Artificial intelligence is like a baby that learns something new every day. It is a computer program based on the workings of the human brain that you can teach to interpret data. With its quick development, scientists uncover more of its potential to recognize and treat serious health conditions. In a recent study of how a computer can speed screening and treatment of eye diseases, computers are proving to be an indispensable tool when it comes to recognizing diabetic retinopathy and macular degeneration.
Teaching To Computer To Be A Specialist
The research published in February 2018 in the journal Cell reports how the technology can be applied to individuals with retinal problems. Dr. Kang Zhang, professor of ophthalmology at Shiley Eye Institute at the University of California, led the study that helped demonstrate that a computer can learn to accurately and reliably identify common eye disorders. He explains that this technology is about using a computer, which does not get tired and can work unsocial hours, to interpret X-ray pictures in the lab.
The goal is for the computer to be as reliable as a specialist who studied medicine and is highly trained in diagnostics and treatment. A specialist goes through decades of practical experience to gain the highest levels of expertise. With AI, however, this ability can be learned within just a few days. The paper follows other recent studies that demonstrate deep-learning programs may have a legitimate place in healthcare.
Dr. Rahul Khurana, a clinical spokesperson for the American Academy of Ophthalmology, says that this tool is tremendously accurate in diagnosing certain types of diseases. This amazing feat is creating some excitement in the field.
Diagnosing Diabetic Retinopathy, Macular Degeneration
In the study, Zhang and his colleagues in Texas, China, and Germany first encoded images of diabetic retinopathy and macular degeneration into the computer. The photos were taken using a technique known as optical coherence tomography. This advanced screening technology uses light waves to take high-resolution, cross-section pictures of the eye. The scans allow doctors to map and measure the retina in detail.
The images are used to help spot prevalent eye conditions like diabetic retinopathy, a complication of diabetes that causes the blood vessels in the retina to swell and leak fluid, and macular degeneration where the macula, a part of the retina, deteriorates. Both are dangerous disorders that can lead to blindness if they are not diagnosed and treated immediately.
Typically, a computer needs millions of images to be accurate at detecting a condition. In Zhang’s research, they utilized an AI-based convolutional neural network that required a much smaller sample of around 200,000 scans.
With a variety of pictures to learn and memorize, the computer becomes more familiar with the map of the eye. It soon recognizes that if a spot is at a certain place, the eye might have macular degeneration. What makes this approach different is that instead of having the computer learn by itself, scientists can declare what it should be looking for. Ultimately, it is about designing a program that makes computers think like a trained specialist.
Within 30 seconds of scanning, Zhang’s software was able to generate a decision on whether a patient should be referred for treatment with 90 percent accuracy. This shows the potential of neural networks in assisting physicians and perhaps even outpace them since they can memorize so much data. Such a program will have uses across the globe, Zhang predicts. With highly industrialized countries like the United States, computers can speed the critical time between signs of disease and treatment.
Referrals and doctor appointments can take several months, and that’s a long time for patients who need treatment as soon as possible. When accurately diagnosed using the AI, a person with possible macular degeneration may be treated within a month.
AI-Assistance Where Specialists Are Few
In resource-poor areas, having computers to assist specialists will help increase the number of patients that can get care. The research team has visited Haiti to assess the utility of the application. The region has a large population of people with diabetes who have a high chance of developing retinopathy, but the ophthalmologists are fewer than 60.
If successful, more patients can gain access to the healthcare system before their symptoms become irreparable. It is estimated that 415,000 people living with diabetes are at risk for diabetic retinopathy. It’s a win-win for both doctors and patients whenever a new and improved technology would allow for faster and better diagnosis. This will make healthcare more accessible to a broader population.
The Challenge Of Trusting Computers
One of the main challenges of implementing an AI-based network in healthcare is getting doctors to trust the computers. In the paper, Zhang and his team also taught the computer to explain its diagnosis, determining the regions of the eye that were considered and were the basis for the system’s conclusion.
It’s important to note that the machine doesn’t just come up with a diagnosis. It elaborates on the diagnosis and recommendation it suggests. This transparency helps physicians to trust the computer more. That way, they will have an idea of why it came with a particular conclusion.
Other Applications For Ai Technology
The potential of AI-based networks in healthcare is staggering, especially in screening for diseases. The research team also showed that the system can distinguish between bacterial and viral pneumonia in children by analyzing X-rays. While bacterial pneumonia requires prompt antibiotic medication to stop serious complications, treatment may not be necessary for viral pneumonia.
There is a variety of medical fields that can greatly benefit from artificial intelligence, and the industry is seeing more applications for this tool. Artificial intelligence, without a doubt, will be managing many jobs in society in the future.
Step by step, people are going to accept that it’s something to help facilitate our lives, not least within areas such as research and medicine. Khurana believes the excitement around the technology is gaining momentum, and we can only look forward to the wonders it can bring in medical care.
