A new smartphone app, CRADLE (ComputeR Assisted Detector LEukocoia), can detect early signs of eye disorders in children - including an aggressive eye cancer called retinoblastoma.
The app was developed and studied by researchers at Baylor University in Texas. Their study found that the Cradle app can help in clinical screenings for leukocoria - an abnormal white reflection from the retina, which is a primary symptom of retinoblastoma.
As well as looking at photos of the children themselves, the app looks at family photos and checks for any traces of abnormal reflections that might indicate leukocoria. As well as screening for retinoblastoma, it can also help parents check for other common eye diseases.
To evaluate the accuracy, sensitivity and specificity of Cradle, the Baylor University researchers showed the app more than 50,000 photographs of children taken before diagnosis.
Conventional screening approaches only detect leukocoria in 8% of cases. But the app correctly spotted it in 80% of the children diagnosed with eye diseases - in children aged two and under, it was even more sensitive than 80%.
More impressively still, it was able to detect leukocoria in photos taken on average 1.3 years before a human doctor managed to diagnose them.
In study findings published in the Science Advances journal, the research team said that Cradle’s algorithm has improved in its ability to identify even very mild cases of leukocoria.
Bryan Shaw, associate professor of chemistry and biochemistry at Baylor University said: “We suspected that the app would detect leukocoria associated with other disorders. We were right.”
As well as retinoblastoma, doctors and parents have used Cradle to detect cataracts, Coats’ disease, refractive error, and myelin retinal nerve fibre layer.
The next step is to reduce false positive results. The research team are now showing Cradle another nearly 100,000 photos to identify features that could help it eliminate false positives..