Hello, and thank you for joining us for today's webinar entitled Unlocking the Radiologist's Role in Alzheimer's Diagnosis and Treatment, sponsored by Eli Lilly and Company. I'm Vittoria Spampinato. I am the Division Director of Neuroradiology at the Medical University of South Carolina. The Radiological Society of North America would like to gratefully acknowledge Eli Lilly and Company, the sponsor of today's webinar. Today, Dr. McConaty, Dr. Raggi, and I will cover an introduction to biomarkers in imaging and amyloid PET. We'll also cover pretreatment MRI evaluation in Alzheimer's disease and neuroradiology of ARIA with CAIT examples. We want this to be an interactive webinar, so please submit your questions in the Q&A panel, and we will address them after our presentations. We will now hear from Dr. John McConaty. Thank you. Okay. Hello, everyone. My name is John McConaty. I'm a Professor of Radiology at the University of Alabama at Birmingham, where I'm the Director of Molecular Imaging and Therapeutics and the Imaging Core Leader for our Alzheimer's Disease Research Center. I'll be starting off with an overview of imaging biomarkers and the diagnosis and treatment of Alzheimer's disease, and we'll focus on amyloid PET in this section. Here are my financial disclosures. Most notably, I have had consulting and research support from Eli Lillyavid, and also I'll be discussing FDA-approved tracers according to their prescribing and approved indication. So in terms of learning objectives, here they are. I'm going to focus on recognizing the value of an early and accurate diagnosis of AD for optimal patient outcomes and the role of PET imaging in the identification of AD pathology. So why is biomarker characterization important in patients being evaluated for Alzheimer's disease? Well, first, there are more and more people with Alzheimer's disease in the United States, with an estimated almost 14 million projected in 2050. This is a worldwide trend, and a lot of this has to do with the aging of the worldwide population and the U.S. population. Another important feature of Alzheimer's disease is there's a long preclinical phase that we can detect with biomarkers, as well as gradual decline once cognitive impairment starts. And probably most game-changing is there are now anti-amyloid antibody therapies that are entering routine clinical use, with Leucanumab approved in 2023 and Denanumab in 2024. And these require MRI, as well as imaging biomarkers, to perform this therapy in the right patient and make sure safety is being optimized. We have a lot of molecular imaging tools for brain imaging relevant to Alzheimer's disease. There are three approved beta-amyloid PET tracers, and that will be the focus of this talk. There is one FDA-approved tau tracer, which is also highly relevant to Alzheimer's disease. We also have other tools, like FDG for glucose metabolism, as well as dopaminergic imaging agents that I'll very briefly touch on, which can be relevant, but aren't the focus of this presentation. So, in terms of the accuracy of clinical diagnosis of AD, when a memory specialist, such as neurologists, diagnoses a patient with probable AD dementia, that's roughly 80% concordant with pathology. However, this doesn't account for patients with possible AD dementia diagnosis due to potentially comorbid factors, and it can be also difficult early on, particularly in first patient encounters, to make a diagnosis. The other issue is there are other relatively frequent causes of dementia, such as dementia with Lewy bodies and prototemporal dementia, where diagnostic accuracy of clinical diagnosis alone is much lower. And finally, and I think very important, is that by the time a confident clinical diagnosis is made, neurodegeneration is often advanced, and this really may not be adequate for selecting patients for therapy, both in the clinical trial setting, as well as in standard of care settings. This is an image showing a kind of common time course of imaging and other biomarkers in Alzheimer's disease, where the biomarker is becoming abnormal on the y-axis as the patient's cognitive status goes from normal to mild cognitive impairment to dementia. And the main points here are that amyloid, measured with PET in the CSF, or more recently with plasma biomarkers, is the earliest phenotypic marker of Alzheimer's disease pathophysiology. And what's important to know is this can be abnormal 10 or even 20 years before patients start developing cognitive impairment. Another key protanopathy, tau, in Alzheimer's disease becomes abnormal later in the time course of ADD pathophysiology and is much more tightly connected to changes and worsening of cognitive status. That's also true for nonspecific biomarkers like FDG glucose metabolism and regional brain volumes measured with MRI. Recently, there's been a new biomarker format that's been proposed for diagnosis of AD. The core one biomarkers focus on the diagnosis of AD pathophysiology, and these focus around amyloid measures and detecting amyloid. The core two biomarkers focus on tau, and these are more for staging, so trying to understand where a patient is along the continuum and progression of Alzheimer's disease pathophysiology. There are other biomarkers for neurodegeneration, for inflammation, vascular brain injury, and alpha-synuclein, which play potential roles in Alzheimer's disease patients or patients being evaluated for Alzheimer's disease, but we'll be focusing primarily on amyloid and to a lesser extent tau. In addition to making the diagnosis, it's important to think of amyloid as a theranostic target for PET, in the sense that we use amyloid PET to decide if a patient has amyloid present or anti-amyloid therapy if the patient has Alzheimer's disease, and for detecting clearance of amyloid from the cortex with therapy. So here are published data from trials with Denanamab, Lecanamab, and Angiocanamab, three anti-amyloid antibodies, and for each of these treatments, you can see over the course of approximately 18 months that compared to the placebo group, we see decline in PET measures of cortical amyloid with all three of these therapeutics, and this clearance is on the order of 60% to 80% in most patients. So we can also use amyloid PET to monitor clearance of amyloid from the brain. In terms of amyloid PET, here are three different patients who underwent amyloid PET. Patient one is a negative case. Patients two and three are positive, and you need to know the interpretation criteria for the particular amyloid PET agent you're using. They're slightly different, but in general, we see with, in positive cases, increasing accumulation of cortical amyloid and lose the gray-white