Finding relocated private elements using "PrivateCreator"

Private elements can be relocated (group number will change) and some popular PACS will absolutely do this to data passing through them. Instead of looking up an absolute (group_number, element_number) in rules, we should allow looking up private elements with a PrivateCreator plus an element_number. Then when processing a dataset you look for a matching PrivateCreator string (e.g. “SIEMENS CSA HEADER”) to find the private group number dynamically.

I’m not sure how this capability should be exposed in the recipes, any thoughts?

[jstorrs]

Edit: nevermind I just realized after posting the URL that I was looking at the 2013 version (somehow comes to the top of Google for me). The latest version forbids it and says if there’s a need for some reason the implementation should use sequences.

https://dicom.nema.org/medical/dicom/current/output/chtml/part05/sect_7.8.html

[jstorrs]

Odd question and I don’t know if anyone here has an answer… is it valid for the same private creator string to be reused for multiple reservations in the same block? i.e. suppose you have:

(0021,0010) LO [MY PRIVATE CREATOR] (0021,0011) LO [MY PRIVATE CREATOR] (0021,0012) LO [MY PRIVATE CREATOR] (0021,0013) LO [MY PRIVATE CREATOR] (0021,1001) LO [VALUE 1] (0021,1101) LO [VALUE 2] (0021,1201) DS [3] (0021,1301) LO [VALUE 4]

I haven’t been able to find anything in the standard that forbids this. But then if you encounter a file like this, I’d flag it as obviously something’s amok. This is obviously going to be a challenge for this sort of thing. I’ll see whether pydicom or dcmtk has thought about this weird case.

I have not encountered this in the wild, I’m just pondering how to handle this. Previously the thought was that (0x0021,“MY PRIVATE CREATOR”,0x01) must obviously be unique and duplicate private creators are forbidden, but I can’t find where/if that’s specified in the standard.

https://dicom.nema.org/dicom/2013/output/chtml/part05/sect_7.8.html