Global Mapping of Genetic Eye Diseases using AI | Dr Nikolas Pontikos

Dr Nikolas Pontikos is an Associate Professor at University College London (UCL) and the founder of Eye2Gene, a pioneering project that leverages AI to diagnose genetic eye diseases from retinal images. His background spans computer science (UCL), bioinformatics (Imperial College), and a PhD in type one diabetes and genetics (University of Cambridge), with an early passion for using technology to improve lives. He transitioned into ophthalmology nearly a decade ago, focusing on the genetic analysis of patients with eye conditions.

The Eye2Gene project, active since 2019, has involved training an AI system to predict the likely genetic cause of a patient's disease directly from their retinal scans, making it a widely accessible, point-of-care tool. While not a replacement for gold-standard genetic testing, Eye2Gene addresses a critical barrier to diagnosis, particularly in regions lacking genetic testing infrastructure.

By "lighting up the map" with identified patients, one of it's aims is to stimulate pharmaceutical companies to develop treatments for rare diseases by demonstrating patient populations. The model has shown an accuracy of 83.9% in published research, significantly outperforming retinal specialists. An interesting discovery has been the AI's ability to identify similar retinal patterns even from genes not used in its training, highlighting its potential for novel scientific insights beyond mere efficiency. Currently, Eye2Gene is deployed as a research tool, with ongoing efforts towards medical device regulation to enable broader clinical use.Dr Pontikos advocates for investing in the "infrastructure, the training, the people behind" AI in healthcare, rather than solely the "shiny car" of new technology. He is deeply mindful of the ethical implications of AI, including data bias and the challenges of "black box predictions," emphasising the need for responsible development and deployment.