Eye scans using AI can make diagnoses
A team of researchers has developed an artificial intelligence (AI) system that uses eye scans to improve the diagnoses of inherited retinal diseases (IRDs). These diseases, which are caused by single-gene disorders affecting the retina, are challenging to diagnose due to their rarity and the involvement of multiple candidate genes.
The team, led by Dr. Nikolas Pontikos from the University College London’s Institute of Ophthalmology and Moorfields Eye Hospital, created an AI system called Eye2Gene. This system can analyze retinal scans and identify the genetic cause of IRDs with higher accuracy than most human experts.
Typically, the identification of the gene responsible for a retinal disease relies on the patient’s phenotype, which is defined using the Human Phenotype Ontology (HPO). The researchers benchmarked Eye2Gene using 130 cases of IRDs with known gene diagnoses, along with whole exome/genome data, retinal scans, and detailed HPO descriptions. They compared the gene scores provided by Eye2Gene with those obtained solely through HPO analysis.
The results showed that Eye2Gene ranked the correct gene higher or equal to the HPO-only score in over 70 percent of cases. The researchers suggest that Eye2Gene could be seamlessly integrated into standard retinal examinations in the future.
Further evaluation of Eye2Gene is needed to assess its performance across different types of IRD patients, ethnicities, imaging devices, and settings. The findings were presented at the annual conference of the European Society of Human Genetics.
