His laboratory issues a one-page report—“that they sometimes don’t read,” Dr. Frykman says—that does indicate how to act on discordant results. “But in addition, we do contact physicians if the results are very discordant,” such as a positive ratio but a negative p-tau217. “That is definitely a discussion we initiate, which I’ve had a few times now.”
Laboratories are still figuring out how to report results, Dr. Tran acknowledges. “Do you give out the actual numbers? Do you give out the individual biomarkers, versus the ratio, or both? That’s all still evolving.”
Individual neurologists will likely have their own preferences, he continues. But those preferences should be developed with help from the laboratory. “Partnerships are not optional. Remember,” Dr. Tran says, “you are your friendly, local, neighborhood pathologist or laboratorian. You need to maintain that visibility and be part of the team.”
(There’s another reason laboratories should forge these partnerships, he says. “You’re going to need a number of samples that span the analytic measurement range of the test.” Highs may be relatively scarce—“fortunately,” he says, “because most people are relatively healthy.” But that means it could take a while to amass enough samples. If clinical colleagues know what the laboratory needs, they’ll be more likely to obtain an extra blood sample for validation purposes.)
The laboratory also needs to work closely with clinicians to define appropriate cutoffs, Dr. Tran says, and clarify, for example, what an FDA-approved cutoff means “versus coming up with your own.” This will become “hugely important as 181 p-tau becomes more widely used. Because now you’re not just educating the superstar neurologist. You’re educating the superstar primary care doctor.”
The cutoffs will also evolve as everyone gains experience with the test, he continues, and as the assays improve. Neurologists may not want an overly sensitive test, Dr. Tran says, because they won’t want to be excessively burdened by false-positives. Primary care doctors, on the other hand, may find higher sensitivities valuable because of the reduced number of false-negatives.
Clinicians will need to make the implications of these tests clear to patients. Drawing on his own experiences with family, friends, and others, Dr. Tran says some are enthusiastic about getting the test. “They think it’s awesome and want the test now—and they’re 30 years old,” he says. Others, often older, express no interest in what the test might reveal. Helping educate clinical colleagues about the test and using it in a shared decision-making context will be invaluable.
In many ways this is the gumshoe work that takes place for most laboratory tests. As Dr. Shaw puts it, “These are basic, basic things. The lab needs to be engaged with clinical colleagues before they set this up, when they set this up, and after they set this up. How is it performing? Is it providing the expected utility? Have any unexpected issues cropped up? Are the sample preanalytics in order?”
Diagnosing Alzheimer’s disease, however, brings added burdens.
“The reality is,” Dr. Tran says, “we don’t want a situation where we tell someone they may have Alzheimer’s,” but the result is a false-positive. The stress on patients and their families can be enormous; moreover, it creates unnecessary inefficiencies in the health care system, including referrals to neurologists, superfluous PET scans, and so on. “Those are the broader costs we have to worry about.”
Dr. Shaw sees a bigger picture as well. Even as he acknowledges the importance of individual laboratories validating tests, working with clinicians to establish cutoffs, and so on, he’s eagerly awaiting the guidance document from the Association for Diagnostics and Laboratory Medicine. “It’s in review,” he says. “It’s coming.”
As Alzheimer’s disease testing moves, figuratively, from spring to summer to fall in the coming years, many hold out hope the tests will address the entire spectrum of disease, from early screening to late-onset disease, as well as the accompanying therapeutic biomarkers.
“Once you have a positive patient you want to stratify them,” Dr. Frykman says. Does the patient have amyloid pathology only? Or also tau pathology? If the patient has Alzheimer’s and tau pathology, how much? What drugs might be suitable? Which aren’t? “The ideal patient currently is the patient who has amyloid but not tau,” he says, adding: “We go through the first phase, which is, Wow, with the blood test, we can actually pick out the needle in the haystack. But going forward we’ll have to be more specific about what is driving the disease.”
Some researchers are suggesting that neurofilament light chain, or NfL, might prove useful in showing disease progression, since it’s a biomarker of active neuroaxonal damage, says Dr. Tran. “Clearly there will be new FDA-approved biomarkers, such as NfL and APOE. Maybe p-tau231 will rear its head. We will also see the introduction of more digital biomarkers, and, as mentioned, AI will no doubt play a role in helping with clinical decision-making, risk-stratification, and, on the research side, perhaps further biomarker discovery.”