Monitoring ARIA in Anti-Amyloid Therapy in Alzheimer’s

Alzheimer’s disease (AD) is the most common form of dementia worldwide. Past treatments have focused on managing symptoms, whereas recent advancements, including anti–amyloid-beta monoclonal antibody (MAB) infusion therapy, target the underlying disease process in early-stage AD by binding to and removing amyloid plaques to slow the rate of cognitive decline. There are potential side effects associated with MAB use, mainly amyloid-related imaging abnormalities (ARIA). This article reviews the importance of early detection of AD and best practices for evaluating and monitoring ARIA in patients on MAB therapy.
Although the exact cause of AD is unknown, there are several hypotheses, including the amyloid cascade hypothesis, which suggests that deposition of amyloid-beta protein “is the causative agent of Alzheimer’s pathology and that the neurofibrillary tangles, cell loss, vascular damage, and dementia follow as a direct result of this deposition.” It is now thought that amyloid-beta accumulation drives AD and potentiates a cycle of further aggregation. Additionally, diffuse and benign plaques develop earlier than neuritic plaques, supporting therapeutic interventions in the beginning stages of AD that focus on decreasing amyloid-beta production or increasing its clearance—such as anti–amyloid-beta MAB therapy.
The aim of most of the current AD treatments is to improve cognitive and behavioral symptoms without altering the disease course. Anti–amyloid-beta MABs, however, are a new approach to treatment that targets underlying amyloid-beta pathology and facilitates the clearance of amyloid-beta aggregates by unbundling the amyloid plaque and engaging microglia and complement activation. Currently, donanemab and lecanemab are the only available anti–amyloid-beta MABs that are approved by the US Food and Drug Administration (FDA). (Aducanumab, the first amyloid-targeting MAB, was discontinued by its manufacturer in January 2024.)
Anti–amyloid-beta MABs are indicated for patients with early-stage AD; therefore, early detection is essential. Research on the role of amyloid-beta in the pathogenesis of AD has led to advancements in early detection of AD, and several imaging and cerebrospinal fluid (CSF) biomarkers have been recognized by the National Institute on Aging and Alzheimer’s Association research framework to assist in diagnosis. Among these, 18F-fluorodeoxyglucose PET is becoming increasingly recognized as a vital tool for earlier diagnosis of neurodegenerative conditions. Moreover, the use of radiotracers targeting AD-specific proteins, amyloid-beta and tau, offer more precise information than conventional structural imaging studies (eg, MRI, FDG PET). Florbetaben F-18, florbetapir F-18, and flutemetamol F-18 have been approved by the FDA for use with brain PET imaging to detect amyloid-beta neuritic plaque density. Another radiotracer, flortaucipir F-18, is currently the only tau tracer approved for PET imaging to detect tau neurofibrillary tangles.
Amyloid-Related Imaging Abnormalities 
ARIA is the most common side effect of anti–amyloid-beta MAB treatment. MABs increase vascular permeability, which can result in inflammation and leakage of proteinaceous fluid and blood products, leading to edema (ARIA-E) and hemorrhage (ARIA-H). Patients with preexisting amyloid pathologic conditions such as AD are more susceptible to vascular extravasation. The extent of increased vascular permeability from MABs depends on various factors, including the severity of existing amyloid angiopathy, the effectiveness of amyloid clearance, and local inflammatory response. Long-term treatment leads to the removal of amyloid from the vessel walls and the reorganization of arterial smooth muscle. 
Pretreatment risk factors for ARIA include preexisting microhemorrhage and siderosis and apolipoprotein E4 carriership. Another major risk factor for ARIA-E and ARIA-H is drug dosage. Most cases of ARIA develop during the titration phase and early in treatment, usually during the first eight doses. The standard dosing protocol for MAB therapy is 1 mg/kg for the first two treatments, 3 mg/kg for the third and fourth treatments, 6 mg/kg for the fifth and sixth treatments, and 10 mg/kg for the seventh and subsequent treatments. Most studies have found a notably higher occurrence of ARIA-E (30%-35%) in the initial stages of treatment compared with ARIA-H (15%-20%) 
MRI is the only imaging study that can detect ARIA. The International Collaboration for Real-World Evidence in Alzheimer’s Disease (ICARE AD) established baseline imaging and serial monitoring guidelines for ARIA. The ICARE AD protocol includes baseline pretreatment MRI, posttreatment MRI before the seventh and 12th infusions, and surveillance MRI every 6-12 months thereafter for up to 5 years. The mandatory sequences include three-dimensional T2-weighted fluid-attenuated inversion-recovery (FLAIR), T2*-weighted gradient-recalled echo (GRE), and diffusion-weighted imaging. The FLAIR sequence is essential for diagnosing ARIA-E, whereas GRE sequences are utilized for monitoring ARIA-H. For all sequence types, a section thickness ≤ 5 mm is recommended, with a minimum scanner field strength of 1.5 T owing to the limited availability of high-field scanners. Although 1.5 T is the minimum standard, centers with access to higher-field MRI scanners (3-T or 7-T magnets) should be consistent with field strength and vendor type. Close collaboration between clinicians and radiologists is essential to ensure proper monitoring and communication. 
