AGS Alzheimer’s Disease: New Diagnostic Technologies and Treatments Podcast

The Role of Neuroimaging in Diagnosis of Alzheimer Disease

• American Geriatrics Society • Season 1 • Episode 3

Share episode

Join Dr. Michael Harper, Professor of Medicine at the University of California, San Francisco and Dr. Kyra O'Brien Assistant Professor of Neurology at the Hospital of the University of Pennsylvania who, co-authored the module titled The Role of Neuroimaging in Diagnosis of Alzheimer Disease along with Jason Karlawish, MD, Professor of Medicine, medical ethics and health policy, and neurology at the University of Pennsylvania Perelman School of Medicine and Co-Associate Director of the Alzheimer’s Disease Research Center, and Co-Director of the Penn Memory Center.

To view a transcript click here then select the transcript tab. 

Michael Harper:

Hello and welcome. My name is Michael Harper. I'm a geriatrician at the University of California, San Francisco. And I'm pleased to be moderating today's podcast, which is part of the American Geriatric Society's new educational curriculum on Alzheimer's disease, new diagnostic technologies and treatments. And we created this curriculum because we know that the landscape around the diagnosis and treatment of Alzheimer's disease is really changing rapidly, and we want to make sure that our members and also our learners, our fellowship trainees, are kind of given the opportunity to kind of keep up to speed with what's changing so that they can provide counsel to their patients who may or may not be considering these new emerging treatments. And so we want to put that all into context for folks. Today I have the pleasure of speaking with Dr. Kyra O'Brien. Dr. O'Brien is a neurologist and an assistant professor in the Department of Neurology at the University of Pennsylvania Pearlman School of Medicine. She specializes in the diagnosis and treatment of older adults with memory disorders as a clinician in the Penn Memory Center. She's also an active researcher whose interests are focused on improving access to care for people living with Alzheimer's disease and related dementias and their care partners with an emphasis on the primary care setting and evaluating the clinical utility and implementation of advances in Alzheimer's disease diagnostics. And that's really why she's here today, because we're going to talk about her module that she co-wrote with Dr. Jason Karlawish, her colleague at the University of Pennsylvania. Their module is the role of neuroimaging in Alzheimer's disease diagnosis. So Dr. O'Brien, welcome.

Kyra O'Brien:

Thank you. Thanks for having me.

Michael Harper:

So although Jason's not here, when he and I began our careers in the 1990s, to make a diagnosis of Alzheimer's disease, we really had to rely on clinical criteria. And the only way to know for sure was if someone had Alzheimer's disease was with autopsy. So in the in the reading 30 years or so, a lot has clearly changed since then in our ability to sort of make a primordial diagnosis of Alzheimer's disease, particularly with the emergence of both fluid biomarkers and now these more advanced neuroimaging modalities, which is actually our subject today. But before we discuss some of those more advanced tools, I thought we'd start by talking about the role MRI plays in identifying the cause or contributors to someone's cognitive impairment. So what can you tell us about MRI and what its role is and what its limitations are?

Kyra O'Brien:

Certainly. Yeah. So if if a patient is identified as having cognitive impairment, they definitely should get an MRI, but an MRI has limitations. It allows us to see the structure of the brain. Sometimes we can see if there is shrinkage in certain regions associated with particular neurodegenerative diseases. We can see evidence of stroke buildup. We can see conditions like amyloid angiopathy, but we can't make a definitive diagnosis using MRI information alone. We also use it just to rule out other things that could be going on. Sometimes you pick up the incidental mass, uh, inflammatory changes. I would say that's rare, but that's another reason to get that kind of scan.

Michael Harper:

And does it help with, you know, obviously the reason we're talking about this subject is in large part because of these potential for treatment now if we identify Alzheimer's disease early in the course. Um but let's say we you see a scan that's not, you know, it's not consistent with Alzheimer's disease, at least not initially, but maybe it shows you some sort of evidence of other causes. Is that someplace? Is it sometimes enough to just sort of stop there?

Kyra O'Brien:

So I would never stop at a normal MRI or a quote unquote normal MRI. MRI changes, particularly shrinkage from neurodegeneration, happens fairly late in the disease course. So you can have a totally normal MRI, but still have the proteins built up in the brain. So again, the absence of structural changes in the brain doesn't mean that you don't have that disease.

Michael Harper:

Okay. So it's a good place to start, but depending upon the situation, it sounds like there may be opportunities to go further. Time to move on and ask you about metabolic PET. I think you know, for a lot of us who are not routinely looking at neuroimaging, these kind of modalities are still a little bit mysterious. I wonder if you could briefly tell us what metabolic PET is and what role it plays in diagnosis.

Kyra O'Brien:

Sure. So metabolic PET, rather than tell us about the brain structure, it tells us about brain activity. Patients get a tagged glucose molecule that we can then see how well it metabolizes that glucose molecule, hence the name metabolic PET. When we think about neurodegenerative diseases, we can see changes in brain activity usually decreases in activity in certain regions of the brain depending on which disease we're dealing with. So Alzheimer's has a particular pattern, Lewy body disease, frontotemporal degeneration, they all have patterns associated with them.

