Pitfalls of Voxel-Based Amyloid PET Analyses for Diagnosis of Alzheimer’s Disease: Artifacts due to Non-Specific Uptake in the White Matter and the Skull
Bibliographic record
Abstract
Two methods are commonly used in brain image voxel-based analyses widely used for dementia work-ups: 3-dimensional stereotactic surface projections (3D-SSP) and statistical parametric mapping (SPM). The methods calculate the Z-scores of the cortical voxels that represent the significance of differences compared to a database of brain images with normal findings, and visualize them as surface brain maps. The methods are considered useful in amyloid positron emission tomography (PET) analyses to detect small amounts of amyloid-β deposits in early-stage Alzheimer's disease (AD), but are not fully validated. We analyzed the (11)C-labeled 2-(2-[2-dimethylaminothiazol-5-yl]ethenyl)-6-(2-[fluoro]ethoxy)benzoxazole (BF-227) amyloid PET imaging of 56 subjects (20 individuals with mild cognitive impairment [MCI], 19 AD patients, and 17 non-demented [ND] volunteers) with 3D-SSP and the easy Z-score imaging system (eZIS) that is an SPM-based method. To clarify these methods' limitations, we visually compared Z-score maps output from the two methods and investigated the causes of discrepancies between them. Discrepancies were found in 27 subjects (9 MCI, 13 AD, and 5 ND). Relatively high white matter uptake was considered to cause higher Z-scores on 3D-SSP in 4 subjects (1 MCI and 3 ND). Meanwhile, in 17 subjects (6 MCI, 9 AD, and 2 ND), Z-score overestimation on eZIS corresponded with high skull uptake and disappeared after removing the skull uptake ("scalping"). Our results suggest that non-specific uptakes in the white matter and skull account for errors in voxel-based amyloid PET analyses. Thus, diagnoses based on 3D-SSP data require checking white matter uptake, and "scalping" is recommended before eZIS analysis.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".