Quantifying Brain Atrophy Using a CSF-Focused Segmentation Approach.

Area of Science:

• Neuroimaging
• Neurodegeneration
• Medical image analysis

Background:

• Brain atrophy, marked by sulcal widening and ventricular enlargement, is key in neurodegenerative diseases like Alzheimer's.
• Current visual and automated methods for atrophy assessment have limitations in subjectivity, variability, and standardization.
• Accurate quantification of brain atrophy is crucial for understanding disease progression and evaluating interventions.

Purpose of the Study:

• To develop and evaluate a novel method for brain atrophy assessment.
• To focus on cerebrospinal fluid (CSF) regions for improved segmentation accuracy.
• To assess stage-specific atrophy patterns and test generalizability across unstandardized datasets.

Main Methods:

• Utilized T1-weighted MRI data from 3,315 (Samsung Medical Center) and 1,439 (other hospitals) participants.
• Evaluated segmentation accuracy using Dice Similarity Coefficient (DSC).
• Calculated W-scores for regions of interest (ROIs) to identify stage-specific atrophy patterns.

Main Results:

• Achieved high segmentation accuracy (DSC > 0.9 for ventricles/hippocampus, > 0.8 for cortical regions).
• Observed significant differences in W-scores across cognitive stages (unimpaired, MCI, AD) for all ROIs (p<0.05).
• Demonstrated generalizability to unstandardized datasets from other hospitals.

Conclusions:

• The novel CSF-focused segmentation method is robust and clinically applicable for brain atrophy assessment.
• Provides a scalable, objective framework for evaluating structural changes across cognitive stages.
• Has potential for broader application in neurodegenerative disease research and clinical practice.