Michael Harper:

And are there, I assume, like MRI, there are limits to what we can tell?

Kyra O'Brien:

Yes. So while metabolic PET scans can give us a glimpse of something abnormal, something neurodegenerative going on in the brain, even before an MRI shows signs of that, they are still not able to confirm a diagnosis. So we still need actual evidence that the proteins that cause that disease are building up in the brain before we can say definitively this is Alzheimer's disease, for example.

Michael Harper:

Maybe we'll come back and do a little bit more comparison, but let's start by now some of these more advanced modalities that are available to us. So what can you tell us about amyloid and tau pets? How do they work and what can they tell us?

Kyra O'Brien:

Sure. So these imaging studies enable us to see deposition of amyloid and tau in the brain. So amyloid pet shows us if there are amyloid plaques, tau pet shows us if there are tau neurofibrillary tangles, and these are the two abnormal proteins specific to Alzheimer's disease. As you said before, previously we only found out if someone had Alzheimer's at autopsy when they actually looked at the brains of these patients and saw those proteins. But now we can see these proteins in living individuals using these scans.

Michael Harper:

How accurate are they?

Kyra O'Brien:

Amyloid PET is considered to be kind of the gold standard diagnostic for confirmation of amyloid deposition. Tau pet, there's, I think a little more work is being done in that area. Usually at the level where we can detect tau is when patients start to have symptoms. But there is some early buildup of Tau that might not be captured by a Tau pet early on. So still some development happening in that area.

Michael Harper:

And if you're caring for someone, you're evaluating someone and it's clear to you that they have mild cognitive impairment or early dementia, where do you start in terms of you still start with a plain MRI? When do you get metabolic PET CT? When do you get these amyloid or tau pets? How do you sort of think about that?

Kyra O'Brien:

Sure. So I usually start by considering what the most likely disease process is based on their clinical symptoms. Get my MRI. The MRI looks both for causes of cognitive impairment. So in particular vascular cognitive impairment, you can see those changes on MRI. And then I think about how this image is going to affect treatment. So for Alzheimer's disease, we now have anti-amyloid therapy. MRI is crucial in that consideration because if patients have too many microbleeds, they're not a candidate for those treatments. So that is another good use of the MRI. But let's say we get the MRI, there's no contraindication to anti-amyloid therapy. That's when I would go for a confirmatory biomarker because you need to make sure they have the target of the therapy or amyloid in their brain. I will say I've started to incorporate fluid biomarkers, blood biomarkers as a first step to try to reduce the number of patients going on to get amyloid PET, because even though it is covered by Medicare, it can still be kind of cost prohibitive for people. And we want to make sure we're not sending people on who are probably going to have a negative amyloid PET scan.

Michael Harper:

So as you probably know, we're going to do a we're we have a separate podcast. We're going to talk about fluid biomarkers. But in that vein, if you have positive fluid biomarkers, do you still then also need to get an imaging test confirming it?

Kyra O'Brien:

So some institutions are actually getting approval for anti-amyloid therapy with fluid biomarkers alone, but they have to be very good, have high sensitivity and specificity. Our center is a little more conservative. We still want to see confirmation of amyloid deposition on the amyloid pet. Other reason we want it is because when we monitor treatment response, you want to see how much of that amyloid has been cleared out of the brain. And so if you have that scan before they ever start therapy, you have something to compare to. Whereas it's hard to compare to a fluid biomarker level.

Michael Harper:

Oh, that's interesting. That's important. So that's so you would still you still think there's utility in getting it even if you have the diagnosis confirmed potentially confirmed by a fluid biomarker. So as you and I are both well aware, there continues to be debate among clinicians and probably, I imagine, among uh investigators about when and how to use these kind of diagnostic tools. Can you sort of alluded a little bit earlier, but can you maybe share some sort of use cases when when you would use these advanced modalities?

Kyra O'Brien:

Sure. So I would first only use it for patients who have cognitive impairment. Um so right now these tests are do have the ability to identify amyloid and tau deposition in the brains of people who are cognitively unimpaired. we call that preclinical Alzheimer's disease. We don't recommend use in that situation because there's nothing that we would treat that individual with. And so it's , you know, not right to do these tests if you can't manage a condition. Um that said, there are research, you know, trials going on right now looking at treatment of preclinical disease with some of these anti-amyloid therapies, so that could change in the future. Um, but for right now, we only test individuals who have cognitive impairment. And then in terms of uh uh treatment eligibility, you some patients still want to know if they have Alzheimer's disease. And you know, it does give us some information on what to expect in the future. Um, so that prognostic information some patients really value, even if they can't receive treatment. Um so for those patients, I might still consider getting an amyloid pet. The Tau pet is a little bit more nuanced. Uh, we're starting to think about using that to figure out how progress someone's Alzheimer's disease is, and therefore how much benefit they might have from anti-amyloid therapies. But those are much newer. We haven't implemented that across the board, and they also don't include Tau pet in the appropriate uh use criteria for these drugs.

