DOI 10.1097/ico.0000000000004223
Automated Tissue Segmentation Software for Assessing Global Area of Cell Damage After Descemet Membrane Endothelial Keratoplasty.
Аннотация
PURPOSE: The purpose of this study was to 1) determine the feasibility of imaging donor corneas during sterile tissue processing for computer vision endothelial cell viability analysis and to 2) compare the accuracy of automated software versus manual segmentation versus operator estimates in measuring percent area of cell damage (%ACD) on trypan-stained, postprocessed Descemet membrane endothelial keratoplasty (DMEK) corneas.
METHODS
Donor corneas were peeled for DMEK and stained with 0.06% trypan blue for 90 seconds and then photographed with a high-resolution microscope camera under sterile conditions in a biosafety cabinet before preloading into an injector or placing in a cornea viewing chamber. The corneas were evaluated for global cell viability based on %ACD using 3 methods: 1) manual Weka segmentation software (ImageJ, National Institute of Health, Bethesda, MD), 2) Kerify (custom-built Aphelion application, ADCIS, S.A., Saint-Contest, France), and 3) technician estimates. The %ACD measured was compared across all 3 modalities using 1-way ANOVA and regression analysis.
RESULTS
Standardized whole graft viability imaging was performed for 50 consecutive transplant DMEK grafts. There was no significant difference in group means between Kerify, manual software, and technician estimates in measuring cell damage (5.8% vs. 5.4% vs. 6.1%, P = 0.24). However, at higher ACD levels, Kerify-using manual Weka segmentation software as the reference standard-was more precise than technician estimates at predicting ACD.
CONCLUSIONS
Our automated ACD computer vision software Kerify saves time, removes bias, and demonstrates good accuracy compared with manual computer segmentation. Kerify may be a powerful new tool for postprocessing tissue evaluation.
Подготовка и перевод публикации
Обработка статьи · этапы, ошибки и управление
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Тип публикации указан по данным источника. Уровень доказательности автоматически не присваивается.