Amyloid PET is often treated as a binary or scalar biomarker: positive or negative, high or low, above or below a Centiloid threshold. That framework is clinically useful, especially for confirming Alzheimer's disease biology and determining eligibility for anti-amyloid therapy. But it compresses a spatially complex disease process into a single global number.

A growing literature suggests that where amyloid binds may carry clinically important information beyond how much amyloid is present. One emerging pattern is posterior-predominant amyloid PET binding, characterized by disproportionate amyloid signal in posterior cortical regions, particularly occipital and parietal cortex.

The key observation is not simply that these patients have more amyloid. In large cohort analyses, posterior-predominant amyloid binding appears to identify a biologically distinct subgroup: more severe clinical impairment, lower APOE-e4 carriership, greater posterior tau burden, posterior cortical thinning, and enrichment for cerebral amyloid angiopathy at autopsy. The largest reported analysis used data-driven amyloid PET topography methods across four cohorts, including IDEAS, LEADS, ADNI, and UCSF, totaling 12,379 cognitively impaired participants.

This matters because amyloid PET is no longer only a diagnostic tool. In the anti-amyloid treatment era, amyloid biomarkers increasingly inform treatment selection, longitudinal monitoring, and risk stratification. If posterior amyloid PET signal is enriched for CAA biology, then amyloid topography may become clinically relevant for estimating vascular amyloid burden and ARIA vulnerability.

The Limitation of Global Amyloid Burden

The Centiloid framework was developed to standardize amyloid PET quantification across tracers, scanners, and analytic pipelines. This has been essential for research harmonization and for clinical use, particularly when amyloid PET is used to support diagnosis or monitor amyloid-lowering therapy.

However, global Centiloid values average signal across composite cortical regions. This approach is useful but anatomically blunt. Two patients may have similar global amyloid burden but very different regional distributions. One may show a more typical diffuse or frontal-predominant Alzheimer pattern; another may show posterior/occipital-predominant uptake. A single scalar value can obscure this distinction.

That distinction may be clinically meaningful. Amyloid PET tracers bind fibrillar amyloid, but they do not perfectly distinguish parenchymal plaques from vascular amyloid deposition. This limitation becomes important in cerebral amyloid angiopathy, where amyloid accumulates in cortical and leptomeningeal vessel walls rather than only in parenchymal plaques. Amyloid PET has been shown to have moderate-to-good diagnostic accuracy for differentiating probable CAA from some control groups, but positive scans remain nonspecific because both Alzheimer's plaques and vascular amyloid can contribute to signal.

The implication is straightforward: a global amyloid-positive scan confirms fibrillar amyloid biology, but topography may help refine what kind of amyloid-positive brain we are looking at.

The Posterior-Predominant Phenotype

Recent data-driven analyses of amyloid PET images have identified a posterior-predominant amyloid-positive subgroup. In the AAIC/Alzheimer's & Dementia report, investigators applied spatial modeling to amyloid PET scans from cognitively impaired participants and identified three broad groups: amyloid-negative, amyloid-positive typical, and amyloid-positive posterior. The posterior group showed predominant occipital binding and replicated across independent cohorts.

The 2026 preprint expanded this work and reported that posterior-predominant amyloid-positive participants were less likely to carry APOE-e4, more severely impaired, had thinner posterior cortex, greater posterior tau PET burden, and higher likelihood of CAA at autopsy. Because this is a preprint, these findings should be treated as important but still provisional pending peer review.

The central clinical point is that posterior-predominant binding does not appear to be merely a "high amyloid" state. It may represent a spatially distinct amyloid phenotype with different associations from the more typical amyloid-positive pattern.

Why the Occipital Lobe Matters

The occipital cortex has long been relevant to CAA. CAA has a posterior predilection in many patients, and occipital amyloid PET signal has been repeatedly discussed as a potential imaging clue to vascular amyloid. However, amyloid PET alone is not diagnostic of CAA because the tracer signal can reflect both vascular and parenchymal amyloid.

