Osprey problems with DICOM files from 3T Siemens Prisma

MRS Software Osprey
1

Dear everyone

We have acquired MRS data using the MEGA-PRESS sequence from CMRR on a 3T Siemens Prisma. We have 3 acquisitions:

  1. Water-supressed MEGA-PRESS (multiple files, e.g. 512 dicom)
  2. Water reference (same MEGA-PRESS sequence, but without water supression) (2 dicom)
  3. Standard PRESS sequence with short TR (and TE) but without water supression (2 dicom)

Osprey seems to assume the same number of files for each data type, although the MEGA-PRESS scan gives us multiple files. I haven’t been able to find any input data specification in the Osprey documentation for dicom.

Any help will be greatly appreciated !

With best wishes
Martin

2

Hi Martin,

Can you share the relevant contents of your job file (i.e. the lines specifying the files, files_ref, and files_w input variables)?

For DICOM, the documentation states “MRS raw data can be exported in DICOM and Siemens RDA format, where each average is saved as a separate .DCM or .RDA file. In this case, you have to enter the full path to the folder containing all single-average files belong to one acquisition.”

This means you must separate the water-suppressed MEGA DICOMs from the water reference MEGA DICOMs from the water reference short-TE PRESS DICOMs.

Cheers,
Georg

3

Dear Georg

Apologies for my tardy reply !

I have 3 acquisitions:

  1. MEGA-PRESS sequence for metabolite data
  2. MEGA-PRESS sequence without water supression (water reference)
  3. PRESS water reference scan with short TE of 30 ms

The metabolite sequence results in 2 dicom files (as well as a larger number of files that I suppose are the individuals samples used to create the 2 averaged metabolite files). The water reference scan results in 2 dicom files and the water reference scan with short TE results in 1 dicom file.

I get following error message from Osprey (supposedly due to different numbers of files):

‘files’ has 2 entries, but ‘files_ref’ has 2 entries, ‘files_w’ has 1 entries

All the best
Martin

4

Can you post the relevant sections of your job file? You need to point the files variables to the folders containing the DICOM data.

If you have the metabolite transients resolved as individual DICOMs, move them all to a folder separate from the averaged metabolite DICOMs and use that. You never want to use data that have been pre-averaged on the scanner - this deprives you of the option to do the vital shot-to-shot alignment.

5

Dear Georg

The files variable is a 256×1 cell array containing the full paths to the dicom files with metabolite data:

files =

256×1 cell array

{'/Volumes/projects/MINDLAB2026_MEG-RMS-SCZ/raw/0005/20260216_152606/MR/014.svs_mpress_metabolites_RES/files/PROJ0590_SUBJ0005_SER014_ACQ00001_IMG00001_263370219333.dcm'}
{'/Volumes/projects/MINDLAB2026_MEG-RMS-SCZ/raw/0005/20260216_152606/MR/014.svs_mpress_metabolites_RES/files/PROJ0590_SUBJ0005_SER014_ACQ00001_IMG00002_263469619334.dcm'}
{'/Volumes/projects/MINDLAB2026_MEG-RMS-SCZ/raw/0005/20260216_152606/MR/014.svs_mpress_metabolites_RES/files/PROJ0590_SUBJ0005_SER014_ACQ00001_IMG00003_263535419335.dcm'}

{ … }

files_ref =

2×1 cell array

{'/Volumes/projects/MINDLAB2026_MEG-RMS-SCZ/raw/0005/20260216_152606/MR/011.svs_mpress_No_Water_Sup/files/PROJ0590_SUBJ0005_SER011_ACQ00001_IMG00001_238902618793.dcm'}
{'/Volumes/projects/MINDLAB2026_MEG-RMS-SCZ/raw/0005/20260216_152606/MR/011.svs_mpress_No_Water_Sup/files/PROJ0590_SUBJ0005_SER011_ACQ00001_IMG00002_238998918794.dcm'}

files_w =

1×1 cell array

{'/Volumes/projects/MINDLAB2026_MEG-RMS-SCZ/raw/0005/20260216_152606/MR/012.svs_Water_short_TE/files/PROJ0590_SUBJ0005_SER012_ACQ00001_IMG00001_955241919069.dcm'}
6

Can you try pointing to the directories rather than the individual files?

The length of the files, files_ref, and files_w variables needs to be identical (1x1 for one single dataset).

Ex:

files     = {'/Volumes/projects/MINDLAB2026_MEG-RMS-SCZ/raw/0005/20260216_152606/MR/014.svs_mpress_metabolites_RES/files/'}
files_ref = {'/Volumes/projects/MINDLAB2026_MEG-RMS-SCZ/raw/0005/20260216_152606/MR/011.svs_mpress_No_Water_Sup/files/'}
files_w   = {'/Volumes/projects/MINDLAB2026_MEG-RMS-SCZ/raw/0005/20260216_152606/MR/012.svs_Water_short_TE/files/'}

edit: Do you have the _RES version of the water reference data as well? It looks like these are stored in an averaged fashion (only 2 files; unless you only have 2 water-unsuppressed transients specified in your protocol).

I would also highly recommend not using paths that contain . characters; these have tripped Osprey up in the past. My recommendation would be to use the BIDS formalism to store your raw data (your future self will thank you as you can easily repurpose pipelines you build).