[Intl_DxMedPhys] GE private DICOM tags for MRI

Steven Shea steven.michael.shea at gmail.com
Thu Aug 14 18:31:43 EDT 2025 

Wanted to add a quick note/helpful info that was passed to me at some point
in the past. DICOM conformance statements from GE (as well as other
vendors) will provide a lot of help when trying to find private tag
meanings or other system parameters. Here are links to GE & Siemens, but
all the vendors have these:

https://urldefense.com/v3/__https://www.gehealthcare.com/products/interoperability/dicom/magnetic-resonance-imaging-dicom-conformance-statements__;!!KGKeukY!yOAt5xr_a2PbQOzzLMbFEjf7N1DsPQkAMobeNZhegxI5-XawjmguzDCeqVyYavtIgjEslkg_yWfr2viPqdpqGhegDPeUi4nm3Ft9EfH9GdewC8fu$ 

https://urldefense.com/v3/__https://www.siemens-healthineers.com/services/it-standards/dicom-conformance-statements-magnetic-resonance__;!!KGKeukY!yOAt5xr_a2PbQOzzLMbFEjf7N1DsPQkAMobeNZhegxI5-XawjmguzDCeqVyYavtIgjEslkg_yWfr2viPqdpqGhegDPeUi4nm3Ft9EfH9GUaEOEuM$ 


You often have to wade through a lot of extraneous information, but it can
be very helpful for finding specific technical info (eg parallel imaging
On/Off, acceleration factor). Below are a mulitple GE DICOM tags that I
find very useful:

0018,0020: ScanningSequence

   -

   Ex: [EP\SE]



0018,0022: ScanOptions

   -

   Ex: [EPI_GEMS\FILTERED_GEMS\ACC_GEMS\PFF]
   -

   ACC_GEMS= Parallel image acceleration
   -

   EDR_GEMS= Extended Dynamic Range



0018,9087: DiffusionBValue (b value for particular image)



0019,105a: scan duration in us



0019,1093: Center Frequency (but not Hz, appears to be 1E8 = 10 MHz)



0019,1094: Transmit Gain

0019,1095: R1

0019,1096: R2



0019,109c: pulse sequence type



0043,102d: post-processing filter settings

   -

   P=PURE
   -

   S=SCIC
   -

      reFINE (PURE) and SCIC are features that improve image uniformity by
      applying a pixel intensity correction. SCIC applies a unique
correction for
      every image, which means that the same correctio43n will not be applied
      even to images acquired at the same slice location. This may lead to
      incorrect results when READY View or any post processing
software including
      subtraction is used. As a result, it is strongly recommended to disable
      SCIC for dynamic studies as well as any acquisition that generates more
      than one image per slice location, for example, DWI or DTI. PURE can be
      used as long as it uses the same PURE calibration series for all
the series
      or images involved with the processing.
      -

      The reFINE (PURE) feature changes the pixel intensity in the image to
      provide a more uniform image. This may lead to altered results when
      CADstream is used in conjunction with PURE for processing MR images. A
      modification to the CADstream difference threshold may be required. It is
      recommended to either disable PURE for dynamic studies or
applying the PURE
      feature uniformly across all the dynamic series.
      -

   s = SCENIC (?? Need to double-check)
   -

   A - Little sharpening, some smoothing: produces some smoothing while
   sharpening a little to suppress noise. This filter provides some edge
   enhancement of structural edges. Candidate images include 512×512 head,
   abdomen, and extremity images.
   -

   B - High sharpening, some smoothing: a high sharpening filter with
   little smoothing. This filter enhances structural edges. Used for high
   signal and low noise 256×256 images. Candidate images include axial heads,
   cervical spines, lumbar spines (small FOV), abdomens, and extremities.
   -

   C - Little sharpening, high smoothing: produces smoothing and very
   little sharpening. This filter suppresses noise and provides smoother
   structural edges. Candidate images include high-resolution and low-signal
   images.
   -

   D - High sharpening, high smoothing: suppresses noise and provides
   smoother regions and sharper structural edges. Used for high-resolution and
   low-signal images.
   -

   E - MR angio filter: produces smoothing of vascular structures without
   introducing black lines, and provides good vessel definition. Used for
   source MIP1s from MRA2 studies.
   -

   F - Some sharpening, noise reduction : produces reduced image noise
   while preserving edges. It also provides edge enhancement. Candidate images
   include musculoskeletal and spine images.
   -

   G - Noise reduction: produces reduces image noise. Typically used with
   neuro and spine images.
   -

   U: provides 5 levels (0-4) of noise reduction and sharpening options.
   These can be individually adjusted and saved to suit your requirements.