differentiation that we normally see in the cerebrum, with the cerebellum as the reference, which in clinical patients should stay free of significant amyloid deposition in the vast majority of cases. In terms of how these perform, they all perform similarly. These are the sensitivity and specificity data from the FDA, the trials that led to FDA approval for these agents. So they perform similarly. They have similar appearances, but again, you want to know the specific diagnostic criteria and undergo the reader training for the agent you're using clinically. This is a little scheme showing how you might think about clinical neuroimaging time course in patients with AD, from diagnosis to anti-amyloid therapy to follow-up. And so, essentially, all patients are going to need an MRI. We'll hear more about that in the subsequent talks, to rule out other causes, structural lesions, and some sort of amyloid biomarker, which may be PET, may be CSF. It could potentially be plasma. There are other options that might be performed in selected patients. For example, tau might be useful for confirming the diagnosis of AD if there's a question. And although this is not part of routine clinical practice, there may be prognostic value in tau PET. Patients with extensive neocortical tau may get less benefit from anti-amyloid therapy. Ioflupine may be useful for distinguishing Alzheimer's disease from dementia with Lewy bodies, both of which dementia with Lewy bodies can be amyloid positive, and will also have dopaminergic deficits, whereas pure Alzheimer's disease does not have a dopaminergic deficit. And finally, FDG. We still perform this in some patients who are being evaluated specifically for AD versus FTD, and in a few cases for problem solving. Once patients start anti-amyloid therapy, they're going to undergo serial MRI, looking for amyloid-related imaging abnormalities, a marker of adverse events related to therapy. And again, you'll hear more about this in the subsequent talks. And at our institution, we are performing amyloid PET typically around 18 months while on therapy to look at amyloid clearance. The concept is we may be able to decide when to stop therapy based on amyloid clearance. This is an evolving area. I think we need more data to know confidently how to use post-therapy amyloid PET. And I'll also admit, I'm not sure exactly how we're going to handle follow-up in terms of repeat amyloid imaging. But I think over the next few years, as research continues, we'll have more guidance in the follow-up period. So, I'd like to tie some of these concepts together with a clinical case. This is—I'm presenting this as it came through the reading room. The history I got was 67-year-old man with mild dementia due to Alzheimer's disease. Here was the amyloid PET going from vertex to cerebellum. And looking at this case, I ultimately thought this was borderline. I would have read it as negative, but there are a few places where the gray-white differentiation may not be as good as I'd like. There is some atrophy. So, it's a little bit of a tricky case. And as I was looking at the chart, I realized that this was actually a follow-up study. This patient had undergone 18 months of anti-amyloid antibody therapy. And here was their pre-therapy image. And I think qualitatively, it's easy to see there's really no gray-white differentiation prior to therapy, but it's clearly substantially better after therapy. And while I can tell you it looks better, one of the challenges in quantifying or characterizing the response to therapy is I don't think I could tell you what percentage was cleared. And this is where quantitation may become quite valuable. I'm showing this in a fairly simplistic way where we have a cortical region of interest as well as a cerebellar region of interest, in this case, the whole cerebellum. And you can see that the SUV ratio, the ratio of activity in the cortex to the cerebellum, has decreased over the two years or 18 months of therapy. We can quantify this and present this as SUVRs. And so in this patient, about 70% of the amyloid cleared. We could also get useful information from the MR, like regional brain volumes, and look for progressive atrophy a little bit on the right in the hippocampus. In order to compare across centers, scintilloid scales have been developed. And this is basically a linear transformation that's used to convert SUVRs to a scintilloid value that should reflect differences in amyloid PET tracers and difference in quantitative processing. And so around a scintilloid value of 40, in general, experienced readers will interpret them as positive. There is a little bit of a gray area between around 25 to 40, which is borderline. And this is actually an interesting patient population as well. And finally, below about 25 to 20, readers are generally going to classify those as negative visually. Going to talk a little bit about Tau. Tau is a more restricted spatial distribution than amyloid. It starts in the transenterinal cortex. And this can occur in normal, healthy, older adults. But once it begins spreading to limbic regions and neocortical regions, this is the pathologic spread seen in Alzheimer's disease. There's currently one approved PET agent for imaging Tau in the human brain. That's for Tau sapir. It works well for neocortical involvement, but it may not have as much sensitivity when Tau only involves limbic regions as it spread. But in those with neocortical Tau, the sensitivity and specificity you can see across five different readers in the trial used for FDA approval is good. In terms of reading it, it's actually done based on an SUV ratio between the cerebellum drawn in two dimensions with a 1.65 cutoff. This is for clinical interpretation. I've used a color scale where if it's below 1.65 fold relative to the cerebellum, it's in grayscale, but it becomes color once it exceeds that threshold. You can see there's some off-target binding in the basal ganglia as well as the choryplexis and brain stem. But you can see within the cortex, there's really no areas that exceed this 1.65 fold threshold. In contrast, here's a clearly positive case where you can see regions of the brain exceeding this threshold in the temporal cortex, occipital cortex, and parietal cortex. One of the reasons I think Tau is quite interesting is it does have strong prognostic value. This is a multi-center study. I believe seven centers contributed data looking at people with normal cognition and whether they progressed to mild cognitive impairment or to dementia over the course of six years. And the major features in terms of the MCI progression, only those that were Tau positive had a high risk of progression. There was