Fatal iatrogenic cerebral β-amyloid-related arteritis in a woman treated with lecanemab for Alzheimer’s disease

Over the past 15 years, passive immunotherapies targeting β-amyloid have been tested to clear plaques, oligomers, and soluble Aβ in Alzheimer’s disease, following earlier active vaccination efforts (e.g., AN1792) that were halted due to encephalitis in a subset of patients. Those prior cases showed accelerated cerebral amyloid angiopathy (CAA), blood–brain barrier disruption, inflammation, and hemorrhage, though mechanisms remain uncertain. With monoclonal antibodies, a related clinical–radiographic syndrome termed amyloid-related imaging abnormalities (ARIA) has been common (up to ~40% in some trials), often asymptomatic, but with poorly defined neuropathology and uncertain mechanistic equivalence to CAA-related inflammation (CAA-ri). This report investigates a fatal case of ARIA after lecanemab (BAN2401) therapy, detailing clinical course, multimodal neuroimaging, and neuropathology to better define the syndrome and its mechanisms.

The paper situates the case within prior β-amyloid immunotherapy experience: AN1792 vaccination led to encephalitis with neuropathologic features including accelerated CAA, inflammation, and hemorrhage. ARIA has been frequently observed across passive antibody trials (e.g., aducanumab, donanemab, lecanemab), typically asymptomatic, but with limited post-mortem correlation in the literature, leaving questions about whether ARIA shares mechanisms with CAA-ri. Regulatory and trial-design history noted that early FDA guidance restricted participants with microhemorrhages due to ARIA risk; later, the Alzheimer’s Association working group coined ARIA (subtypes ARIA-E for edema/effusion, ARIA-H for hemorrhage). APOE4 genotype and presence of CAA are established risk factors for ARIA, and homozygous APOE4 patients have not clearly benefited in subgroup analyses in lecanemab and donanemab phase 3 trials. The authors highlight that standard clinical MRI can underestimate CAA burden relative to high-field post-mortem imaging and histopathology and advocate for using established Boston criteria for CAA in screening.

Design: Single-patient case report with clinical, in vivo MRI, post-mortem MRI, gross pathology, histology, immunohistochemistry (IHC), and tissue clearing with light-sheet microscopy. Ethics: Conducted under Vanderbilt University Medical Center IRB oversight (OSCAR protocol 10578) with consent for autopsy, tissue donation, and publication. In vivo clinical imaging: Pre-treatment MRI (FLAIR and T2*) at trial enrollment and post-treatment MRI after neurological deterioration; amyloid PET (18F-florbetaben) and tau PET (flortaucipir) were obtained before open-label extension. Post-mortem MRI: Coronal slabs embedded in 1% agarose; SWI and FLAIR at 3.0T (Philips Ingenia) and 7.0T (Philips Achieva), 32-channel head coils. Representative parameters: 3T SWI 3D GRE, 4 echoes (first echo ~7.2 ms), TR ~31 ms, 0.6×0.6×2 mm; 3T 3D FLAIR TE 271 ms, TR 4800 ms, TI 1650 ms, 1 mm isotropic; 7T SWI 3D GRE, 7 echoes (first echo 0.8 ms), TR ~64 ms, 0.65×0.65×0.9 mm; 7T 3D FLAIR TE 280 ms, TR 3925 ms, TI 1375 ms, 0.8 mm isotropic. Histology and IHC: Sixteen tissue blocks examined (representative six highlighted). Stains included H&E; β-amyloid (e.g., RBT-4A), phospho-tau (AT8), CD68 (macrophages/microglia), CD4/CD8 (T cells), IBA1 (microglia), ATP6V0A1 and PLD3 (dystrophic neurite markers), methoxy-X04/Thiazine Red for amyloid/aggregates. Confocal imaging via Zeiss LSM 710 (20×/63× objectives). Tissue clearing and light-sheet microscopy: CLARITY hydrogel embedding, SDS clearing at 37 °C to transparency, staining with tomato lectin and Thiazine Red, refractive index matching (TDE), imaged on light-sheet microscope; processing with Imaris. Statistics and reproducibility: Descriptive single-case analysis, no power calculation; no randomization/blinding; findings representative of examined tissue blocks. Data availability: All shareable data in manuscript/supplement; de-identified imaging available on request.

