Lecanemab Effects on Aβ Plaques

A Whole-Brain Light-Sheet Microscopy Analysis

Spatial distribution · Treatment efficacy · Vascular relationships

March 2026

Outline

  1. Background — Alzheimer's disease, Lecanemab & ApoE
  2. Dataset — Cohort, imaging modality, data structure
  3. Analysis Pipeline — Overview of 5-notebook workflow
  4. Methods — Regional, treatment & vascular analyses
  5. Results — Two cohorts: Early (3–6 mo) and Late (12–15 mo) treatment

Background

Alzheimer's Disease & Aβ Plaques

Amyloid-β (Aβ) plaques are a hallmark pathological feature of Alzheimer's disease, accumulating in the brain parenchyma and around cerebral vasculature.

Key questions:

  • Which brain regions accumulate the most plaques?
  • Does Lecanemab (anti-Aβ immunotherapy) reduce plaque burden?
  • Does ApoE genotype (ApoE3 vs ApoE4) modulate treatment response?
  • What is the spatial relationship between plaques and blood vessels?

Experimental model:

  • Humanized mouse model of Alzheimer's disease
  • hAppNL-F-hMAPT-APOE4 and hAppNL-F-hMAPT-APOE3

Treatment arms:

Group Description
PBS Vehicle control
Lecanemab Anti-Aβ immunotherapy

Readout: Instance-level Aβ plaque segmentation via whole-brain 3-D light-sheet microscopy

Why Light-Sheet Microscopy?

Whole-brain 3-D imaging at near-cellular resolution enables:

  • Unbiased spatial mapping of every Aβ plaque
  • Registration to a common template (CCFv3 / ABAv3)
  • Quantification of vessel proximity via CD31 vascular segmentation

Voxel resolution: 1.6 × 1.6 × 2.75 µm

Signed Distance Transform (SDT):

A plaque's sdt_CD31 value encodes its distance to the nearest vessel wall:

SDT value Location
< 0 Inside vessel lumen
0 – 5 µm Near vessel surface
> 5 µm Far from vasculature

Vessel calibre estimated as:
diam_vessel ≈ diam_plaque + 2|SDT|

Dataset

Two Cohorts: Early vs Late Treatment

Early cohort (data_early.parquet)

  • Lecanemab injections at 3–6 months
  • Light-sheet imaging at 12 months

Late cohort (data_late.parquet)

  • Lecanemab injections at 12–15 months
  • Light-sheet imaging at 15 months

Each cohort is analysed independently; figures are saved as
fig_early_*.png and fig_late_*.png

Per-cohort plaque metrics:

  • nvoxels — voxel count (min: 200)
  • plaque_vol_ml — volume (µL)
  • equiv_diam_um — equivalent spherical diameter

Spatial coordinates:

  • pos_x / y / z — raw image space
  • template_x / y / z — CCFv3 template space

Demographics:

  • subject — individual animal ID
  • genotype — ApoE3 / ApoE4
  • sex — M / F
  • treatment — PBS / Lecanemab

Analysis Pipeline

Five-Notebook Workflow

import_data.ipynb
   │  Load raw data · Derive metrics
   │  → data_early.parquet  +  data_late.parquet
   ▼
regional_dataframe.ipynb
   │  Classify plaque–vessel proximity (per dataset)
   │  Aggregate to per-subject × per-region ROI summaries
   │  → roi_data_early.parquet  +  roi_data_late.parquet
   ├──▶ regional_analysis.ipynb
   │       Atlas heatmaps · ROI rankings · Fold-change maps
   │       → fig_{early|late}_atlas_*.png  etc.
   ├──▶ treatment_effect_analysis.ipynb
   │       Two-way ANOVA · Post-hoc comparisons · ECDFs
   │       → fig_{early|late}_boxplots_*.png  etc.
   └──▶ vessel_spatial_analysis.ipynb
           SDT distributions · Proximity fractions · Vessel calibre
           → fig_{early|late}_sdt_ecdf.png  etc.

