Partial volume correction of multiple inversion time arterial spin labeling MRI data.

Area of Science:

• Neuroimaging
• Medical Physics
• Radiology

Background:

• Cerebral blood flow (CBF) estimation using arterial spin labeling (ASL) is often inaccurate due to partial volume (PV) effects, where voxels contain both gray matter (GM) and white matter (WM).
• Previous PV correction methods for ASL have utilized local linear regression to separate GM and WM signals.

Purpose of the Study:

• To propose and evaluate a new partial volume (PV) correction method for multi-inversion time arterial spin labeling (ASL).
• To improve the accuracy and spatial resolution of cerebral blood flow (CBF) estimates in gray matter (GM).

Main Methods:

• A novel PV correction method was developed for multi-inversion time ASL, integrating PV estimates with a spatially regularized kinetic curve model analysis.
• The method leverages differences in ASL signal kinetics between GM and WM, alongside PV estimates, for signal separation.
• The approach was validated using both simulated and real neuroimaging data.

Main Results:

• The proposed PV correction method demonstrated comparable GM CBF correction accuracy to existing linear regression techniques.
• The new method successfully preserved greater spatial detail in the resulting CBF images.
• On real data, corrected GM CBF values showed independence from GM PV, indicating successful correction.
• A significant increase in mean GM CBF, ranging from 69-80%, was observed after applying the correction.

Conclusions:

• The novel PV correction method effectively improves the accuracy of GM CBF estimates from multi-inversion time ASL.
• This technique offers enhanced spatial resolution compared to previous methods.
• The findings suggest that this method can lead to more reliable and detailed assessments of cerebral blood flow in neuroimaging studies.