Ai In The Diagnosis Of Corneal Disease

Artificial intelligence (AI) has been shown to be a promising tool in the diagnosis of ophthalmic and ocular disease, and to improve referral patterns and diagnoses in Ophthalmology. Our purpose is to review and evaluate the applications of artificial intelligence (AI) in the diagnosis of keratoconus, secondary ectasia, dry eye, corneal dystrophies and other corneal disorders.

Illinois Eye and Ear Infirmary, University of Illinois College of Medicine

We reviewed the literature of AI in ophthalmology spanning between 2000 to 2022. Google Scholar was used to interrogate the results of “Artificial Intelligence” and “[disease state]” ([disease state] including keratoconus, secondary ectasia, dry eye, and corneal dystrophies) as the search parameters. We broke down the time periods into 5-year intervals (with 2020 through 2022 in a separate 3-year interval).  We tabulated the percent increase for every interval and caluclated the annual growth rates. We also tabulated the advantages and limitations of AI in these conditions.  

The annual growth rate ranged from 30-150%.  For keratoconus, AI-related articles were 14, 39, 89, 313 & 981 for the period 2000-04, 2005-09, 2010-14, 2015-19 & 2020-22, respectively, showing an annual growth rate of 32%. For dry eye, the count was 32, 49, 106, 277 & 783 for the same periods, respectively (average annual growth rate: 82%). For Fuchs Dystrophy, the count was 1, 1, 4, 10,& 43 for the same periods, respectively (average annual growth rate: 31%). Challenges included creating standardized processes for high quality usable images, reduction of time, expense, and variations in reporting and conduct of studies. Additional information will be presented on the advantages & limitations of AI in the diagnosis of corneal disease. 

AI is a promising aid in the diagnosis of keratoconus and other eye conditions but does come with limitations. The utility of AI in corneal disease was shown to improve imaging techniques and quality through standardization as well as enhance patient care through increased diagnostic performance and disease predictions.