Talking about Tau: The Role of Tau PET Scans in Alzheimer’s Research and Care

Image

Accreditation Statement

In support of improving patient care, the University of Wisconsin–Madison ICEP is jointly accredited by the Accreditation Council for Continuing Medical Education (ACCME), the Accreditation Council for Pharmacy Education (ACPE), and the American Nurses Credentialing Center (ANCC) to provide continuing education for the healthcare team.

Credit Designation Statements

For more information about continuing education credit for this episode, visit the ICEP online learning portal at https://ce.icep.wisc.edu/dementiamatters2025/taupetscan

The accreditation for this course expires 4/7/2026. After this date, you will no longer be able to access the course or claim credit.

Claiming Credit

An ICEP account is required to claim credit and to complete the episode evaluation form. Create an account

To claim continuing education credits using the texting function:

• At the end of the podcast episode, you will receive information about texting in a code
  with your cell phone.

To claim continuing education credits through the ICEP online learning portal.

• Visit https://ce.icep.wisc.edu/dementiamatters2025/taupetscan. Go to the Register tab, click the Begin button and under the Course Progress section, click on Attendance Code -> Start and enter the text code given at the end of the podcast episode.

To complete the episode evaluation:

• Go to the Register tab, click the Begin button and under the Course Progress section, click on Episode Evaluation -> Start

Learning Objectives

As a result of participation in this educational activity, members of the healthcare team will:

• Define the role of tau in the amyloid hypothesis
• Describe the role of tau PET scans in Alzheimer’s disease detection
• Describe the challenges for using tau PET scans in clinical settings

References

• Braak H, Braak E. Neuropathological stageing of Alzheimer-related changes. Acta Neuropathol. 1991;82(4):239-259. doi:10.1007/BF00308809
• Jack CR, Andrews JS, Beach TG, et al. Revised criteria for diagnosis and staging of Alzheimer's disease: Alzheimer's Association Workgroup. Alzheimer's Dement. 2024; 20: 5143–5169. https://doi.org/10.1002/alz.13859
• Klunk WE, Engler H, Nordberg A, et al. Imaging brain amyloid in Alzheimer's disease with Pittsburgh Compound-B. Ann Neurol. 2004;55(3):306-319. doi:10.1002/ana.20009

Transcript

Intro: I'm Dr. Nathaniel Chin, and you're listening to Dementia Matters, a podcast about Alzheimer's disease. Dementia Matters is a production of the Wisconsin Alzheimer's Disease Research Center. Our goal is to educate listeners on the latest news in Alzheimer's disease research and caregiver strategies. Thanks for joining us.

Accreditation Intro: This episode is approved for continuing education credits for physicians, physician assistants, nurses, and other members of the healthcare team through the Interprofessional Continuing Education Partnership at UW–Madison. At the time of this recording, Dr. Gil Rabinovici discloses the following relevant financial relationships with ineligible companies. He is a contractor with Eli Lilly and Company, F. Hoffmann-La Roche, Johnson & Johnson Health Care Systems Inc., Novo Nordisk, Merck. Dr. Rabinovici also discloses grants with GE Healthcare, Life Molecular Imaging, Avid Radiopharmaceuticals, Inc. Genentech USA, Inc. I, Dr. Nathaniel Chin, have the following relevant financial relationships with the ineligible company as a consultant: NewAmsterdam Pharma and Eli Lilly and Company. Information on how to claim credit will be shared at the end of this episode. Additional continuing education information is provided in the show notes.

Dr. Nathaniel Chin: Welcome back to Dementia Matters. A key focus of Alzheimer’s disease research is early detection, as the earlier it is detected the more effective interventions and treatment can be. One indicator of Alzheimer’s is a buildup of the protein tau in the brain. This protein can be detected with a tau PET scan, which uses a radioactive tracer to detect the protein in the brain. Today, I’m joined by Gil Rabinovici, one of the foremost researchers into this technique. He’s the director of the UCSF Alzheimer’s Disease Research Center, study chair of the Imaging Dementia-Evidence for Amyloid Scanning (IDEAS) and New IDEAS studies, as well as co-principal investigator of both the Alzheimer’s Network for Treatment and Diagnostics (ALZ-NET) and the Longitudinal Evaluation of Alzheimer’s Disease Study (LEADS). He’s also an associate editor for the Journal of the American Medical Association and has authored over 350 peer-reviewed publications. Gil joins us today to tell us more about tau and the tau PET scan, its importance in the field of Alzheimer’s diagnosis, and the potential challenges surrounding its use. Gil, welcome to Dementia Matters.

Dr. Gil Rabinovici: Thank you, Nate. It's really a pleasure to be here. I'm a big fan of the show and thrilled to be on.

Chin: I am looking forward to picking your brain on a very complex topic, and I know you can explain it for us in easy-to-understand terms. So first, Gil, I want to go a little bit into you. How did you get into this field of Alzheimer's disease and other causes of dementia, and then specifically into imaging techniques like PET scans?

