Most blood tests for Alzheimer's disease ask one question: how much?
How much amyloid is floating in the bloodstream. How much tau. How much of the proteins associated with the disease's characteristic brain plaques and tangles.
A team at Scripps Research is asking a different question: *what shape*?
Their answer, published in *Nature Aging* in late February 2026, has produced a blood test with 93% accuracy in distinguishing healthy individuals from those with mild cognitive impairment or Alzheimer's disease — one that may work months or years before clinical symptoms begin.
**The Shape of Disease**
The biological insight at the heart of this work is called proteostasis — the body's system for maintaining the correct three-dimensional structure of proteins. When that system is under stress, proteins begin to fold differently. They become less "open," their structural signatures shifting in ways that can be measured in blood.
What the Scripps team found is that in people with Alzheimer's disease, three specific proteins show distinctive shape changes long before conventional biomarkers — or symptoms — appear:
- **C1QA**: a protein involved in immune signalling and the clearing of dead cells from the brain - **Clusterin**: associated with protein folding support and amyloid clearance - **Apolipoprotein B**: related to vascular and lipid health
None of these proteins were previously considered Alzheimer's biomarkers. They're not amyloid. They're not tau. They're the structural maintenance crew — and their shape, it turns out, tells a story about whether the brain's clearance system is functioning normally.
**93% Accuracy — and It Correlates with MRI**
In a validation study, the structural scores produced by measuring these three proteins were compared against cognitive test results and brain MRI scans measuring atrophy.
The correlation was strong. Patients with higher structural abnormality scores also showed greater brain volume loss on imaging and lower scores on cognitive assessments. The test wasn't just detecting a blood signature — it appeared to be tracking the actual progression of disease in the brain.
At 93% accuracy in differentiating healthy controls from Alzheimer's patients, the test outperforms several existing blood biomarker panels and approaches the sensitivity of the far more expensive and invasive PET scans currently used for definitive diagnosis.
"We're not measuring quantity. We're measuring conformation," said lead researcher Professor Sonia Bhatt of the Bhatt Laboratory at Scripps. "That turns out to give you a different — and in some ways earlier — window into what's happening."
**Earlier Diagnosis, Better Outcomes**
The timing of Alzheimer's diagnosis matters enormously. Current treatments — including the recently approved lecanemab and donanemab — work better the earlier they're administered, ideally before significant neuronal loss has occurred. Most people are currently diagnosed after symptoms have already disrupted daily life, by which point a substantial proportion of damage is done.
A test that works in pre-symptomatic stages could change when treatment begins — and potentially change what treatment can achieve.
The research team noted that the structural score correlated most strongly with the early-stage patients — those with mild cognitive impairment rather than full Alzheimer's. This is consistent with the hypothesis that protein shape change is an early-stage phenomenon that normalises (paradoxically) as disease advances and cellular machinery shuts down entirely.
**What Comes Next**
The Scripps test is not yet available clinically. Validation in larger, more diverse cohorts is required, and regulatory approval pathways are being explored. The NIH provided funding for the research and has flagged the proteostasis approach as a priority area for further investigation.
Separately, researchers at UC San Diego published a study in *JAMA Network Open* this month showing that elevated blood levels of phosphorylated tau 217 (p-tau217) could predict a woman's risk of developing dementia up to 25 years before symptoms appear — another dimension of the rapidly expanding field of blood-based Alzheimer's detection.
Taken together, these studies suggest that the coming years may see a shift in how Alzheimer's is detected: from a late-stage clinical diagnosis to an early-stage blood panel, administered as part of routine health screening.
For the 55 million people currently living with dementia worldwide — and the hundreds of millions who are statistically at risk — that shift cannot come fast enough.
*Sources: Nature Aging (February 27, 2026) · Scripps Research · ScienceDaily (March 12, 2026) · NIH News · Neuroscience News · news-medical.net*