slightly more progression in amyloid positive Tau negative compared to amyloid negative Tau negative, but not a lot of difference. And similarly, progression to all-cause dementia was by far and away highest in those who had neocortical Tau on PET. So, there's potential prognostic information from Tau, and over time, we may be making some of the anti-amyloid treatment decisions based on Tau status. Briefly, reporting amyloid and Tau PET findings, I think it's important to describe the brain regions that have abnormal amyloid or Tau, often done on a low-bar basis. It's important to apply the individual amyloid and Tau tracer reading criteria for standard of care reporting. And there are training modules available for each of these agents, and be sure to do the one for the agent you're going to be using. The primary evaluation currently is positive or negative, so it's a positive or negative amyloid or Tau scan. And remember that amyloid and Tau PET, it's a biomarker. It is part of the overall diagnosis, but with the imaging alone, I can't use an amyloid PET to say a person has Alzheimer's disease. They have Alzheimer's disease pathophysiology, but I can't say in isolation that an amyloid PET, a positive amyloid PET scan means a patient has dementia due to Alzheimer's disease. They could have a comorbidity. It could be dementia with Lewy bodies. So, it's important, but it's only part of the overall diagnosis. I think if you're performing particularly amyloid PET in the setting of therapy, you should strongly consider adding quantitative data, either SUV ratios or senoloids. There are commercial vendors that have software that can provide these, and I think it'd be useful both for diagnosis, but more importantly, for quantitating clearance after therapy. I think it's also important to convey uncertainty when appropriate. You don't want to equivocate on every study, but there are borderline cases. There are cases where there's technical artifacts, and at least at my institution, my colleagues in neurology and the memory clinic appreciate it when we flag cases and say, well, there's flag cases that are not straightforward. And finally, if you can, understanding how you're referring physicians or using the PET data is going to make your report more effective. So, to summarize, brain MRI and amyloid PET play key roles in the diagnosis of Alzheimer's dementia and modern anti-amyloid therapy. And amyloid biomarkers, with PET being an important one of these biomarkers, are key to deciding if anti-amyloid therapy is appropriate. We didn't talk much about this, but do remember plasma and CSF biomarkers are alternative methods for amyloid, tau, and other biomarkers relevant to AD. The plasma biomarker field in particular has really made a lot of progress, and I think we're going to see evolution in using plasma versus PET as biomarkers over time. Quantitation is playing an increasingly key role in clinical amyloid PET, and other molecular imaging biomarkers, particularly tau PET, may play a growing role in staging Alzheimer's disease and providing prognostication for therapies. So, I appreciate your attention. I'll now turn the webinar over to our next speaker, Vittoria Spampinato. Thank you. Hello, everyone. I'm Vittoria Spampinato. I'm a professor of radiology at the Medical University of South Carolina, where I also serve as the vice chair of faculty development and the neuroradiology division director. I'm going to talk about pretreatment MRI evaluation, and these are my disclosures. So, why is the pretreatment MRI important? In order to determine treatment eligibility, patients need to meet certain inclusion and exclusion criteria, and MRI plays a key role in treatment eligibility. I'm going to talk about the steps and the procedures that we need to put in place to provide the imaging required for these applications, including MRI protocols. I'm going to talk next about MRI criteria for treatment eligibility and risk factors. And finally, I'm going to talk about MRI interpretation and result communication to our clinical colleagues. So, according to the FDA label, a recent-betan MRI is needed to establish the safety of the anti-amyloid immunotherapy. This MRI will also serve as a baseline for comparison with the subsequent MRIs that are performed for safety monitoring. But how many MRIs are we talking about? The circles in the schematic indicate the schedule of the infusions, while the arrows indicate the schedule of the baseline MRI and of the subsequent MRIs that are performed for safety monitoring. So, the importance of the baseline MRI is clear. Each patient, at a minimum, will receive four to five MRIs, not taking into account any additional MRIs that may be needed if the patient develops symptoms or treatment complications occur. So, in light of the importance of the baseline MRI, the fact that each patient will receive multiple MRIs, we will certainly see an increase in MRI requests in our practices, and we are already seeing that. In fact, at my home institution, which is a mid-size academic institution in South Carolina, we saw in the past 12 months already a 60% increase in the number of studies that are performed for indications such as MCI, dementia, and memory loss. So, what are the steps that we need to set up for this patient? Many experts recommend to develop a seamless workflow integrating newly developed MRI orders in our medical records that are specific to this patient population, and these orders should be linked to MRI protocols on the scanners, possibly with the same name for consistency, and this could be linked in turn to standardized reports to facilitate communication of reports of results that is clear and effective. So, with respect to the protocol, we probably should set up our baseline MRI protocol to be the same as the protocol for our general dementia application so that we don't need to repeat unnecessarily MRIs multiple times before treatment onset. So, but what are the key sequences that we need to include in the baseline MRI protocol? As a key task of the radiologist is to identify intracranial hemorrhage, we need to make sure to use a high-quality blood-sensitive sequence. And in clinical trials, T2 star GRE sequences were employed because of the ease of standardization for that sequence across different institutions and vendors. So, T2 star GRE is the key sequence. It needs to be optimized, and for optimization of the sequence, I refer the audience to two great white papers by the ASNR group led by Dr. Cogswell. And Dr. Raggi, our speaker, is also an author on both publications. Next, we need a FLIR sequence, which is great both to characterize any abnormal findings on the baseline MRI, but also