Rabinovici: Sure. I think my interest in the brain and neuroscience really began, like for many people, with a personal connection. Right before I was ready to go to college, a good friend of mine was in a terrible car accident and sustained a traumatic brain injury and was left with a really dense amnesia where he couldn't remember really basic details about his life and past events. He made a remarkable recovery, and his stories about his recovery really captured my interest. His memories came back to him, not slowly, but in big floods. He would describe coming back from the hospital to the room where he had basically lived since he was five years old and just having a flood of memories of different things that happened in the room come back to his consciousness. He would pick a book off his bookshelf and have no recollection of having read the book, and just reading through the first few pages the entire plot would come back to him. I just found this so fascinating. Like a lot of future neurologists, I think I started to read Oliver Sacks and got interested in brain behavior relationships. Eventually, through some really terrific teachers and mentors like Bob Sapolsky and Bruce McIver when I was an undergraduate at Stanford and Marcel Mesulam when I was a medical student at Northwestern, I got very, very interested specifically in how different brain diseases can impact core elements of our human experience, like memory and language abilities, spatial and visual abilities. I was fortunate enough to do my neurology residency at UCSF under the mentorship of Dr. Bruce Miller, who's really been a major luminary in the field of behavioral neurology. Bruce got me very, very interested in these different neurodegenerative diseases. When I was a resident, I was in the neurology library and I picked up a journal called Annals of Neurology and happened to read a paper by Bill Klunk and colleagues, which described the first amyloid PET scans in living patients with Alzheimer's disease. This really captured my imagination. The ability to detect this core element of pathology that I had been taught could only be really determined at the time of death–I thought it had enormous potential for understanding the disease better and also for helping with clinical care. So again, just through serendipity and luck, Bill Jagust had just moved back from UC Davis to UC Berkeley and we were discussing a collaboration with him around doing amyloid PET scans with the first tracer called PiB. Bruce suggested that I take leadership of this project and here we are, 20 years later. So really, serendipity, luck and right place, right time, mostly.

Chin: I mean, that's an incredible story, Gil. It starts out kind of sad, and I appreciate you sharing about your friend. That does seem very tragic. Although, of course, a positive outcome in that he started to recover. Thank goodness, young brains seem to do well. But then, it kind of does seem like the importance of mentorship, you know, you did encounter a lot of great mentors and they helped you figure out what it is you want to do. It is hard for me to believe that it was an article that pointed you in this direction, but I mean, it is your story, Gil. An article pointed you here and now you're a leader, a thought leader, in the field of tau PET imaging.

Rabinovici: Yeah, luckily the pager didn't go off just as I was opening the journal or we might be in a different multiverse.

Chin: (Laughs) Okay, so before we dive into tau PET scans, can you explain what tau is and then how it fits into the amyloid cascade hypothesis, which focuses on the buildup of proteins like amyloid as the main cause for Alzheimer's disease?

Rabinovici: Yeah, sure. Tau is a protein that's normally present in neurons. In healthy neurons, it binds to microtubules, which are a key element of the cytoskeleton or structure of the cell. They help maintain the cell of the neuron. They help get nutrients and energy and other biological factors transported back and forth within the neuron. Tau probably has many other functions, actually. It's involved in cell signaling pathways, in the development of the brain, and the response of the brain to different insults or injuries. What happens in Alzheimer's disease is that this normally present protein of tau starts to undergo some modifications. There are phosphate moieties or phosphorylation, also methylation that occurs on this protein. As a result, it disassembles or leaves the microtubules and starts to accumulate. Initially, it forms soluble deposits that we call oligomers. These are just sticky tau, different amounts of tau sticking to each other and floating around but eventually it starts to form structures that we can detect under the microscope, and these are the neurofibrillary tangles that Alzheimer himself described in his first case report as something remarkable that he saw in the neurons. This tau buildup is occurring in all of us as we get older, especially in the part of the brain called the entorhinal cortex, which is in the temporal lobe and is really important for memory. It seems to be almost inevitable for us to develop some of these tangles in the entorhinal cortex, but it doesn't seem to be too bad in terms of the impact that it has on our memory and function. What happens with Alzheimer's disease is you create this toxic cocktail that includes not only tangles, but also the amyloid plaques and the different forms of sticky amyloid called oligomers. It seems that if you're just accumulating plaques in the brain or just accumulating tangles you're probably going to function pretty well, but it's the combination of those two that really forms a cascade of events that eventually leads to dysfunction of neurons and synapses, degeneration of synapses and cell connections, and this translates into cognitive decline and ultimately into functional decline and dementia. It is that synergy between amyloid and tau that seems to be the toxic recipe for Alzheimer's disease. There are other elements involved in this biology, like inflammation in the brain that might play a role in propagating these different proteins or in the predisposition to accumulate them. It's a pretty complicated biology, but tau for sure plays a central role.

Chin: So knowing that, though, isn't that kind of a bad name to say amyloid cascade hypothesis because it is this combination of amyloid and tau?

Rabinovici: I think this comes from really the genetic findings, which tell us that all the mutations that lead to familial autosomal dominant heritable forms of Alzheimer's disease in humans are involved genes that are directly involved in the processing of the amyloid protein. If you express these genes in mice, the mice get plaques and eventually they get mouse-heimer’s disease. They develop memory decline and functional changes. It's interesting that even though tau and tangles are very central, as we'll discuss, to the evolution of Alzheimer's disease, people who have mutations in the tau protein itself don't develop Alzheimer's disease. They develop a familial form of a different type of dementia called frontotemporal dementia. This is a disease that primarily affects the frontal and temporal lobes of the brain and presents clinically not with memory loss, but actually with changes in personality, behavior, social behavior, emotional behaviors or with changes in language abilities. So, again, mutations in the tau protein cause frontotemporal dementia. Mutations in genes associated with the amyloid protein cause Alzheimer's disease. But in Alzheimer's disease, the accumulation of both these proteins is central to the evolution of the disease.

