Eye2Gene published in Leading Artificial Intelligence Journal

Eye2Gene's paper Next‑generation phenotyping of inherited retinal diseases from multimodal imaging with Eye2Gene showcases the power of artificial intelligence in revolutionising genetic diagnosis for inherited retinal diseases (IRDs).

Rare eye conditions like inherited retinal diseases (IRDs) can be difficult to diagnose through genetic testing. These disorders are usually caused by changes in a single gene and are one of the main causes of blindness in children and working-age adults around the world. More and more of these conditions are now the focus of clinical trials, and some treatments have already been approved. But to access these therapies, a genetic diagnosis must be made early enough—something that still remains a major challenge.

Eye2Gene’s deep-learning algorithm has been trained on the largest dataset of IRD scans available, allowing it to achieve expert-level accuracy in predicting 36 of the most common IRD genes. This algorithm has been deployed online and externally validated on data provided by four different clinical centres. By making expert-level diagnosis more widely available — beyond just the few specialist centres where it is currently available globally - Eye2Gene can speed up the long and often difficult process of getting a genetic diagnosis.

A key factor behind Eye2Gene’s success has been their innovative use of federated learning—a privacy-preserving machine learning technique they worked on during their time on the CDI Impact Accelerator programme. The Eye2Gene team implemented and externally validated this framework across clinical centres in the UK, Germany, and Australia. Federated learning allowed the model to be trained on distributed datasets from different hospitals without sharing sensitive patient data, overcoming major privacy and data governance challenges.

The impact of Eye2Gene’s research is profound, highlighting how collaboration, cutting-edge AI, and secure data sharing can transform rare disease diagnosis and patient care on a global scale. A huge congratulations to the entire Eye2Gene team! We’re incredibly proud to support your impact-driven innovation that’s paving the way for faster, more accurate diagnoses and better outcomes for patients with inherited retinal diseases.