to monitor for ARIA, as we'll see in the next presentation. And FLIR could be a 3D FLIR with multi-flare reformats, which is employed by many practices, but it could be also a high-quality 2D FLIR. And we also need DWI to evaluate for any acute infarction. In the baseline MRI, also a 3D, T1, and a T2 could be employed. One more comment about blood-sensitive sequences. Although the GRE is the key sequence, many practices also add an SWI or SWAN that has a better spatial resolution and is more sensitive to blood products. The drawback is that it's more difficult to standardize. So what is the timeline of the recent baseline MRI? According to the FDA label, the MRI could be performed within 12 months prior to the beginning of treatment, but most experts agree on recommending an MRI to be performed within three to six months prior to the onset of treatment. A CT scan is not adequate to evaluate for findings that will determine patient eligibility and certainly not adequate to monitor for treatment complications. How about field strength? 3D FLIR is certainly recommended, and 1.5D FLIR can be adequate if the patient cannot get a 3D MRI. Any field strength less than 1.5 is not acceptable. Efforts should be made to be consistent with MRI protocol field strength and scanner vendor from scan to scan. And I wanted to add that our neurologists find it useful to schedule all the MRIs for the year, starting with the baseline MRI and the routine treatment monitoring MRI all at once at the same imaging center so that on one end the patient has guaranteed access for those routine MRIs, but also for consistency of protocols. Our societies have worked with the vendors, with multiple vendors of the MRI scanners to develop high quality protocols that are standardized. And these protocols can be actually found on the ASNR website and in some cases may also be available in the protocol tree of our scanners. So let's now talk about appropriate use recommendation and what are the findings that we are looking for in the baseline MRI. So the most important task for the radiologist is to identify and characterize intracranial hemorrhage. With respect to microhemorrhages, we need to remember the rule of five. Any cases where there are five or more microhemorrhages will be excluded from treatment. Patients with less than five microhemorrhages can be considered for inclusion. Superficial sclerosis and any microhemorrhage described as a bleed greater than 10 millimeter are also a reason for exclusion. In this case, on the left, we see signal dropout in the central and post-central circles due to superficial sclerosis. And we also see in the right frontal lobe an area of micro hemorrhage. Cerebral myeloid angiopathy related inflammation is also a diagnosis that leads to exclusion from treatment, also because of the increased risk of bleeding. This was the case of a patient that presented with cognitive decline, seizures, and a headache. The MRI revealed a large area of hyperintensity on the flare in the subcortical white matter extending to the deep white matter, as well as other areas of subcortical signal abnormality in the posterior brain, as well as innumerable microhemorrhages in a lower distribution. The diagnosis of CARI was made, and the patient was treated. And you can see, as on the follow-up exam, the findings have almost completely resolved. It's very important in the Bayesian MRI to describe accurately any findings of ischemic stroke, because patients with more than two lacunar infarcts or patients with territorial infarcts are excluded from treatment. It is also important to describe and characterize white matter hyperintensities, and especially it's important to provide a grading of the severity into mild, moderate, severe. Many people use the Fazikas scale that map into these three categories. This is important because patients with Fazikas grade 3 or severe white matter hyperintensities are not eligible for treatment. You can see in this example at the bottom, there is confluent white matter hyperintensity in the deep white matter, and this is a sign of severe small vessel disease. As we scan more and more patients for these applications, we often run into incidental findings, and it's important to know the implications of those. Any vascular malformation, including cavernous malformation or intracranial aneurysm, are reason for exclusion due to the risk of bleeding. Patients with intracranial masses are not eligible for treatment. The exception are subcentimeter meningiomas, seen in this case in the left frontal lobe, left frontal region. Patients with small meningiomas could be considered potentially for inclusion. All right, after the overview of the findings that we're looking for, I'm going to talk about standardized reporting of the screening MRI. It is highly recommended for the Bayesian MRI to employ a structured report. This report is like a checklist where all the items that will determine eligibility are spelled out, and they're listed in this table. So we need to specifically mention microhemorrhage and their number, and we'll summarize the number in the impression of the report based on these three categories, 0 to 4, 5 to 9, or at least 10. We need to comment on whether superficial siderosis is absent or present. We need to quantify white matter hyperintensity for our neurology colleagues, and we need to characterize findings of ischemic stroke. The report will also have a narrative describing any other acute or chronic findings, any incidental findings that are important. But how can we get started with a structured report? Our societies have worked on the development of very good structured reports. There is a structured report available for the Bayesian MRI, as well as for the subsequent safety monitoring MRI available at the website listed here. In this slide, I'm just sharing a snapshot of the structural report for the Basan MRI. And notice how there is a specific field to fill out what blood-sensitive sequence was employed and what was the field strength. And that will be helpful for comparison with future examinations. In conclusion, advances in the treatment of Alzheimer's disease have started and will continue to impact our imaging services. Radiologists certainly play a key role in helping making important treatment-concerning decisions. And my call to action is to work with your practices to ensure that high-quality MRI protocols are implemented. As radiologists, our task is to systematically review the MRIs with knowledge of the inclusion-exclusion criteria, make sure that all the pertinent findings are described, and then utilize standardized reporting for clear communication of the results. More and more, radiologists