Researchers have developed a prototype smartphone application that allows parents to detect the first signs of eye disorders in children. The tool is particularly effective for detecting a retinoblastoma, a cancerous tumor of the retina, from a specific symptom: a white reflection of the pupil.
Smartphone App to detect diseases of the eye in children
Smartphone applications dedicated to health are flourishing on the market. Researchers at Baylor University came up with the idea of creating one to help parents detect the early signs of various eye diseases in their children, such as retinoblastoma, an aggressive eye cancer. This application named CRADLE (ComputeR Assisted Detector Leukocoia) works by detecting with software a clinical sign called leucocoria, which is a white reflection in the pupil. As explained by the National Union of Ophthalmologists of France, it is visible at the beginning only under certain lighting and in some directions of the glance then it becomes permanent.
He adds that this sign must imperatively make an examination of the fundus after dilation. The study, published in the journal Science Advances , found that the application was effective in increasing the clinical screening of leucocoria, allowing parents to screen their children more often and more effectively throughout their growth. The researchers determined its sensitivity and accuracy by analyzing more than 50,000 photographs of children taken before their diagnosis. For those with ocular disorders, CRADLE was able to detect leucocoria in 80% of children, in photos taken on average 1.3 years before the official diagnosis.
A tool to detect several pathologies of the eye
However, the effectiveness of a conventional examination is more limited with signs of retinoblastoma via detection of leukocoria in only 8% of cases. The sensitivity of CRADLE in children aged 2 years and under exceeded 80%, “threshold considered by ophthalmologists as the standard of reference for sensitivity for similar devices,” says the lead author of the study Professor Micheal Munson. The effectiveness of the application is explained by the fact that it works from pictures of children . The more these are numerous and taken in different places, the more possibilities exist for the light to be reflected on the ocular lesions, whatever their location in the eye.
As the application’s algorithm has become more sophisticated, its ability to detect even mild cases of leucocoria has improved. “We wanted to be able to detect all the nuances and intensities of leucocoria. As a parent of a child with retinoblastoma, I am particularly interested in detecting traces of leukocoria that appear as “gray” pupils that are difficult to detect with the naked eye, “adds the researcher. Initially, the application was to be used primarily to identify retinoblastoma, one of the most common cancerous tumors in children under 5 years of age.
The latter reaches the cells of the retina but is treated particularly well, if it is diagnosed in time. According to the Institut Curie, “the precocity of diagnosis determines the implementation of treatments that best preserve vision. The treatments are more and more effective and differ depending on the shape of the tumor, its location, its volume and the age of the child. But the researchers found that the application could be used to detect other eye diseases as myelin, a cataract and Coats’ disease (a serious disease of the small vessels of the retina). The next step will be to add features to reduce false positive detections.