Roche’s New Alzheimer’s Blood Test Is Strongest at Ruling Out

The FDA last week cleared Roche and Eli Lilly’s Elecsys pTau217 Alzheimer’s blood test for people 55 and older who already have memory or thinking complaints. The assay runs on Roche machines already sitting in U.S. labs and returns a positive, negative, or intermediate read on amyloid pathology, the sticky plaque tied to Alzheimer’s disease.

It is the fourth Alzheimer’s blood test the agency has cleared in a little over a year. A low result can spare someone a brain scan. A high or middle result still tends to send that person toward the PET imaging and spinal taps the blood draw was supposed to skip, and the clearance does not change the sleep, movement, hearing, and blood-pressure work researchers still tie to a large share of dementia risk.

Roche’s Blood Test Reads Amyloid From One Tube

Roche said on Aug. 24 that Elecsys Phospho-Tau (217P) Plasma, developed with Eli Lilly and Company, is the first FDA-cleared single-biomarker blood test that can both rule amyloid pathology in and rule it out, using the same cutoffs in primary care and in specialty clinics. It is meant for adults 55 and older with signs, symptoms, or complaints of cognitive decline, not for people who feel fine and want a peek at their future.

Carole Ho, M.D., executive vice president and president of Lilly Neuroscience, said the pairing was built to put one blood test where most patients first mention a cognitive complaint, then keep that same test usable when a neurologist takes over. Dan Malarek, president and CEO of Roche Diagnostics North America, said the goal is to bring the workup closer to the patient. Roche did not publish a price.

The practical hook is installed base. Roche designed the assay for more than 4,500 cobas laboratory instruments already in U.S. clinical labs, so hospitals are not being asked to buy a new box. Alzheimer’s Disease International has estimated that 75% of people living with dementia remain undiagnosed. Researchers at University College London have put the average wait from first symptoms to a dementia diagnosis at 3.5 years.

THE SCALE OF THE DIAGNOSIS GAP

• Age gate: The cleared use is adults 55 and older who already have cognitive complaints, not a screening test for birthday checkups.
• Machine count: Roche says the assay can run on more than 4,500 cobas instruments already in U.S. labs.
• Undiagnosed share: Alzheimer’s Disease International has estimated that 75% of people living with dementia never get a diagnosis.
• Time to a name: University College London research Roche cited puts the average wait from first symptoms to a dementia diagnosis at 3.5 years.

Jared R. Brosch, M.D., a neurologist at Indiana University Health, said PET scans and spinal-fluid tests can confirm amyloid in people with cognitive impairment, yet they are costly, invasive, and hard to reach outside specialty centers. A plasma tube that travels with routine lab work is the opening Roche is selling. It is also where the limits start.

Positive, Negative and the Gray Zone Between Them

The cleared test does not spit out a yes-or-no Alzheimer’s diagnosis. Roche said it returns positive, intermediate, and negative categories for the likelihood of amyloid pathology, and that those results have to be read with the exam, the history, and other tests. The company also said safety and effectiveness have not been established for predicting whether dementia will develop, for other neurologic diseases, or for watching how a drug is working.

WHAT THE THREE RESULT BINS MEAN

• Negative: Low likelihood of amyloid pathology, which should push the clinician to look for other causes of the thinking change.
• Positive: High likelihood of amyloid pathology, which still sits inside a broader Alzheimer’s workup rather than ending it.
• Intermediate: An indeterminate band where Roche says further testing should be considered before anyone treats the result as an answer.

That middle bin is the part the victory lap leaves out. Blood amyloid tests in clinic populations often leave a sizable share of people in an indeterminate zone. The Global CEO Initiative on Alzheimer’s Disease has used a 15% to 20% indeterminate band as a reference for a typical clinical group. Roche has not published the indeterminate rate for the cleared Elecsys pTau217 cutoffs.

A 2026 study of a prototype pTau217 immunoassay in five cohorts (2,148 people) shows why a single cutoff is a blunt tool. Against amyloid PET, mean plasma pTau217 in cognitively impaired volunteers was 0.835 pg/mL in amyloid-positive people (n=394) and 0.361 pg/mL in amyloid-negative people (n=144). A cutoff below 0.189 pg/mL produced a negative predictive value of 92.51% in impaired people and 98.60% in unimpaired people. Sensitivity was 98.98% in the impaired group, while specificity fell to 29.17%. Positive predictive value was 76.48% in impaired people and 23.86% in unimpaired people. Those figures describe a prototype rule-out screen, not the two-cutoff label the FDA cleared, and they are why a three-bin design exists at all.

