Using AI to detect paediatric eye conditions from smartphone images

New research published in JAMA Network Open has examined the potential of an artificial intelligence (AI) model in identifying paediatric eye conditions from smartphone images.

The model was built using images from 476 children with myopia, strabismus and ptosis gathered at Shanghai Ninth People’s Hospital.

The study participants were photographed without spectacles using a smartphone at a distance of around 4cm from the patient’s face.

All participants were assessed by an ophthalmologist and received a diagnosis of myopia, strabismus or ptosis.

The AI model had an accuracy of 80% in detecting myopia and strabismus from smartphone images. Its accuracy in detecting ptosis was 92%.

The researchers highlighted that the technology could assist the equitable allocation of limited medical resources.

“These results suggest that [the AI model] can assist families in screening children for myopia, strabismus, and ptosis, facilitating early identification and reducing the risk of visual function loss and severe problems due to delayed screening,” the authors shared.