Spatiospectral image processing workflow considerations for advanced MR spectroscopy of the brain.

Area of Science:

• Neuroimaging
• Metabolic measurements
• Non-invasive imaging

Background:

• Magnetic resonance spectroscopy (MRS) traditionally had limited clinical utility, often reduced to single-voxel analysis for lactate detection in neonates.
• Advancements in scanner technology and MRS sequences have increased data complexity, introducing spatial and spectral dimensions.
• Challenges exist in accessing and manipulating MRS data across diverse scanners, formats, and software standards.

Purpose of the Study:

• To address the need for standardized image processing in advanced MRS.
• To describe a generalized framework for manipulating spatially- and spectrally-variant MRS data.
• To facilitate innovation in MRS data analysis for neuroimaging researchers.

Main Methods:

• Developed a framework for MRS data manipulation, generalizing across voxel, spectral, and metabolite levels.
• Ensured compatibility with established neuroimaging standards and multiple imaging sites.
• Integrated the framework with LCModel, a quantitative MRS peak-fitting platform.

Main Results:

• Demonstrated the advantages of the proposed workflow with examples from recent publications and new data.
• Provided a characterization of MRS image processing considerations.
• Showcased a method for robustly accessing and manipulating MRS data.

Conclusions:

• The proposed framework lowers the barrier to entry for MRS data processing.
• Standardized processing enhances the utility of complex MRS data for neuroimaging research.
• Facilitates wider adoption and innovation in the field of quantitative magnetic resonance spectroscopy.