Description

1. Synopsis / Summary

Design: Comparative imaging study.

Population: 60 participants:

Group A: 20 patients with suspected Alzheimer's disease (AD) without type 2 diabetes (T2D)

Group B: 20 patients with suspected AD + T2D

Group C: 20 healthy age-matched controls

Procedures

Dynamic \[11C\]PiB positron emission tomography (PET) after acetazolamide: early phase as perfusion/vascular function marker; late phase as β-amyloid load.

Dynamic \[18F\]FDG PET with heart scans and prebolus method: glucose transport and cerebral metabolic rate of glucose (CMRglc).

Magnetic resonance imaging (MRI) (if not available within 5 years and clinically acceptable): assessment of white matter lesions (Fazekas score) and vascular markers.

Neuropsychological testing (if not already available clinically).

Primary endpoint: Volume and overlap of abnormal regions across imaging biomarkers (with diabetes status considered).

Secondary endpoints: Correlation between cognitive index score and biomarker volumes.

Timeline: First subject first visit Sep 2025; last subject last visit Dec 2026; end of trial Sep 2027. 2. Background and Rationale

Dementia prevalence is rising; Alzheimer's disease and vascular dementia account for the majority of cases, often co-existing ("mixed dementia"). AD is characterized by early β-amyloid deposition and later neurodegeneration, with hypometabolism and perfusion abnormalities in characteristic parietotemporal regions. T2D is associated with cognitive impairment and increased dementia risk, potentially via impaired insulin signaling and altered brain glucose handling.

A key methodological barrier has been robust, clinically feasible quantification of glucose transport across the blood-brain barrier (BBB). Prior work suggests blood-brain glucose transport is sensitive to AD-related changes and may respond to metabolic interventions. Conventional glucose transport estimation requires arterial cannulation and high temporal resolution. We have developed a non-invasive approach using an image-derived input function with heart scanning and a prebolus strategy (method manuscript in preparation), enabling estimation of CMRglc and glucose transport in the same FDG session.

For vascular reserve, acetazolamide challenge probes endothelial-mediated vasodilation. We will use early-phase \[11C\]PiB as a perfusion proxy (validated against FDG/perfusion measures in prior studies).

Overall aim: Map and compare regional extent and overlap of quantitative imaging biomarkers in suspected AD with and without T2D versus healthy controls. 3. Hypotheses

Patients with suspected AD are amyloid-positive on \[11C\]PiB PET.

Aside from amyloid load, biomarkers show broadly similar regional patterns within subjects, but differ in affected volume.

Glucose transport impairment affects a larger cortical volume than glucose metabolism impairment (CMRglc).

Perfusion during vasodilation shows larger affected volume than glucose metabolism.

White matter lesion burden co-localizes with affected regions and differs by diabetes status. 4. Objectives and Endpoints 4.1 Primary objective

Map neurodegenerative PET biomarkers and vascular MRI markers in suspected AD and compare the volume and overlap of regional involvement across biomarkers (accounting for T2D).

4.2 Secondary objectives

Relate imaging biomarker volumes to cognitive performance.

4.3 Primary endpoint

Volume and overlap of:

Reduced glucose transport from FDG PET

Reduced CMRglc from FDG PET

Increased amyloid load from late \[11C\]PiB PET (standard uptake value ratio, (SUVR) \> 1.5, cortex vs cerebellum)

Reduced perfusion during vasodilation from early \[11C\]PiB PET

White matter lesions (none / some / confluent; incl. Fazekas scoring)

4.4 Secondary endpoints

Association between cognitive index score and volumes of the above abnormalities.

4.5 Exploratory endpoints

Associations between imaging volumes and metabolic/clinical measures (body mass index (BMI), waist, fasting glucose/insulin measures, etc.). 5. Methods 5.1 Design

Comparative cross-sectional imaging study with prespecified group comparisons (A vs B vs C).

5.2 Schedule

Procedures are conducted over 2-3 study visits. 6. Trial Population 6.1 Recruitment

Patients (A, B): Identified through the memory clinic at Herlev or referrals to clinical brain FDG PET. A health professional asks permission for research contact; written information is provided. Research staff contacts the patient ≥3 days later to screen and offer an information meeting.

Controls (C): Recruitment via approved advertisements (TrialTree, forskningnu.dk, local media, posters, social media) with safeguards against tagging/comment disclosure.

6.2 Inclusion criteria

Written informed consent before any study procedures

Age 60-90 years

Group A/B: clinically suspected AD

Group B: documented T2D

Group C: no significant somatic/psychiatric disease

Ability and willingness to comply with protocol

6.3 Exclusion criteria

Suspicion of non-AD dementia

Significant brain disease unrelated to dementia

Medication affecting perfusion or glucose metabolism (e.g., GLP-1 receptor agonists; metformin/insulin allowed in Group B)

Active cancer treatment

Alcohol/drug abuse history

Severe claustrophobia

Pregnancy/breastfeeding

MRI contraindications (metal fragments, pacemaker, incompatible implants, etc.)

