Code system related peculiarities in codes.xml
21 views
Skip to first unread message
Andreas Schnabl
Apr 10, 2024, 8:02:51 AM4/10/24
to IHE MHD Implementors
Hi,
while testing my MHD server against the FHIR Toolkit, I have noticed two code system related peculiarities in the codes.xml file delivered with it (and XDS Toolkit):
1.) In the following line
<Code code="1.2.840.10008.5.1.4.1.1.88.59" display="Key Object Selection Document" codingScheme="1.2.840.10008.2.6.1" system="http://dicom.nema.org/resources/ontology/DCM"/>
the code system URL "http://dicom.nema.org/resources/ontology/DCM" is given, which, according to [1], identifies the DCM coding scheme. However, in the same line, the coding scheme OID "1.2.840.10008.2.6.1" is given, which, also according to [1], identifies the DCMUID coding scheme. Looking through the actual tables for DCM [2] and DCMUID [3], I can find the definition of the given code "1.2.840.10008.5.1.4.1.1.88.59" at [3] (DCMUID), but not at [2] (DCM).
So am I correct in concluding that this code lives just in the DCMUID scheme, and thus in disregarding the given URI (http://dicom.nema.org/resources/ontology/DCM)?
Side question: if I am correct, then what would be the correct URI representing the DCMUID coding scheme? Just "urn:oid:1.2.840.10008.2.6.1"? Or does there exist a canonical URL for it? At least at [4], I couldn't find any entry about DCMUID, just about DCM (DICOM Controlled Terminology).
2.) The following two lines
<Code code="urn:ihe:iti:xdw:2011:eventCode:open" display="Open Workflow" codingScheme="1.3.6.1.4.1.19376.1.2.3" system="urn:oid:1.3.6.1.4.1.19376.1.2.3"/>
<Code code="urn:ihe:iti:xdw:2011:eventCode:closed" display="Closed Workflow" codingScheme="1.3.6.1.4.1.19376.1.2.3" system="urn:oid:1.3.6.1.4.1.19376.1.2.3"/>
<Code code="urn:ihe:iti:xdw:2011:eventCode:closed" display="Closed Workflow" codingScheme="1.3.6.1.4.1.19376.1.2.3" system="urn:oid:1.3.6.1.4.1.19376.1.2.3"/>
use the OID "1.3.6.1.4.1.19376.1.2.3" for their code systems together with the URI "urn:oid:1.3.6.1.4.1.19376.1.2.3". However, later in the codes.xml file, I can find the same OID paired with the URI "http://ihe.net/fhir/ihe.formatcode.fhir/CodeSystem/formatcode" for many formatCode definitions.
When I look through the code system definition at [5], I can find the various format codes in the list of values, but not the two XDW event codes listed above. However, I can find references to these event codes (using the same codingScheme OID 1.3.6.1.4.1.19376.1.2.3) in the definition of XDW constraints on XDS metadata at [6].
This leaves me wondering how to interpret these code systems. Should I consider the OID 1.3.6.1.4.1.19376.1.2.3 to be an identifier for one single code system, or are there actually differences between the meanings of the OID's uses in [5] and [6]?
David Clunie
Apr 10, 2024, 9:55:42 AM4/10/24
to ihe-mhd-im...@googlegroups.com, Elliot Silver, Harry Solomon, Boone, Keith W (GE Healthcare), Brad Genereaux, John Moehrke, Jonathan Whitby
Hi Andreas
If the code is intended to be a concept that is a DICOM KOS object, you want to use
the DICOM SOP Class UID as the code from the DCMUID coding scheme that is defined
by the list of SOP Class UIDs in DICOM PS3.5 Annex A, then the attribute for
"system" is definitely incorrect, and as you imply, should be something like
<Code code="1.2.840.10008.5.1.4.1.1.88.59" display="Key Object Selection Document" codingScheme="1.2.840.10008.2.6.1" system="http://dicom.nema.org/resources/ontology/DCMUID"/>
since use of
system="http://dicom.nema.org/resources/ontology/DCM"
is definitely wrong, because that refers only to the DCM codes in PS3.16 Annex D.
However, AFAIK, nobody has formally defined the value for "system" for DCMUID, e.g.,
something like:
system="http://dicom.nema.org/resources/ontology/DCMUID"
even though the OID is registered [1], etc.
Since FHIR doers not list DCMUID (as distinct from DCM) as a system in their list [2]
I am reluctant to add it to DICOM PS3.16 Table 8-1 as a "FHIR:" (Coding.system)"
reference until/if they do.
However, in the interim, I have added a folder on the NEMA server so that at least
the proposed/implied DCMUID system resource URL resolves, and contains a redirect
to [3].
There was a long discussion 5 or 6 years ago among FHIR imaging folks about whether
or not DICOM UIDs were different than any other kind of OID, and whether or not they
should be from system "urn:ietf:rfc:3986" and prefixed with "urn:oid:...", but that was
more in relation to transient OIDs (such as those identifying study, series, instance)
rather than well-known OIDs defined by PS.6 that constitute a particular "value set"
of OIDs. I think it was resolved in favor of the RFC 3986 approach, and per [4] there
is an argument this should be:
<Code code="urn:oid:1.2.840.10008.5.1.4.1.1.88.59" display="Key Object Selection Document" codingScheme="2.16.840.1.113883.4.873" system="urn:ietf:rfc:3986"/>
or to be DICOM-specific:
<Code code="urn:oid:1.2.840.10008.5.1.4.1.1.88.59" display="Key Object Selection Document" codingScheme="1.2.840.10008.2.6.1" system="urn:dicom:uid"/>
though the latter goes beyond [4] since that list does not contain an OID for
system="urn:dicom:uid" and as I said, DCMUID was never intended for transient UIDs.
