Related Experiment Video
Updated: Dec 4, 2025
14:27
Identification of Disease-related Spatial Covariance Patterns using Neuroimaging Data
Published on: June 26, 2013
16.0K
Principal-Component Analysis-Based Measures of PET Data Closely Reflect Neuropathologic Staging Schemes
Ganna Blazhenets1, Lars Frings2, Arnd Sörensen2
1Department of Nuclear Medicine, University of Freiburg Medical Center, Freiburg, Germany ganna.blazhenets@uniklinik-freiburg.de.
Summary
Alzheimer's Disease Conversion-Related Patterns (ADCRPs) derived from PET scans accurately reflect neuropathologic findings. These patterns, particularly amyloid-β (Aβ)-ADCRPs, show strong correlation with Alzheimer's disease (AD) progression and can predict AD neuropathologic change.
Area of Science:
- Neuroimaging and Neuropathology
- Biomarker Validation in Alzheimer's Disease
Background:
- Voxel-based principal-component analysis identifies patterns in glucose metabolism and amyloid deposition linked to Alzheimer's disease (AD) conversion from mild cognitive impairment (MCI).
- Alzheimer's Disease Conversion-Related Patterns (ADCRPs) require validation against definitive neuropathologic findings.
Purpose of the Study:
- To validate ADCRPs, specifically 18F-FDG PET-derived neurodegeneration patterns and amyloid-β (Aβ) PET-derived deposition patterns, against post-mortem neuropathologic assessments.
- To assess the diagnostic performance of ADCRPs and conventional PET measures in identifying intermediate-to-high Alzheimer's Disease Neuropathologic Change (ADNC).
Main Methods:
- Analysis of patients from the Alzheimer's Disease Neuroimaging Initiative with available autopsy, 18F-FDG PET, and Aβ PET data.
- Comparison of pattern expression scores (PESs) of 18F-FDG-ADCRP and Aβ-ADCRP with Braak tangle stage and Thal amyloid phase, respectively.
- Evaluation of volume-of-interest (VOI)-based PET measures and diagnostic performance using receiver-operating-characteristic (ROC) curves.
Main Results:
- Significant correlations were found between 18F-FDG-ADCRP PES and Braak stage (ρ > 0.48), and between Aβ-ADCRP PES and Thal phase (ρ > 0.66).
- The combination of 18F-FDG-ADCRP and Aβ-ADCRP PES achieved an AUC of 0.98 for identifying intermediate-to-high ADNC.
- VOI-based PET measures also showed significant associations with Braak stage and Thal phase, with combined measures yielding an AUC of 0.90.
Conclusions:
- 18F-FDG-ADCRP PES is a valid measure of neurodegeneration that closely corresponds with tau pathology (Braak stage).
- Aβ-ADCRP PES serves as a reliable biomarker for underlying amyloid pathology, correlating strongly with Thal phase.
- The combination of ADCRPs demonstrates superior performance in predicting ADNC compared to individual PET measures.
