Google DeepMind claims its AI can analyse eye scans in seconds for signs of 50 diseases

'Trained' DeepMind AI can accurately diagnose 50 eye conditions from OCT scans in seconds

Google's London-based DeepMind artificial intelligence unit claims to have built an artificial intelligence system that can analyse eye scans in seconds and diagnose signs of more than 50 different diseases and conditions.

The system was developed in a partnership with Moorfields Eye Hospital, after consultant ophthalmologist Pearse Keane contacted DeepMind to explore how the two organisations could work together on two conditions that cause sight loss: diabetic retinopathy and age-related macular degeneration (AMD).

Together, these affect more than 625,000 people in the UK and more than 100 million people worldwide.

Eyecare professionals today use optical coherence tomography (OCT) scans to help diagnose patients. These produce 3D images that are difficult to interpret without training. Because they take so long to analyse manually, there can be significant delays between the scan, diagnosis and treatment.

The results of the research have been published in Nature Medicine.

DeepMind claims that it can detect ‘the features of eye diseases' in seconds, as well as highlighting patients who may need to be prioritised for urgent treatment.

The challenge of using such a system in healthcare is that existing artificial intelligence techniques give no insight into their decision-making: the data goes in at one end and the answer comes out of the other.

This is referred as the AI ‘black box' and is a significant block to the wider clinical use of such systems.

DeepMind claims that it has combined two different neural networks to help solve this problem.

The segmentation network analyses the optical coherence tomography (OCT) scan to create a map of the eye and any damage, which professionals can use to see what the system is ‘thinking'.

Meanwhile, the classification network analyses the map to present diagnoses and referral recommendations.

Importantly, the system delivers its recommendations as a percentage. Clinicians can use this to judge the AI's confidence in its plans.

There are still strict clinical trials and regulatory approval for the system to pass before it can be used in practice. If that does happen, it could be deployed worldwide on many different types of eye scanner; not only the one that it was trained on at Moorfields.

"We set up DeepMind Health because we believe artificial intelligence can help solve some of society's biggest health challenges, like avoidable sight loss, which affects millions of people across the globe," said Mustafa Suleyman, co-founder and head of Applied AI at DeepMind Health.

He continued: "These incredibly exciting results take us one step closer to that goal and could, in time, transform the diagnosis, treatment and management of patients with sight-threatening eye conditions, not just at Moorfields, but around the world."