Improved Prediction of Amyloid-β and Tau Burden Using Hippocampal Surface Multivariate Morphometry Statistics and...

Area of Science:

• Neuroimaging
• Biomarker Discovery
• Alzheimer's Disease Research

Background:

• Alzheimer's disease (AD) is characterized by amyloid-beta (Aβ) plaques and tau tangles, alongside neurodegeneration (ATN framework).
• Current detection methods for Aβ/tau pathology include invasive cerebrospinal fluid analysis, costly positron emission tomography (PET), and developing blood biomarkers.
• There is a need for non-invasive, widely accessible methods to quantify Aβ and tau pathology.

Purpose of the Study:

• To develop a non-invasive, widely available structural magnetic resonance imaging (MRI)-based framework.
• To quantitatively predict amyloid and tau measurements using MRI data.

Main Methods:

• Utilized MRI-based hippocampal multivariate morphometry statistics (MMS) features.
• Applied a Patch Analysis-based Surface Correntropy-induced Sparse coding and max-pooling (PASCS-MP) method with ridge regression.
• Evaluated the framework on amyloid PET/MRI and tau PET/MRI datasets from the Alzheimer's Disease Neuroimaging Initiative.

Main Results:

• The developed PASCS-MP framework demonstrated accurate prediction of amyloid and tau measurements.
• Predicted values were closer to actual measurements compared to traditional morphometry features like hippocampal surface area, volume, and shape.
• The framework effectively bridges hippocampal atrophy with amyloid and tau pathology.

Conclusions:

• The MMS-based PASCS-MP method is an efficient tool for assessing Alzheimer's disease burden and progression.
• This non-invasive MRI approach can aid in monitoring treatment effects.
• The framework offers a promising alternative for widespread clinical application in AD diagnostics.