Chin: You answered one of my questions, which was, ‘can you have this abnormal tau protein without amyloid?’ And the answer is, yes, and it's just a different disease which we are looking at when we think about dementia but it is a different presentation, a different condition. In your answer to my first question, you brought up methylation, phosphorylation. And I think that's really important because in the news, we hear a lot about phosphorylated tau and there's this new test. It's not really that new, but it's new in blood, p-Tau, p-Tau 217. Is this the same thing in blood that you look at in a tau PET scan or are they different?

Rabinovici: They're related, they're not exactly the same. The fact that there are modifications in the tau protein allows us to detect it in very specific ways. For example, when a neuropathologist is looking at brain tissue, they're actually staining the tissue with antibodies that are detecting these abnormal phosphorylation sites on tau, and that's how they pick up tangles and other forms of tau accumulation. In the cerebrospinal fluid and now in the blood, we can detect fragments of tau that have these phosphorylation changes, and these seem to be very specific to Alzheimer's disease. We can, for example, in the blood or spinal fluid measure increases in the concentrations of tau that's phosphorylated at position 217 or position 181 in the protein. When you detect those elevations, those seem to be very specific for Alzheimer's disease tau in particular, not other forms of tau that you see in other brain diseases. In the tangles, there are accumulations of the tau protein. Actually, this 217 or 181 site might actually get spliced out so it might not actually be present in the tangle itself. That's the reason why it's present in the fluids like the spinal fluid or the blood. What we're detecting with the PET scans is something a little different. The PET tracers are small molecules that are labeled with a radioisotope and they bind to folding pockets within aggregated tau protein. As tau forms these tangles, it forms fibrils that are actually folded structures. What the PET tracers do is they bind in those folds to the tangles if they're present in the brain and then they set out a radioactive signal. That's what the scanner is detecting is actually that radioactivity. The PET scans are detecting tangles, not specific phosphorylation sites in tau, and they're detecting accumulated tau, whereas in the blood or spinal fluid we're really measuring one tau protein at a time that is altered in this way through phosphorylation or other changes to the protein structure.

Chin: So given your role, too, in the IDEAS study and now what you're talking about with  tau PET, how do you see PET scans playing a role in Alzheimer's disease detection? And then, in that answer, how are amyloid and tau PET scans actually different in what they're measuring?

Rabinovici: Right, so when the neuropathologist, just to go back to kind of the roots of our field and how the the disease was diagnosed–even when I was a med student and a resident reading these articles, the only way to define Alzheimer's was to look for these accumulated proteins, the plaques and the tangles in brain tissue. Then in 2004, I read this article that was a huge breakthrough. It was an imaging technique, amyloid PET, that allows us to light up plaques in living people. What that showed us is actually we can detect these plaques not only in people who suffer from dementia due to Alzheimer's, but actually we can detect the plaques accumulating 15, 20 years before people have even the very earliest symptoms, so very early detection of these biological changes. The tau PET now allows us to detect that second piece of the disease, the neurofibrillary tangles–that other protein, the tau protein–that's depositing in the brain and light it up and show us where in the brain it is and how much has accumulated. So now, just with imaging two separate brain scans, we're able to actually replicate what pathologists could do by detecting plaques and tangles in the brain but not after someone has passed away, but in living people. This has really revolutionized our ability to study the natural history of the disease, see how amyloid and tau accumulation are related to each other, how together they lead to changes in brain structure and function, and ultimately how the combination of the two, even in people who have no symptoms and are cognitively unimpaired, really create a very heightened risk for people developing cognitive changes in the next few years. The research use of these scans has been really transformative, I think, in our understanding of the evolution of the disease in humans and ultimately in our ability to do drug trials and test different pharmacological interventions that are targeting the plaques or the tangles and understanding what that's doing to the biology of the disease. Clinical practice has lagged behind. Even though amyloid PET scans have been available for over 20 years in research studies, it's only recently that we've started to use them, I think, more frequently in clinical practice, often to determine if people have plaques in order to determine if they may be candidates for new therapies that are specifically targeting the plaques or if we need that clarity in helping to guide their diagnosis and treatment plan, even if they're not interested or aren't candidates for some of these new therapies. Tau PET scans were really developed first in 2013, 12 years ago, and the clinical use has been minimal I would say. Even though one of the tau PET techniques that was developed is FDA approved since 2020 for clinical use–this is tau PET tracer called flortaucipir, marketed under the name TAUVID–it really isn't available to most clinical PET centers, and most clinicians are not able to use it in their clinical practice. Tau PET, which has very strong correlations with disease progression and symptoms, its use has been limited largely to research settings but I do hope that that will change because I do think that there is potential for tau PET to help inform our diagnosis and treatment of patients.

Chin: When you think of tau PET when it comes to a person's clinical symptoms or just clinically, can you speak to the importance of the regional distribution or the spread of that protein in comparison to how we tend to think of amyloid as if you have it or you don't have it?

