The Blood Test That Decides Whether Leqembi Has a Ceiling

Every two weeks, a patient returns for another infusion of Leqembi, the Alzheimer's drug Biogen shares with Eisai. That rhythm is the easy part of the treatment. Before the first dose, a doctor must prove the brain is actually carrying amyloid, usually with a PET scan or a spinal tap. Then, month after month, the patient is imaged again to watch for the brain swelling Leqembi itself can cause. The apparatus that keeps the treatment safe is the same apparatus keeping it slow: certainty is invasive, costly, and repeatable.

A small single-hospital study published in the journal Alzheimer's & Dementia is evidence that the repeat scans may not be needed at all — that a simple blood draw can track whether the drug is working. For a stock whose fortunes increasingly rest on one medicine, the question is what that costs, and who pays it.

The mechanism the study hands investors

At Korea University Guro Hospital, researchers followed 153 patients with early Alzheimer's disease who were receiving lecanemab. They measured a blood protein called p-tau217 — a marker of the tau tangles that pile up alongside amyloid — at the start of treatment, at three months, and at six months. Levels fell noticeably by three months, dropped fastest between three and six months, then flattened out. But patients split into two camps: 29 showed a large drop in the marker, while 52 showed a smaller one. A year later, the large-drop group was losing cognitive function markedly more slowly, and the difference on the standard CDR-SB dementia scale was statistically strong (p < 0.001). Patients with hypertension tended to respond less.

The authors call this the first evidence from a real-world clinic that a patient's trajectory on a blood test during the first six months predicts how they will fare later — not a one-time snapshot, but the direction of change. That is the finding's real weight. It does not prove Leqembi works; it suggests that whether it is working can be read cheaply, out of a few tubes of blood, instead of repeated scans. Exactly the kind of observation that matters if blood-based markers ever replace the confirmation ritual that currently gates and burdens the drug.

The drug Biogen can't afford to stay slow

The science prices itself once you see Biogen's books. Biogen carries a 50% share of Leqembi's net revenue and costs, and Leqembi has become the answer to a franchise that is shrinking. In the first quarter of 2026, Biogen's Alzheimer's collaboration revenue was $59.5 million — up 80% from $33.0 million a year earlier — on Leqembi global in-market sales of $168 million, up 74%, of which $86 million came from the U.S. Yet total company revenue rose just 2% to $2.48 billion, and Biogen guided full-year 2026 revenue to decline by a mid-single-digit percentage, with falling multiple-sclerosis sales offset only by growth products like Leqembi.

So the economics do not just tolerate a cheaper monitor; they need it. Every scan required before a patient starts — and every reason a patient or caregiver hesitates at the cost and burden — is a drag on the adoption curve that Biogen's declining base is counting on.

Who actually receives the check

Here is the uncomfortable part for anyone who wants this study to be a Biogen catalyst. A cheaper blood test does not put a single extra dollar on Biogen's own line. The p-tau217 tests are being built by the diagnostics industry — Roche's Elecsys platform, developed with Eli Lilly, has an FDA breakthrough-device designation, as does Quanterix's Simoa test; ALZpath sells its own assay — not by Biogen or Eisai. The Alzheimer's Association has already issued clinical-practice guidance for blood-based biomarker tests. What cheaper monitoring actually does is remove a cost that today lands on patients and the health system: repeated PET scans and spinal taps. Remove enough of that cost and you press on the throttle of patient starts and adherence — and Biogen, holding half the collaboration, collects its share downstream.

The bill for believing too hard

The honest count against all of this: only 81 of the 153 patients had complete marker data, the work comes from a single hospital, and the cognitive follow-up runs one year. There is no agreed threshold yet for what "responding" means — the authors say larger, multicenter studies with longer follow-up are required to validate the patterns. A single-center Korean cohort does not tell an American prescriber what to do. On the day it lands, this study changes no reported number and no guidance.

Set BIIB against that. The stock trades near $217 as of mid-September, up roughly 23% this year and close to its 52-week high, so investors are already paying for the Leqembi story to accelerate. A small study like this is not that acceleration; it is evidence the mechanism underneath it is still alive — that every marker letting a doctor monitor Leqembi with a blood draw instead of a scan is a small bet on a higher ceiling and a thinner bill for whoever funds the treatment.

Leqembi's future, in other words, turns less on the science of clearing amyloid and more on whether the cost of proving the treatment is working keeps falling. This study supports the direction of that fall. It does not yet prove the fall arrives at scale — and Biogen's shareholder, holding half the economics and most of the patience, remains the one paying for the wait until it does.