Reproducibility of Longitudinal Changes in Cortical Thickness Determined by Surface-Based Morphometry Between...

Area of Science:

• Neuroimaging
• Radiology
• Biomedical Engineering

Background:

• Accelerated MRI techniques reduce scan times but may compromise signal-to-noise ratio and image quality.
• The impact of accelerated imaging on the reproducibility of longitudinal brain structure analysis, particularly cortical thickness, remains unclear.

Purpose of the Study:

• To assess the reproducibility of longitudinal cortical thickness changes using surface-based morphometry.
• To compare non-accelerated and accelerated T1-weighted MRI in healthy elderly, mild cognitive impairment (MCI), and Alzheimer's disease (AD) populations.

Main Methods:

• Retrospective study analyzing 2-year interval MRI data from 50 healthy elderly, 54 MCI, and 8 AD patients.
• Cortical thickness was estimated using FreeSurfer's longitudinal stream.
• Image quality was visually assessed by three radiologists.

Main Results:

• Reproducibility, measured by intraclass correlation coefficient (ICC), varied across groups and brain regions (overall mean ICC=0.45).
• Mean ICC was 0.33 for healthy elderly, 0.49 for MCI, and 0.31 for AD patients.
• Within-subject variability in cortical thickness changes was similar in healthy elderly (1.3%/year) and MCI (1.3%/year) but higher in AD patients (1.7%/year). Image quality was comparable across groups.

Conclusions:

• Reproducibility of longitudinal cortical thickness measurements differs between non-accelerated and accelerated MRI.
• The variability in reproducibility is influenced by disease status (MCI, AD) and specific brain regions.
• Accelerated MRI shows potential for longitudinal studies but requires careful consideration of disease-specific effects.