Rabinovici: Yeah, it's really true. Amyloid tends to pop up in large areas of the cortex of the brain, the surface of the brain, almost as this multifocal process. When you look at the scans, you can really see that because in most scans, the amyloid is either really diffusely present throughout the brain surface or it's not present at all. They're just–maybe 15% of scans or so show more of a regional distribution to the pathology. The tangles tend to progress in a stereotypical fashion. They progress from these memory areas in the temporal lobe and to other cortical areas in the back of the brain, the lateral temporal and parietal lobes. Ultimately, as disease advances they go into the frontal lobes, so they're very helpful for disease staging in the sense that if you see more tau in the brain on a PET scan and more distributed tau in the PET scan, that means the disease is more biologically advanced. In fact, this is represented in the revised diagnostic and criteria staging criteria that were put forth by the Alzheimer's Association last year, where amyloid PET is really used to diagnose the presence of amyloid and Alzheimer's but tau PET is available to stage the disease to see how much it has progressed. Some of the data from the clinical trials suggest that that might be very important in the sense that people who are treated with amyloid-removing antibodies and have plaques removed through these antibodies, the people who have lower stages of tau spread when they enter the trial tend to derive more clinical benefits. Conversely, if people entered the trial for an anti-amyloid antibody and they had widespread tau in the brain when they entered the trial, those individuals really derived much less clinical benefit yet were still exposed to some of the risks of the treatment. Tau PET, even in clinical trials that we're doing now, might help us with disease staging. Just like you determine cancer treatments based on staging–what treatment is available depending on how far the cancer is spread and where it is–it might turn out that the specific interventions that we recommend in Alzheimer's disease are also very largely driven by the stage of the disease. A technique like tau PET can help us determine who is likely to benefit from a specific intervention like removal of amyloid and who may be beyond that point and may be unlikely to benefit and still be susceptible to risks of treatment. The other really interesting thing is that where the tau is is very strongly correlated to the types of symptoms that patients have. If you see a lot of tau in memory areas in the brain, in the medial temporal lobes, those patients tend to have memory problems. And it's even more than that. If they have tau in the left medial temporal lobe, which is the more verbal medial temporal lobe, they have trouble with verbal memory. People who have tau in their right medial temporal lobe might have more trouble with visual memory or spatial memory, so it really is that specific. Patients who have a lot of difficulty with language functions and have aphasia as part of their Alzheimer's tend to have tau in language areas in the left hemisphere. Patients who have visual and spatial problems tend to have more tau in visual networks in the occipital lobes of the brain. So it's not just how much tau but where the tau is in the brain that is very strongly related to the types of symptoms patients have. This can be very useful from a clinical point of view in the sense that if you have a patient who suffers from unusual symptoms. For example, they have trouble with their visual and spatial function. Say you get an amyloid PET scan and the PET scan shows there's amyloid everywhere in the brain. Well, they have plaques and they probably have Alzheimer's disease biologically, but is it responsible for these unusual symptoms? Well, now if you do a tau PET and you see those visual processing areas in the brain lighting up, you're pretty convinced that it is, in fact, the tangles of Alzheimer's disease that are responsible for dysfunction of those brain areas and the symptoms the patient has. From a clinical point of view, there really is a strong correlation that can be very helpful to doctors caring for patients and for patients and families to understand what is causing their symptoms based on the location of the tau.

Chin: Yeah, I was just going to say, Gil, there's a lot of personal meaning to that. People want to know, what is causing this? What is causing my experience? Even if you can't treat it and cure it, what exactly is the reason for that? It seems to me you can actually map that with this impressive imaging, so there'd be just a lot of utility for the patient and the family.

Rabinovici: I agree. That has definitely been our clinical experience. You know, at UCSF, we have been disclosing research tau PET results to patients with Alzheimer's disease as part of some of our studies in our Alzheimer's Disease Research Center and also as part of some national studies like the early onset AD or LEAD study that you mentioned. Our experience has been that the tau PET piece of this is very important. When I can point to a location on the brain and say, “This area in the brain is really important for remembering words and names, and this high, bright red signal that I'm pointing to are tangles in those areas, and you have a lot of tangles in those areas. That is the reason why you're having so much difficulty remembering names or words.” That can be really powerful and, I think, empower patients and families to really understand their disease in a way that many people seek to do when they go and seek specialty care and memory clinics or when they become part of research studies to try to help advance our knowledge of these diseases. They're also very interested in understanding what's going on with their brain in addition, of course, to being motivated to contribute to research overall.

Chin: I have a couple questions for you about returning these results. Before I get to those, I did have a couple questions related to the timeline because I do feel like you describe a nice biological timeline of amyloid. Some time goes by, maybe this is where more inflammation is involved. There's tau. Some time goes by. There's brain cell death, there's synaptic dysfunction. You also mentioned that it can be 10 to 20 years before–we can identify amyloid 10 to 20 years before someone starts having symptoms. Where does tau fit into that? Like, when you start seeing that first early signal of tau PET in someone who's unimpaired, doesn't have symptoms, do you have a general idea of when they might start developing symptoms? Just a general idea?