are getting engaged in multidisciplinary review of clinical imaging findings to discuss interesting cases or complicated cases. And our neurology colleagues certainly appreciate our input on imaging findings. And with that, I would like to thank you all for joining the webinar. And I would like to acknowledge my outstanding neuroradiology and neurology colleagues at the Medical University of South Carolina. And now, our next speaker is Dr. Cyrus Raji. And he's going to talk about the neuroradiology of ARIA. Thank you. Hello. My name is Dr. Cyrus Raji. I'm an associate professor of radiology and neurology at the Washington University Malinckrodt Institute of Radiology. I'm also the director of Neuromagnetic Resonance Imaging at Barnes-Jewish Hospital and the associate leader of the imaging core of the Knight Alzheimer Disease Research Center. These are my disclosures. I'll note that, in particular, I do consult to Eli Lilly and also to Hyperfine. I'll be mentioning some related products there in terms of the anti-amyloid drugs and some examples of hyperfine imaging. And with that, let me get into the learning objectives, which is to define amyloid-related imaging abnormalities and the related therapies, provide case examples of ARIA to review some of the protocol considerations. And I do think, in conjunction with some of the excellent talks that we've seen, also give a little bit of repetition in terms of what the template of ARIA reporting will look like. Amyloid-related imaging abnormalities, or ARIA, refers to characterized complications of anti-amyloid immunotherapy for Alzheimer disease. And what we saw in the really nice specific case example from Dr. McConathy is that when anti-amyloid treatments work, what they do is they clear amyloid from the brain. And these examples from amyloid imaging on PET demonstrate that, with placebo, there'll be a persistence of amyloid plaque. With treatment, there'll be a clearance of that plaque. And since plaque is an important hallmark of Alzheimer's pathology and part of the pathophysiology of the disease, then understanding this change is important. But what we also note is that there are some complications that happen from the clearance of this plaque. And so when they're characterized, they're characterized in terms of ARIA-E and ARIA-H. This is from the FDA label of Ferdinandimab. And what we're seeing here is that, when we talk about ARIA types, we really talk about them in terms of edema, which is what the E stands for, and the H, which is what the hemorrhage stands for. Now, it's important to understand that with edema, nested within that definition is that the flare hyperintensity that you'll see can be confined to the cortex or subcortical areas of white matter, but also it can be as subtle as a single sulcus. And so it's possible to have ARIA-E in a single sulcus, and we'll show some examples of that shortly. And then of course, when we go from mild to moderate to severe ARIA-E, it's important to understand that it comes down to the measurement. And so, you know, with mild ARIA-E, we have a flare hyperintensity essentially less than five centimeters. With moderate ARIA-E, it is five to 10 in one location or multiple sites less than 10. In severe ARIA-E, we see that the size of the sulcal or edema in the cortex is greater than 10 centimeters. And that's the definition of ARIA-E. ARIA-H, we talk about microhemorrhages and superficial pseudorhosis. And so mild is essentially less than or equal to four new incident microhemorrhages. Moderate is five to nine new incident microhemorrhages. And then severe is greater than 10. With superficial pseudorhosis, you can have mild ARIA-H superficial pseudorhosis, which with as little as one focal area of superficial pseudorhosis. As I'll show you from the case examples, these can be quite subtle. And of course, when we go to severe ARIA-H superficial pseudorhosis, that's only greater than two new areas of superficial pseudorhosis. So it is an important consideration to note. And I think that as we kind of move forward, we'll see some examples of that as well. So starting with examples from the literature, and I start with examples from the literature for a couple of reasons. One is to show you that we've been aware of ARIA for a long time, but also to note that the original clinical trials and investigations used comparatively older imaging techniques. And so 2D flare images showing ARIA-E as we see in this example from the literature, as well as gradient echo sequences for ARIA-H. And so some individuals will assert that we should use whatever is consistent with the trial. Some will assert that using more modern techniques like 3D flare or SWI will allow us to catch more abnormalities. In terms of our own practice, when it comes to ARIA, we use a little bit of both. And so we use both susceptibility weighted images as well as gradient echo, because it turns out that there can be some fake outs for susceptibility weighted images like vessels and profile. And so gradient echo can be a useful problem solving sequence as well. The other advantage of susceptibility weighted imaging is that we typically get them in thinner cuts. And so it's easier to do reformats. And so that's just some of the protocol considerations that we'll review near the end. What we're seeing here are examples from the literature from a paper that I authored showing examples of what ARIA-H cerebral microblades and superficial sclerosis would look like, as well as ARIA-E with multiple areas of edema on the right on the flare image. And these are not subtle examples. I mean, in terms of the middle panel showing the superficial sclerosis, we can see obvious areas of a lot of abnormality. And so we do see that there. And then sometimes you can have more comparative examples. And the key to this picture, also from one of our prior papers, is that ARIA-E typically resolves. And so a baseline scan shows no edema kind of near the higher parts of the vertex. And then of course, after a few doses, we see the development of focal ARIA-E that resolves on follow-up when the dose is discontinued. Of course, differential considerations for hemorrhagic lesions and microblades in particular can include cerebral amyloid angiopathy and hypertension. This is especially common in older populations. And what we see is an example on this slide where you can have inflammatory cerebral amyloid angiopathy that mimics ARIA, that can look like ARIA. And so for instance, we see here on these panels, multiple microblades in the lower distribution. Some of them are actually looking more like larger hemorrhages. We can see these flare hyperintensities associated enhancement on post-contrast