The Effect of Segmentation Method on Medial Temporal Lobe Subregion Volumes in Aging

Area of Science:

• Neuroimaging
• Neurodegeneration
• Computational anatomy

Background:

• Early Alzheimer's disease (AD) is linked to medial temporal lobe (MTL) volume reductions.
• Manual segmentation protocols (e.g., Olsen-Amaral-Palombo) are time-consuming.
• Cognitive status (Montreal Cognitive Assessment scores) correlates with MTL atrophy in healthy older adults.

Purpose of the Study:

• To evaluate the utility of Automatic Segmentation of Hippocampal Subfields (ASHS) for detecting volumetric differences in MTL subregions.
• To compare an ASHS atlas trained on the OAP protocol (ASHS-OAP) with manual segmentation and another ASHS atlas (ASHS-PMC).

Main Methods:

• Trained ASHS on the OAP protocol to create the ASHS-OAP atlas.
• Compared volumetric measures from ASHS-OAP and ASHS-PMC with manual segmentation.
• Analyzed volumetric differences in MTL subregions in healthy older adults with varying cognitive status (MoCA scores).

Main Results:

• ASHS-OAP replicated manual segmentation findings for anterolateral entorhinal cortex and perirhinal cortex.
• Both ASHS-OAP and ASHS-PMC detected early neurodegeneration signs in CA2/CA3/DG.
• ASHS-OAP and ASHS-PMC yielded different volumes for most regions but identified similar trends in neurodegeneration.

Conclusions:

• Automated segmentation using ASHS-OAP is a viable alternative to manual methods for detecting cognitive status-related volumetric differences in MTL.
• Further development is needed for a unified and harmonized MTL segmentation atlas.
• Automated methods show potential for efficient early detection of neurodegenerative changes.