Rabinovici: Yeah, so I just want to separate two things, which is when we can detect the tau versus when it's actually depositing in the brain. Let me talk first about the detection. It's true that we can typically first detect plaques in people and only detect the tangles with tau PET later on. When that occurs, the appearance of the tangles on PET scans is much closer to the onset of symptoms. If amyloid is detectable 15 or 20 years on average before people have symptoms, the tangles are detectable maybe more like six to eight years before people develop symptoms. In fact, we know that if people have detectable amyloid and detectable tau in the brain, about 50 percent of those individuals will progress to have at least mild cognitive impairment within about three to four years. So it's not 100 percent of people, but it's a very high proportion of people who have both of these pathologies on PET that are going to develop clinically significant cognitive impairment not in the far future, but in the relatively near future. There are a lot of things that can modify those timelines in individual people. Probably there are differences in each of our brains in terms of our resilience to having the pathology and how much we can compensate for the pathology before we develop symptoms. One of the major factors in that is, probably, what are some other pathologies that are developing in the brain that we can't detect with PET scans or spinal fluid tests or blood tests yet? Things like Lewy bodies–just starting to be able to detect those on spinal fluid. TDP-43–I know you had David Wolk present and he talked about LATE disease so that's a really important contributor to memory loss, especially in older people that we can't yet measure with imaging or fluid. Those are probably going to modify how much tau pathology you can have in your brain before you start to develop symptoms. Neither of the scans are detecting the first plaque that ever appears in the brain or the first tangle that ever appears in the brain. If the pathologists are looking at a brain after someone has passed away, they will probably identify these pathologies in areas where PET can't yet detect them. For example, if we're seeing tangles on a PET scan in the temporal lobes of the brain, the pathologist might be able, if the patient passes away, to see the tangles already have spread to the frontal lobe. There's just not enough of them there yet for us to detect them with imaging. I really think of what we see with imaging as the tip of the iceberg. There's probably quite a significant amount of pathology in any brain area before it starts to pop up on the PET scan and become detectable.

Chin: That was my second question for you is accuracy, or how good these scans are, because I think a lot of our research participants and patients will say, “Are we at the beginning of this process? Is the technology really–is it rudimentary or are we at a point where this is as good as we're going get?” And then it's just a matter of finding other diseases. I mean, where do you feel like we are in the spectrum?

Rabinovici: You know, Nate, we're probably somewhere in the middle. We're not detecting the very earliest pathology. We are starting to detect pathology in that middle range or intermediate range on the neuropathological scales. It's in that range that the pathology starts to become clinically meaningful to patients. And so the way that pathologists stage tau is using a system called Braak staging, developed by Braak and Braak, which takes advantage of the stereotypical progression of tangles and classifies the spread of tangles based on where they are in the brain at stages one through six. The PET scans are probably starting to pick up the tau pathology around Braak stage four, so somewhere in the middle of that scale. That's the stage where tau pathology starts to become clinically meaningful, where most people or the average person will have some mild cognitive impairment with that level of tau. More reliably, the PET scans are detecting Braak stages five to six, which are very strongly associated with clinically meaningful cognitive impairment. So we're not picking up the very earliest pathology; we're probably picking it up somewhere in the middle. It would be great if we had the tools to pick it up even earlier, but I think we are starting to pick it up at the stage where it becomes clinically meaningful.

Chin: Gil, everything you've said about the importance of tau PET, but also the unique perspective in the cases that you've described, where do you see tau PET scans being used clinically? Who's ordering these? What circumstances do you think they're going to provide the most utility?

Rabinovici: Right now, there is one FDA-approved tau PET tracer that you can order for clinical use, and that is the tracer flortaucipir, which is marketed under the brand name TAUVID. However, clinical access to the tracer is very limited. Even though it's being used quite extensively in research studies and in clinical trials, we, as clinicians treating patients, have very limited access to it. I'm hoping that will change because I do believe there is a clinical role for tau PET. I think it falls into two broad categories. The first is staging of disease. We know that amyloid reaches an early plateau, by which I mean it kind of saturates the binding even before people sometimes meet criteria for mild cognitive impairment, and certainly by the time they have dementia. Tau PET is more dynamic throughout the disease course, and the more intense the binding and the more widespread the binding in the brain, the later the disease stage, the more the tau tangles have progressed through the brain, that's telling us something important about the progression of disease. In the new criteria that have been put forth by the Alzheimer's Association for diagnosis and staging of AD, tau PET is really used for staging the disease. There are four categories. Category zero is if someone has a positive amyloid biomarker but doesn't show any evidence of tau on tau PET. Stage A is when there is binding on the tau PET, but it's limited to those memory areas in the medial temporal lobe. Then stage C is when it has spread into the cortex, but there's still only a moderate intensity in the cortex. Finally, stage D, which is the most advanced stage, is when there's high uptake of the tau PET tracer in the neocortex. So why is this important? Well, at least in one of the clinical trials of the amyloid-removing antibodies, a clinical trial called TRAILBLAZER-ALZ 2 that looked at the antibody donanemab, patients who had low to medium tau burden when they entered the trial had a better clinical response than patients who had higher uptake of tau PET. When we're trying to inform patients and discuss the risks and benefits, for example, of an amyloid-lowering therapy, knowing how far the tangles have advanced in the brain might help us predict who might derive more versus less clinical benefit from the treatment and that might make a big difference, especially in people who might be higher risk for side effects based on their genetics or their clinical stage. And so I think that is one use case where tau PET could help us right now in making and giving patients the best possible counseling about whether they are likely to benefit from a treatment and whether that benefit might outweigh the risks of the treatment. The other broad use case is people who have atypical clinical presentations. For example, they don't present with memory loss but they have trouble with their language abilities or even sometimes their visual or spatial abilities. We know that sometimes Alzheimer's can present in unusual ways. In fact, the primary language problem or a primary visual problem can be the primary manifestation of Alzheimer's but it's unusual. This is a case where tau PET can really help because the location of the tau PET uptake correlates very strongly with the brain areas that are affected and the symptoms of the disease. If you have someone, for example, who has visual problems and you do a tau-pet scan and it shows really high uptake in those visual areas in the occipital levels of the brain, I think that really can convince you, and ultimately the patient, that the symptoms are very much due to Alzheimer's disease. However, if you don't see much tau PET uptake or if it's not in those areas, you might think that there may be a different brain disorder that's contributing to those symptoms. That might really make a difference in how you ultimately diagnose and manage the patient. I think for these unusual cases and then sometimes–especially in borderline cases in patients who are being evaluated for the novel amyloid antibody therapies, those would be the most immediate use cases that I can see for tau PET in the clinic.