imaging when it is available. And the important thing to note here is that when you're trying to consider ARIA development, it's important to note whether or not the patient was on immunotherapy. This is absolutely fundamental. And it's not always intuitive that you'll know that. At our institution, we very rigorously track individuals who are on therapy. So we know who's getting these drugs, but in a given facility or a new site doing these therapies, when there are presentations for ARIA that have associated clinical sequelae, they may present as stroke cases or encephalopathy situations. And you may not know the history, going into looking at the case. So in older individuals, especially with a history of dementia or memory problems, it's important to keep it on the differential and importantly, follow up with the clinical team if there is an actual use of immunotherapy as that's key in making this diagnosis. Of course, when looking at microhemorrhages, it's important to note that microhemorrhages have a differential diagnosis. And so when they cluster around the basal ganglia or thalamus, they are typically more consistent with a hypertensive etiology. When they're in a diffuse low bar distribution cortically, that typically happens with cerebral amyloid angiopathy. And when they're asymmetric or diffuse at the gray matter, white matter boundary, that can be consistent with diffuse axonal injury or traumatic brain injury in particular. So that's all literature cases. Let's get right into actual clinical cases from our institution. And we have about 200 individuals on active treatment right now. So we have a lot of cases. There are about three to five scans a day that I see in the reading room when I'm on service. And so we do see ARIA and we do observe it. The rate of ARIA is depending on what papers you're talking about around 20 to 30% of cases. And so you will see these in practice. And so that's why there's so much dosing and imaging that's performed. And what we see here is that in this pre-dosing example, we see that there are white matter hyperintensities consistent with small vessel ischemic disease. And I would say that if I were grading this, it'd be like a Physicus II, so maybe like a mild or moderate grade. And then with post-dosing, we see this development of ARIA-E, which is moderate in distribution. And on cessation of dosing, we do see resolution for the most part of the ARIA-E. However, there is some dilatation of the adjacent ventricle, which is consistent with volume loss that's been reported in individuals who are getting anti-amyloid therapies. This is another example from the same case showing this kind of biooccipital distribution of moderate ARIA-E that develops. Again, there were no evidence of edema on the prior study, but we do see them develop in the occipital lobes and they resolve or at least improve partially on follow-up imaging. In terms of microhemorrhages and siderosis, we can see on the predosing scans on the susceptibility-weighted images on the far left that there were no areas of hemorrhage or siderosis. We see these areas of hemorrhage and siderosis develop on the post-dosing. When the dose was stopped, we do see that there's still a development and progression of severe ARIA-H, both in terms of the microbleeds as well as the superficial siderosis seen here in the kind of left parietal lobe. In the same case, we see also examples of superficial siderosis developing in the cerebellum. Again, predosing, no evidence of siderosis, and we see quite extensive siderosis develop on the post-dosing scans and they persist after cessation of dosing. And so with ARIA-E, as I was saying earlier, you should expect there to be improvement after cessation of dosing, but the hemorrhages do persist as well as the siderosis. And that was a blatant example in terms of the case. I wanna show you some subtle examples, credit to my colleague, Dr. Cammie Besinger, who has really pioneered this area. And what we can see here is that this is one area of superficial siderosis. And the reason I make the arrow so big and obvious is that it's really hard to see. It's almost kind of like an eyelash, but we can see that replicated on reformats. And so multi-planar reformats, susceptibility-weighted imaging are quite helpful and we see this kind of sole area of the superficial siderosis, which again would make it mild ARIA-H, superficial siderosis, or we'd have it. Again, another example of subtle superficial siderosis kind of here in the parietal lobe on the left, as well as in the occipital lobe and then develops on the contralateral side. And so when these areas progress, they can be very subtle. So careful scrutiny of the images is absolutely essential in picking up these findings. This is an example of a siderosis This is an example of what a sulcal effusion looks like with RAE. And so again, you see these subtle sulcal effusions and it even looks kind of like it would be a subdural, but again, we see the adjacent sulcus affected. And again, it can be quite subtle and very easy to miss. These are other examples of scans showing sulcal effusions. Again, very subtle case right here, showing this kind of anterior temporal lobe sulcal effusion. It kind of looks like a subdural, but again, the sulcal effusion is kind of the giveaway right here. It does resolve. And this is an example where you can actually see the sulcal effusion on mobile MRI. Not that that's what we recommend, but it does show that if you're getting the mobile MRI for other reasons, typically for stroke, you may also see manifestations of RAE on mobile MRI protocols. And so to the extent that practices are using these techniques, that might be something to be aware of. Again, another example of no RAE, a mild distribution on the follow-up dosing, and then again, moderate on progression that can also be seen on mobile MRI imaging. Now, what do we do in individuals who have RAE or RAH? It basically depends on if they're asymptomatic and if it's mild. And if they're asymptomatic, dosing might be continued. If it's mild, dosing may or may not be continued, depending on how conservative your neurology team is. And with moderate levels of imaging findings or clinical presentation, dosing will typically be suspended. And this is just showing you whether you're talking about, you know, what Canamap label or Donanamap, it's essentially the same set of recommendations. These labels are very similar in terms of how they present the imaging findings of RA as well as the related recommendations. And of course, we've published on this in 2022 and most recently in 2024. In terms of the protocol and pyramid, the biologists say that standardized sequences are