Roche and Lilly Wager Alzheimer’s Care on a Blood Test

The FDA cleared Roche’s Elecsys pTau217 Alzheimer’s blood test on August 24, 2026 for people 55 and older with cognitive decline. Built with Eli Lilly, the assay uses one plasma protein to sort who is likely, unlikely, or still unclear for amyloid plaque.

Lilly already sells Kisunla, an amyloid-clearing antibody that cannot start until plaque is confirmed. Roche is running its own antibody in late-stage trials. The clearance is the top of that funnel.

A Single Protein Now Rules Amyloid in and Out

Elecsys Phospho-Tau (217P) Plasma measures phosphorylated tau 217, a blood signal that tracks amyloid burden in the brain. Roche says it is the first and only FDA-cleared single-biomarker blood test that supports both rule-in and rule-out of amyloid pathology, using the same validated clinical cutoffs across settings.

Results come back as positive, intermediate, or negative. Family doctors and specialists are supposed to read those bins with the same numbers, then decide who needs a referral, a PET scan, or a pause. The test is not a stand-alone Alzheimer’s diagnosis.

Roche Diagnostics USA posted the clearance the same morning.

We’ve received FDA clearance for the Elecsys® pTau217 blood test, which can help bring clinical clarity to Alzheimer’s assessment. With validated cutoffs across primary and specialty care, we’re expanding diagnostic access so providers can act earlier: https://t.co/9z2VqgvzrM pic.twitter.com/42fpRcCTU3

— Roche Diagnostics USA (@RocheDiaUSA) August 24, 2026

An estimated 75% of people living with dementia remain undiagnosed, Roche said, citing Alzheimer’s Disease International’s World Alzheimer Report on undiagnosed dementia. People who do get a diagnosis often wait years after symptoms start. PET scans and spinal taps can confirm amyloid, and both remain hard to reach outside specialty centers.

THE NUMBERS ROCHE IS BANKING ON

• Intended age: Adults 55 and older with signs, symptoms, or complaints of cognitive decline.
• Installed base: More than 4,500 cobas laboratory instruments already in U.S. labs, so sites do not need a new machine.
• Result bins: Positive, intermediate, and negative likelihood of amyloid pathology.
• Europe: The same assay received a CE mark in May 2026 before the U.S. file.

Dan Malarek, president and CEO of Roche Diagnostics North America, said the assay can bring evaluation closer to patients and give clinicians more confidence about the next step. Family doctors are the first stop for most memory complaints, a point already laid out in the primary-care rollout of the same assay.

Lilly Needs That Amyloid Answer for Kisunla

Carole Ho, M.D., executive vice president and president of Lilly Neuroscience, put the collaboration in the clinic where the complaint starts, not only in memory centers.

Lilly’s collaboration with Roche on Elecsys pTau217 was driven by a shared commitment to bringing one blood test into both primary care, where most patients with cognitive complaints are first seen, and specialty care, helping close the gap between symptoms and answers.

Carole Ho, M.D., President of Lilly Neuroscience, Roche announcement

Kisunla (donanemab-azbt) is indicated for Alzheimer’s disease and should be started in mild cognitive impairment or mild dementia, the group studied in trials. The label tells prescribers to confirm the presence of amyloid beta pathology before the first infusion.

That confirmation used to mean a PET scan or a spinal tap. Lilly’s clinician pages now walk primary-care doctors through blood biomarker tests such as P-tau217 and cite Global CEO Initiative guidance that plasma P-tau217 used with a clinical exam can produce cost savings of up to 81% versus PET and up to 60% versus cerebrospinal fluid tests.

Those pages also say patients with mild cognitive impairment and amyloid positivity are 5 times more likely to progress than people who are amyloid negative, and that 95% of surveyed U.S. adults would want a simple medical test for Alzheimer’s-related biomarkers if early symptoms appeared. A donanemab work group had already said high-performing, FDA-cleared blood tests would become appropriate for establishing eligibility, while warning against blood-only decisions before that bar was met.

pTau217 has been running as a lab-developed send-out for months. The FDA file is what lets a family doctor order a standardized cutoff that a health plan can recognize, and what lets Lilly argue the amyloid question can be asked before a specialty wait list.

Why the Test Cannot Steer Antibody Dosing

The Elecsys pTau217 file does not let clinicians track whether Kisunla or any antibody is working, and it does not predict who will develop dementia. Roche’s intended-use language still requires other diagnostic tools, and Lilly’s Kisunla label still uses amyloid PET to decide when to stop infusions. A positive blood bin is a sorting tool, not a prescription.