Chin: Yeah, Gil, as you're explaining that, I just had a clinical case of someone with mild cognitive impairment. We know that mild cognitive impairment is such a broad stage and you can be in early, in the middle, the late of it. This person said to me, “Well, you're telling me I have elevated amyloid and I'm thinking about the risks and the benefits. I just don't know. And I wish you could tell me if I'm going to be that person who responds and I'm going to gain this year of having more function. Or if I'm not going to respond, I don't really want to go into the infusion center every other week.” And they just were really struggling with it. We wanted that extra piece of information. So I completely understand that this could be really helpful in helping someone make the best informed decision.

Rabinovici: It's early days. I mean, there's really only one trial where tau PET was used for stratifying patients and outcomes but there are signals of this in other amyloid trials, even though not everyone had tau PET. For example, people who were in the Clarity AD trial of lecanemab, some of them had tau PET and it tells a very similar story, which is that the people that were in really early stages of tau spread are the ones who seem to have benefited the most. I think that this is something that we may be using in our clinical decision-making in the future if this technology is clinically available and reimbursed.

Chin: But like you just said, though, tau PET isn't just a binary–you have it, you don't have it–which is sort of what we've been saying, at least lately, with amyloid. In your experience, because you are sharing this with your research participants, how does this process go? What are you exactly saying to your research participants when you're returning these results to them?

Rabinovici: Yeah, we return both the amyloid and the tau PET results to research participants who are cognitively impaired. That's one caveat is that everyone to whom we're disclosing these results has cognitive impairment, and it's just a question of is it Alzheimer's or is it something else? With that in mind, we disclose both the amyloid and the tau PET results. We start with the amyloid and then go on to the tau. We preface it by explaining, really, that Alzheimer's disease is defined by these two protein deposits, the plaques and the tangles, and that we've done a brain scan to try to evaluate each of those processes. Then we show individuals–before we show them their own scan–we show them examples. For amyloid, for example, a prototypical scan of someone who has non-elevated amyloid where all the binding of the tracer is in the white matter and then a prototypical example of someone with elevated amyloid where you see a lot of uptake throughout the cortex. Then after showing them those examples, we show them their scan and we point out, well, your scan looks more like scan A, which is the non-elevated, or scan B, which is the elevated. I think that helps them really understand what they're looking at and gives them the context they need. Seeing that image tends to be very powerful. It's not just that we tell them what we saw, but we actually show it to them. Somehow our brains are very biased to believe what we see and so when you see that image, I think that can have a really powerful effect for patients and families. Usually it's very empowering in the sense that they've come to our research study seeking an answer to what is going on with their brain. This is part of the reciprocity of us providing them with that answer. With tau PET, we go one step further sometimes, especially if people have unusual symptoms, but even not. We point to different brain areas and we say, “You see all that red signal here in the temporal lobe. This is an area of the brain that's really important for recent memory, for short term memory. And so you've told us that you have a lot of problems with your recent memory. In fact, seeing all these tangles really makes sense because this is the brain area that seems to be giving you problems. This is in fact where we're seeing the tangles.” And so, making that direct link for the patient and family about neuroanatomy and clinical symptoms can be helpful. As long as it's broached in an approachable way, I think people retain it. They understand it and it helps them understand their own disease better, which I think is always empowering for patients.

Chin: And you might have mentioned this earlier, but do you tell patients with mild cognitive impairment about the studies that have shown the power of having tau when it comes to sort of the general predictive ability that a certain percentage of people who have this tau are likely to progress to having dementia? Or is that something that still stays in the research space that you just don't think is needed at this point?

