AI System Development to Identify Neuro-Ophthalmological Conditions from Eye Photos in Emergency Departments

What this Trial Is About

Researchers are studying how artificial intelligence AI can improve diagnosis of neuro-ophthalmological and neurological conditions affecting the optic nerves and brain using fundus images. The study focuses on enhancing an AI algorithm to better identify a broader range of retinal and optic nerve problems, especially in emergency departments where quick and accurate diagnosis is critical. This observational research aims to support emergency physicians who may lack ophthalmology specialists by providing a reliable AI-based diagnostic tool. Participants in this study are patients presenting to the emergency departments of the Fondation Adolphe de Rothschild Hospital. The study will collect fundus photographs from these patients to build a database and test the AI algorithms ability to distinguish different eye and optic nerve diseases. The AIs performance will be evaluated using measures such as sensitivity, specificity, accuracy, and area under the curve AUC. There are no drug or device interventions this is an observational study. During the study, participants will provide fundus images and give consent to be included. Researchers will analyze these images to assess the AI algorithms diagnostic accuracy by comparing predictions to actual conditions. The primary outcome is the proportion of correct predictions by the AI at 30 days. Secondary outcomes include detailed sensitivity and specificity for each eye disease. The study is expected to continue until October 2027, with ongoing monitoring of AI performance and patient data collection.

Development of a Deep Learning System for Identification of Neuro-Ophthalmological Conditions on Color Fundus Photographs in Emergency Department