Posterior-predominant amyloid PET binding may therefore be best understood as a risk-enrichment signal, not a stand-alone diagnostic marker. In the right clinical context, posterior amyloid uptake should increase suspicion for CAA overlap, especially when MRI shows lobar microbleeds, cortical superficial siderosis, convexity subarachnoid hemorrhage, posterior-predominant white matter disease, or enlarged centrum semiovale perivascular spaces.

This is particularly relevant because CAA and Alzheimer's disease frequently overlap. Pathology-based estimates suggest that CAA is common in Alzheimer's disease, and imaging-based markers may underestimate its true prevalence.

APOE: A Useful but Incomplete Risk Axis

One of the most interesting findings is that posterior-predominant amyloid binding appears to be associated with lower APOE-e4 carriership compared with typical amyloid-positive binding.

This is somewhat counterintuitive. APOE-e4 is strongly associated with Alzheimer's disease risk, amyloid positivity, CAA, and ARIA risk during anti-amyloid therapy. But the posterior-predominant phenotype suggests that APOE-e4 may not explain all clinically relevant amyloid biology.

Several interpretations are plausible:

• APOE-e4 may preferentially drive a more typical diffuse or anterior/frontal-heavy amyloid distribution.
• Posterior-predominant binding may reflect partially APOE-independent mechanisms, including vascular amyloid deposition, impaired perivascular clearance, regional vascular vulnerability, or non-APOE genetic modifiers.
• The posterior phenotype may be biologically heterogeneous, capturing patients with atypical Alzheimer's disease, CAA overlap, posterior cortical atrophy, mixed vascular-neurodegenerative disease, or posterior tau vulnerability.

The practical takeaway is that APOE remains important, but it should not be treated as the only axis of treatment risk or disease heterogeneity.

Clinical Correlates

Posterior-predominant amyloid binding appears to track with more severe clinical impairment, despite not necessarily having higher global amyloid burden. This suggests that topography may capture clinically meaningful biology missed by global quantification.

Clinically, this phenotype may be relevant in patients with:

• posterior cortical atrophy;
• visuospatial dysfunction;
• visual processing complaints;
• executive dysfunction;
• slowed processing speed;
• dyspraxia or corticobasal features;
• mixed AD/vascular presentations;
• MRI features suspicious for CAA.

It may also help explain patients who appear "more impaired than expected" based on hippocampal volume, memory-predominant models, or global amyloid burden alone.

Tau and Posterior Network Degeneration

Amyloid burden alone does not explain cognitive phenotype. Symptoms correlate more closely with tau distribution, neurodegeneration, synaptic dysfunction, and network-level failure.

The posterior-predominant amyloid group appears to have greater posterior tau PET burden and posterior cortical thinning. This raises an important mechanistic question: does posterior amyloid promote posterior tau deposition, or does it identify a brain already vulnerable to posterior tau propagation?

Several models are possible:

• Local permissive model: posterior amyloid creates a regional environment that facilitates tau aggregation and neuroinflammation.
• Network vulnerability model: posterior cortical networks are selectively vulnerable in atypical Alzheimer phenotypes.
• Vascular-neurodegenerative model: CAA-related vascular dysfunction amplifies posterior cortical injury.
• Shared vulnerability model: regional transcriptomic, metabolic, or vascular factors predispose posterior cortex to both amyloid signal and tau-mediated neurodegeneration.

From a clinical standpoint, the key implication is that posterior amyloid should prompt careful assessment of posterior tau, posterior atrophy, visuospatial function, and CAA risk.

Implications for Anti-Amyloid Therapy

Anti-amyloid monoclonal antibodies have changed the clinical meaning of amyloid PET. A positive amyloid biomarker may now lead to treatment consideration, serial MRI surveillance, ARIA counseling, and long-term risk monitoring.

ARIA risk is influenced by multiple factors, including APOE genotype, baseline microhemorrhages, cortical superficial siderosis, CAA-like imaging findings, antibody type, dose/titration, antithrombotic exposure, and early treatment phase. Appropriate use guidance for anti-amyloid therapies emphasizes careful MRI screening and exclusion of patients with imaging findings that suggest elevated hemorrhagic risk, including more than four microbleeds or cortical superficial siderosis.

Posterior-predominant amyloid PET binding should not currently be treated as a contraindication to anti-amyloid therapy. The evidence is not yet sufficient for that. However, it should probably be treated as a signal that warrants more careful review.

