DeepMind’s artificial intelligence can now spot key signs of eye disease like world’s leading retina specialists, with an error rate of only 5.5%. This system analyze retinal scans and spot symptoms of more than 50 eye sight-threatening diseases. DeepMind AI also recommends proper treatment to patients regarding the eye diseases it have.
The Artificial intelligence is able to identify serious conditions such as wet age-related macular degeneration (AMD), which can lead to blindness unless treated quickly. It can quickly examine optical coherence tomography (OCT) scans and make diagnoses with the same accuracy as human clinicians.
A team at DeepMind, based in London, created this AI. DeepMind was founded in 2010 and soon it was acquired by Google in 2014.
To fully train and develop the AI, DeepMind made partnership with researchers at London’s Moorfields Eye Hospital NHS Foundation Trust and the University College London Institute of Ophthalmology.
This systems follows an algorithm and mathematical set of rules to analyse optical coherence tomography (OCT) -high resolution 3D scan of the back of the eye. From the 3D scans this new system identifies eye diseases. As of now, it can detect around 50 eye diseases and researchers claim that the accuracy rate of detection is 94%.
The results, published in the journal Nature Medicine.
“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,” Mustafa Suleyman, co-founder and head of applied AI at DeepMind, said.
“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.”
This AI first learned how to first identify the different anatomical elements of the eye (a process known as segmentation) and then recommend clinical action based on the various signs of diseases that the scans show.
The AI system works on the principle of a human-like logical approach for analyzing the highly complex OCT scans of a patient’s retina. Scientist at DeepMind has perform thousands of scans to train the AI “how to read the scans”. Scientist train the AI over 15,000 OCT scans from some 7,500 patients to spot 10 key features of eye disease.
It uses near-infrared light, which is bounced off of the interior surface of the eye. This is done to create a 3D image of the eye tissues, after the system will evaluate the 3D tissue map, it make a decision about what the diseases might be and how urgently they need treatment. The scan takes around 10 minutes.
DeepMind AI doesn’t give a single answer for each diagnosis, it gives several possible explanations about the diseases. It also shows how it has labeled the parts of the patient’s eye, giving doctors an opportunity to spot faulty analysis.
DeepMind’s system uses two separate networks. Both the networks has trained independently so that any freak error will be overruled by the majority. The first one is named as segmentation network which converts the raw OCT scan into a 3D tissue map with clearly-defined, color-coded slices. The second one is named as classification network which analyzes the 3D tissue map and makes decisions about what the diseases might be and how urgent they are for referral and treatment.
DeepMind co-founder Mustafa Suleyman says the firm and Moorfields’ are now planning on using this AI in clinical trials and will attempt to get a final product approved by regulators.
Have a great day !
Read also, you may like it :