The experimental 15 protein panel can flag whether a patient has Alzheimer's alongside Parkinson's, Lewy body dementia (a related memory and movement disorder) or frontotemporal dementia (which often shows up first as personality or language changes), combinations that single…
When a patient walks into a memory clinic with what looks like Alzheimer's disease, the diagnosis is usually treated as a single answer. Autopsies keep telling a different story: most of those brains carry more than one neurodegenerative disease at the same time, and the second or third condition is usually missed while the patient is alive.
A new research-grade blood panel, reported in the May issue of Alzheimer's & Dementia, is built around that gap. The test measures 15 blood proteins and uses a machine-learning pattern finder to flag which of four major dementias a person has, and in what proportion: Alzheimer's disease, the most common cause of memory loss; Parkinson's disease dementia and dementia with Lewy bodies, two related conditions that combine memory loss with movement symptoms and fluctuating alertness; and frontotemporal dementia, a group of disorders that attack the brain's frontal and temporal lobes and often show up first as personality or language changes rather than memory loss.
In a validation cohort, the panel correctly identified whether a person had more than one of those diseases 92.3% of the time, well above the roughly 50/50 chance level that would apply if clinicians were guessing blind, according to the Science News report on the study.
That mix matters because each of the four diseases calls for a different care plan, and a misdiagnosis in any direction can mean the wrong treatment path. A patient whose symptoms come from a Lewy body or frontotemporal process can be told they have Alzheimer's, and the actual disease driving the decline keeps progressing unrecognized. The protein panel is the first research tool designed to surface all of those possibilities from a single blood draw.
The shift also exposes how narrow current blood testing has been. The only FDA-cleared blood test for Alzheimer's detects that disease alone, and several un-cleared Alzheimer's tests do the same, according to the Science News write-up of the study. None of them can see Parkinson's, Lewy body, or frontotemporal pathology at all, so a "negative" Alzheimer's result does not mean a patient has no dementia. It just means they don't have that one. Researchers quoted in the same piece said the four diseases are more complex than initially thought and overlap more than expected, which is why a single-disease view has consistently undercounted mixed cases.
The caveats are real. The 92.3% figure comes from a single study, on a single validation set, and the test is not FDA-cleared, meaning it is currently a research tool, not a clinical one. Independent replication in a broader, general-population cohort has not yet been published. The test was developed on samples from specialty memory clinics, which may not generalize to the broader population where symptoms are messier and overlap is harder to disentangle. Dementia affects more than 6 million people in the United States and is the seventh leading cause of death worldwide, according to the same Science News report, so the eventual clinical stakes are large.
What to watch next: whether the panel holds up in a general-population cohort, and whether any of the four diseases it detects end up calling for a treatment that would have been missed under the old single-disease testing regime. A diagnostic that finally names what is actually in the brain is only useful if the health system can act on the answer.