In a specialty memory clinic, posterior-predominant amyloid PET binding should prompt:

• detailed review of SWI/GRE MRI sequences;
• explicit assessment for strictly lobar microbleeds;
• assessment for cortical superficial siderosis;
• review for convexity subarachnoid hemorrhage or prior lobar hemorrhage;
• attention to posterior white matter disease;
• review of antiplatelet, anticoagulant, and thrombolytic exposure;
• APOE genotyping when considering anti-amyloid therapy;
• careful counseling about CAA overlap and ARIA uncertainty.

The practical point is not that PET topography replaces MRI safety screening. It does not. Rather, posterior amyloid topography may identify patients in whom MRI review and treatment counseling should be especially rigorous.

A Practical Reporting Framework

Amyloid PET reports should ideally move beyond "positive/negative" and global burden alone. A more clinically useful report in the treatment era would include:

• Amyloid status: positive, negative, or equivocal.
• Quantitative burden: Centiloid or tracer-specific quantitative value.
• Topographic pattern: typical diffuse, posterior-predominant, frontal-predominant, asymmetric, or otherwise atypical.
• Occipital comment: whether occipital uptake is disproportionate.
• CAA relevance: posterior/occipital-predominant uptake may be associated with CAA but is not diagnostic by itself.
• Recommended correlation: MRI susceptibility markers, clinical syndrome, APOE genotype, tau biomarkers, and vascular risk profile.

This would better align amyloid PET interpretation with how these scans are now being used clinically: not just to diagnose Alzheimer's biology, but to support treatment decisions.

Bottom Line

A global amyloid PET value answers one question:

How much fibrillar amyloid signal is present?

Posterior-predominant amyloid PET binding asks a more nuanced question:

What kind of amyloid-positive brain is this?

That distinction may matter. Posterior-predominant amyloid binding appears to identify a subgroup with greater posterior network involvement, lower APOE-e4 carriership, more posterior tau burden, and higher likelihood of CAA pathology. The finding is not yet ready to function as a treatment exclusion criterion, but it is highly relevant to diagnostic formulation, MRI review, and anti-amyloid therapy risk counseling.

In the era of disease-modifying therapy, amyloid PET should no longer be interpreted only as a global burden score. Topography matters.

Key References

• Giorgio J, et al. Robust associations with posterior predominant amyloid PET binding. Alzheimer's & Dementia. 2025.
  Data-driven amyloid PET analysis across four cohorts totaling 12,379 cognitively impaired participants; identified a posterior-predominant amyloid-positive subgroup associated with clinical severity, lower APOE-e4 carriership, posterior tau, and CAA at autopsy.
• Giorgio J, et al. Characterisation of posterior predominant amyloid PET binding across multiple cohorts. bioRxiv. 2026.
  Expanded preprint analysis of posterior-predominant amyloid PET binding; important but not yet peer-reviewed.
• Charidimou A, Farid K, Baron J-C. Amyloid-PET in sporadic cerebral amyloid angiopathy: A diagnostic accuracy meta-analysis. Neurology. 2017.
  Shows amyloid PET may help rule out CAA when negative, but positive scans are nonspecific because tracers detect both vascular and parenchymal amyloid.
• Theodorou A, et al. Cerebral amyloid angiopathy and amyloid load distribution. 2025.
  Systematic review/meta-analysis examining global and regional amyloid PET uptake patterns in CAA.
• Collij LE, et al. Centiloid recommendations for clinical context-of-use from the AMYPAD consortium. Alzheimer's & Dementia. 2024.
  Establishes clinical context for Centiloid quantification as an adjunct to visual amyloid PET assessment.
• Rabinovici GD, et al. Updated appropriate use criteria for amyloid and tau PET. 2025.
  Provides expert recommendations for amyloid and tau PET use in the evolving diagnostic and therapeutic Alzheimer landscape.
• Farrar G, et al. Expert opinion on Centiloid thresholds suitable for initiating anti-amyloid therapy. 2025.
  Discusses practical Centiloid thresholds for anti-amyloid therapy decision-making in early Alzheimer's disease.