A research team led by Professor Jianping Jia at Xuanwu Hospital of Capital Medical University has developed a blood test that detects the seeding activity of amyloid beta - the protein aggregation process at the heart of Alzheimer's disease - and reported diagnostic accuracy above 90% in an independent validation cohort. The findings were published online in the Chinese Medical Journal on June 16, 2026.
The method relies on real-time ultrasonic protein misfolding cyclic amplification, or PMCA, applied for the first time to plasma amyloid beta seeding activity. Misfolded amyloid beta behaves somewhat like a prion: a small amount of misshapen protein can coax normal amyloid beta into the same abnormal conformation, propagating the damage. PMCA exploits that behavior in a test tube, amplifying trace amounts of misfolded protein until they can be measured. According to Jia, the ultrasonic version detects amyloid beta oligomers at concentrations as low as one femtomole, better than traditional shaking-based approaches. Cavitation from the ultrasound creates spherical interfaces and energy that encourage fibril formation, while localized heat and shear forces drive misfolding, and the addition of fluorescence tracking allows the aggregation to be watched in real time, with results in under 24 hours.
That matters because confirming Alzheimer's still generally requires a lumbar puncture for cerebrospinal fluid or a PET scan - one invasive, the other expensive and unavailable in many places. Blood biomarkers have made real progress in recent years, but relatively little of that work has targeted amyloid aggregation seeding activity itself rather than protein levels. A blood draw that reflects the underlying disease process could widen access to accurate diagnosis, particularly at the mild cognitive impairment stage when treatment decisions and planning are most consequential.
The study used a two-stage design: 120 participants in a discovery phase to establish preliminary performance, then 429 participants in an independent validation cohort that included cognitively normal individuals, people with MCI due to Alzheimer's, people with Alzheimer's dementia, and people with non-Alzheimer's dementia. Including that last group is important, since a useful test must distinguish Alzheimer's from other causes of cognitive decline, not simply separate the impaired from the healthy.