Examining the performance of AI in identifying corneal infections

A new meta-analysis by researchers from the UK, US and Singapore has examined the performance of deep learning (DL) in identifying infectious keratitis.

The research, which was published in eClinicalMedicine, reviewed 35 studies related to DL for the diagnosis of infectious keratitis.

The studies examined through the review involved a total of 136,401 corneal images and more than 56,000 patients.

The researchers determined that DL may have good diagnostic accuracy for infectious keratitis – and comparable performance to ophthalmologists.

They emphasised that the results highlight the potential clinical value of DL as a medical aid in real-world settings.

Senior author of the study and consultant ophthalmologist, Dr Darren Ting, of the University of Birmingham, highlighted that the findings are particularly promising for regions where access to specialist care is limited.

“Our study shows that AI has the potential to provide fast, reliable diagnoses, which could revolutionise how we manage corneal infections globally,” he said.