Automatic Classification of Slit-Lamp Photographs by Imaging Illumination.

PURPOSE: The aim of this study was to facilitate deep learning systems in image annotations for diagnosing keratitis type by developing an automated algorithm to classify slit-lamp photographs (SLPs) based on illumination technique. METHODS: SLPs were collected from patients with corneal ulcer at Kellogg Eye Center, Bascom Palmer Eye Institute, and Aravind Eye Care Systems. Illumination techniques were slit beam, diffuse white light, diffuse blue light with fluorescein, and sclerotic scatter (ScS). Images were manually labeled for illumination and randomly split into training, validation, and testing data sets (70%:15%:15%). Classification algorithms including MobileNetV2, ResNet50, LeNet, AlexNet, multilayer perceptron, and k-nearest neighborhood were trained to distinguish 4 type of illumination techniques. The algorithm performances on the test data set were evaluated with 95% confidence intervals (CIs) for accuracy, F1 score, and area under the receiver operator characteristics curve (AUC-ROC), overall and by class (one-vs-rest). RESULTS: A total of 12,132 images from 409 patients were analyzed, including 41.8% (n = 5069) slit-beam photographs, 21.2% (2571) diffuse white light, 19.5% (2364) diffuse blue light, and 17.5% (2128) ScS. MobileNetV2 achieved the highest overall F1 score of 97.95% (CI, 97.94%-97.97%), AUC-ROC of 99.83% (99.72%-99.9%), and accuracy of 98.98% (98.97%-98.98%). The F1 scores for slit beam, diffuse white light, diffuse blue light, and ScS were 97.82% (97.80%-97.84%), 96.62% (96.58%-96.66%), 99.88% (99.87%-99.89%), and 97.59% (97.55%-97.62%), respectively. Slit beam and ScS were the 2 most frequently misclassified illumination. CONCLUSIONS: MobileNetV2 accurately labeled illumination of SLPs using a large data set of corneal images. Effective, automatic classification of SLPs is key to integrating deep learning systems for clinical decision support into practice workflows.

The effect of social determinants of health on severity of microbial keratitis presentation at a tertiary eye care hospital in Southern India.

Purpose: Understanding the association between social determinants of health (SDoHs) and microbial keratitis (MK) can inform underlying risk for patients and identify risk factors associated with worse disease, such as presenting visual acuity (VA) and time to initial presentation. Methods: This was a cross-sectional study was conducted with patients presenting with MK to the cornea clinic at a tertiary care hospital in Madurai, India. Patient demographics, SDoH survey responses, geographic pollution, and clinical features at presentation were collected. Descriptive statistics, univariate analysis, multi-variable linear regression models, and Poisson regression models were utilized. Results: There were 51 patients evaluated. The mean age was 51.2 years (SD = 13.3); 33.3% were female and 55% did not visit a vision center (VC) prior to presenting to the clinic. The median presenting logarithm of the minimum angle of resolution (logMAR) VA was 1.1 [Snellen 20/240, inter-quartile range (IQR) = 20/80 to 20/4000]. The median time to presentation was 7 days (IQR = 4.5 to 10). The average particulate matter 2.5 (PM2.5) concentration, a measure of air pollution, for the districts from which the patients traveled was 24.3 µg/m3 (SD = 1.6). Age- and sex-adjusted linear regression and Poisson regression results showed that higher levels of PM2.5 were associated with 0.28 worse presenting logMAR VA (Snellen 2.8 lines, P = 0.002). Patients who did not visit a VC had a 100% longer time to presentation compared to those who did (incidence rate ratio = 2.0, 95% confidence interval = 1.3-3.0, P = 0.001). Conclusion: Patient SDoH and environmental exposures can impact MK presentation. Understanding SDoH is important for public health and policy implications to mitigate eye health disparities in India.