Hello everybody,

we are working on a novel PET device, and want to allow the users to select the radionuclide, and radiopharmaceutical from lists when they scan patients.

We were thinking about using CID 4021, and CID 4020 for that purpose, but want to abstract the different coding schemes from the users.

To do so, the lists would have to be free of duplicates across different coding schemes, which on first glace they generally are (with a single possible exception of MSH C000591008). I understand that this is not guaranteed, but maybe it is done by the committee nevertheless?

It would also be of interest to only allow the users selections of combinations that make sense. I can se how the names of the radiopharmaceuticals have the radionuclide in them, but I'm not sure it allows a unique mapping to the list in CID 4020 since a coding scheme is involved again.

Does anybody have some input or experience in this regard? Any answer, however short, will be greatly appreciated.

Best regards and have a nice evening,
Alon Dolev

All of the entries in CID 4021 PET Radiopharmaceutical already include the nuclide, as you point out, so the user does not need to make a separate selection.

One can pre-configure an appropriate CID 4020 entry to include in Radionuclide Code Sequence (0054,0300) whenever a CID 4021 entry for inclusion in Radiopharmaceutical Code Sequence (0054,0304) is selected. This mapping should indeed be unique (i.e., there is only one (77004003, SCT, "^18^Fluorine") from CID 4020 to use when the radiopharmaceutical from CID 4021 is (35321007, SCT, "Fluorodeoxyglucose F^18^"). If you find any inconsistencies, let me know.

That a coding scheme is involved is irrelevant - DICOM regards codes (and the coding schemes they are drawn from) a representation of concepts, and the selection of a particular radiopharmaceutical "concept", with which the corresponding nuclide "concept" is "pre-coordinated" (as the terminology folks say) is definitive. Do not be concerned about mixing and matching coding schemes. I.e., one can use the SCT code for the "^18^Fluorine" concept when the radiopharmaceutical code is from a different scheme, such as DCM (e.g.., for (126503, DCM, "Flubatine F^18^")). This may not be ideal practice from a terminology/ontology purist's perspective but is the price we pay for not having a single comprehensive grand unified coding scheme of everything..

For SNOMED CT codes one might be able to automate this to some extent. E.g., if you follow (35321007, SCT, "Fluorodeoxyglucose F^18^") to its definition (e.g., using the SCT browser, "http://snomed.info/id/35321007"), you will see it is "defined as" being a "modification of" (77004003, SCT, "^18^Fluorine"). Though "modification of" is not a very specific relationship, at least one can extract the mapping mechanically from the SNOMED-supplied release files (https://confluence.ihtsdotools.org/display/DOCRELFMT/SNOMED+CT+Release+File+Specifications).

PS. I will fix the C000591008 flortaucipir/Tauvid duplicate, which is a mistake, obviously. Thanks for pointing that out.

Hi David,

thank you very much for your competent answer and explanations. I have discussed it internally, and we are going to follow your advice and implement the list as you suggested. So far, we have not found any inconsistencies.

I was wondering: the radio-pharmaceutical (0054,0304) is type 3, but the radionuclide (0054,0300) is type 2 in the PET IOD. Is there a use case when one would set the radionuclide, but not the radio-pharmaceutical? The only thing that came to my mind would be a solid phantom?

Thanks again, and have a nice day

Alon Dolev