Artificial Neural Network for Automated Keratoconus Detection Using a Combined Placido Disc and Anterior Segment Optical Coherence Tomography Topographer.

Purpose: To assess the efficacy of an automated program for keratoconus and keratoconus suspect detection based on corneal measurements provided by a combined Placido disc and anterior segment optical coherence tomography (OCT) topographer. Methods: In a multicentric cross-sectional study, an artificial neural network (ANN) was created using 6677 eyes from an equal number of patients (classified as 2663 normal eyes, 1616 keratoconus eyes, 210 keratoconus suspect eyes, 1519 myopic postoperative eyes, and 669 abnormal eyes). Each group was randomly divided into a training set (70% of the dataset) and a validation set (the remaining 30%). A multilayer perceptron network with a backpropagation learning algorithm was developed for the study. Indexes used to train the ANN were based on curvature and elevation of both the anterior and posterior corneal surfaces and the new corneal OCT indexes-based on corneal, stromal, and epithelial thicknesses. Results: For keratoconus detection, our ANN showed an accuracy of 98.6%, precision of 96%, recall of 97.9%, and F1-score of 96.9%. For keratoconus suspect detection, our ANN showed an accuracy of 98.5%, precision of 83.6%, recall of 69.7%, and F1-score of 76%. Conclusions: Compared to previous literature, the addition of new OCT-based epithelial and stromal thickness indexes improves ANN detection capacity of keratoconus suspect eyes. For already stablished keratoconus our ANN detection capacity is excellent, but equivalent to previous evidence without incorporating such new OCT-based indexes. Translational Relevance: OCT-based epithelial and stromal thickness indexes improve ANN detection capacity of keratoconus on its early stages.

Long-Term Effect of Anterior Chamber Depth on Endothelial Cell Density in Patients With Iris-Fixated Phakic Intraocular Lenses.

PURPOSE: To evaluate the long-term endothelial cell density (ECD) change and its correlation with preoperative anterior chamber depth (ACD) and aqueous depth (AQD) in patients with iris-fixated phakic intraocular lenses (pIOLs). METHODS: A total of 90 eyes from 57 patients who underwent pIOL implantation were retrospectively enrolled. Correlation between ACD and endothelial cell loss (ECL) was investigated. Optimal cut-off values for ACD and AQD were assessed. RESULTS: The average follow-up period was 11.8 ± 2.0 years (range: 9.1 to 17.3 years). Compared to the baseline data, the mean ECD change was -894 ± 732 cells/mm2 (range: -2,281 to 389 cells/mm2). The mean ECD change from baseline was -53.5% ± 19.1% (range: -75.1% to -5.2%) in eyes with preoperative ACD of 3.20 mm or greater -26.0% ± 26.6% (range: -74.1% to 0.9%) in eyes with ACD of 3.21 to 3.49 mm, and -5.2 ± 8.7% (range: -21.3% to 16.8%) in eyes with ACD of 3.50 mm or greater (P < .001). No eyes with ACD of 3.50 mm or greater had significant endothelial cell loss (SECL), whereas 84% of the eyes with ACD of 3.20 mm or less experienced SECL (P < .001). There was a significant negative correlation between ECL and ACD (r = -0.70, P < .001) and AQD (r = -0.65, P < .001). Receiver operating characteristic curve analysis revealed that ACD of 3.35 mm provides 84% sensitivity and 88% specificity and AQD of 2.75 mm provides 88% sensitivity and 81% specificity for preventing SECL. CONCLUSIONS: Smaller ACD and AQD are significantly correlated with more ECL. Minimum ACD of 3.35 mm or AQD of 2.75 mm are recommended for better long-term endothelial safety. [J Refract Surg. 2019;35(8):493-500.].