A study released on Tuesday reveals that artificial intelligence (AI) models are on par with eye care specialists in diagnosing infectious keratitis (IK), a major cause of corneal blindness globally.
The research highlights the potential of AI and deep learning in enhancing healthcare services.
Infectious keratitis, commonly referred to as corneal infection, leads to approximately 5 million cases of blindness worldwide and contributes to around 2 million cases of monocular blindness each year, particularly impacting individuals in low- and middle-income countries.
Conducted by researchers at the University of Birmingham in the UK, the study involved a meta-analysis of 35 studies that employed deep learning models for diagnosing infectious keratitis.
Published in eClinicalMedicine, the results indicated that the AI models achieved diagnostic accuracy comparable to that of ophthalmologists. While ophthalmologists demonstrated 82.2% sensitivity and 89.6% specificity, the AI models recorded 89.2% sensitivity and 93.2% specificity.
Dr. Darren Ting, a Consultant Ophthalmologist at the University of Birmingham, stated, “Our study shows that AI has the potential to provide fast, reliable diagnoses, which could revolutionize how we manage corneal infections globally.” He noted the AI models could be particularly beneficial in areas with limited access to specialized eye care, potentially reducing the incidence of preventable blindness.
The AI models were also effective in distinguishing between healthy eyes, infected corneas, and the various causes of IK, such as bacterial or fungal infections.
However, the researchers urged caution in interpreting the findings due to the image-based analysis, which did not account for potential correlations within individuals. They emphasized the need for more diverse data and further external validation to enhance the reliability of these models for clinical applications.