Michael Harper:

So here's a question I think that a lot of my colleagues who are sort of say general geriatricians might want to know. So, how do I need to be able to interpret these studies? Or should I rely on the neuroradiologists who are reading these? How do I think about this as I'm a you know, as sort of a general geriatrician who's potentially advising patients about whether or not to get these studies and then how to how to interpret the results for them?

Kyra O'Brien:

Sure, that is a great question. So, and and it's even something we struggle with at our institution, um, particularly with metabolic amyloid and tau pet reads. Um you have to have a trusted neuroradiologist who really specializes in this area. Um so, you know, we have one or two in our in our institution who we run every scan by. I tell my patients, you know, don't worry about what the read is that shows up in your portal. I'm gonna talk with the radiologist and we'll let you know what they think. Um so it I definitely think there needs to be more training on the radiology side around these scans and how to interpret them. Um there are uh there is a push for amyloid pet uh to start publishing centiloid values in the reports. Um so this is you know a standardized scale for the level level of amyloid. Um but even then, depending on the program that you use to derive the centeloid value, it could be way off. I've had plenty of scans where they say it's negative, but on visual read it's clearly positive still. So find a trusted neuroradiologist.

Michael Harper:

And is there a sort of a typical error? Is it more typically an under-read or an overread, or can it really sort of be either?

Kyra O'Brien:

It can go either way.

Michael Harper:

Okay. So you really need to rely on someone who has looked at a lot of these and knows what they're knows what they're seeing, and and not to rely just on that report that you get unless it's coming from that person whom you trust.

Kyra O'Brien:

Exactly. Same for MRI reads. I always look at my own images and have built up some skills and looking for microhemorrhages and things that might disqualify someone from anti-amyloid therapy, but I still run it by my neuroradiology friend.

Michael Harper:

I want to get back because I think I'm I'm still having a little um lack of clarity about when the when you might get a metabolic PET, or if you had an MRI, you had the right clinical scenario, you had an MRI that wasn't clearly showing some other cause. Would you skip right to the amyloid pet, or is there a reason to still get metabolic PET first? I'm just trying to understand the sequencing here.

Kyra O'Brien:

Sure. So metabolic PET can be helpful if there is a question of is there a neurodegenerative process or not? So we have a lot of patients who have uh comorbid psychiatric conditions, severe depression, anxiety. And the question is, is that what's causing their cognitive impairment or is there actually something neurodegenerative here? So we'll often use uh metabolic pet in that situation. If there's something atypical about the clinical picture, like it doesn't clearly fit with Alzheimer's disease, we'll use it there. Um right now we also don't have specific biomarkers for frontotemporal degeneration, Lewy body disease. So we rely a little more on metabolic pet to look for patterns consistent with those diseases. Um and sometimes, you know, I'll have a patient who the plasma biomarker makes it seem like it's Alzheimer's disease, and then I get the amyloid pet and it's actually negative. And uh and then I might reach for the metabolic pet in that situation to try to get more information about what might be going on.

Michael Harper:

Oh, that's great. Well, thank you. That's all the formal questions I have. I wanted to know if there's any take-home points that you want to sort of highlight for us.

Kyra O'Brien:

Sure. I think what I most often see is patients will receive like a head CT for structural brain imaging, but we really should be pushing towards getting MRIs on everyone because the CT doesn't tell us enough, especially now in the era of anti-amyloid therapy. We we really rely on that better resolution structural imaging.

Michael Harper:

I want to follow that up. Um are we speaking about specifically people with early stage disease, either MCI or early stage dementia? Because I'm imagining folks who are, let's say, older, lots of chronic comorbidities, have more present with a more advanced stage, maybe pretty typical for a clinical syndrome that we recognize, Alzheimer's, Lewy body. Do you still think that those folks require MRI?

Kyra O'Brien:

So that's a great question. And I always come back to how is this study going to change management? And so I do see patients who are in the severe stages of dementia and they've never had a scan, but I'm like, is a scan going to help here really? And so I don't get anything for them. So I would say if you feel like the scan is going to change your clinical management, then I would probably go for MRI if they can tolerate it.

Michael Harper:

Yeah, that's a really great place to end. I think we always need to think about how what we're going to do as a next step may impact what we do moving forward. Sometimes it's for prognostic information, sometimes it's for treatment decisions, uh, and sometimes uh, you know, maybe for other reasons. But Dr. Brian Karlawish, I really want to thank you for pulling together this module with your partner, Dr. Karlawish, uh, and for taking the time to be with me today to talk about this and to help guide our geriatrician colleagues in thinking about how to use these tools uh in the care of their patients. So thank you very much.

Kyra O'Brien:

You're welcome. It's my pleasure.