Rabinovici: We talk about prognosis. Based on the knowledge that we have now, I think we're pretty rudimentary in saying if you have both amyloid and tau PET signals, that makes it more likely that you're going to progress to dementia in the next few years compared to being positive just on one scan or the other. I think we have the data now that shows that pretty clearly, even for cognitively unimpaired people to whom we're not yet disclosing these results. I think the data are pretty strong that those that have both amyloid and tau PET signals are at quite high risk, maybe about 50 percent of them will progress to developing clinically relevant cognitive impairment, so at least MCI if not dementia, within the next three to four years. Tau PET can probably tell us even more. It tends to be the most powerful prognostic tool that we have–more than amyloid, more than MRI, sometimes more than the memory test that you do at baseline–in who is likely to progress but I don't think that we yet have the clinical tools to provide more resolution than just elevated or non-elevated. I hope that the staging system that was proposed by the Alzheimer's Association–we should test it out and see if, even giving people stage A, B, C, or D of tau PET, we can help give them more specific and precise prognostic information than just elevated or non-elevated. I think that still needs to be shown prospectively, but I think that might be a very good methodology for giving patients really something they almost universally are seeking once they have diagnosis, which is prognosis, and something that we, even as a dementia specialist, are not very good at at providing patients. Even though they really want it, we often are very careful about what we say, and rightfully so. We don't yet have the tools to be able to give people more precise information about what's likely to happen in the next one year, two years, three years. For obvious reasons, patients and families really need that information for care planning. I think tau PET could be one of the instruments that could in the future help us with this and do a better job of this.

Chin: And so given how impactful this could be, what are some of the challenges that come up with tau PET scans and why are we not seeing it used more broadly in the clinical setting?

Rabinovici: Tau PET scans are a little more difficult to read than amyloid PET scans. The FDA approved methodology for reading them is quite labor-intensive actually on the nuclear medicine physician or the radiologist who's trying to interpret the scan. You have to draw regions in the brain and re-scale the images. It's a little complicated. I think it's a tool that most clinicians have not been exposed to yet. I don't think it's insurmountable, but I think we need to develop easier to implement methodologies for interpreting these scans. I think quantification will really help. As opposed to amyloid where a global measure like a centiloid value, which tells you roughly how much amyloid is in the brain–for tau PET, we're really going to want regional values. How much tau PET is in those memory areas in the medial temporal lobe? How much is in the parietal lobe or occipital lobe or frontal lobe? I think developing those tools is going to be important. And then there's just availability and access. The companies, for whatever reason, have not prioritized developing a distribution network for commercial radiopharmacies. A lot of centers, even major academic centers and even in major urban areas, there's just not really access yet to the radio tracers. There's the whole issue of cost and third-party payer reimbursement, which still has to be sorted. It's not obvious whether Medicare, for example, will cover tau PET. For amyloid PET, they made a universal decision in 2013. They said, we're only covering it as part of research studies that will help us develop evidence about clinical utility. In 2023, they rescinded that and now amyloid PET is covered by Medicare in many areas, though it's a bit contractor and geography dependent. For tau PET, there's never been a national coverage decision, so it really is going to be up to the individual Medicare contractors and Medicare Advantage plans and private insurance plans to determine whether they will cover it. I would argue that if it could tell you if someone is unlikely to benefit from a very expensive treatment like an amyloid antibody treatment that may also have significant resource needs around side-effect monitoring, it would make sense to cover the tau PET–you identify people a priori who are unlikely to respond–but those types of economic decisions haven't yet been made and policy decisions around coverage. That's another challenge that I foresee but hopefully we're moving. We're making progress. I'd say we're making rapid progress towards seeing some of these biomarkers getting into clinical practice now, which is great. That applies to the blood and I think much more so the PET scans. At our center at UCSF, we're ordering something like 30 to 35 amyloid PET scans a month. Usually the clinical question is around treatment eligibility for new therapies, but not always. There are other clinical scenarios where amyloid PET can be helpful even if the patient is not interested in treatment or is not a candidate for treatment. It's just great to see us being able to translate some of these amazing tools that have been a cornerstone of our research for so long and actually taking them also into our clinic when we wear our clinician hat and seeing how they can benefit patient care.

Chin: Well, you mentioned research, Gil, so I just have two questions for research to wrap things up. Do you think tau PET scans are going to be used more often in clinical trials, given their potential in helping stratify who might respond the best or other potential use cases?

Rabinovici: You know, Nate, I really think it would be wise to use tau PET in clinical trials, both trials that are targeting amyloid or other elements of the disease and certainly drugs that are targeting the tau tangles themselves. There's more and more different approaches to trying to target the tangles in trials. I think it makes sense to stratify patients using the PET scan and to see whether people respond. For example, they may only respond in the early stages of tau accumulation and a treatment might be very impactful at that stage, but if you treat patients who have very advanced tau in the brain, the treatment may no longer be useful. We'll never know unless we actually characterize the tau PET binding in those people. I'm a little bit worried because what I'm seeing as I listen to different clinical trials that are being developed now is there's a desire to do everything with blood biomarkers. Blood biomarkers are great. They're much more accurate than I ever would have predicted a few years ago, but there's one thing they–at least the markers that we have now–are not very good at, which is staging disease and telling us how much the tangles have progressed. If you're looking at one of these phosphorylated taus that we talked about earlier–say you're looking at p-Tau 217–you can have, say, an elevated measure and it doesn't actually tell you if the patient is in very early stages of tau PET spread or very late stages. I hope that we're not going to take a step back and try to use blood biomarkers for cost-saving in a trial in a way that they really aren't supposed to be used and in a way that scientifically doesn't make sense. We're going to lose a lot of information. I mean for so many years, we did these clinical trials blinded to biology. We only diagnosed people based on symptoms. We learned that that is really not a very effective method for advancing drug development. And so I'm hoping that at least at the clinical trial stage, people will utilize this tool wisely in order to determine which drugs are working and worth pursuing and which ones are not, but time will tell.