Single ‘Alzheimer’s blood tests’ can be fraught with challenges, and this new test is no different.

Dr. Richard Isaacson, director of research, Institute for Neurodegenerative Diseases

Isaacson said the new test has a sufficient negative predictive value, meaning a low result is fairly good at correctly ruling Alzheimer’s disease out, and that the positive predictive value is not great. He warned that false positives are a problem in clinic, especially with a test that sorts people into positive, negative, and indeterminate groups. Trouble handling the sample, kidney disease, or a viral illness can scramble the read, he said. In his own practice he runs pTau217 and amyloid blood tests to see how people respond to treatment and to diet, exercise, sleep, and social changes, rather than as a one-shot diagnosis.

The Four FDA-Cleared Alzheimer’s Blood Tests

Elecsys pTau217 arrives in labs that already run other Alzheimer’s blood assays. Fujirebio’s Lumipulse G pTau217/β-Amyloid 1-42 Plasma Ratio was the first FDA-cleared blood test to aid Alzheimer’s assessment, in 2025. Roche’s own Elecsys pTau181 test, which measures a different phosphorylated tau, was cleared the same year as a rule-out tool. On Aug. 20, the FDA cleared C2N Diagnostics’ PrecivityAD2 test, which the Alzheimer’s Association called the third blood-based biomarker test for the disease, and the first cleared down to age 40.

FDA-CLEARED ALZHEIMER’S BLOOD TESTS

In the FDA study of 499 adults, a positive Lumipulse result matched amyloid PET or spinal-fluid tests 91.7% of the time, and a negative result matched 97.3% of the time, with fewer than 20% of results landing indeterminate. Labcorp, which already offers pTau181, cites a 97.9% negative predictive value for that earlier Roche assay when the job is to say amyloid is unlikely. Roche’s claim for pTau217 is different: one protein, one set of cutoffs, and both a rule-in and a rule-out in primary and specialty care, with high agreement versus amyloid PET. The company has not put sensitivity, specificity, or predictive values for those cleared cutoffs in its public statement.

PET and Spinal Taps Do Not Go Away

Rachel Buckley, an associate professor of neurology at Harvard Medical School, has said blood tests that measure pTau217 strongly predict the buildup of beta-amyloid plaques. Those plaques can collect years before memory loss, even when people are in their 30s and 40s. As amyloid rises, tau tangles gather inside neurons. In some diseases such as frontal lobe dementia, tangles can build without amyloid, which is one reason a tau blood number cannot be treated as a complete map of the brain.

High amyloid does not always lead to dementia, and tau in the brain does not dictate later impairment, Buckley said. That biology is why a blood tube is a clue, not a verdict. Traditional confirmation still uses amyloid PET or cerebrospinal fluid from a spinal tap, both of which remain the reference Roche measured against.

In July 2025 the Alzheimer’s Association released its first clinical practice guideline on blood-based biomarker tests. Those recommendations apply to people with objective cognitive impairment who are already in specialized memory care, not to a general primary-care panel. Tests that reach 90 percent sensitivity and 75 percent specificity can be used to triage, meaning a negative result rules Alzheimer’s pathology out with high probability and a positive result still needs confirmation by CSF or amyloid PET. Tests that reach 90% for both sensitivity and specificity can stand in for PET or CSF. The panel reviewed 49 observational studies and 31 assays and did not endorse a brand. It also said a blood test should not be drawn before a full clinical evaluation, and that many commercial tests do not meet those bars.

Maria C. Carrillo, Ph.D., the Association’s chief science officer, said the guideline is meant to help clinicians use blood tests consistently. Roche is pitching Elecsys pTau217 into primary care as well as specialty clinics, which is a wider door than the Association’s first guideline was written to cover. A family doctor can now order a plasma pTau217. A positive or intermediate result still hands the patient back to the same scarce PET slots and lumbar punctures, plus the wait that goes with them.

Who Should Skip a pTau217 Blood Test

The intended-use line is doing a lot of work. Roche limited clearance to people 55 and older who already have cognitive complaints. That matches the Association’s warning that these assays were studied in people with impairment, not in people who are aging and curious.

We do not recommend blood tests in individuals without symptoms; until there is a disease modifying therapy available for asymptomatic individuals, a physician would not currently change their clinical guidance based on pTau217 levels alone in this population.

