The aims of this study were 1) to compare area of cell loss (ACL) on trypan staining with ACL on specular imaging and 2) to evaluate the use of automated software for measuring ACL on trypan staining.Methods: Donor corneas with transplant-grade endothelium were mechanically injured with an 18-gauge cannula and a Fogla deep anterior lamellar keratoplasty dissector tip to create an easily identifiable "bullseye" pattern of cell death. Each cornea was then stained with trypan blue 0.06% for 90 seconds and imaged at 2• magnification. ACL on staining was measured using manual (ImageJ, National Institute of Health, Bethesda, MD) versus automated software tools (custom-built Aphelion macro, ADCIS, S.A., Saint-Contest, France). The bullseye was then imaged using specular microscopy, and ACL was measured by tracing the dead cell borders. ACL was then compared between both modalities.Results: Eleven donor corneas were evaluated. Both manual (0.42 mm 2 ) and automated (0.45 mm 2 ) measurements of ACL after trypan staining underestimated mean ACL on specular imaging (0.54 mm 2 ) (P , 0.01). However, on regression analysis, there was a good predictive correlation between automated trypan measurements and specular imaging (R 2 = 0.99, residual SE = 0.0044, P , 0.01). When ACL on specular imaging was measured by tracing cell nuclei along the margin of injury (rather than cell borders) (0.45 mm 2 ), there was no statistically significant difference between specular and automated trypan measurements (P = 0.95).Conclusions: Trypan-assisted automated measurements of ACL correlated well with ACL on specular imaging, suggesting that automated software may be a useful tool for evaluating endothelium in donor corneas.

Purpose: To evaluate the ex vivo safety and efficacy of using paired peripheral incisions to achieve a triple scroll conformation that facilitates unscrolling in Descemet membrane endothelial keratoplasty (DMEK). Methods: The safety of adding paired peripheral incisions to DMEK grafts was evaluated by assessing endothelial cell loss (ECL) and risk of tearing. ECL was measured using calcein-AM staining after incisions. The risk of tearing was evaluated by comparing incision lengths before and after simulated DMEK surgery using cadaveric eyes. Efficacy was evaluated by comparing the scrolling pattern and the width of grafts with different incision lengths (0.0 mm, 0.5 mm, and 1.0 mm). Surgical unscrolling times in simulated DMEK surgery by a novice DMEK surgeon were evaluated to determine whether incisions facilitate unscrolling in DMEK surgery. Results: The mean ECL after adding incisions was 0.78% ± 0.23%. There was no significant change in incision length after simulated DMEK surgery (P = 0.6). In donor grafts aged less than or equal to 65 years, 60% (6/10) achieved a stable triple scroll with 0.5 mm incisions and 80% (8/10) achieved a stable triple scroll with 1.0 mm incisions. In donor grafts aged greater than 65 years, 0% (0/4) achieved a stable triple scroll. Mean graft width increased significantly after forming a triple scroll (5575 μm ± 1128 μm) compared with baseline (1563 μm ± 428 μm) (P < 0.001). In the hands of a novice DMEK surgeon, the mean unscrolling time was significantly shorter with incisions (2.61 min ± 1.41 min) versus without incisions (5.44 min ± 3.17 min) (P = 0.02). Conclusions: Paired peripheral incisions are safe and effective for inducing a triple scroll in DMEK grafts with donor age less than or equal to 65 years. Adding incisions may facilitate unscrolling for inexperienced DMEK surgeons.