Alzheimer’s Is a Delivery Problem: A Neurosurgery-Enabled Evidence Map of Focused Ultrasound Blood-Brain Barrier Opening as a CNS Delivery Platform

Philadelphia, PA

Introduction: Disease-modifying therapies for Alzheimer’s disease (AD) provide modest benefit in selected early-AD populations, but their use is limited by safety risks (e.g., amyloid-related imaging abnormalities [ARIA]) and intensive monitoring. A key reason many therapies underperform is that the blood–brain barrier (BBB) acts like a protective “filter,” allowing only a small fraction of systemically delivered drugs, especially large molecules, to reach the brain. MRI-guided focused ultrasound, paired with intravenously administered microbubbles, can temporarily and locally “loosen” the BBB at a chosen brain region, then allow it to reseal. This approach is being studied as a noninvasive delivery platform that could increase brain exposure where it is needed most and support next-generation AD therapeutic development.

Objective: To map the human FUS-BBBO AD evidence base and characterize target regions, treatment schedules, safety, and endpoint maturity (cognition, amyloid PET, fluid biomarkers).

Methods: We performed a structured evidence map of human studies evaluating FUS-BBBO in AD. Extracted variables included target region, treatment schedule (single vs repeated), BBB opening/closure confirmation, and outcomes across four domains: safety/adverse events, cognition, amyloid PET, and fluid biomarkers (CSF/plasma/EV). Endpoint reporting was summarized as a matrix and scored using an Endpoint Maturity Score (0–4; one point each for safety, cognition, PET, and fluid biomarkers). Studies were categorized into Target-Engagement Tiers: Tier 0 (BBBO confirmation only), Tier 1 (BBBO + PET change), Tier 2 (Tier 1 + fluid biomarker signal), Tier 3 (Tier 2 + durability ≥6–12 months). Findings were synthesized descriptively without meta-analysis.

Results: Ten studies (N = 63 participants) met inclusion. Targeting clustered by region: hippocampus/entorhinal (2), frontal lobe (6), multifocal (2), and default mode network (1). Protocols were predominantly repeated-session regimens (9/10), with one single-session portable system study. When specified, MRI-confirmed BBB closure occurred within 24–72 hours. No procedure-related life-threatening events were reported; where described, imaging changes (e.g., transient edema or susceptibility effects) were self-limited and resolved on follow-up imaging.

Endpoint reporting was heterogeneous: safety was reported in 10/10 studies; cognition in 8/10; amyloid PET in 7/10; and fluid biomarkers in 2/10. The median Endpoint Maturity Score was 3/4, reflecting common reporting of safety/cognition/PET but limited biomarker integration. Tier 1 target-engagement evidence (BBBO + PET change, including treated-versus-contralateral or regional PET differences) was present in 7/10 studies, whereas Tier 2 evidence (Tier 1 + fluid biomarkers) was limited (2/10).

Conclusion: Human FUS-BBBO studies in AD consistently demonstrate reversible BBB permeability changes with an encouraging early safety profile across multiple targeting strategies, but efficacy interpretation is limited by endpoint heterogeneity and sparse biologic confirmation. Only a minority incorporate fluid biomarkers, and follow-up windows vary widely, limiting dose–response and durability assessment. Standardized reporting of exposure (target/volume/interval), harmonized PET ROI methods (including treated-versus-contralateral comparisons), and a minimum biomarker set would enable cross-trial comparability and accelerate platform-ready designs for combination therapy testing.