[CODE SUBMISSION] Parametric maps and Spectroscopy Tool (PST)

MRS Software
1

Dear all,

I’d like to share our open-source Parametric Maps and Spectroscopy Tool (PST) for MATLAB, designed for Philips and Siemens 2D MRSI/MRS data (currently supporting .SDAT and .RDA formats, twix is coming soon).

The tool provides the following functionality:

  • Extraction of the MRSI slab from a reference anatomical image and from quantitative maps (e.g., GM, WM, CSF, qT1, qT2, H2O map, B1 field, QSM, etc.). A blurring model is applied to the qMRI slabs to account for the limited point spread function (PSF) of MRSI.
  • Selection of voxels of interest and generation of individual MRSI voxel masks. These masks are multiplied with the corresponding quantitative map slabs to obtain voxel-wise qMRI values.
  • Processing of the selected voxels in LCModel. A range of LCModel parameters can be modified directly from the GUI, and the resulting PostScript (PS) output can be visualized within the GUI.
  • Optional: Addressing the chemical shift displacement artifact (CSDA) for a user-defined ppm offset from the spectral center frequency. This feature has been validated for Philips data and is currently a work in progress for Siemens data. For 2D MRSI, the slab position is recalculated along the slice-selective direction. For the VOI in e.g., 2D PRESS, 2D sLASER and single-voxel MRS, corrections are applied in all three spatial directions. The shifted VOI/SV is displayed in white. The qMRI values are then calculated for each user-defined chemical shift, either for individual MRSI voxels or for the single voxel.

The current version has been tested on Windows with MATLAB R2024a and R2026a. Voxel composition analysis should also be supported on macOS and Unix/Linux systems, as it relies solely on MATLAB code, whereas LCModel functionality on these platforms is a work in progress.

The tool uses SPM for NIfTI file operations and incorporates numerous routines from FID-A, Gannet, and Osprey for MRS loading and geometry processing. Many thanks to the developers of these outstanding toolboxes!

The Tool is available on GitHub:

Thank you, and any feedback is welcome!

Best,
Andy

SUBMISSION.md (603 Bytes)

2 Likes