A fully automated method for quantifying and localizing white matter hyperintensities on MR images.

Area of Science:

• Neuroimaging
• Medical Image Analysis
• Neurology

Background:

• White matter hyperintensities (WMH) are common in the elderly and linked to neuropsychiatric disorders like dementia and depression.
• Previous WMH assessments relied on subjective global ratings, limiting precision.
• Late-onset late-life depression (LLD) is associated with a significant WMH burden.

Purpose of the Study:

• To develop and validate an automated method for quantifying and localizing WMHs using MRI.
• To compare the automated method's WMH quantification with manual ratings.
• To investigate WMH burden differences between LLD patients and elderly controls.

Main Methods:

• Adapted a fuzzy-connected algorithm for automated WMH segmentation.
• Employed demons-based image registration for automated anatomic localization using the Johns Hopkins University White Matter Atlas.
• Validated the method on MRI data from 11 LLD subjects and 8 elderly controls.

Main Results:

• The automated WMH quantification strongly correlated with manual ratings (P<0.0001).
• LLD subjects exhibited a significantly greater WMH burden than controls, confirmed by both methods.
• The automated method demonstrated a larger effect size, indicating higher specificity.

Conclusions:

• The automated method provides accurate and reliable quantification and localization of WMHs.
• This technique enhances the specificity of WMH measurement, particularly in neuropsychiatric conditions like LLD.
• Automated WMH analysis is crucial for understanding their role in neuropsychiatric disorders, especially with growing neuroimaging databases.