This preprint reports a blinded, MRI–histopathology correlative study in 146 5xFAD mice across seven cohorts, demonstrating that weekly anti-Aβ antibodies induce ARIA-like lesions detectable by clinically relevant T2w/T2* MRI sequences at 9.4T. The central finding is a dose- and grade-dependent relationship between meningovascular inflammation (MVI) and ARIA-E-like events: odds ratios of 44 (Grade 2), 342 (Grade 3), and 3430 (Grade 4) versus Grades 0–1, with 99% sensitivity and 74% specificity for MVI detecting ARIA on MRI.
The model's strongest translational claim is that differential ARIA incidence across antibodies (3D6 ~100%, aducanumab ~50%, gantenerumab ~25%, aglycosylated variants 0%) parallels clinical rankings, with bapineuzumab highest in patients . The aglycosylated controls and TfR-shuttle validation (ATV:Aβ low ARIA; trontinemab <5% clinically) provide independent effector-function and delivery-variable support.
Weekly dosing produced near-100% ARIA incidence—far exceeding clinical rates (e.g., 24.9% ARIA-E in gantenerumab Phase 3 ), making the model more useful for rank-ordering than absolute risk prediction. 5xFAD is not a robust CAA model. All authors are Biogen employees/shareholders , creating sponsor bias risk for positive reporting. Microinfarcts with restricted diffusion were observed in mice but have not been confirmed in humans—the authors acknowledge this gap honestly. No female mice were used, limiting sex-dependent APOE4 interaction analysis.
Know what changed, what holds up, and what remains uncertain. Every Friday. No ads.