5 Questions About Tau-PET as a Precision Medicine Tool in Alzheimer Disease


Alzheimer disease (AD) progression varies widely among patients. But new research may aid in predicting a patient’s atrophy. 

Findings of a new study1 suggest that tau-positron emission tomography (PET) signal can predict brain atrophy at later stages for individuals in the early symptomatic stages of AD. The researchers reached this conclusion by analyzing longitudinal magnetic resonance imaging (MRI) scans acquired at the time of PET and 15 months later.

Neurology Consultant asked the study’s lead author, Renaud La Joie, PhD, a postdoctoral fellow in the Weill Institute for Neurosciences at the University of California, San Francisco, about what the findings may mean for the future of tau-PET as a precision medicine tool in the prediction of AD progression as well for the future design of new therapies.

NEUROLOGY CONSULTANT: You hypothesized that tau is a key driver of neurodegeneration in AD. In what ways do your results support your hypothesis?

Renaud La Joie: Our data shows that the tau burden measured with Flortaucipir-PET is predictive of how much brain atrophy is going to happen in the near future (ie, how fast the brain is shrinking over time). The data also showed that the regional distribution of tau pathology is variable across patients and greatly matters because the topography of tau precisely forecasts which part of the brain is going to shrink in the future. Brain degeneration—as measured by the shrinkage we can see using MRI—really follows tau (measured with PET).

This data definitely support the hypothesis that tau drives neurodegeneration locally, though our work is by definition observational, so our ability to infer causal relationships is limited.


MRI scans
Tau-PET brain scans (green) in patients with early clinical-stage AD accurately predict the location of brain atrophy measured by MRI 1 to 2 years later (magenta). Amyloid-PET imaging (blue) does not predict the location of either tau or future brain atrophy. Image courtesy: Rabinovici Lab / UCSF


NEURO CON: How do you suggest clinicians implement the study’s findings into their everyday practice?

RL: The current findings cannot be immediately translated into clinical practice for two main reasons. First, tau-PET does not have US Food and Drug Administration (FDA) approval for clinical practice at the moment; it solely is a research tool. Second, our research showed that tau-PET can forecast how much the brain is going to shrink in the near future—and, specifically, where. However, we did not have enough data to test whether tau-PET could help predict what matters the most to the patient: their upcoming clinical decline, not only in terms of how fast the decline is going to be, but also what kind of decline should be expected, such as memory deficits or language problems. This is what we are looking at now. Based on the strong literature on the tight relationship between cortical atrophy in specific areas and cognitive deficits, we definitely expect tau-PET to help with clinical prognosis, but we have to properly test this.

NEURO CON: How can your research on the association between age, tau levels, and atrophy help clinicians better understand how to manage AD?

RL: Existing literature indicates that early symptom onset is usually associated with a more aggressive disease; for instance, patients with early onset symptoms usually have a steeper cognitive and clinical decline than patients who develop symptoms in later life. Here, we showed that this difference between early and late-onset AD can be at least partly explained by the fact that younger patients have much more tau pathology in their brains.

The other thing we saw is that, in older patients, the distribution of tau at baseline was not as good at predicting future brain atrophy as it is in younger patients. Structural brain changes still tend to happen in tau-rich regions in older patients, but the overlap was more striking in younger patients. We think this reflects the likely presence of copathologies in older patients—the presence of other proteinopathies or vascular injury that also contribute to brain damage and cognitive deficits. In older patients, atrophy is probably the result of multiple etiologies, and that is why we cannot predict it so well if we only consider tau pathology. This is a good reminder that in older patients, mixed etiologies are probably involved and that effective treatments will likely have to target multiple pathways to be fully effective. 

NEURO CON: One of your goals is to develop noninvasive brain imaging tools that would allow clinicians to see whether the location of tau buildup early in the disease predicts later brain degeneration. How has this study helped you in achieving this goal? What further work is needed to fully achieve the goal?

RL: The present study is a good proof of concept that tau-PET can help forecast future brain degeneration. But this was a small study conducted among 32 patients recruited from an academic center. We need to replicate this finding in much larger populations and in more representative samples to ensure the validity of our approach for larger-scale use.

NEURO CON: What might these findings mean for the development of new AD therapies and the future of patient care? 

RL: We think this has several applications. First, it highlights the role of tau as a main driver of local neurodegeneration (as observed with structural MRI) at this symptomatic stage. In turn, it really emphasizes the need to target tau in this population.

Next, we think that our data show a potential role of tau-PET in the design of clinical trials. The amount of tau-PET signal at screening could help stratify patients, because it is likely that patients with different amounts of tau pathology at baseline will have different clinical trajectories; the patients may potentially benefit differently from a drug. Then, tau-PET could help design more individualized ways to monitor disease progression and detect a drug effect. Two patients with different tau-PET patterns at baseline (eg, left cortical vs very medial/inferior temporal patterns) will have different patterns of brain atrophy and will likely develop different cognitive deficits (eg, predominant language vs memory). With tau-PET, we could be able to better predict what the disease course is going to be for a given patient and see if he or she progressed as expected.


  1. La Joie R, Visani AV, Baker SL, et al. Prospective longitudinal atrophy in Alzheimer’s disease correlates with the intensity and topography of baseline tau-PET. Sci Transl Med. 2020;12(524). doi:10.1126/scitranslmed.aau5732.