Reproducibility: All notebooks executed via pixi run run_notebooks
using the pinned Pixi environment (pixi.lock)

Methods

Methods: Data Preprocessing

Step 1 — Plaque size filtering

  • Minimum size threshold: 200 voxels (~1,408 µm³ at 1.6 × 1.6 × 2.75 µm voxel size)
  • Maximum: 1×10⁻⁴ mL (removes implausibly large objects)

Step 2 — Plaque–vessel proximity classification

Category Criterion
Inside vessel SDT < 0
Near vessel 0 ≤ SDT < 5 µm
Far from vessel SDT ≥ 5 µm

Step 3 — Regional aggregation

  • Each plaque mapped to an ABA v3 atlas region
  • Per-subject, per-region counts → plaque density (plaques/mm³)
  • Proximity fractions computed per region per subject

Methods: Regional Analysis

Objective: Identify brain regions with the highest Aβ burden and characterise spatial patterns.

Approaches:

  • ROI ranking — Top regions by mean plaque density across subjects
  • Multi-metric heatmap — Normalised [0, 1] comparison of density, diameter, and vessel-fraction across top ROIs
  • Treatment fold-change — Log₂(Lecanemab / PBS) per ROI
  • Atlas heatmaps — Max-intensity projections of density, fold-change, mean diameter, and intravascular fraction painted onto the 3-D Allen Brain Atlas volume (3 orthogonal views)
  • ROI density boxplot — Per-subject density distributions across top regions
  • ROI proximity fractions — Fraction of plaques near/inside vessels per region

Methods: Treatment Effect Analysis

Objective: Quantify the effect of Lecanemab treatment on Aβ plaque burden and size, modulated by ApoE genotype.

Statistical framework:

Test Purpose
Two-way ANOVA Treatment × Genotype interaction
Tukey HSD post-hoc Pairwise group comparisons
Mann-Whitney U Within-genotype PBS vs Lecanemab
ECDF plots Distribution-level comparisons

Results — Early Cohort

(Lecanemab 3–6 mo, imaging at 12 mo)

[Early] Regional Analysis

[Early] Regional Plaque Burden

Top brain regions ranked by mean Aβ plaque density — early cohort.

[Early] ROI Density Boxplot

Per-subject plaque density distributions across top regions — early cohort.

[Early] Multi-Metric Regional Heatmap

Each metric is normalised [0, 1] per column — early cohort.

[Early] ROI Proximity Fractions

Fraction of plaques near/inside vessels per region — early cohort.

[Early] ROI Treatment Fold-Change

Log₂ fold-change (Lecanemab / PBS) per ROI — early cohort.

[Early] Atlas-Level Plaque Density — All Subjects

Max-intensity projection of mean plaque density — early cohort.

[Early] Atlas-Level Density by Group

Plaque density heatmaps by treatment × genotype — early cohort.

[Early] Treatment Fold-Change Atlas

Log₂ fold-change (Lecanemab / PBS) painted onto atlas volume — early cohort.

[Early] Atlas Mean Plaque Diameter

Mean equivalent plaque diameter painted onto atlas volume — early cohort.

[Early] Atlas Fraction Inside Vessel

Fraction of intravascular plaques painted onto atlas volume — early cohort.

[Early] Treatment Effect Analysis

[Early] Treatment Effects — Subject-Level Metrics

Subject-level summary metrics by treatment × genotype — early cohort.

[Early] Violin — Plaque Size by Treatment & Genotype

Per-plaque size distributions by treatment × genotype — early cohort.

[Early] Treatment × Genotype Interaction

Mean ± SEM interaction plots for key metrics — early cohort.

[Early] Plaque Count Stratified by Genotype

Per-subject plaque counts stratified by genotype — early cohort.

[Early] ECDF — Plaque Size by Genotype

Empirical CDF of per-plaque size by genotype — early cohort.

[Early] Plaque Size Distribution

Histogram of per-plaque size distributions by genotype — early cohort.

