handling dynamic PET data? - DICOM

This is a discussion on handling dynamic PET data? - DICOM ; I'm preparing software to read dynamic pet data from dicom files created by our Siemens biograph16 PET/CT system. I've found some things I don't understand and haven't been able to decide if the problem is with my code, or with ...

+ Reply to Thread
Results 1 to 4 of 4

Thread: handling dynamic PET data?

  1. handling dynamic PET data?

    I'm preparing software to read dynamic pet data from dicom files
    created by our Siemens biograph16 PET/CT system. I've found some
    things I don't understand and haven't been able to decide if the
    problem is with my code, or with the DICOM data, or with my mind. Can
    anyone suggest an existing piece of software that is able to handle and
    manipulate dynamic and gated pet series data?

    Mike


  2. Re: handling dynamic PET data?

    Mike wrote:
    > I'm preparing software to read dynamic pet data from dicom files
    > created by our Siemens biograph16 PET/CT system. I've found some
    > things I don't understand and haven't been able to decide if the
    > problem is with my code, or with the DICOM data, or with my mind. Can
    > anyone suggest an existing piece of software that is able to handle and
    > manipulate dynamic and gated pet series data?
    >
    > Mike


    Mike -

    I cannot recommend some existing software, but I also have been trying
    to work with DICOM dynamic PET data. I have examples of dynamic PET
    data from Philips and GE scanners. (I thought that the Siemens
    biograph wasn't capable of creating dynamic images?). What problems
    are you having? I have questions about various time fields in the file
    and how they are used, specifically Series Time, Acquisition Time and
    Frame Reference Time. Do you know anything about that?

    Thanks, Ginny


  3. Re: handling dynamic PET data?

    >>>>> "vrogers" == vrogers writes:

    > I cannot recommend some existing software, but I also have
    > been trying to work with DICOM dynamic PET data. I have
    > examples of dynamic PET data from Philips and GE scanners.
    > (I thought that the Siemens biograph wasn't capable of
    > creating dynamic images?).


    Hello Ginny,

    The biograph dynamic software is available. We've been testing
    it in beta form for a while now and it has recently been moved to
    it's first released version.

    > What problems are you having? I have questions about
    > various time fields in the file and how they are used,
    > specifically Series Time, Acquisition Time and Frame
    > Reference Time. Do you know anything about that?


    My original problem was confusion, which seems to have been
    resolved. I've talked with engineers at Siemens and with various
    others about the dynamic pet series, so I know how it is
    implemented by Siemens/CTI. First of all, there is a known
    inconsistency in the standard. Section C.8.9.1.1.5 Decay
    Correction states both that all images in a series are decay
    corrected to the same time and that, if Decay Correction
    (0054,1102) is START, then each image is decayed to the
    Acquisition Time (0008,0032). But then the figure in
    C.8.9.1.1.11 Temporal Relationships of Images in a PET Series,
    shows that each frame has it's own acquisition time. I'm told
    this will be fixed in the next version of the DICOM standard.

    Inconsistancies in the standard aside, in the real world, the
    biograph decay corrects all frames to the Series Time
    (0008,0031), which is the time that the database entry is created
    for the series. In all the tests that I've looked at, the Series
    Time, the manual start, the start of first time slice and
    Acquisition Time of the first frame are coincident. This decay
    correction is just like what is done on the HR+ and as documented
    in the HR+ manuals.

    Note that I was cautioned that it is possible to build an
    acquisition protocol that includes a delay between pressing the
    manual start and the start of the first time slice though. Watch
    for that if you are syncronizing with additional real world
    events, like injection time or blood sampling.

    I would be happy to trade sample data with you. I have a couple
    of series that are images of a cylindrical phantom containing a
    known quantity of 11C. This series made it easy for me to check
    my code - I'll get you a copy if you want it. If you have
    similar samples from other scanners that you are willing to
    share, I'd like to get a copy. In fact, I'll set up a web page
    on our server so we can have a central place to put information
    about dynamic PET series.

    Regards, Mike

    --
    Michael A. Miller mmiller3@iupui.edu
    Imaging Sciences, Department of Radiology, IU School of Medicine

  4. Re: handling dynamic PET data?

    Thanks Mike -
    The inconsistency in the standard regarding the acquisition time, which
    can (should) be different for each frame, and the statement that all
    frames are decay corrected to the same time -- acquistion time -- is
    the very thing I have been struggling with. We have done plenty of
    studies at our site where the acquisition start time for frame 1 plus
    the frame durations do *not* completely describe the time sequence of
    the study (ie, there are gaps between one or more frames) and the fact
    that the acquisition time was the same for each frame (which is what I
    have observed in the headers) did not make sense to me.

    I would like to get a copy of your biograph dynamic phantom images, and
    would be happy to share dynamic phantom scans (FDG) from GE and Philips
    scanners.

    Ginny


+ Reply to Thread