Classification of Alzheimer's disease based on hippocampal multivariate morphometry statistics

Area of Science:

• Neuroimaging
• Biomarker Discovery
• Neurology

Background:

• Alzheimer's disease (AD) and mild cognitive impairment (MCI) involve progressive cognitive decline.
• Hippocampal morphometry using magnetic resonance imaging (MRI) are key markers for AD and MCI.
• Multivariate morphometry statistics (MMS) quantifies surface deformations for robust hippocampus evaluation.

Purpose of the Study:

• To assess if hippocampal surface deformation features can classify AD, MCI, and healthy controls (HC).
• To investigate MMS as a potential imaging biomarker for early AD detection.

Main Methods:

• Explored differences in hippocampus surface deformation among AD, MCI, and HC groups using MMS analysis.
• Utilized hippocampal MMS features with support vector machine (SVM) for binary and triple classification.
• Analyzed selective patches for detailed feature extraction.

Main Results:

• Identified significant hippocampal deformation differences across the three groups, particularly in the hippocampal CA1 region.
• Achieved good performance in binary classifications (AD/HC, MCI/HC, AD/MCI).
• The triple-classification model yielded an area under the curve (AUC) of 0.85, demonstrating strong diagnostic potential.

Conclusions:

• Significant hippocampal deformation is evident in individuals with AD, MCI, and HC.
• Hippocampal MMS serves as a sensitive imaging biomarker for the early, individual-level diagnosis of AD.
• The findings support the use of MMS in clinical settings for neurodegenerative disease assessment.