AI Used to Identify Undiagnosed Patients with Immune Disorder through EHRs

Rare Daily Staff

Researchers were able to identify people suspected of having an undiagnosed, rare immunodeficency through the use of artificial intelligence to review electronic health records.

In a study published in Science Translational Medicine, researchers at UCLA Health and elsewhere, reported on the use of PheNet, a machine learning algorithm that could identify patients who may have common variable immunodeficiency (CVID), an inborn error of immunity characterized by antibody deficiency and impaired B cell responses.

PheNet is designed to scour patient electronic health record data and rank individuals according to their likelihood of having CVID. The researchers’ retrospective analysis suggested that the method could help diagnose many individuals earlier than standard clinical methods, accelerating their time to a diagnosis and treatment.

“Patients who have rare diseases may face prolonged delays in diagnosis and treatment, resulting in unnecessary testing, progressive illness, psychological stresses, and financial burdens,” said Manish Butte, a UCLA professor in pediatrics, human genetics, and microbiology/immunology who cares for these patients in his clinic at UCLA. “Machine learning and other artificial intelligence methods are making their way into health care. Using these tools, we developed an approach to speed the diagnosis of undiagnosed patients by identifying patterns in their electronic health records that resemble those of patients who are known to have the disorders.”

CVID can go undiagnosed for years after symptom onset because of its rarity and the fact that symptoms vary greatly from person to person. It makes people susceptible to infection, autoimmunity, and autoinflammation. Symptoms are often mistaken as being caused by more common disorders. CVID is caused by a mutation to a single gene, but more than 60 genes have been implicated to date. Because of that, there is not a single causal mechanism and there are no genetic tests to provide a definitive diagnosis.

Butte and Bogdan Pasaniuc, a professor of computational medicine, human genetics, and pathology and laboratory medicine at UCLA David Geffen School of Medicine, led a team that developed PheNet. Both are co-senior authors of the journal article.

PheNet learns phenotypic patterns from verified CVID cases and uses this knowledge to rank patients by likelihood of having CVID.

“Our own patients report experiencing years to decades of symptoms before they were referred to our immunology clinic,” Butte added. “With PheNet, dozens of patients could have been diagnosed one to four years earlier than they were, and by bringing patients to care years earlier, we should be able to reduce their costs and improve their health outcomes.”

When the research team applied PheNet to the UCLA electronic health record data comprising millions of patient records and followed up with a blinded chart review of the top 100 patients ranked by the system, they found that 74 percent were deemed probable to have CVID.

Based on these preliminary data, Butte and Pasaniuc successfully competed to receive $4 million of National Institutes of Health funding, which allows them to apply their AI in the real world.

They started by validating PheNet with more than 6 million records of patients from disparate medical systems in the University of California Data Warehouse and at Vanderbilt Medical Center in Tennessee. A collaboration led by Butte to have specialists see the patients identified by the algorithm was launched with the immunology clinics at University of California campuses in San Diego, Irvine, Davis, and San Francisco.

“We show that artificial intelligence algorithms such as PheNet can offer clinical benefits by expediting the diagnosis of CVID, and we expect this to apply to other rare diseases, as well,” said Pasaniuc. “Our implementation across all five University of California medical centers is already making an impact. We are now improving the precision of our approach to better identify CVID while expanding to other diseases. We will also plan to teach the system to read medical notes to glean even more information about patients and their illnesses.”