Eye disease glaucoma is the leading cause of irreversible blindness worldwide. Early treatment can often prevent this, but it is difficult to diagnose. In the Netherlands, the correct diagnosis is made on time in only half of the cases, in many other countries this percentage is significantly lower.
35 percent better diagnosis
Three years ago, Hans Lemij (glaucoma specialist at Het Oogziekenhuis Rotterdam) and Koen Vermeer (former director and currently advisor at the Rotterdam Ophthalmic Institute) started a plan to use AI to detect glaucoma. That has now been achieved. The current version of the algorithm can diagnose glaucoma 35 percent better. That means non-specialists can do just as well as doctors who specialize in glaucoma. This can have a major impact worldwide.
Ethnicity doesn’t matter
The algorithm is fed with photos of more than a hundred thousand eyes. The scientists used a representative mix of ethnicities for this. As a result, the AI application works equally well for all patients, regardless of their origin. A similar algorithm previously created in China was found to work well only for Chinese eyes.
Photo at optician
The research team expects that the algorithm will greatly improve the early diagnosis and treatment of glaucoma. In the future, for example, an optician could take a picture of the eye, after which the computer would indicate within a few minutes whether the dreaded eye disease may be present.
<!–
–>
–
