According to Moorfields Eye Hospital NHS Foundation Trust in the UK, over 285 million people across the global live with some sort of sight loss. If those patients’ eye conditions had been diagnosed and treated earlier, most of their sight could have been saved. With a new artificial intelligence (AI) tool recently developed, over 50 eye diseases can be detected readily in the near future.

Since 2016, DeepMind Health, a world-leading AI research company owned by Google, has been partnering with London's Moorfields Eye Hospital NHS Foundation Trust and the University College London Institute of Ophthalmology. They worked together on developing different machine learning systems with AI technology to solve avoidable sight loss problems.

In the study, the research team input about 15,000 historic and de-personalised eye scans into an AI database every day. Then, five machine-learning systems automatically learned to analyse the highly complex Optical Coherence Tomography (OCT) scans of patient retinas, and triage those with sight problems into four clinical categories: urgent, semi-urgent, routine and observation only. The visualised data provided by the AI systems helps clinicians interpret and explain the results more easily and accurately.

With a number of trials and errors, a human-like AI system was eventually developed whereby more than 50 different eye diseases can be detected with 94% accuracy and patients provided with correct referral decisions. The positive result of this pioneering research demonstrates the potential sight-saving impact of AI for patients and signifies an important breakthrough in the field of medical science.

In the next step, the researchers will carry out more clinical trials with this credible AI system and explore how this technology can be deployed to improve patient care in practice.