[Early] Vessel Spatial Analysis

[Early] SDT ECDF — Distance to Vessel Wall

ECDF of signed distance transform (SDT) to nearest vessel — early cohort.

[Early] Plaque–Vessel Proximity Fractions (Stacked)

Stacked bar charts of categorized plaque–vessel proximity — early cohort.

[Early] Plaque–Vessel Proximity Fractions

Grouped bar charts of plaque–vessel proximity fractions — early cohort.

[Early] Proximity × Treatment Interaction

Interaction plots for proximity fractions by treatment × genotype — early cohort.

[Early] Vessel Calibre ECDF

ECDF of estimated vessel calibre for intravascular plaques — early cohort.

[Early] Vessel Calibre by Subject

Estimated vessel calibre per subject — early cohort.

[Early] Vessel Diameter Bins

Distribution of plaques across vessel diameter bins — early cohort.

[Early] Spatial Vessel Proximity

Spatial mapping of plaque–vessel proximity across the brain — early cohort.

Results — Late Cohort

(Lecanemab 12–15 mo, imaging at 15 mo)

[Late] Regional Analysis

[Late] Regional Plaque Burden

Top brain regions ranked by mean Aβ plaque density — late cohort.

[Late] ROI Density Boxplot

Per-subject plaque density distributions across top regions — late cohort.

[Late] Multi-Metric Regional Heatmap

Each metric is normalised [0, 1] per column — late cohort.

[Late] ROI Proximity Fractions

Fraction of plaques near/inside vessels per region — late cohort.

[Late] ROI Treatment Fold-Change

Log₂ fold-change (Lecanemab / PBS) per ROI — late cohort.

[Late] Atlas-Level Plaque Density — All Subjects

Max-intensity projection of mean plaque density — late cohort.

[Late] Atlas-Level Density by Group

Plaque density heatmaps by treatment × genotype — late cohort.

[Late] Treatment Fold-Change Atlas

Log₂ fold-change (Lecanemab / PBS) painted onto atlas volume — late cohort.

[Late] Atlas Mean Plaque Diameter

Mean equivalent plaque diameter painted onto atlas volume — late cohort.

[Late] Atlas Fraction Inside Vessel

Fraction of intravascular plaques painted onto atlas volume — late cohort.

[Late] Treatment Effect Analysis

[Late] Treatment Effects — Subject-Level Metrics

Subject-level summary metrics by treatment × genotype — late cohort.

[Late] Violin — Plaque Size by Treatment & Genotype

Per-plaque size distributions by treatment × genotype — late cohort.

[Late] Treatment × Genotype Interaction

Mean ± SEM interaction plots for key metrics — late cohort.

[Late] Plaque Count Stratified by Genotype

Per-subject plaque counts stratified by genotype — late cohort.

[Late] ECDF — Plaque Size by Genotype

Empirical CDF of per-plaque size by genotype — late cohort.

[Late] Plaque Size Distribution

Histogram of per-plaque size distributions by genotype — late cohort.

[Late] Vessel Spatial Analysis

[Late] SDT ECDF — Distance to Vessel Wall

ECDF of signed distance transform (SDT) to nearest vessel — late cohort.

[Late] Plaque–Vessel Proximity Fractions (Stacked)

Stacked bar charts of categorized plaque–vessel proximity — late cohort.

[Late] Plaque–Vessel Proximity Fractions

Grouped bar charts of plaque–vessel proximity fractions — late cohort.

[Late] Proximity × Treatment Interaction

Interaction plots for proximity fractions by treatment × genotype — late cohort.

[Late] Vessel Calibre ECDF

ECDF of estimated vessel calibre for intravascular plaques — late cohort.

[Late] Vessel Calibre by Subject

Estimated vessel calibre per subject — late cohort.

[Late] Vessel Diameter Bins

Distribution of plaques across vessel diameter bins — late cohort.

[Late] Spatial Vessel Proximity

Spatial mapping of plaque–vessel proximity across the brain — late cohort.

Slides built with Marp — source: presentation.md