search

QIICR / dcmqi

dcmqi (DICOM for Quantitative Imaging) is a free, open source C++ library for conversion between imaging research formats and the standard DICOM representation for image analysis results
https://qiicr.gitbook.io/dcmqi-guide/
BSD 3-Clause "New" or "Revised" License
228 stars 62 forks source link

Consider initializing General Reference module #382

Open fedorov opened 5 years ago

fedorov commented 5 years ago

"General Reference" module is "U", but has the advantage of being able to describe purpose of reference, unlike "Referenced Series Sequence" (which is "C") that is initialized right now.

Purpose of reference might come very handy for the situations where multiple series are used in derivation of a parametric map. In the case of DCE analysis, this can be the case quite often: AIF mask might be defined in a separate SEG series, and the AIF image itself might be separate from the image acquisition covering the area of interest.

fedorov commented 5 years ago

Pointed out by Ralf Floca and Ina Kompan at DKFZ

rfloca commented 5 years ago

Thanks for opening that issue.

I think in this context it is also valuable to discuss the usage of "0008,2112 Source Image Sequence" and "0008,1140 Referenced Image Sequence".

In context of DCE analysis, I would use the first to reference the image that was reale used for the fit (so could be the derived concentration image when using e.g. eTofts or 2CXM) and the second sequence would be always ther original signal image that was acquired.

I think it is important to be able to document these differences and the elements seem well suited to model this, If we use a proper purpose encoding.

I also second you thoughts about multiple series in context of AIF. :)