Diffusion sensitivity of 3D-GRASE in arterial spin labeling perfusion

Area of Science:

• Magnetic Resonance Imaging (MRI)
• Neuroimaging
• Quantitative Physiology

Background:

• Arterial Spin Labeling (ASL) is a non-invasive MRI technique for measuring cerebral blood flow (CBF).
• 3D Gradient and Spin Echo (GRASE) readouts are commonly used in ASL but can be affected by diffusion and flow artifacts.
• Understanding these artifacts is crucial for accurate CBF quantification, especially in small brain structures.

Purpose of the Study:

• To evaluate the impact of true diffusion and flow-related pseudodiffusion on CBF quantification using ASL with 3D-GRASE.
• To assess the role of pseudodiffusion as a source of bias in ASL-based CBF measurements.
• To investigate methods for mitigating CBF estimation bias in ASL.

Main Methods:

• Augmented Extended Phase Graph (aEPG) algorithm developed to model pseudodiffusion in 3D-GRASE acquisitions.
• Simulations and in vivo experiments in 13 healthy subjects were conducted.
• CBF estimation bias was assessed for various postlabeling delays, crusher gradients, and segmentation factors.

Main Results:

• In vivo experiments showed systematic underestimation of gray matter (GM) CBF using segmented 3D-GRASE.
• CBF underestimation reached up to 25% with low segmentation (1PAR × 2PE) but was lower with higher segmentation (4PAR × 2PE or 8PAR × 2PE).
• Pseudodiffusion accounted for the majority of the CBF estimation bias, exceeding the contribution of T1/T2 relaxation and true diffusion.

Conclusions:

• Flow-related pseudodiffusion introduces significant estimation bias in ASL-based CBF quantification with 3D-GRASE.
• Increasing segmentation factors in 3D-GRASE ASL acquisitions can substantially reduce this bias.
• Accurate CBF quantification with ASL-GRASE requires careful consideration and mitigation of pseudodiffusion effects.