important. And of course, you want the field strength to be consistent and ideally the same MRI and vendor if you can. Of course, you can use 2D or 3D methods for flare or gradient echo or susceptibility-weighted imaging. And again, in our practice, we use both for the hemorrhage imaging. And of course, it's depending on your scanner capability as well as your throughput. And these are examples, again, of reporting templates that show the ways of reporting, you know, RAH and RAE that are available through ASNR. And with that, I would like to thank you for your time. Well, great. I'd like to thank the other presenters for great presentations. I think now we will move into the question and answer portion of this presentation. So everyone's a little bit shy. We don't have any questions yet. So maybe I'll put this to you, Cyrus. You know, and Dr. Spampanato, if you have anything to add, I'd be very interested to hear what you have to say. You know, we had the same kind of issue with orders for brain MRI, understanding if these were done specifically for on therapy versus just general indications. And in order to be sure that the reporting was done in a format that made RA assessment match the label in terms of decision making, any recommendations or anything you found useful at your institution for making that work better? Like in terms of asking referring physicians to specify, you know, the reason is the patient's on anti-amyloid therapy. Yeah, I'm happy to take a first crack at that. At our institution, we built a distinct RA protocol that in conjunction with vendors is now becoming increasingly available on protocol trees. We have a distinct order set for RA, you know, follow-up. We call it like MRI Alzheimer's treatment. So when it populates our list, it shows up with those words. So we know why we're getting that. And since we have a lot of patients who are well-characterized and tracked by our neurology clinicians, many of them work in the ADRC, it's a pretty seamless process. And we're in close communication so that whenever we do see these abnormalities, we immediately let them know, even if they're mild. And of course, they really appreciate that too. So that's more or less the broad strokes of how we've structured it here. Thanks. And I can comment on that as well. So we have a similar setup for our patients that are known to be considered for treatment. I find that the challenge are the patients that do get an MRI, but they are not actively being considered for treatment. And then the neurologist will call us because, you know, the dictation does not include all the pertinent negative of the findings. So that's why I think moving forward, it would be useful to kind of use the same pipeline for our dementia patients and for the baseline treatment consideration. Great. Now, thanks to both of you. We do have a couple of questions coming through. So the first one is, which is more common, ARIA-H or ARIA-E? Cyrus, do you want to field that one? Depending on the rates, they're similar. You know, they overlap in terms of the trial. Some trials have found that ARIA-E, you know, is as high in rate as like the, you know, 30 plus percent. But some trials have found similar results with ARIA-H. Some have found it to be lower, as low as like maybe six to 8% between the two. In our practice, we found that both are quite common, although I would say that we tend to, you know, find more ARIA-H. And I think the important thing to know with ARIA-E is that of the two, ARIA-E is more clinically consequential in terms of correlating to adverse outcomes in the clinical trials. However, ARIA-H is potentially more clinically impactful if you have situations where you're concerned about stroke and there's interest in giving thrombolytics, especially when there's acute presentation where individuals with ARIA are mimicking stroke situations. And so those are some kind of considerations that we've encountered in our practice. Thanks. Dr. Spampanato, did you have anything you wanted to add? No. Okay. So the next question, I'll answer this one. So how does the clinical, how does clinical dementia correlate with reduction amyloid on imaging? And, you know, what's the time course of, you know, the reduction in dementia? And so, yeah, to be clear, you know, really this is slowing the time course of progression in, you know, the trials that have been done with nanomab and licanumab. We're not, the patients are not having improved cognition. Their rate of decline is slowing. And, you know, this is happening over the course of months. You know, you're not going to see an instant effect. And, you know, I do think that's one of the challenges is while I do think this availability of effective disease modifying therapies is a huge step forward, you know, the results are not as good as we would like. Some of it may be if we can implement these treatments earlier before, earlier in the disease course, that may be a way to improve the efficacy. There also may be other factors, you know, potentially other treatment targets that may further improve patient's outcome. But, yeah, I mean, I wish these were wonder drugs where people were having improving cognition. We're not there yet. If either of you have anything you'd like to add, feel free. Okay. The next question, are patients scanned with pretreatment, post-treatment PET in addition to MRI, or is PET not used as much? So I'll take a stab at this. So, you know, everyone needs MRI before and after anti-amyloid therapy. PET is one of the options for showing that there's an amyloid biomarker positivity. You need that. You could do CSF. If a patient has CSF amyloid, that might be sufficient. At our institution, we are choosing to do PET even in patients who have positive plasma biomarkers. And it could be, and I think this is an open question, and if either of you have opinions, I'd be interested to hear them. You know, I think there's probably stronger data in terms of phase III trials with amyloid PET. And I think there's also the idea that we can follow PET over time with treatment. It's not clear how the plasma biomarkers will use. So we're typically performing amyloid PET or CSF before treatment. But that is not a requirement. You know, if you have, but you do need an amyloid biomarker that is positive that you believe in. Cyrus or Vittoria, anything you'd like to add? We've certainly seen in our practice an increase in the lumbar punctures being prescribed for this application. So that's another area of neuroradiology that is being affected. But overall, I believe at our institution, amyloid PET is the main exam that is utilized. I'll just add that from the imaging standpoint, I do agree with your