ARIA-E is the most common side effect of MAB treatment and refers to parenchymal edema and/or sulcal effusion. Findings from the EMERGE and ENGAGE trials showed the highest risk for ARIA-E is in the initial 3 and 6 months of treatment, with a significant decrease in risk after the first 9 months of treatment. Parenchymal edema appears on T2-weighted FLAIR sequences as cortical-subcortical areas of hyperintensity with mild gyral swelling and mass effect. In addition, FLAIR sequences may reveal exudates as hyperintense signal and absence of CSF signal suppression. ARIA-E most commonly develops in the occipital lobes but also occurs in the frontoparietal lobes and the cerebellum. Edema and effusion may coincide, typically in the same area. The severity of ARIA-E depends on the area involved. Several grading scales, such as the Barkhof Grand Total Scale (BGTS), 3-point Severity Scale of ARIA-E (SSAE-3), and 5-point Severity Scale of ARIA-E (SSAE-5), are used to determine ARIA-E severity and guide clinicians in deciding whether MAB therapy should be continued, adjusted, or discontinued. In both SSAE grading systems, a single region with a maximum diameter < 5 cm is considered mild whereas > 10 cm is graded as severe. Studies, including the Clarity-AD trial, showed that the majority of ARIA-E events (approximately 80%) were asymptomatic and cleared up with dose adjustment. When symptoms did occur, headaches were the most frequently reported. 
ARIA-H is identified by parenchymal microhemorrhages and/or superficial siderosis and has an incidence of 15%-20%. Microhemorrhages are more prevalent than superficial siderosis and appear as small, rounded areas of signal loss measuring < 1 cm on the gradient sequences. Superficial siderosis appears as curvilinear areas of signal loss along the brain’s surface. Though siderosis is less frequent than microhemorrhage, it is more severe and often leads to discontinuation of MAB therapy. Careful assessment of the baseline number of microhemorrhages is critical to accurately grade severity. ARIA-H is graded as mild when four or fewer microhemorrhages are present or when there are two or fewer focal areas of superficial siderosis; it is considered severe when there are 10 or more microhemorrhages or more than two areas of superficial siderosis. Gradient and susceptibility-weighted sequences (SWI) are ideal for identifying ARIA-H. T2*-weighted GRE is still recommended based on clinical trials, but SWI is also widely available and offers higher spatial resolution. 
ARIA-E and ARIA-H often occur together due to the increased vascular permeability that permits fluid and heme substances to traverse the compromised vessel wall. Some studies suggest that ARIA-E almost always occurs to some extent with ARIA-H; however, ARIA-H does not always occur with ARIA-E. The decision to continue patients with ARIA on MAB therapy depends on proper grading and monitoring. Because ARIA-E is usually asymptomatic and self-limiting, most patients may continue or temporarily discontinue MAB therapy. The decision to continue patients with ARIA-H on MAB therapy depends on the severity and whether the condition remains stable over time. 
Because of overlapping pathophysiology, conditions such as inflammatory cerebral amyloid angiopathy (CAA) can mimic ARIA on MRI. Inflammatory CAA comprises two subtypes: CAA-related inflammation (CAA-RI) and amyloid-beta–related angiitis (AB-RA). CAA-RI, an autoimmune-related spontaneous type of ARIA, develops secondary to the targeting of amyloid-beta protein deposits in vessel walls and is indistinguishable from ARIA on imaging studies, with the only distinction being a clinical history of anti–amyloid-beta MAB use. In addition to inflammatory CAA, posterior reversible encephalopathy syndrome (PRES), progressive multifocal encephalopathy, subacute infarcts, and vasculitis also should be considered in the differential diagnosis. Among these conditions, ARIA-E closely resembles PRES in that both are reversible and tend to develop in the occipital lobes with petechial hemorrhages.
In conclusion, anti–amyloid-beta MAB therapy is becoming more prevalent in managing early-stage dementia in AD. Therefore, early detection is key to identifying patients who would potentially benefit from this treatment. It is important, however, that patients undergoing MAB therapy be informed of and surveilled for the side effects associated with this therapy, particularly ARIA-E and ARIA-H.