[MINC-users] dcm2mnc for Siemens Biograph mCT 64 PET/CT scanner

Thu Nov 12 03:39:23 EST 2015

Hi Bert,

1. I've uploaded two DICOM tar.gz files to BIC. PM sent.

2. As far as I checked, time-width gets lost completely after running
mincconcat for an "irregular-regular" pair. Some important tools we use
regularly (e.g. TAC extraction) don't work, due to lack of time-width.

Thank you again for your help. I appreciate it very much.

Best regards,
Yoshi

On Mon, Nov 9, 2015 at 11:59 PM, Robert D. Vincent <
robert.d.vincent at mcgill.ca> wrote:

> Hi Yoshi,
>
> 1. We read the frame times from specific DICOM fields, so it is likely that
> the DICOM itself is recording the mid point. Again, if you can provide a
> scan, we can change the behavior of the program to record the initial frame
> time.
>
> 2. This could be easily done, but it should not make any actual difference.
> Does mincconcat work properly with the regular/irregular dimensions?
>
> 3. You are getting "Argument list too long" when running dcm2mnc? This
> should be easily avoidable. You can generally just specify a directory name
> for a set of DICOM files:
>
> dcm2mnc <input-directory> <output_directory>
>
> Alternatively, you can read a list of files from standard input:
>
> find <input-directory> -name "*.dcm" | dcm2mnc -stdin <output_directory>
>
>     -bert
>
>
> On Mon, Nov 9, 2015 at 9:31 AM, Yoshitaka KUMAKURA <
> yoshi.kumakura at gmail.com
> > wrote:
>
> > Thank you very much, Andrew and Vladimir.
> >
> > Here is my progress.
> >
> > 1)    I've installed dcm2mnc version 2.01.01 (built  2015), and found it
> > working much better. Now, MINC files converted from DICOM of Biograph
> > PET/CT can be viewed on MNI-register, and they look good, though I
> haven't
> > checked voxel values in detail yet.
> >
> > In terms of recording the time dimension, I see dcm2mnc behave
> differently
> > from ecattominc. Frame MID times are extracted by dcm2mnc. But, this is a
> > bit confusing for old Siemens PET users, because frame START times were
> > extracted by (the good old days) ecattominc. We already have a lot of
> tools
> > customized for that.
> >
> > <mincheader after dcm2mnc>
> > time = 15, 45, 75, 105, 135, 165, 210, 270, 330, 390, 450...
> > time-width = 30, 30, 30, 30, 30, 30, 60, 60, 60, 60, 60, 180, 180, 180,
> > 180...
> >
> > So, I would be even more grateful if it could be like:
> > time = 0, 30, 60, 90....
> >
> > Is there any good idea to solve or avoid the issue of ambiguity in
> > recording "time" between dcm2mnc and ecattominc? I am also wondering
> where
> > this has come from.
> >
> > 2)    Our PET experiment is consisted of two PET/CT scans on the same
> day.
> > They are separated by a 30-min scanning break, and need to be
> concatenated.
> > The first scan's frame width is incremental, like typical dynamic PET
> > scans. The second scan's frame width is constant, 360sec (x 10 frames).
> > Some header lines that seem to be related to this issue are shown below.
> >
> > SCAN1(incremental frame time-width)
> > double time(time) ;
> > time:start = 15. ;
> > time:spacing = "irregular" ;
> > time:alignment = "start_" ;
> >
> >  time = 15, 45, 75, 105, 135, 165, 210, 270, 330, 390, 450, 570,
> >     750, 930, 1110, 1350, 1650, 1950, 2250, 2550,.....(truncated)
> >
> >  time-width = 30, 30, 30, 30, 30, 30, 60, 60, 60, 60, 60, 180, 180, 180,
> > 180,
> >     300, 300, 300, 300, 300, 300, 300, 300 ;
> >
> >
> > SCAN2 (360 sec x 10 fr)
> > double time(time) ;
> > time:start = 179. ;
> > time:spacing = "regular__" ;
> > time:alignment = "start_" ;
> >
> >
> >  time = 179, 539, 899, 1259, 1619, 1979, 2339, 2699,.....(truncated)
> >
> > Here, time-with is unavailable in the 2nd scan's header. It would be very
> > nice if MINC experts could implement an option that forces dcm2mnc to
> work
> > in the "irregular" mode, so that the constant time-width would be
> > explicitly recovered in the 2nd scan's MINC header. In this way, MINC
> uses
> > like myself can feel more comfortable, when concatenating two dissimilar
> > files (after time adjustment, decay correction, head motion correction
> etc,
> > of course).
> >
> > 3)    Sometimes, stuck in "Argument list too long". Please help me learn
> > how to use find and xargs commands properly...
> >
> > Best regards,
> > Yoshi
> >
> > On Fri, Nov 6, 2015 at 1:07 AM, Vladimir S. FONOV <
> > vladimir.fonov at gmail.com>
> > wrote:
> >
> > > Hello Yoshi,
> > >
> > > I had a similar situation some time ago, I found that I had to rely on
> > > dcmdump from dcmtk first to split all the DICOM files into individual
> 3D
> > > volumes, then run dcm2mnc on each set independently, and finally
> > > concatenating them with mincconcat along time dimension.
> > >
> > >
> > > On 15-11-05 12:50 AM, Yoshitaka KUMAKURA wrote:
> > >
> > >> Dear MINC experts,
> > >>
> > >> We would like to use MINC tools to analyze dynamic PET files of a
> > Siemens
> > >> Biograph mCT 64 PET/CT scanner in an affiliated hospital in
> Copenhagen.
> > We
> > >> are not new to MINC, because we are former BIC researchers (currently
> in
> > >> Albert Gjedde's lab). But, we are a small PET research group with no
> > >> computer scientist.
> > >>
> > >> We have a favor to ask. We've tried dcm2mnc to convert a set of
> dynamic
> > 4D
> > >> PET Dicom files to a minc files, and found that it just makes a 3D
> minc
> > >> file, missing the 4th time dimension. Has anybody encountered the same
> > >> issue? Is a dcm2mnc utility for the scanner available? We are in great
> > >> need
> > >> of the minc converter customized for the PET/CT scanner. We would
> > >> appreciate it very much, if you could find us a solution.
> > >>
> > >
> > >
> > > --
> > > Best regards,
> > >
> > >  Vladimir S. FONOV ~ vladimir.fonov <at> gmail.com
> > > _______________________________________________
> > > MINC-users at bic.mni.mcgill.ca
> > > http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
> > >
> > _______________________________________________
> > MINC-users at bic.mni.mcgill.ca
> > http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
> >
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>