AI breakthrough expected in nine months

London’s Moorfields Eye Hospital has announced its large project, in collaboration with the tech giant, which started nine months ago, is on track.Eventually, its general-purpose algorithm will be able to review optical coherence tomography (OCT) scans and diagnose conditions just like an eye specialist.{{quote-A:R-W:450-I:2-Q: The algorithm will eventually be able to detect eye conditions such as age-related macular degeneration and diabetes-related eye disorders. -WHO:Dr Pearse Keane, Consultant Ophthalmologist at Moorfields and Clinician Scientist at the University College London}}Dr Pearse Keane, a consultant ophthalmologist at Moorfields and clinician scientist at the University College London, said the aim of the project is to take advantage of AI in ophthalmology. Keane said AI is an ideal “test bed” in harnessing the technology for the healthcare practice.The project is aimed at reducing unnecessary referrals to the National Health Service (NHS) on OCT scans and allow clinicians to focus on urgent cases. Keane said Moorfields averaged more than 600,000 eye patients per year, with 3,000 OCT scans each week.The success of the OCT scan test in diagnosing eye probls has been responsible for a substantial influx of outpatients for the NHS, as any slight deviation from normal standards usually results in an urgent referral.The large increase in outpatients has led to the search for streamlined processes for effective service delivery by clinicians.The algorithm will eventually be able to detect eye conditions such as age-related macular degeneration and diabetes-related eye disorders. Much of the time spent in preparation for the program was the consolidation of all 1.2 million 3D OCT scans.Once the scans have been properly collated, they will then be wiped-off along with any patient-related profiles and information, before going through a final screening for anything they have overlooked.After the OCT scans have been cleared it will then be shared anonymously with DeepMind for the algorithm program to check all images.Keane explained that deep learning algorithms “will derive new clinical and scientific insight from the data” and start using “novel deep learning approaches to try and figure out information about the path of physiology, the natural history, and prognosis of these conditions.”