1. Differentiation of Active Corneal Infections from Healed Scars Using Deep Learning
- Author
-
Tiwari, Mo, Piech, Chris, Baitemirova, Medina, Prajna, Namperumalsamy V, Srinivasan, Muthiah, Lalitha, Prajna, Villegas, Natacha, Balachandar, Niranjan, Chua, Janice T, Redd, Travis, Lietman, Thomas M, Thrun, Sebastian, and Lin, Charles C
- Subjects
Biomedical and Clinical Sciences ,Ophthalmology and Optometry ,Eye Disease and Disorders of Vision ,Clinical Research ,Eye ,Good Health and Well Being ,Algorithms ,Area Under Curve ,Cicatrix ,Corneal Ulcer ,Deep Learning ,Eye Infections ,Bacterial ,Eye Infections ,Fungal ,False Positive Reactions ,Humans ,Photography ,Predictive Value of Tests ,ROC Curve ,Retrospective Studies ,Sensitivity and Specificity ,Slit Lamp Microscopy ,Wound Healing ,Artificial intelligence ,Corneal scar ,Corneal ulcer ,Deep learning ,Infectious keratitis ,Clinical Sciences ,Opthalmology and Optometry ,Public Health and Health Services ,Ophthalmology & Optometry ,Ophthalmology and optometry - Abstract
PurposeTo develop and evaluate an automated, portable algorithm to differentiate active corneal ulcers from healed scars using only external photographs.DesignA convolutional neural network was trained and tested using photographs of corneal ulcers and scars.ParticipantsDe-identified photographs of corneal ulcers were obtained from the Steroids for Corneal Ulcers Trial (SCUT), Mycotic Ulcer Treatment Trial (MUTT), and Byers Eye Institute at Stanford University.MethodsPhotographs of corneal ulcers (n = 1313) and scars (n = 1132) from the SCUT and MUTT were used to train a convolutional neural network (CNN). The CNN was tested on 2 different patient populations from eye clinics in India (n = 200) and the Byers Eye Institute at Stanford University (n = 101). Accuracy was evaluated against gold standard clinical classifications. Feature importances for the trained model were visualized using gradient-weighted class activation mapping.Main outcome measuresAccuracy of the CNN was assessed via F1 score. The area under the receiver operating characteristic (ROC) curve (AUC) was used to measure the precision-recall trade-off.ResultsThe CNN correctly classified 115 of 123 active ulcers and 65 of 77 scars in patients with corneal ulcer from India (F1 score, 92.0% [95% confidence interval (CI), 88.2%-95.8%]; sensitivity, 93.5% [95% CI, 89.1%-97.9%]; specificity, 84.42% [95% CI, 79.42%-89.42%]; ROC: AUC, 0.9731). The CNN correctly classified 43 of 55 active ulcers and 42 of 46 scars in patients with corneal ulcers from Northern California (F1 score, 84.3% [95% CI, 77.2%-91.4%]; sensitivity, 78.2% [95% CI, 67.3%-89.1%]; specificity, 91.3% [95% CI, 85.8%-96.8%]; ROC: AUC, 0.9474). The CNN visualizations correlated with clinically relevant features such as corneal infiltrate, hypopyon, and conjunctival injection.ConclusionsThe CNN classified corneal ulcers and scars with high accuracy and generalized to patient populations outside of its training data. The CNN focused on clinically relevant features when it made a diagnosis. The CNN demonstrated potential as an inexpensive diagnostic approach that may aid triage in communities with limited access to eye care.
- Published
- 2022