I attended a very interesting early morning session involving the “Artificial Intelligence Group” at the 18th Euretina Congress in Vienna on Thursday the 20thSeptember 2018.
The title of the symposium was “Innovations in Ophthalmology Spotlight on Artificial intelligence Applications” it was chaired by Tariq Aslam a Consultant from Manchester Royal Eye Hospital.
It was a review of recent advances in AI in Ophthalmology imaging and it highlighted the amazing potential of these machines to efficiently analyse images and provide accurate diagnosis with predictions on disease progression.
An AI system developed by researchers at Moorfields Eye Hospital NHS foundation trust, DeepMind health and the University College London Institute of Ophthalmology can recommend the correct referral decision for a wide range of eye disease such as age related macular degeneration and diabetic eye disease. It is currently going through robust trials and needs to pass regulatory standards. This technology may be widespread available in as little as 5 year’s time.
There is also a potential for this technology to run virtual clinical trials. Already AI technology in Manchester has been able to predict the outcome of an existing trial. This opens the opportunity to provide individualised treatments as the technology can predict which treatment and regime can provide the most optimal response based on a patients starting OCT image.
The technology has an ability to self-learn and after been given over one million OCT images it has been able to find associations which human observers have not been able to do. Amazingly the AI technology from Deepmind has the ability predict the age, sex and a relatively accurate blood pressure estimate with only the OCT image. The researchers do not know what features on the image that the technology is using to make these predictions.
There are always concerns regarding new technology but AI is here to stay and the benefits are likely to outweigh the risks. There are opportunities to address some of the worlds inequalities in health care provision by automated AI mass screening programmes that only require a patient to access a device that takes a photo. At risk patients can be quickly identified without the need for vast expensive networks of human screeners.
However I am not going to be out of a job anyway soon. These systems will help make retinal specialists better able to provide the best treatments to our patients.