Insight
Mechanism evidence quality varies more than stage does
Competitive decks in this field rank programmes by phase, because phase is the one field every database populates. In non-invasive blood-brain barrier delivery, phase turns out to be close to uninformative about whether anyone has published how the asset crosses.
The distribution
Every row in the market map carries a mechanism-confidence value alongside its stage. HIGH means a company primary document, a granted or published patent, a regulator label, or a peer-reviewed paper. MEDIUM means an inference from a clinical or pharmacological signal, stated as an inference. LOW means secondary coverage only.
Across 69 in-scope rows: 42 HIGH, 17 MEDIUM, 10 LOW.
Those 10 LOW rows are not early or obscure programmes. One of them holds an approval.
Four programmes, in ascending order of stage
| Asset | Stage | Mechanism confidence |
|---|---|---|
| ALIA-1758, AbbVie | Phase 1 complete | LOW |
| ABL301, ABL Bio | Phase 1 complete | LOW |
| trontinemab, Roche | Phase 3 recruiting | HIGH |
| verenafusp alfa, AO GENERIUM | Approved, Russia | LOW |
ClinicalTrials.gov NCT06406348, NCT05756920, NCT07169578 and NCT06475404.
Sorting that table by stage puts the least-documented mechanism at the top. Sorting it by evidence puts it at the bottom. A competitive map that reports only the first column is reporting the wrong thing.
What a HIGH row costs to establish
Roche's transferrin receptor attribution for trontinemab needs two documents read together, and neither is sufficient alone. Niewoehner and colleagues establish that the Brain Shuttle module binds the transferrin receptor and that monovalent binding is required, since bivalent binding routes the construct to lysosomes. A separate Roche release then establishes that trontinemab is built on that module.
Niewoehner et al., Neuron 2014; Roche press release, 7 July 2026.
Three Roche and Genentech shuttle patent families exist and they name different mechanisms. One identifies no receptor anywhere in the document, its peptide motifs having been selected phenotypically by in vivo phage display in rats. Another claims LRP8. Only the Brain Shuttle work names the transferrin receptor.
WO2015124540A1, F. Hoffmann-La Roche AG, priority February 2014; WO2011091304A1, Genentech Inc and F. Hoffmann-La Roche AG, filed January 2011.
The practical consequence is worth stating plainly. A patent bearing a sponsor's name is not evidence about that sponsor's other assets, and holding a shuttle patent family says nothing about which shuttle any particular asset uses.
What a LOW row looks like
AbbVie acquired Aliada Therapeutics in December 2024 and describe ALIA-1758 as an anti-pyroglutamate amyloid beta antibody using a novel blood-brain barrier crossing technology. No receptor is named. No Phase 1 safety data has been disclosed. The Phase 1 is registered as completed.
AbbVie press release, 11 December 2024; ClinicalTrials.gov NCT06406348.
That is not a criticism of AbbVie, who are under no obligation to publish a mechanism. It is a statement about what a reader is entitled to conclude, which is very little.
Whole classes sit at one tier
All 10 intranasal programmes sit at MEDIUM. The registered route is documented in every case, but nose-to-brain transport is asserted for the class rather than demonstrated per programme. The 10 also collapse to two agent families, foralumab and intranasal insulin, which is not what a count of 17 registry records suggests on first reading.
Foralumab: ClinicalTrials.gov NCT06489548, NCT06868628, NCT06292923, NCT07688239 and NCT06879067. Intranasal insulin: NCT00438568, NCT01547169, NCT01767909, NCT01595646, NCT05081219, NCT01436045 and NCT02503501, among others.
All six exosome programmes sit at LOW. Two of the six state a route in the registered record, both intranasal. Four state no route at all, and none is confirmed intravenous.
ClinicalTrials.gov NCT04388982, NCT07638813, NCT07554872, NCT05152394, NCT07105371 and NCT06598202.
Why this column exists
A confidence tier is cheap to compute and expensive to omit. Without it, an approved asset and a Phase 1 asset with no published mechanism read as though the field has told you the same amount about each, and a reader who is deciding where to spend attention has been handed the wrong map.
Keeping the tier as a column rather than a footnote also means it survives the trip from research file to slide, which is the point in the process where this kind of qualification usually gets lost.