Gadolinium effect on thalamus and whole brain tissue segmentation

Area of Science:

• Neuroimaging
• Medical Image Analysis

Background:

• Accurate segmentation of subcortical gray matter (GM) is crucial for neurological research.
• The influence of Gadolinium-based contrast agents (GBCAs) on automated segmentation algorithms is not well-established.

Purpose of the Study:

• To evaluate the effect of GBCAs on automated segmentation of the thalamus and whole brain tissue.
• To compare segmentation accuracy using different automated techniques with and without GBCAs.

Main Methods:

• MRI scans from 74 multiple sclerosis patients (3DT1-weighted, with and without GBCAs) were analyzed.
• Manual thalamic segmentation served as the gold standard.
• Automated segmentation was performed using volBrain and FSL-Anat; DICE scores compared manual and automatic results.

Main Results:

• Significant differences in DICE scores were observed between enhanced and unenhanced T1 images for both automated methods (p < 0.001).
• Automatic tissue segmentation achieved a mean DICE of 81.5%, with lateral ventricles at 74.2%.
• Automatic thalamic segmentation showed high similarity (FSL-Anat: 91.7%, volBrain: 90.7%).

Conclusions:

• GBCAs significantly affect subcortical GM segmentation.
• The observed changes are subtle and potentially acceptable for region-based analyses in perfusion or diffusion imaging.