Rachel Buckley, associate professor of neurology, Harvard Medical School

Isaacson’s caution about kidneys and viral illness belongs on the same list. Phosphorylated tau in plasma can move for reasons that have nothing to do with plaques, so a person with reduced kidney function or an acute infection is a poor candidate for a one-number decision. Sample handling errors produce the same mess. A clinician who orders the test and then treats a lone elevated pTau217 as Alzheimer’s is using the assay outside what Roche says it is for.

There is also the question people keep asking once a primary-care blood test exists: if there is no cure in the exam room, why know. Disease-modifying amyloid drugs exist for some patients who already have confirmed pathology and who meet narrow trial-style criteria. Most people who mention a word-finding slip at a checkup will not walk out with one of those infusions. What they can walk out with, if the result is positive or stuck in the middle, is a longer specialist queue and a label they cannot act on except through the unglamorous work of blood pressure, sleep, hearing, and movement. Easier testing does not, by itself, create a treatment that primary care can start that afternoon. It can move the fear earlier, unless a clearly negative result is allowed to stop the hunt.

Fourteen Risks a Blood Draw Cannot Change

Isaacson said diet, exercise, sleep, and social contact, plus treatment of insulin and cholesterol problems when needed, have lowered amyloid and tau in patients who stick with the plan. That is the part of this story a fitness and wellness reader can use on the same day as the FDA notice, and it does not require a cobas instrument.

The 2024 Lancet Commission on dementia, led by Gill Livingston, a professor of psychiatry of older people at University College London, and written with 26 other specialists, estimated that about 45% of dementia cases worldwide could theoretically be prevented or delayed by tackling 14 modifiable risks across life. The 2020 edition had 12 factors and a figure near 40%. The 2024 update added high LDL cholesterol, tied to about 7% of cases from midlife, and untreated vision loss in later life, tied to about 2%. On July 15, 2026, the World Health Organization issued updated risk-reduction guidance that used the same ceiling, that up to 45 percent of dementia risk traces to factors people and health systems can change. The 45% figure is a population estimate, not a promise to any one patient.

THE 14 MODIFIABLE DEMENTIA RISKS

• Early life: Less education, which the commission links to lower cognitive reserve later on.
• Midlife hearing: Untreated hearing loss, still one of the largest single contributors in the model.
• Midlife LDL: High LDL cholesterol from around age 40, the 2024 addition tied to about 7% of cases.
• Midlife vessels: High blood pressure, obesity, diabetes, and smoking, which damage the same circulation the brain depends on.
• Midlife injury and alcohol: Traumatic brain injury and heavy drinking, both with direct effects on brain tissue.
• Later-life mood and movement: Depression, physical inactivity, and social isolation.
• Later-life environment: Air pollution, plus untreated vision loss, the other 2024 addition, at about 2% of cases.

Livingston has said it is never too early or too late to act. Laura Nisenbaum, interim chief science officer at the Alzheimer’s Drug Discovery Foundation, has pointed to exercise, diet, social engagement, cognitive training, and control of vascular and metabolic risks as the cluster of habits tied to that prevention estimate. None of those habits is a substitute for a workup when memory is already slipping. They are what remains available before a blood test, after a negative result, and during the long wait that still follows a positive or intermediate one.

Labcorp and Quest Will Draw Blood This Fall

Access is the piece Roche can actually ship. Labcorp said on Aug. 24 that it will offer Elecsys pTau217 nationwide in the coming months, with draws in a doctor’s office or at more than 2,200 patient service centers. Dr. Brian Caveney, Labcorp’s chief medical and scientific officer, said the path from a cognitive complaint to an evaluation has been long and uncertain, and that a nationwide blood test is meant to shorten that path. Labcorp already runs Elecsys pTau181 and the Lumipulse pTau217/beta-amyloid 42 ratio, so this is another assay in a menu, not a first toe in the water.

Quest Diagnostics said the same day that it will add an AD-Detect laboratory service based on the FDA-cleared Elecsys pTau217 assay for physicians and trial collaborators in the fourth quarter of 2026, then fold the Roche test into future AD-Detect panels in 2027. Michael K. Racke, M.D., a neurologist and Quest’s senior medical director for neurology, called blood-based biomarker testing a new standard for guiding diagnosis and treatment decisions. Quest draws at about 2,000 patient-service sites, in offices, and through mobile phlebotomy, and a clinician still has to order the test. Separately, Quest said it would launch its own multi-biomarker AD-Detect panel (amyloid beta 42/40, pTau217, and ApoE) at the end of August, a lab-developed test with a 10% indeterminate rate in company research. That panel is not the FDA-cleared Elecsys assay.