THE LABEL’S HARD STOPS

• Not stand-alone: Results must be read with other diagnostic tools and clinical information.
• No dementia forecast: Safety and effectiveness have not been established for predicting development of dementia or other neurologic conditions.
• No drug monitoring: The assay is not cleared for monitoring the effect of therapeutic products.
• No silent screening: The intended use is people 55 and older who already have signs, symptoms, or complaints of decline, not people with no symptoms.

Kisunla still needs a recent baseline brain MRI, then further MRIs before the 2nd, 3rd, 4th, and 7th infusions, because amyloid-related imaging abnormalities can be serious and can be fatal. Testing for ApoE ε4 status should be performed before treatment to inform ARIA risk; about 15% of Alzheimer’s patients are homozygotes and face a higher rate of those events. Dosing may stop when amyloid plaques fall to minimal levels on PET imaging. None of those steps moved into a plasma tube.

The Gray Zone That Still Needs a PET Scan

The intermediate bin is the part of the brochure that still looks like 2024. Roche says the test can distinguish people with a high or low likelihood of amyloid pathology and those who should be considered for further testing. Further testing, in that sentence, still means PET or cerebrospinal fluid.

Jared R. Brosch, M.D., a neurologist at Indiana University Health, said PET imaging and cerebrospinal fluid biomarkers can confirm amyloid in people with cognitive impairment, and that those approaches may be costly, invasive, and not readily accessible. Blood-based biomarkers, he said, may let clinicians evaluate more patients earlier. He did not say they retire the scanner.

A positive bin is not a treatment plan. Antibody drugs still need the right clinical stage, an MRI safety screen, and a confirmed amyloid answer, and many people who test positive will not qualify. A blood result without that path still turns every lost key into a scare, which is why the “come back when it is worse” delay was never a kindness and is also why a lab slip cannot be the whole conversation.

Roche did not disclose a price. Payers that still want PET or a spinal tap before Kisunla will treat the blood test as a triage step, which is useful and is not the same as replacing the workup.

Labs Can Run It Without Buying New Hardware

The commercial bet only works if the draw is ordinary. Roche designed Elecsys pTau217 for cobas instruments already sitting in hospital and reference labs. Labcorp said it will offer the test nationwide in the coming months and that, once a clinician orders it, patients can have blood drawn in a physician’s office or at more than 2,200 patient service centers.

Brian Caveney, Labcorp’s chief medical and scientific officer, said the best advances should make the diagnostic process simpler, more accessible, and more consistent. Quest Diagnostics has also said it plans to offer the test. Labcorp already runs Roche’s Elecsys pTau181 assay and Fujirebio’s Lumipulse pTau-217/Beta-Amyloid 42 ratio, so this clearance lands in a menu, not a vacuum.

Fujirebio’s Lumipulse became the first FDA-cleared Alzheimer’s blood test in 2025. Roche’s pTau181 assay was cleared the same year. C2N Diagnostics’ PrecivityAD2 test received clearance in the days before Roche’s announcement. Roche’s claim for pTau217 is narrower and more commercial: one biomarker, both directions, the same cutoffs in primary care and specialty care, on machines labs already own.

HOW CLINICS SORT AMYLOID NOW

Labcorp says Elecsys pTau217 offers performance comparable to cerebrospinal fluid testing and PET imaging. Roche’s own language is “high agreement with amyloid PET.” Neither statement is a promise that an intermediate result can be ignored.

Roche Is Testing Its Own Brainshuttle Antibody

Roche did not build a diagnostics-only franchise. The same announcement lists trontinemab, an experimental Brainshuttle amyloid-beta antibody designed for enhanced brain delivery and now in two phase 3 studies in early Alzheimer’s disease, and nivegacetor, an experimental oral gamma-secretase modulator in phase 2. The diagnostics column already holds pTau217, pTau181, and two cerebrospinal fluid ratios.

Lilly’s near-term prize is a wider pool of people whose amyloid status is known while they still sit in mild disease, which is the only window Kisunla’s label uses. Roche’s longer prize is the same blood sort feeding its own late-stage antibody if those trials hold. PET still decides when an infusion has done enough. MRI still watches for ARIA. The new object in the room is a plasma tube that can be drawn in a clinic that never owned a PET camera.

Roche and Lilly now have a primary-care blood sort sitting in front of the amyloid question Kisunla already requires in writing.

Disclaimer: This article is news reporting and analysis of an FDA-cleared laboratory test and related Alzheimer’s drugs, and it is for information only. It does not constitute medical advice, a diagnosis, or a recommendation to seek, skip, or interpret any blood test, PET scan, spinal tap, or antibody therapy. Readers should consult a qualified physician, neurologist, or geriatrician before acting on memory concerns or any lab result. Figures, intended uses, prices, and availability reflect company statements and labeling as of September 2, 2026 and may change as labs launch the assay and as payers set coverage rules.