Deformable model reconstruction of the subarachnoid space

Area of Science:

• Neuroimaging
• Medical Image Analysis
• Neuroscience

Background:

• The subarachnoid space, a meningeal layer, contains cerebrospinal fluid and trabeculae.
• Quantifying subarachnoid space volume and thickness is crucial for understanding neurodegenerative disease pathogenesis.
• Accurate measurement is needed for comparisons between healthy individuals and patients.

Purpose of the Study:

• To present an automatic method for reconstructing the subarachnoid space with subvoxel accuracy.
• To evaluate the efficacy of this novel method in clinical studies.

Main Methods:

• Utilized a nested deformable model for subarachnoid space reconstruction.
• Initialized the model using the convex hull of the brain's outer surfaces (cerebrum, cerebellum, brainstem).
• Employed T1-weighted and T2-weighted MRI data with region forces to guide the model.

Main Results:

• The proposed automatic method achieves subvoxel accuracy in subarachnoid space reconstruction.
• Demonstrated comparable or improved accuracy when validated against semi-automatic and multi-atlas methods.
• A pilot study showed the method's utility in differentiating normal pressure hydrocephalus patients from healthy controls.

Conclusions:

• The developed automatic method provides accurate quantification of subarachnoid space volume and thickness.
• This technique holds promise for advancing research in neurodegenerative diseases and hydrocephalus.
• The method offers a robust tool for clinical neuroimaging analysis.