A team of researchers has harnessed the power of artificial intelligence (AI) to develop an eye scan capable of improving the diagnosis of inherited retinal diseases (IRDs). These diseases, which affect the retina and result from single-gene disorders, are notoriously challenging to identify due to their rarity and the multitude of candidate genes involved.
Dr. Nikolas Pontikos, a group leader at the University College London’s Institute of Ophthalmology and Moorfields Eye Hospital, and his team have created Eye2Gene, an AI system that can detect the genetic cause of IRDs from retinal scans. Dr. Pontikos emphasized the system’s superiority in accuracy compared to most human experts, stating, “Identifying the causative gene from a retinal scan is considered extremely challenging, even by experts. However, the AI is able to achieve this to a higher level of accuracy than most human experts.”
Typically, the identification of the gene associated with a retinal disease is guided by the patient’s phenotype, as defined using the Human Phenotype Ontology (HPO). To evaluate Eye2Gene’s performance, the team conducted a benchmark study using 130 IRD cases with known gene diagnoses, complete with whole exome/genome data, retinal scans, and detailed HPO descriptions. Comparing the HPO gene scores with Eye2Gene gene scores, they found that Eye2Gene provided a rank for the correct gene that was equal to or higher than the HPO-only score in over 70 percent of cases.
Looking ahead, Eye2Gene holds the potential to be seamlessly integrated into standard retinal examinations. Dr. Pontikos envisions the software being embedded directly into the retinal imaging device. However, further evaluations are necessary to determine the system’s performance across different IRD patient types, ethnicities, imaging devices, and settings. The team presented their findings at the annual conference of the European Society of Human Genetics, emphasizing the need for additional assessments and validation of Eye2Gene’s capabilities.
Post Your Comments