Eye Hospitals partnered with DeepMind, one of the world’s leading AI services companies. Through the partnership, researchers hope to use one million anonymous retinal images to train artificial intelligence (AI) in the automated diagnosis of optical coherence tomography (OCT) images.
OCT images are complex and take a long time for doctors to evaluate, which affects how quickly patients can obtain a formal diagnosis and initiate treatment. However, the open-access algorithm can be used to automatically diagnose eye diseases such as one-day age-related macular degeneration and diabetic retinopathy, reduce patient waiting time and prevent blindness. Moorefields research team does not need to learn code because AI is produced through user-friendly deep learning software.
The technique is now proven to match the accuracy of expert ophthalmologists and optometrists and generate the right referral information. The diagnostic capabilities of AI are benchmarked against doctors’ decisions at Moorefields Eye Hospital, demonstrating its real-world application.
How does it work?
Say you have 1,000 photos of cats and 1,000 dogs, and you want to train AI. All you have to do is create a folder containing all the images and a spreadsheet that categorizes them, “Image One Cat, Image Two Dog” and so on. Upload it to the system, there is an application programming interface (API) that you can use.m That was a game-changer for me. People like me who are researching AI but not computer scientists can start doing it ourselves and democratize applications.
How is it applied in ophthalmology?
We only use it in research projects. I think it is still a few years away for use in patient care and a lot of extra work is needed. We wonder, for example, whether we can train the AI algorithm to look at a photo and see if a patient can qualify for a clinical trial.
We started by obtaining five publicly available medical image data sets. We got photos of skin like moans and bruises, we got adult and child chest x-rays, we had retinal scans from diabetic eye disease. We have trained an algorithm using Google’s open-source AI platform, AutoML, and for some datasets, we have good results with no state-of-the-art coding experience.
It is receiving a lot of attention worldwide from doctors in every medical specialty. I am essaying this kind of hunger among doctors. We have thousands of ideas, we don’t have the practical facilities to translate them into reality, and that probably opens up that possibility.
What is the most exciting project you can think of right now in this area?
One of the hottest topics not only in ophthalmology but also in health care and outside of health care is called the Generative Adversary Network (GAN). GANs are neural networks that can simulate any data set distribution. What it means is that they can effectively create synthetic images, sounds or videos. They can create photos of people who don’t exist, politicians can create some fake videos that say things they don’t say, deep fakes and the like. FaceApp, the Russian AI Face Editor, which makes you look old, made me look like my daddy, it made my kids think they were grand.
However, that technology actually has legitimate health care applications. If you are studying a rare disease and you do not have much data, it is possible to increase your data using it. You can train your neural network using synthetic images, and then you don’t have to worry about privacy issues and data protection.