AI Predicts Causative Genes of Hereditary Retinal Diseases: "More Accurate Than Specialists in Clinical Trials"
Predicting Candidate Genes Using Only Fundus Photographs and OCT
Clinical Trials Demonstrate Effectiveness
On August 14, Severance Hospital announced that a research team led by Professors Jinwoo Han and Seokho Byun from the Department of Ophthalmology at Severance Hospital, and Professor Junwon Lee from the Department of Ophthalmology at Gangnam Severance Hospital, has developed an artificial intelligence (AI) system that predicts the causative genes of hereditary retinal diseases using only fundus photographs and optical coherence tomography (OCT) images. The team validated its clinical usefulness through clinical trials. The study results were recently published in the international journal "Nature Medicine."
Professors Jinwoo Han and Sukho Byun, Department of Ophthalmology, Severance Hospital; Professor Junwon Lee, Department of Ophthalmology, Gangnam Severance Hospital. Severance Hospital
View original imageHereditary retinal diseases are caused by defects in more than 300 genes and can lead to blindness. Accurately identifying the causative gene is crucial for determining treatment strategies and prognosis, as well as for genetic counseling and selecting patients eligible for gene therapy.
However, gene diagnosis requires a series of steps including ophthalmic examinations such as fundus and OCT imaging, next-generation sequencing (NGS), and multidisciplinary consultations, which can take years to complete. It is reported that in more than 40% of patients, the causative gene cannot be identified.
The research team developed "Retina4IRD," an AI-based clinical decision support system that utilizes fundus photographs, OCT images, and clinical information such as the patient's age, sex, family history, age of onset, and disease duration.
The system was trained using data from 1,843 patients (3,376 eyes) with genetically confirmed hereditary retinal diseases, collected from nine medical institutions in Korea, China, and Poland. Retina4IRD presents a ranked list of candidate causative genes across 17 gene categories, including genes for which treatments are available or undergoing clinical trials. In internal validation, the 'TOP-5' accuracy rate, which measures the inclusion of the true causative gene among the top five candidates, reached 90.4%.
A randomized controlled clinical trial was conducted with 295 participants to verify practical clinical applicability. Patients were divided into an AI-assisted diagnosis group (148 patients) and a specialist-only diagnosis group (147 patients). After comparing diagnosis results, accuracy was evaluated against next-generation sequencing results.
As a result, the primary endpoint—TOP-5 accuracy in gene prediction—was 88.5% in the AI-assisted group, higher than the 67.3% achieved by the specialist-only group. The 'TOP-1' accuracy, predicting the single most likely gene, was also higher in the AI-assisted group at 37.8%, compared to 22.4% in the specialist-only group.
The research team explained that this is the first case in the world to prospectively verify, through a randomized controlled clinical trial, the clinical usefulness of an AI system that predicts causative genes using only routinely conducted fundus photographs and OCT in ophthalmology practice.
Hot Picks Today
"Two Spend 500,000 Won in a Flash, Yet Wait Two Hours?" ... 'Dreamland' Shunned by Kids and Foreign Tourists
- The Era of Getting 200 Million Won After 30 Years Is Over... A Japanese Bank's "Perpetual Department Head" Who Just Clocked In Cries Out
- Spain Hit by Its Worst Tragedy... Lee Kangin, Moreno and Real Madrid Join Soccer World in Mourning
- They Said, “Make It Your Home for Life” - Then Suddenly Kicked 200 Passengers Off... What Happened to the Around-the-World Cruise?
- "130 Rice Bowls Sold Out in a Day"... 'Unfiltered Face' Adds 8 Million Won to Night Market Sales: China's Bungeunwoo
Professor Jinwoo Han stated, "Retina4IRD is not a tool to replace physicians in making a diagnosis, but rather serves as a pre-screening tool that narrows down the likely causative genes before conducting genetic tests. It is expected to help reduce unnecessary testing costs and diagnosis time, and facilitate early identification of patients for whom timely intervention, such as gene therapy, is critical."
© The Asia Business Daily. All rights reserved. Unauthorized AI training and use prohibited.