Last updated March 1, 2018 at 11:13 am
Researchers in the US have built an artificial intelligence tool that can diagnose and suggest treatment for eye diseases with 95% accuracy.
In the quest for faster, more accurate medical diagnoses, medical researchers and doctors are increasingly looking at artificial intelligence. The latest field to set their eyes on AI is ophthalmology, with the announcement of a new artificial intelligence-powered diagnosis tool for retinal diseases.
The new program was not only able to diagnose macular degeneration and diabetic macular oedema, two leading causes of blindness, but also rate their severity and indicate whether they needed referral to a specialist.
“Macular degeneration and diabetic macular oedema are the two most common causes of irreversible blindness but are both very treatable if they are caught early,” says senior author Kang Zhang from University of California San Diego.
“Artificial intelligence (AI) has huge potential to revolutionize disease diagnosis and management by doing analyses and classifications involving immense amounts of data that are difficult for human experts — and doing them rapidly,”
Diagnosing disease
In the study, the researchers compared the diagnoses from the computer with those from five ophthalmologists who reviewed the same scans.
“With simple training, the machine could perform to the level of a well-trained ophthalmologist. It could generate a decision on whether or not the patient should be referred for treatment within 30 seconds and with more than 95% accuracy,” Zhang said.
In the US, where the study took place, diagnosing and treating retinal diseases normally involves visiting a general medical doctor or an optometrist, then a general ophthalmologist, and finally a retina specialist. This long process requiring multiple referrals can waste time and resources and delay effective treatment.
“Having an automated diagnosis could enable patients who would benefit from treatment to see a specialist and get that treatment much sooner and change outcomes,” added Zhang.
Using AI-based diagnoses tools will also increase the access to health care for people outside urban areas.
“Deciding how and when to treat patients has historically been handled by a small community of specialists who require years of training and are concentrated mostly in urban areas. In contrast, our AI tool can be used anywhere in the world, especially in the rural areas” said Zhang.
“This is important in places like China, India, and Africa, where there are relatively fewer medical resources.”
Thousands of optical coherence tomographs (OCT) form the iris and pupil of a human eye. OCTs are a non-invasive scan commonly used in clinics to diagnose diabetic retinopathy and macular degeneration. Credit: Daniel Kermany, Guangzhou Medical University and Kang Zhang, UC San Diego Health
Giving AI its intelligence
The platform looked at more than 200,000 optical coherence tomography (OCT) images, collected by optometrists, which use light to image the layers of the retina.
Similar to humans, AI learns through experience – seeing multiple images of healthy and diseased eyes and working out different characteristics which make the retina diseased or not.
Other AI have used machine learning to study retinal images, but the authors of the new study say their platform goes a step further by using a technique called transfer learning. This is a type of machine learning where knowledge gained solving one problem is applied to a different disease area. This allows the AI system to learn effectively with a much smaller dataset than traditional methods, and closer mimics human learning.
In addition to making a medical diagnosis, this AI platform also can make referral and treatment recommendations, which is another step that goes beyond previous studies.
The researchers also used occlusion testing, during which the AI showed which areas of the image it was considering important when reaching its diagnosis.
“Machine learning is often like a black box, where we don’t know exactly what is happening,” Zhang explained. “With occlusion testing, the computer can tell us where it is looking in an image to arrive at a diagnosis, so we can figure out why the system got the result it did. This makes the system more transparent and increases our trust in the diagnosis.”
Useful for more than just eye diseases
The scientists also used their tool to look wider than just eye diseases.
The program was applied to diagnosing childhood pneumonia, a leading cause of death in children under the age of 5. For this, they used machine analyses of chest X-rays.
The computer was able to determine the difference between viral and bacterial pneumonia with greater than 90% accuracy.
Viral pneumonia is treated mainly by controlling the symptoms while the body naturally rids itself of the virus. Bacterial pneumonia, on the other hand, is a more serious health threat and requires immediate treatment with antibiotics. By being able to differentiate between the two types, the AI could assist doctors in ensuring the correct and most effect treatment was given.
The researchers have suggested the AI could be similarly effective in diagnosing other diseases and assisting in correct treatment, such as telling the difference between malignant or benign cancers.
To do this, they have open-source released their data and program, hoping other groups will be able to improve, refine and develop its potential.
With more data being input to improve the AI, improvements in computational power, and more experience of researchers and doctors using AI tools such as this, artificial intelligence has the potential to revolutionise medical diagnosis in the future, leading to better care for everyone.
The study has been published in Cell
Video courtesy of Kang Zhang et al, and Cell