Europe is already on the other side of the launch. The test received CE mark certification in May and, according to Roche’s diagnostics timeline, was launched there in July. The U.S. clearance is nine days old. The draws, for most Americans, are still ahead.

THE BLOOD TEST CLEARANCE CALENDAR

1. May 16, 2025: The FDA clears Fujirebio’s Lumipulse plasma ratio test, the first Alzheimer’s blood assay in the country.
2. 2025: The FDA clears Roche’s Elecsys pTau181 test as a rule-out tool, later cited by Labcorp at a 97.9% negative predictive value.
3. May 2026: Elecsys pTau217 receives Europe’s CE mark, with a July launch on the continent.
4. July 29, 2025: The Alzheimer’s Association publishes its first blood-biomarker guideline for specialty memory care.
5. Aug. 20, 2026: The FDA clears C2N’s PrecivityAD2 test for adults as young as 40, the third U.S. Alzheimer’s blood test.
6. Aug. 24, 2026: The FDA clears Elecsys pTau217, and Labcorp and Quest say they will put it on their menus.
7. Fourth quarter 2026: Quest plans a nationwide AD-Detect service built on the Roche assay, with panel use in 2027.

A negative pTau217, read with a careful exam, can end an Alzheimer’s scare without a spinal needle. A positive or intermediate result still points to the old, slower tools, and to the same hearing aids, walks, blood-pressure pills, and social calendar that were worth doing before anyone invented this tube.

Frequently Asked Questions

Can Elecsys pTau217 Track Whether an Alzheimer’s Drug Is Working?

No. Roche said the safety and effectiveness of Elecsys pTau217 have not been established for monitoring the effect of therapeutic products, or for predicting whether dementia or another neurologic condition will develop. The cleared job is to help assess the likelihood of amyloid pathology in a symptomatic adult 55 or older, alongside the rest of the workup.

Do You Need a Doctor’s Order for the Roche pTau217 Test?

Yes. Labcorp and Quest both describe a clinician order, then a draw in an office, a patient-service center, or, for Quest, a mobile phlebotomy visit. The Alzheimer’s Association’s blood-test guideline also says a blood biomarker should not be obtained before a comprehensive clinical evaluation, and that results should be interpreted in that clinical context.

How Is pTau217 Different From Roche’s pTau181 Test?

Both are plasma immunoassays on Roche cobas instruments, but they measure tau phosphorylated at different sites. Elecsys pTau181, cleared in 2025, was positioned as a primary-care rule-out, and Labcorp cites a 97.9% negative predictive value for that assay. Elecsys pTau217, cleared Aug. 24, 2026, is Roche’s single-biomarker test for both rule-in and rule-out, with positive, intermediate, and negative bins and the same cutoffs in primary and specialty care.

Can a Positive pTau217 Result Mean a Different Dementia?

It can point at amyloid that is not the whole story, and a tau blood number can move for reasons outside Alzheimer’s disease. Buckley has noted that in frontal lobe dementia, which damages executive function more than memory, tau tangles can accumulate without amyloid. Kidney disease and viral illness can also confuse plasma pTau217, which is why Roche says the test is not a stand-alone diagnosis.

Does a Negative Result Mean You Will Never Develop Alzheimer’s?

No. A negative result means amyloid pathology is unlikely at the time of the draw, which is useful for deciding what not to chase next. Roche has not established the test for predicting future dementia. Plaques can still form later, and other brain diseases can still cause thinking changes even when this assay is low.

Is Elecsys pTau217 the Same as Quest’s AD-Detect Panel?

No. Quest plans to offer an AD-Detect service that uses the FDA-cleared Elecsys pTau217 assay in the fourth quarter of 2026. The company also launched a separate lab-developed AD-Detect panel that combines amyloid beta 42/40, pTau217, and ApoE and that, in Quest research, had a 10% indeterminate rate. That panel is Quest’s own mix, not the Roche IVD the FDA just cleared.

Disclaimer: This article is news reporting and analysis of an FDA-cleared diagnostic test and related dementia research. It is for information only and is not medical advice, a diagnosis, or a recommendation to take or skip any test, scan, drug, diet, or exercise plan. Talk with a qualified physician, neurologist, or geriatrician before ordering cognitive testing or changing medicines or daily habits because of a possible Alzheimer’s diagnosis. Figures, clearance terms, and lab launch dates reflect company statements and public-health sources as of Sept. 2, 2026, and may change.