   Eg: P = PURE, D=smoothing, sharpening



0043,1039: b value for DWI sequences

   -

   Eg: [1000000800\8\0\0] --> b value of 800


0043,0x1081: GE Coil Name

   -

   Ex: [C-GE_HEAD]



0043,1083: Parallel Imaging Factors (Acceleration)

   -

   Eg: [0.5\1] --> 2x acceleration in phase-encode

0043,1084: Additional Parallel Imaging Data

   -

   Eg: [10000\11\-1\0\\YES\ARC]
   -

   Eg: [10000\0\-1\0\\NO\ ] --> no parallel imaging

0043,10B6: Multiband Parameters

0043,10CA: Recon DL Strength

   -

   Numerical representation like 0.30(/0.50/0.75)\ Literal representation
   like LOW(/MEDIUM/HIGH)0.30\LOW, 0.50\Medium , or 0.75\High


Best regards, Steve

Steven Shea, PhD
Associate Professor, MR Scientist
Department of Radiology & Medical Imaging. Loyola Medicine


On Wed, Aug 6, 2025 at 11:01 AM <
intl_dxmedphys_wd_osu_list-request at lists.osu.edu> wrote:

> From: Bob Kobistek <bob at rjkmedphys.com>
> To: Bob Kobistek <bob at rjkmedphys.com>, International Diagnostic Medical
> Physics List <intl_dxmedphys_wd_osu_list at lists.osu.edu>
> Cc:
> Bcc:
> Date: Tue, 5 Aug 2025 21:57:53 +0000
> Subject: Re: [Intl_DxMedPhys] GE private DICOM tags for MRI
> Ignore this. 0018,0022. I must be getting old. It was right in front of
> me. I was looking for a private tag, but this is a public tag. Time for
> some vodka then a nap. Robert J. Kobistek, MS, FACR, DABR, MRSE(MRSC™)
> Medical Physicist RJK
>
> Ignore this. 0018,0022. I must be getting old. It was right in front of
> me. I was looking for a private tag, but this is a public tag.
>
>
>
> Time for some vodka then a nap.
>
>
>
>
>
> Robert J. Kobistek, MS, FACR, DABR, MRSE(MRSC™)
>
> Medical Physicist
>
> RJK Medical Physics, Inc.
>
> 440-463-7879
>
>
>
> *From:* Intl_dxmedphys_wd_osu_list
> <intl_dxmedphys_wd_osu_list-bounces+bob=rjkmedphys.com at lists.osu.edu> *On
> Behalf Of *Bob Kobistek via Intl_dxmedphys_wd_osu_list
> *Sent:* Tuesday, August 5, 2025 5:36 PM
> *To:* International Diagnostic Medical Physics List <
> intl_dxmedphys_wd_osu_list at lists.osu.edu>
> *Subject:* [Intl_DxMedPhys] GE private DICOM tags for MRI
>
>
>
> Hi folks I’m trying to fill out an ACR Phantom Data form for a GE MRI
> scanner. I neglected to record the scan options for the clinical scans.
> Does anybody know how to pull them from the private DICOM tags? Nothing
> jumps out at me, but
>
> Hi folks
>
>
>
> I’m trying to fill out an ACR Phantom Data form for a GE MRI scanner. I
> neglected to record the scan options for the clinical scans. Does anybody
> know how to pull them from the private DICOM tags? Nothing jumps out at me,
> but I know they have to be encoded in there somewhere because they display
> on the images when viewed on the AWS.
>
>
>
> Thanks.
>
>
>
>
>
> Robert J. Kobistek, MS, FACR, DABR, MRSE(MRSC™)
>
> Medical Physicist
>
> RJK Medical Physics, Inc.
>
> 440-463-7879
>
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osu.edu/pipermail/intl_dxmedphys_wd_osu_list/attachments/20250814/2a9813fc/attachment.html>

More information about the Intl_dxmedphys_wd_osu_list mailing list