Missing Magnitude Files

walkerped · May 6, 2021, 2:18pm

I have a dataset that included a scan using a siemens (skyra scanner) fm2d2r sequence for computing a field map. When I run dcm2niix on the dicoms, I get a phase difference file with the suffix e2_ph.nii containing a single volume, but no magnitude files are output. The data originally was converted using an older version of dcm2niix (v1.0.20180518), but I get the same result running it through a new version (v1.0.20210317). How can I get the magnitude images for this scan?

Here is the portion of the dicom header with scan acquisition information for the scan in question:

0018 0015 6 [1410 ] // ACQ Body Part Examined//BRAIN
0018 0020 2 [1424 ] // ACQ Scanning Sequence//GR
0018 0021 2 [1434 ] // ACQ Sequence Variant//SP
0018 0022 0 [1444 ] // ACQ Scan Options//
0018 0023 2 [1452 ] // ACQ MR Acquisition Type //2D
0018 0024 8 [1462 ] // ACQ Sequence Name//*fm2d2r
0018 0025 2 [1478 ] // ACQ Angio Flag//N
0018 0050 2 [1488 ] // ACQ Slice Thickness//3
0018 0080 4 [1498 ] // ACQ Repetition Time//500
0018 0081 4 [1510 ] // ACQ Echo Time//6.86
0018 0083 2 [1522 ] // ACQ Number of Averages//1
0018 0084 10 [1532 ] // ACQ Imaging Frequency//123.25903
0018 0085 2 [1550 ] // ACQ Imaged Nucleus//1H
0018 0086 2 [1560 ] // ACQ Echo Number//2
0018 0087 2 [1570 ] // ACQ Magnetic Field Strength//3
0018 0088 2 [1580 ] // ACQ Spacing Between Slices//3
0018 0089 4 [1590 ] //ACQ Number of Phase Encoding Steps//144
0018 0091 2 [1602 ] // ACQ Echo Train Length//0
0018 0093 4 [1612 ] // ACQ Percent Sampling//100
0018 0094 4 [1624 ] //ACQ Percent Phase Field of View//100
0018 0095 4 [1636 ] // ACQ Pixel Bandwidth//400
0018 1000 6 [1648 ] // ACQ Device Serial Number//45407
0018 1020 14 [1662 ] // ACQ Software Version//syngo MR D13C
0018 1030 18 [1684 ] // ACQ Protocol Name//fMRI_field_mapping
0018 1200 18 [1710 ] // ACQ Date of Last Calibration//20090304\20090304
0018 1201 28 [1736 ] // ACQ Time of Last Calibration//123723.000000\123723.000000
0018 1251 4 [1772 ] // ACQ Transmitting Coil//Body
0018 1310 8 [1784 ] // ACQ Acquisition Matrix// 144 0 0 144
0018 1312 4 [1800 ] // ACQ Phase Encoding Direction//COL
0018 1314 2 [1812 ] // ACQ Flip Angle//35
0018 1315 2 [1822 ] // ACQ Variable Flip Angle//N
0018 1316 16 [1832 ] // ACQ SAR//0.05794959537002
0018 1318 2 [1856 ] // ACQ DB/DT//0
0018 5100 4 [1866 ] // ACQ Patient Position//HFS

Chris_Rorden · May 7, 2021, 11:34am

When setting up a gradient echo field map on a Siemens scanner, make sure you select to store both magnitude and phase images.
For users of Siemens VA…VE systems (like your VD13C), I would also strongly recommend saving field map data as mosaics: Siemens reuses instance numbers (0020,0013) across echoes, any many PACS systems assume these are identical images so they overwrite images. Depending on your PACS and storage system, it is also possible that phase maps overwrote your magnitude images: to check this export the images directly off your console to a USB drive. I want to be clear that users of Siemens XA systems should NEVER save images in mosaic format.
Scientists working with Siemens scanners have access to Siemens Research Collaboration Managers. They have encyclopedic knowledge of these systems and are an invaluable resource for resolving issues and tuning performance.