Chin: You actually answered my second question, which is about tau drug therapy. It would make sense that–and I presume we are working on and developing drugs to target tau–that tau PET would be needed for those though. Doesn't that make sense, Gil?

Rabinovici: I think absolutely. It depends a little bit on the mechanism of the drug. We actually heard last year, for the first time, an antibody that's directed at the mid-domain of tau that was able to slow the progression of tau PET. That's the first example that I'm aware of a tau targeting antibody–so now an antibody that's really designed to target the tangles, not the plaques, that is slowing the progression of tau by intercepting, if you will, the toxic tau species as they're trying to spread through the brain. I think that's incredibly important in demonstrating target engagement. There's another exciting approach that's currently in clinical trials with antisense oligonucleotides, basically drugs that are infused into the spinal fluid–but I think soon will be infused into blood–that prevent the cells from translating the tau protein, from even creating the tau protein. With a very small sample–so a lot of caveats around interpreting these results–in the phase one trial of one of these antisense oligonucleotides or ASOs, they actually showed that the treatment was able not only to stabilize or slow progression of tau PET, but actually to reduce that tau PET signal. Now, whether that translates into a clinical benefit still remains to be seen, but I absolutely agree that tau PET is going to be part and parcel of trying to determine whether these different tau-targeting drugs are actually working and engaging the target and understanding who is responding to different types of drugs and who is not.

Chin: Well, I guess with that, Gil, if there's anything else, where do you hope research into tau and tau PET scans is going to be five years from now? If we do this podcast five years from this moment, where do you think we are?

Rabinovici: Well, like you, Nate, I'm a clinician, so I'm really hoping that we will have more research that will inform our clinical practice around tau PET and that, if we have this conversation in five years, we'll be able to use this as part of our clinical armamentarium, not in every patient necessarily but in select use cases. Developing methodologies for interpreting the scans in clinical practice, quantifying the scans in clinical practice, maybe helping with those prognostic questions that patients and families are really interested in; that would be one direction. As we discussed, the PET tracers are pretty good at picking up moderate to severe pathology but, of course, we want to do better and we really want to pick up the very earliest changes as tau is accumulating in the brain. I'm very hopeful that in five years we'll have even more sensitive PET tracers that are able to pick up that earlier pathology. And though we've mostly talked about Alzheimer's disease in this podcast, we alluded to the fact that tau is so central in many other progressive brain diseases–those related to head trauma, the chronic traumatic encephalopathy that we hear about in professional football players, people in fighting sports, military veterans. It's so important for the frontotemporal dementia syndromes that impact changes in personality, behavior and language. It’s really important for some of the atypical Parkinsonian syndromes like progressive supranuclear palsy, corticobasal degeneration. These are different forms of tau. And so the tau PET tracers that work in Alzheimer's disease don't work in these other tau disorders very well. There are now specific efforts to design imaging technologies that will allow us to pick up these different forms of tau and that will be very important for establishing diagnosis and developing clinical trials for those patient populations. That would be the third direction, which would be to really develop imaging tracers that allow us to image tau across the broad spectrum of this family of tau disorders, not just in Alzheimer's disease.

Chin: Well, it sounds like a very exciting future, Gil. Thank you. Thank you for being on the podcast, and we'll have you on before that five year mark, so I appreciate your time today.

Rabinovici: Thanks so much, Nate. Thank you. It's been a pleasure.

Accreditation Outro: As a reminder, continuing education credit is available for this episode through the Interprofessional Continuing Education Partnership at University of Wisconsin-Madison. To claim credit, you can text this code: NEMBOP at this number 608-260-7097. Again the number is 608-260-7097 and text this code: NEMBOP.  Your feedback is important to us. To complete an evaluation form for this episode, see the show notes.

Outro: Thank you for listening to Dementia Matters. Follow us on Apple Podcasts, Spotify or wherever you listen. Or tell your smart speaker to play the Dementia Matters podcast. Please rate us on your favorite podcast app. It helps other people find our show and lets us know how we're doing. If you enjoy our show and want to support our work, consider making a gift to the Dementia Matters Fund through the UW Initiative to End Alzheimer's. All donations go toward outreach and production. Donate at the link in the description. Dementia Matters is brought to you by the Wisconsin Alzheimer's Disease Research Center at the University of Wisconsin–Madison. It receives funding from private, university, state and national sources, including a grant from the National Institutes on Aging for Alzheimer's Disease Research. This episode of Dementia Matters was produced by Caoilfhinn Rauwerdink and edited by Alexia Spevacek. Our musical jingle is Cases to Rest by Blue Dot Sessions. To learn more about the Wisconsin Alzheimer's Disease Research Center, check out our website at adrc.wisc.edu. That's adrc.wisc.edu and follow us on Facebook and Twitter. If you have any questions or comments, email us at Dementia Matters at medicine.wisc.edu. Thanks for listening.