Howto select the right IOD for writing dicom files

[vda...@gmail.com]

Hi,

I'm involved in a project where XRay images are acquired around an object and then a 3D CT image is produced.
This kind of system is sometimes refers as C-arm based cone-beam CT.

I've been asked to save all the images i.e. all the projection images + the 3D image in DICOM.

Say I have 200 CArm images + 1 3D CT image, my first idea was to save 200 DICOM file of IOD = DX (fluoroscopy)
and one image of IOD = CT or enhanced CT (I'm not clear with the better choice yet).

But I'l not sure that
- IOD = DX is the good choice, it could also be XRF IOD.
- it is really a good idea to save 200 DX (or XRF) images because these images do not have any "diagnosis" nor therapeutic meaning.
I would prefer to store them in a multiframe file but this file must provide me a way to store some parameters (CArm angles, etc... 4 parameters in all)

So what is the best way, wrt to Dicom std, to save both the 'raw' images and the 3D CT image ?
Which IOD should I prefer ?

Thanks for any help/advice...

Vincent

[vda...@gmail.com]

Hi guys,

I continue my search for the best way to store the information and I come to the following conclusion:

1) storing 2D image: DX IOD is not the good one, It should be either XA or XRF, probably XA is a better choice.

2) storing the 2D images used for the 3D reconstruction: I can used XA or XRF to store the images independently (all the image in one series) but if I want to store all the image in one file, RawData IOD is a good candidate, maybe the better one..

3) 3D image: X-Ray 3D Angiographic Image IOD Description seems to me to be a good candidate.

I would really appreciate any advices to choose the right storage IOD from dicom 'gurus' :)

Thanks

V

[Bas Revet]

Op vrijdag 12 april 2013 16:57:09 UTC+2 schreef vda...@gmail.com het volgende:

Hi Vincent,

You are on the right track. The 3D X-Ray storage SOP Class was created to store 3D reconstructions. If that SOP Class can not handled further in the workflow chain the best alternate is 'mis-using' the CT SOP Class as a series of slices to represents the volume.

It common practice to do so. Somebody is working to get cone-beam CT included in CT but I do not know what the status on top of my head.

Regards,
Bas (bas....@philips.com).


Bas Revet. Interoperability Architect @ Philips Healthcare - Best, Netherlands

[David Clunie]

I would suggest that you also store the projection images, using
the Enhanced XA IOD preferably; you may not think they are useful
now, but people will find a use for them, and it is good to have
the option to store them in a standard format, even if most users
configure it off.

In Mammo tomo, we have found that users do sometimes want to see
the projections, especially to detect motion as a cause for poor
reconstructed image quality, and also CAD in mammo is done on
the projections. Independent vendors may well also want to provide
3rd party reconstruction algorithms.

David

[vincent...@minmaxmedical.com]

Hi David,

thanks for your advice.
I planned to store the projection images using RawData IOD (was in balance with Enhanced XA IOD).
I will have a closer look to Enhanced XA IOD.

V

> > Bas.