BY THOMAS R. COLLINS
NEW ORLEANS—A blood test that measures four key proteins combined with an algorithm for clinical risk factors helped differentiate psychogenic non-epileptic seizures (PNES) from epileptic seizures with a high degree of accuracy, according to findings presented here at the American Epilepsy Society annual meeting.
In preliminary findings, the test — developed and presented by officials from Cognizance Biomarkers, a subsidiary of Evogene, Inc. — distinguished PNES from epileptic seizures with 87 percent sensitivity and 100 percent specificity; the area under the curve (AUC) was 0.98 when administered within 24 hours of an event, researchers said.
"What we're doing that's different from the approach that other people have taken is that we're looking at a multi-protein signature response, as opposed to looking at one protein to say yes or no" to differentiate PNES, said John M. Gledhill, PhD, director of research and development at Cognizance Biomarkers.
If the test pans out, clinicians could use it as a tool to improve misdiagnoses, Dr. Gledhill said, noting that PNES patients account for the largest category of people misdiagnosed as having epilepsy. The test is now being assessed in a larger patient pool, he said.
Researchers reviewed data from patients in an epilepsy monitoring unit, 31 of whom had epileptic seizures and 25 of whom had PNES. They had been admitted for surgical evaluation or to get a definitive diagnosis. They drew blood samples each morning and within 24 hours of a clinical event captured by video electroencephalography (EEG).
The research team assayed 51 proteins associated with neuroinflammation, and, through a machine-learning process, identified four proteins — TRAIL, ICAM-1, MCP-2, and TNF-R1 — that were best able to differentiate epileptic seizures from PNES. Dr. Gledhill said that the precise mechanisms that make these four proteins the most useful are not yet known. But with the four proteins, the test achieved a sensitivity of 81 percent and 94 percent specificity.
When the researchers factored in risk factors associated in the literature with PNES — such as major depressive disorder, physical or emotional trauma, and multiple allergies — the test improved to an AUC of 0.98.
"That's really what is going to be important for this test — using all available information," Dr. Gledhill said.
Dr. Gledhill projected the test could come to market in 2021.
Selim R. Benbadis, MD, FAAN, professor of neurology at the University of South Florida, said the data from the test are "intriguing."
He said the test would have reasonable feasibility but noted that the increasing accessibility of video EEG could call its usefulness into question.
In their poster, researchers introduced their findings by saying that video EEG "requires inpatient hospital stays," but Dr. Benbadis said that is no longer the case.
"You can do ambulatory home EEG video studies, and if I can have that, it would be a lot more powerful than any blood work," he said.
He also noted that, in their presentation, researchers described the biomarkers as helping to "triage suspected epileptic seizure patients for expert evaluation," suggesting it may ultimately be a kind of screen before video EEG.
"If it's going to be a screen, how does it compare in cost with prolactin and creatine kinase?" he asked, referring to two other tests that can sometimes be effective. "Is it really worth it? I'm not sure."
"This test might have an adjunctive role, but I'm a little bit skeptical about it. I don't think it's going to change our diagnostic pathway….but it's intriguing."
The research was conducted with grant support from the National Institutes of Health. Dr. Gledhill receives a salary from Evogene, Inc. Dr. Benbadis has received honoraria from LivaNova, Eisai, Greenwich, Lundbeck, Neuropace, and Sunovion Pharmaceuticals.
LINK UP FOR RELATED INFORMATION
AES Abstract 3.094: Gledhill JM, Brand E, St. Clair R, et al. Predictive blood test for psychogenic nonepileptic seizures: Post hoc assessment of plasma biomarkers and risk factors.