Researchers have succeeded in using artificial intelligence to predict patients who need treatment to confirm their cornea and maintain their eyesight.
The study was conducted by researchers from the Murfields Eye Hospital of the National Health Services Corporation in London, and the University College of London in the United Kingdom, and was presented on September 14th during the forty -third conference of the European Association for the Ain Battle Surgeons and vision correction, and the Yurrick Alert website was written about it.
The research focused on people with conical cornea, a visual weakness that usually develops in adolescents and youth, and tends to exacerbate with puberty.
This disease affects up to one person in every 350 people, and in some cases the condition can be treated with contact lenses, but in other cases the condition deteriorates quickly, and patients may need a corneal transplant if they are not treated, and the only way to determine who needs treatment is to monitor patients over time.
“In cone -corneal patients, the cornea – the front window of the eye – highlights abroad, and one treatment called the cross -link can stop the development of the disease, and when it is conducted before the emergence of permanent scars, the cross -binding often prevents the need for corneal cultivation.”
Ultraviolet UV and vitamin B2 (riboflavin) drops is used to strengthen the cornea, and it is successful in more than 95% of cases.
“Doctors are not currently able to predict patients whose condition will develop and need treatment, and who will be stable with monitoring only. This means that patients need frequent monitoring over many years, and the cross -link is usually conducted after the development of the disease occurs.”
The study included a group of patients who were referred to the Morvies Eye Hospital of the National Health Services Corporation for conical corneal evaluation, including wiping the front part of the eye using consensual optical tomography to examine its shape.
The researchers used artificial intelligence to assess patient eyes, as well as other data, and to successfully predict patients who need immediate treatment and who can continue to monitor.
Using artificial intelligence researchers enabled the classification of two -thirds of patients to a low -risk group, that is, those who do not need treatment, and the other third to a high -risk group, that is, those who need a rapid treatment with cross -binding. When including information from a second visit to the hospital, the algorithm was able to successfully classify up to 90% of patients.