Reduced cross-scanner variability using vendor-agnostic sequences for single-shell diffusion MRI.

Area of Science:

• Medical Imaging
• Neuroscience
• Biophysics

Background:

• Diffusion MRI (dMRI) is crucial for mapping white matter structure.
• Inter-scanner variability complicates multi-center studies and data reproducibility.
• Developing vendor-neutral sequences is essential for harmonizing dMRI data.

Purpose of the Study:

• To develop a vendor-neutral dMRI pulse sequence using the open-source Pulseq platform.
• To reduce inter-scanner variability of dMRI measures between different MRI vendors.
• To enhance the reproducibility of neuroimaging studies.

Main Methods:

• Implemented a standard EPI-based dMRI sequence in Pulseq.
• Tested sequences on Siemens Prisma and GE Premier scanners.
• Evaluated variability using phantom and human subject data, analyzing fractional anisotropy (FA) and mean diffusivity (MD).

Main Results:

• Pulseq sequence reduced phantom variability by over 2.5x across vendors.
• In-vivo results showed markedly reduced variability with Pulseq.
• Pulseq maintained scan-rescan repeatability and reduced FA/MD variability by over 50% in cortical/subcortical regions.

Conclusions:

• The Pulseq diffusion sequence effectively reduces cross-scanner variability for phantom and in-vivo data.
• This harmonization benefits multi-center neuroimaging studies.
• Improved reproducibility of neuroimaging research is a key outcome.