It seems that _how_ this works is still unclear, if it works at all. Once validated, it seems plausible that a device/trained animal could replicate the results, assuming its not like this guy: https://en.wikipedia.org/wiki/James_Harrison_(blood_donor)
Well, since it seems likely there's now a scent-based marker, we could train dogs to sniff Alzheimer's out, but we'd still want to know more about the molecule and biochemistry for treatment purposes.
Why not just train a dog to smell it?