- Clinical course: 79-year-old woman with ~4-year mild progressive memory impairment; enrolled in CLARITY-AD placebo, then open-label lecanemab (1.0 mg/kg q2 weeks) for 3 infusions. Headache occurred ~1 hour after each infusion lasting 1–2 days. After third dose, developed worsening confusion (“brain fog”) and a focal seizure with secondary generalization; remained non-responsive. Treated with antiepileptics and IV methylprednisolone 1 g daily ×3 for suspected ARIA; concurrent heparin infusion for new atrial fibrillation. Suffered aspiration leading to sepsis and multiorgan failure; died 5 days after admission. - In vivo imaging: Pre-treatment MRI showed grade 2 age-related white matter changes (ARWMC scale), 4 cortical microhemorrhages on T2* and white matter disease consistent with probable CAA by Boston criteria. Post-treatment MRI showed multifocal vasogenic edema in bilateral temporal, parietal, and occipital lobes (ARIA-E) and increase in microhemorrhages to >30 (ARIA-H). No major acute infarct on DWI/ADC; no contrast enhancement. Pre-open-label PET showed amyloid and tau tracer retention. - Post-mortem MRI: At 3T and 7T, far more extensive microhemorrhagic changes than in vivo, prominent in temporal, parietal, and occipital regions; superficial siderosis; high-field imaging detected additional microbleeds. Optical clearing and backlighting confirmed numerous microhemorrhages. - Neuropathology: Moderate Alzheimer’s disease neuropathologic change (A3, B2, C3; Braak IV; frequent neuritic plaques). Severe CAA with β-amyloid deposition in meningeal vessels and penetrating arterioles, numerous microaneurysms, perivascular lymphocytic infiltrates, abundant macrophages/activated microglia (CD68+/IBA1+), occasional multinucleated giant cells, and arteriolar degeneration with fibrinoid necrosis. Extensive microhemorrhages in parenchyma and leptomeninges. - Plaque dynamics: Approximately 21% of plaques appeared “cleared” (excess dystrophic neurites without central amyloid staining); ~24% showed minimal amyloid staining; some regions with intense microglial activation (IBA1). - Genetics: APOE ε4/ε4 homozygosity confirmed. - Overall interpretation: Clinicopathologic picture consistent with severe, fatal ARIA associated with lecanemab, closely resembling CAA-related inflammation; death likely due to severe cerebral amyloid-related inflammation with widespread microvascular injury.

The case links lecanemab exposure to a fulminant ARIA syndrome characterized by multifocal vasogenic edema and widespread microhemorrhages with severe CAA and intense perivascular inflammation (macrophages/microglia, T cells) and fibrinoid necrosis. These neuropathological and imaging features resemble CAA-related inflammation, suggesting shared mechanisms (e.g., vascular Aβ clearance overwhelm and consequent perivascular inflammatory response). The findings address uncertainties about ARIA pathology by providing detailed neuroimaging–neuropathologic correlation, showing that clinical MRI underestimates microhemorrhage burden relative to high-field post-mortem imaging and histology. Clinically, the case underscores that ARIA, while often asymptomatic, can be severe and fatal, particularly in patients with probable CAA and APOE4 homozygosity. The authors argue for rethinking terminology to distinguish symptomatic/severe ARIA, emphasize risk stratification (Boston CAA criteria, APOE genotyping), and recommend more rigorous MRI screening protocols (≥3T, susceptibility-weighted imaging with thin slices) to improve CAA detection and mitigate risk. They also highlight the limited benefit in APOE4 homozygotes observed in phase 3 subgroup analyses, suggesting avoidance of treatment in this genotype. Understanding the molecular mechanisms driving vascular inflammation during Aβ immunotherapy is identified as critical to improving safety.

This case report provides comprehensive clinical, imaging, and neuropathological evidence that severe, fatal ARIA can occur after lecanemab, with features overlapping CAA-related inflammation. It highlights limitations of standard clinical MRI in detecting the extent of microhemorrhages, the elevated risk associated with pre-existing CAA and APOE ε4/ε4 genotype, and the need for refined patient selection. The authors recommend: (1) screening for CAA using Boston criteria with standardized high-quality MRI (≥3T SWI); (2) APOE genotyping before treatment; and (3) enhanced clinician education on ARIA recognition and management. Future research should focus on defining cellular and molecular mechanisms of ARIA, validating improved screening and monitoring protocols, clarifying plaque and neurite dynamics during/after Aβ clearance, and rigorously assessing risk–benefit in high-risk genotypes.

Single-patient case report without controls limits generalizability and causal inference. Imaging, histology, and outcome assessments were descriptive, with no randomization or blinding. Clinical management (e.g., anticoagulation for atrial fibrillation) may confound the course. Findings may not represent the full spectrum of ARIA. Standard clinical MRI underestimated microhemorrhage burden compared to post-mortem imaging.