Eye care professionals may benefit from the integration of artificial intelligence in diagnosing infectious keratitis (IK), a major contributor to corneal blindness across the globe. A recent study has found that deep learning models are capable of matching the accuracy of human diagnosticians in identifying this infection.
Eye care professionals may benefit from the integration of artificial intelligence in diagnosing infectious keratitis (IK), a significant contributor to corneal blindness worldwide. A recent study indicates that deep learning models are comparably accurate in identifying this infection.
A meta-analysis featured in eClinicalMedicine was led by Dr. Darren Ting from the University of Birmingham. He collaborated with a global team to review 35 studies that employed Deep Learning (DL) models for diagnosing infectious keratitis.
The study revealed that the AI models achieved diagnostic accuracy levels similar to those of ophthalmologists, showing a sensitivity of 89.2% and a specificity of 93.2%, while ophthalmologists displayed 82.2% sensitivity and 89.6% specificity.
With over 136,000 corneal images analyzed in total, the authors assert that these findings reinforce the potential application of artificial intelligence in clinical environments.
According to Dr. Darren Ting, the senior author of the study, a Fellow at Birmingham Health Partners (BHP), and a Consultant Ophthalmologist at the University of Birmingham:
“Our research highlights that AI could offer rapid and trustworthy diagnoses, which could transform the management of corneal infections worldwide. This is especially encouraging for areas with limited access to specialized eye care, potentially alleviating the incidence of preventable blindness globally.”
The AI models also demonstrated proficiency in distinguishing between healthy eyes, infected corneas, and various underlying causes of IK, including bacterial and fungal infections.
Although these findings underscore the promise of deep learning in healthcare, the study’s authors stressed the importance of gathering more varied data and conducting additional external validation to enhance the dependability of these models for clinical implementation.
Infectious keratitis, an inflammation of the cornea, impacts millions of individuals, particularly in low- and middle-income nations where access to specialized eye care is often restricted. As AI technology advances and becomes increasingly integral in medicine, it may soon emerge as a vital tool in the fight against corneal blindness around the world.