For controls: significant cardiovascular disease or diabetes (atrial fibrillation acceptable)

6.4 Informed consent

Consent is voluntary; withdrawal is allowed anytime without consequences for clinical care. Participants are offered up to 7 days to consider participation and may bring a relative/friend to the information meeting. After trial completion, participants will be informed of overall results; individual results will be handled with respect for the right not to know. Permission may be requested for future contact regarding follow-up trials.

6.5 Demography and clinical data

DOB, sex, height/weight/BMI, waist, blood pressure/pulse, smoking/alcohol history, medication list, diabetes history (if applicable), relevant clinical cognitive test results.

6.7 Biobank

No research biobank is created.
7. Study Procedures 7.1 \[18F\]FDG PET

Visit \~2 hours; replaces routine clinical FDG PET where applicable. 4-hour fasting. Venous access for tracer injection. Dynamic brain scan (\~55 min) with heart scans before and after for non-invasive input function. Total injected activity \~200 MBq split (prebolus + bolus). Plasma glucose measured if needed.

7.2 \[11C\]PiB PET + acetazolamide

Visit \~2 hours. Venous access. Acetazolamide 1 g in 50 mL saline infused over 5 min. After \~10 min, dynamic brain \[11C\]PiB PET (\~60 min), activity 100-500 MBq.

7.3 MRI (if needed)

If brain MRI is \<5 years old and clinically acceptable with no suspected major change, MRI is not repeated. Otherwise \~20 min MRI including sequences relevant for vascular markers and white matter lesions. No contrast.

7.4 Neuropsychological testing

Up to 60 min in quiet setting; attention, memory, learning, language, processing speed tests. 8. Safety, Risks, and Discomfort

Long dynamic PET sessions may be tiring; staff monitor participants throughout. Adverse events from FDG, PiB and MRI are rare and typically mild. Acetazolamide may cause headache, tingling, increased diuresis, or altered taste of carbonated drinks; serious events are very rare.

Radiation: FDG (\~4.8 mSv), PiB (\~3.4 mSv), CT (\~3 mSv), total \~11.2 mSv (approx. \<3 years background radiation in Denmark). Many patients would otherwise have routine FDG PET clinically. Only participants ≥60 years are included.

Incidental findings: Scans are not intended as diagnostic exams. If potentially clinically significant incidental findings are suspected, a radiologist will be consulted; participant and GP will be informed where relevant. Participants who do not wish to be informed of incidental findings with potential health implications will not be scanned. 9. Data Handling / GDPR

Data stored for 15 years; thereafter transferred to "Privacy" in anonymised form. Data stored coded in accordance with GDPR and Danish Data Protection rules and registered with the Danish Data Protection Agency. Paper documents destroyed after completion of processing (planned by Dec 2027).

Access to medical records occurs only after signed consent and is used for eligibility assessment, safety, and analysis variables (diagnoses, medication, relevant labs, prior imaging, cognitive tests). Monitoring authorities may access data under confidentiality obligations. 10. Statistical Plan 10.1 General

Descriptive statistics including missingness and reasons. Analyses are non-blinded. Deviations from analysis plan will be documented in publications.

10.2 Image analyses (key outcomes)

\[11C\]PiB early phase: first 10 min reconstructed as "vascular/perfusion" image; outcome = regional uptake relative to cerebellar grey matter; z-scores vs controls.

\[11C\]PiB late phase: 40-60 min reconstructed for amyloid SUVr; cortical volume with SUVr \> 1.5 recorded.

\[18F\]FDG heart IDIF: VOI in LV cavity; prebolus/bolus ratio applied; late point from last heart scan; plasma conversion via population-based ratio.

\[18F\]FDG brain: dynamic modelling (two-compartment); lumped constant 0.76; plasma glucose included. Outcomes: CMRglc and K1-based glucose transport estimate.

MRI: neuroradiologist rates Fazekas score, regional white matter lesions, and other vascular findings.

10.3 Sample size

Exploratory. n=20 per group) accounts for \~10% drop-out, expected \~15% differences in affected volumes, power 90%, alpha 5%, and \~14% variation. Participants unable to complete study visits may be replaced.

11\. End of Trial / Discontinuation

Ethics Committee informed within 90 days after completion. Participants may discontinue at any time; discontinuation does not affect clinical care.

12\. Administrative Procedures

Insurance: Covered under Danish patient insurance law. Finances/COI: Investigator-initiated; no investigators have financial stake. Operating expenses covered by involved departments.

13\. Publication

Data owned by Region Hovedstaden. Results will be published regardless of outcome in peer-reviewed journals, following Vancouver authorship guidelines.