Computer-Aided Discrimination of Glaucoma Patients from Healthy Subjects Using the RETeval Portable Device

Abstract

Glaucoma is a chronic, progressive eye disease affecting the optic nerve, which may cause visual damage and blindness. In this study, we present a machine-learning investigation to classify patients with glaucoma (case group) with respect to normal participants (control group). We examined 172 eyes at the Ophthalmology Clinic of the “Elpis” General Hospital of Athens between October 2022 and September 2023. In addition, we investigated the glaucoma classification in terms of the following: (a) eye selection and (b) gender. Our methodology was based on the features extracted via two diagnostic optical systems: (i) conventional optical coherence tomography (OCT) and (ii) a modern RETeval portable device. The machine-learning approach comprised three different classifiers: the Bayesian, the Probabilistic Neural Network (PNN), and Support Vectors Machines (SVMs). For all cases examined, classification accuracy was found to be significantly higher when using the RETeval device with respect to the OCT system, as follows: 14.7% for all participants, 13.4% and 29.3% for eye selection (right and left, respectively), and 25.6% and 22.6% for gender (male and female, respectively). The most efficient classifier was found to be the SVM compared to the PNN and Bayesian classifiers. In summary, all aforementioned comparisons demonstrate that the RETeval device has the advantage over the OCT system for the classification of glaucoma patients by using the machine-learning approach.

Artificial Intelligence in Glaucoma: A New Landscape of Diagnosis and Management

“…Telemedicine technologies are now capable of distinguishing between various eye diseases, with successful applications including the detection of glaucoma within diabetic populations [41]. AIpowered applications can analyze patient-reported data or measurements from take-home devices, such as a portable tonometer, VF assessment (Eyecatcher), or ERG measurements (RETeval), the best models of which were able to achieve an accuracy of 93% in classifying glaucomatous versus non-glaucomatous eyes [42][43][44]. These data can be used to monitor disease progression or treatment efficacy, reducing the need for frequent in-person visits.…”