Post-acquisition processing confounds in brain volumetric quantification of white matter hyperintensities

Area of Science:

• Neuroimaging
• Medical Image Analysis
• Gerontology

Background:

• Magnetic resonance imaging (MRI) studies on aging populations show significant variability in white matter hyperintensities (WMH) volumetric quantification across different research sites, even with identical acquisition techniques.
• The underlying causes of this variability in WMH measurement have not been thoroughly investigated.

Purpose of the Study:

• To identify and address sources of variability in the post-acquisition processing of 3D FLAIR MRI sequences.
• To develop a standardized protocol for WMH volumetric quantification with minimal inter-rater variance for improved data comparability.

Main Methods:

• Analysis of 3D FLAIR sequences from aged subjects to pinpoint post-acquisition processing variables affecting WMH quantification.
• Exploration of factors such as image center selection, software versions, thresholding, and manual editing.
• Development of a standardized protocol to minimize variability in WMH volume measurements.

Main Results:

• Post-acquisition processing can introduce substantial variance (up to 15%) in WMH volume quantification, independent of scan acquisition quality.
• Controlling for identified sources of variability, including image center, software, thresholding, and manual editing, resulted in a protocol with less than 0.5% inter-rater variance.
• The developed protocol significantly enhances the reliability of WMH volumetric data.

Conclusions:

• Standardizing post-acquisition processing is crucial for reducing variability in MRI-based WMH quantification.
• Implementing controlled protocols minimizes quantitative image processing variance, enabling reliable cross-site and cross-study comparisons in aging research.