Artificial Intelligence Could Help Identify Rare Diseases Earlier
A groundbreaking new study published in Science Translational Medicine suggests that artificial intelligence (AI) may be able to identify patients at risk of developing rare diseases years before they would typically be diagnosed.
Researchers developed an AI program that successfully identified individuals at high risk of a rare immune disorder, known as common variable immunodeficiency (CVID). Out of a group of 100 people identified as high risk by the AI program, 74 were found to likely have the disorder.
Lead researcher Dr. Manish Butte from the University of California, Los Angeles, highlighted the potential benefits of using AI to speed up the diagnosis and treatment of rare diseases. Patients with rare diseases often face delays in diagnosis, leading to unnecessary testing, progressive illness, and financial burdens.
CVID disorders, which affect about 1 in 25,000 people, can cause antibody deficiencies and impaired immune responses. The study showed that for every year a diagnosis is delayed, there is an increase in infections, antibiotic use, hospitalizations, and missed days of work and school.
The research team has received $4 million in funding from the National Institutes of Health to further study the AI program in real-world settings. Senior researcher Bogdan Pasaniuc emphasized the potential clinical benefits of AI in expediting the diagnosis of rare diseases, not just limited to CVID but potentially applicable to other rare conditions as well.
The implementation of AI algorithms like PheNet, developed by the researchers, is already making an impact across all five University of California medical centers. The team plans to improve the precision of their approach, expand to other diseases, and teach the system to extract more information from medical notes to enhance patient care.