was identified in ECD determination between LM and both HY and SM, as well as in CV% between SM and HY. The HY method demonstrated lower inter-operator data variability than LM and higher evaluability rate than SM. In short, the presented novel HY method enables a simplified quantitative assessment of corneal endothelial parameters by integrating the benefits of both SM and LM while mitigating their respective drawbacks. The presented method could address data collection issues related to corneal quality in research and take the first step toward unifying endothelial evaluations in EBs that use different microscopy techniques.