presentation that quantifying centelloids could be particularly impactful in tracking changes and relating them to cognitive improvements. I am curious to know if there are particular software packages that you would suggest for that. I know that certain vendors are available through GE, but I don't know if there are any third-party platforms that could be utilized for that purpose, as there is a lot more interest, even from patients, some of whom are study participants, and they're more savvy of knowing about centelloids. There has been a lot of inquiry that I've encountered to know this information. Yeah, there's a number of commercial vendors. You know, so GE and MIM are one. Hermes is another. I think a lot of the major image-viewing software companies, if they don't already have one, are already developing them. So there are options, and there's also some third-party which are more focused just on assessment of amyloid and neuroimaging for dementia. So there are a lot of options. I think some of it, I think they all can work well. I think a lot of it depends on either what you're using now or what fits best in your clinical workflow. Next question we have is, are micro-images complications of the treatment? Cyrus, do you want to field that one? Yeah, yeah, sure. Yeah, REH microblades are imaging manifestations of treatment complications. I think that in terms of the fact that a lot of amyloid deposits in blood vessels as well as in the brain parenchyma, there is a theory that has propagated in the field that this is to some extent expected as amyloid clearance is taking place and that as a result, there would be some rate of REH expected. I will add that it is more likely to happen in individuals who not only have some of the imaging, you know, manifestations of likely-developed REH, but also ApoE4 positive status, the genetic risk allele of Alzheimer's disease. And so that is something that is noted as well in terms of individuals. It's also less likely for these therapies to be beneficial in individuals who are on the older side, so 85 and older. And so that's another consideration when, you know, understanding who benefits from these therapies as well. Thanks. The next question is, what's the percentage of patients developing intracranial hemorrhage after taking anti-amyloid therapy? And is there a correlation between intracranial hemorrhage and microbleeds, such as the number or the location of microbleeds? Vittoria, do you want to field this one? Yes. So the question is about what's the percentage of patients? They're getting, I think, intracranial hemorrhage, I think they mean beyond microbleeds is how I interpret the question. Yeah. So with accurate, the reason why we perform all the pre-treatment evaluation is to minimize the occurrence of a serious complication like intracranial hemorrhage. For example, somebody was touching on the fact that, you know, patients may present with symptoms of stroke. So important to flag the patient, know that they are on anti-amyloid immunotherapy so that there is no risk for them to receive thrombolytics. That's very dangerous and has led to at least one fatality in the clinical trials. But if the patients are screened appropriately, the risk of serious intracranial hemorrhage is relatively low. There is a risk for sure, but what we see more often are the micro hemorrhages as a result of complication. Cyrus, if you want to add anything based on your experience, I have not seen any in my practice, any macro hemorrhage as a complication. Yeah, thank you, Victoria. I'm only aware of it happening in situations where thrombolytics are applied for these cases where there might be suspected stroke. And I think that, you know, where you're going to encounter that quite a bit is, you know, when individuals come in for getting, you know, CTA and CT perfusion for large vessel occlusion. And, you know, you may encounter abnormalities that are pretty counterintuitive. So for example, we had a case where it was suspected large vessel occlusion, and it wasn't the case that they actually had any. The team was very interested in trying to give, you know, TPA if they could. And we said, well, no, there's no large vessel occlusion. But there was a lot of hypoperfusion, you know, on the CT perfusion, you know, in the bioccipital lobes. And this person ended up having, you know, PRESS, which is another differential consideration for Arian because PRESS can also have, you know, similar flare hyperintensities as well as micro hemorrhages. But this person also had dementia. So we were thinking, okay, are they on the anti-amyloid therapy? And it turns out that they weren't. And so these are some of the kind of diagnostic questions you may have to feel in cases that were previously a lot simpler, in addition to having to track microblades more carefully, because I think historically, they probably weren't as scrutinized. And now, of course, by necessity, with these new protocols of therapy, they should be. Great, thanks. I think we have time for one more question. Is that true? I don't see us getting the hook yet. So last question, maybe a speculative one, but anyone have comments about the lymphatic system? It's ties to Alzheimer's disease and how radiology and imaging might be involved in the future. I hesitate to speculate. I think that's a very good question. And I think actually the door is wide open to understand, to further looking at the neuroimmune system as well as lymphatics to enhance and understand better these therapies. I don't know that I have anything specific that I think is near ready for prime time, but Cyrus and Vittoria, if you have thoughts, please share. It's definitely an exciting field how these two areas may interplay, but certainly it's open for new development and research. I would just add briefly that these techniques do have the potential of identifying individuals who could be at risk of these treatment complications before you see the blatant imaging complications, especially with blood-brain barrier permeability, which is really what lymphatic imaging speaks to. But again, it'll be some time before that's clinically deployed. Well, great. Well, thank you. We appreciate the audience. Thank you for joining and participating in today's webinar. You will receive a notification email from RS&A next week when the recording is available in the RS&A Online Learning Center. And please be sure to check the RS&A website for upcoming educational events and webinars. Thanks again. It was a pleasure to have the chance to talk with you.

Alzheimer's diagnosis

imaging biomarkers

amyloid PET

MRI

anti-amyloid therapies

Leucanumab

Denanumab

ARIA

cognitive impairment

neurodegeneration