Deep learning model to identify homonymous defects on automated perimetry.

BACKGROUND: Homonymous visual field (VF) defects are usually an indicator of serious intracranial pathology but may be subtle and difficult to detect. Artificial intelligence (AI) models could play a key role in simplifying the detection of these defects. This study aimed to develop an automated deep learning AI model to accurately identify homonymous VF defects from automated perimetry. METHODS: VFs performed on Humphrey field analyser (24-2 algorithm) were collected and run through an in-house optical character recognition program that extracted mean deviation data and prepared it for use in the proposed AI model. The deep learning AI model, Deep Homonymous Classifier, was developed using PyTorch framework and used convolutional neural networks to extract spatial features for binary classification. Total collected dataset underwent 7-fold cross validation for model training and evaluation. To address dataset class imbalance, data augmentation techniques and state-of-the-art loss function that uses complement cross entropy were used to train and enhance the proposed AI model. RESULTS: The proposed model was evaluated using 7-fold cross validation and achieved an average accuracy of 87% for detecting homonymous VF defects in previously unseen VFs. Recall, which is a critical value for this model as reducing false negatives is a priority in disease detection, was found to be on average 92%. The calculated F2 score for the proposed model was 0.89 with a Cohen's kappa value of 0.70. CONCLUSION: This newly developed deep learning model achieved an overall average accuracy of 87%, making it highly effective in identifying homonymous VF defects on automated perimetry.

Effect of Collagenase A on Descemet Membrane Endothelial Keratoplasty Scroll Tightness.

PURPOSE: The scrolling properties of the Descemet membrane endothelial keratoplasty (DMEK) graft are essential for surgical success. Currently, there is limited knowledge on what dictates the tightness of the DMEK scroll. The purpose of this study was to determine the impact of temperature and protein digestion on DMEK graft scroll tightness. METHODS: For the temperature experiment, a total of 28 eyes were used for this study. Scrolls in the cold group were kept at 4°C while scrolls in the hot group were kept at 37°C. Scroll width was recorded at the 5-, 15-, and 30-minute mark. For the protein digestion experiment, a total of 18 eyes were exposed to collagenase A (10 CDU/mL) in Optisol solution. Scroll width was recorded at the time points of 1, 3, 5, 10, and 20 minutes. RESULTS: The results of the temperature experiment did not yield any statistically significant changes in the mean scroll width of the DMEK scrolls across both temperature ranges and observation times. For the protein digestion experiment, the mean scroll width grew from 1.85 mm to 2.13 mm from the beginning of the experiment until the final observation at 20 minutes. This is a 14.7% change over 20 minutes with a P value (<0.001), exemplifying a statistically significant change in scroll width. CONCLUSIONS: Temperature did not have any significant effect over scroll tightness, but scroll tightness decreased with collagenase exposure.