A new AI tool has sparked hope in the fight against premature blindness. Scientists from around the world have developed an advanced artificial intelligence system that utilises deep learning. This innovative technology has the capability to pinpoint infants who are susceptible to retinopathy, a condition that, if neglected, could lead to vision loss.
Read on to discover more about Retinopathy of Prematurity, and how the new ground-breaking AI tool may help to prevent premature blindness.
What is Retinopathy of Prematurity (ROP)?
ROP is a condition that can affect infants born prematurely. It is known to contribute towards the vision loss of approximately 50,000 children worldwide.
In a preterm infant the blood vessels in the retina have not yet fully developed. Babies are screened by ophthalmologists in the neonatal unit to check how these blood vessels are growing. For some infants, there is an abnormal growth of vessels and this can lead to retinal detachment and blindness if it is not treated in-time. Over time, there has been an evolution of our understanding of this condition and guidelines have been developed that identify which babies need screening, and at what point a baby needs treatment.
You can’t see the signs of retinopathy just by looking at a baby’s eyes. The only way to spot it is to screen infants individually using a special headlight, lens and a doctor trained in screening for this condition. However, thanks to the new AI tool, detecting the condition could soon become much easier.
New AI tool to detect retinopathy
A research team has designed an advanced deep learning AI model to detect retinopathy. A total of 7,414 eye images from 1,370 newborns treated at Homerton Hospital, London, were used to train the AI. The hospital caters to a diverse ethnic and socioeconomic population, which was beneficial to the study as the condition varies across different ethnic groups.
Once the AI tool had been trained, it was tested using an additional 200 images. Its performance was compared to that of experienced ophthalmologists. The tool was also tested on datasets from the US, Brazil, and Egypt. Results showed that the AI was as proficient as senior paediatric ophthalmologists in detecting potentially blinding retinopathy.
Even though the tool was primarily optimised for a UK demographic, its effectiveness on other continents is encouraging. Developed as a code-free deep learning platform, the tool can be fine-tuned for new settings even by those who don’t have any prior coding knowledge.
Options for treating ROP include laser surgery or injections into the eye. These innovative techniques work by targeting and eliminating the abnormal blood vessels, thereby preventing additional deterioration of vision.
Laser surgery is a sophisticated procedure. It requires the precision and expertise of a well-trained ophthalmologist and potentially saves a patient from vision loss.
Mr. Imran Jawaid is a highly qualified and experienced ophthalmologist with a proven track record in treating retinopathy using laser surgery. His vast experience, combined with his compassionate approach to patient care, ensures that you will receive the best possible treatment. Schedule an appointment with Mr Jawaid today.