I.e., codingScheme="1.2.840.10008.2.6.1" and system="urn:dicom:uid" are not really
the same thing.
So I am not sure that either of those suggestions for <Code/> is really correct.
The history of DCMUID is that it was added in DICOM Supplement 101 to handle encoding
of SOP Class (not instance) UIDs in CDA (HL7 V3 "Act") references, as a code system [5].
This usage persists in the current PS3.20 (CDA report), at least in the examples if
not the normative text [6].
There has also been a discussion in FHIR about a ValueSet for DICOM storage SOP Class
UIDs (a subset of all well-known UIDs that corresponds to [7]), which may or may not
be relevant/helpful for this discussion.
Since they may not be on the ihe-mhd-implementors list, I have cc'd a few folks involved
in the earlier FHIR imaging discussion, who may (or may not) still be interested in
this (and be aware that replies only to the list may not go to them).
Probably a subject for HL7 IISIG DICOM WG 20 to consider (again).
David
1. http://oid-info.com/get/1.2.840.10008.2.6.1
2. http://terminology.hl7.org/external_terminologies_csmr.html
3. http://dicom.nema.org/medical/dicom/current/output/chtml/part06/chapter_A.html
4. https://www.hl7.org/fhir/R4/identifier-registry.html
5. ftp://medical.nema.org/medical/dicom/final/sup101_ft.pdf
6. https://dicom.nema.org/medical/dicom/current/output/chtml/part20/sect_C.5.2.html
7. https://dicom.nema.org/medical/dicom/current/output/chtml/part04/sect_B.5.html#table_B.5-1
On 4/10/24 8:02 AM, Andreas Schnabl <a.sc...@synedra.com> wrote:
> Hi,
>
> while testing my MHD server against the FHIR Toolkit, I have noticed two code system related peculiarities in the codes.xml file delivered with it (and XDS Toolkit):
>
> 1.) In the following line
>
> <Code code="1.2.840.10008.5.1.4.1.1.88.59" display="Key Object Selection Document" codingScheme="1.2.840.10008.2.6.1" system="http://dicom.nema.org/resources/ontology/DCM"/>
>
> the code system URL "http://dicom.nema.org/resources/ontology/DCM" is given, which, according to [1], identifies the DCM coding scheme. However, in the same line, the coding scheme OID "1.2.840.10008.2.6.1" is given, which, also according to [1], identifies the DCMUID coding scheme. Looking through the actual tables for DCM [2] and DCMUID [3], I can find the definition of the given code "1.2.840.10008.5.1.4.1.1.88.59" at [3] (DCMUID), but not at [2] (DCM).
>
> So am I correct in concluding that this code lives just in the DCMUID scheme, and thus in disregarding the given URI (http://dicom.nema.org/resources/ontology/DCM)?
>
> Side question: if I am correct, then what would be the correct URI representing the DCMUID coding scheme? Just "urn:oid:1.2.840.10008.2.6.1"? Or does there exist a canonical URL for it? At least at [4], I couldn't find any entry about DCMUID, just about DCM (DICOM Controlled Terminology).
>
...
If the code is intended to be a concept that is a DICOM KOS object, you want to use
the DICOM SOP Class UID as the code from the DCMUID coding scheme that is defined
by the list of SOP Class UIDs in DICOM PS3.5 Annex A, then the attribute for
"system" is definitely incorrect, and as you imply, should be something like
<Code code="1.2.840.10008.5.1.4.1.1.88.59" display="Key Object Selection Document" codingScheme="1.2.840.10008.2.6.1" system="http://dicom.nema.org/resources/ontology/DCMUID"/>
since use of
system="http://dicom.nema.org/resources/ontology/DCM"
is definitely wrong, because that refers only to the DCM codes in PS3.16 Annex D.
However, AFAIK, nobody has formally defined the value for "system" for DCMUID, e.g.,
something like:
system="http://dicom.nema.org/resources/ontology/DCMUID"
even though the OID is registered [1], etc.
Since FHIR doers not list DCMUID (as distinct from DCM) as a system in their list [2]
I am reluctant to add it to DICOM PS3.16 Table 8-1 as a "FHIR:" (Coding.system)"
reference until/if they do.
However, in the interim, I have added a folder on the NEMA server so that at least
the proposed/implied DCMUID system resource URL resolves, and contains a redirect
to [3].
There was a long discussion 5 or 6 years ago among FHIR imaging folks about whether
or not DICOM UIDs were different than any other kind of OID, and whether or not they
should be from system "urn:ietf:rfc:3986" and prefixed with "urn:oid:...", but that was
more in relation to transient OIDs (such as those identifying study, series, instance)
rather than well-known OIDs defined by PS.6 that constitute a particular "value set"
of OIDs. I think it was resolved in favor of the RFC 3986 approach, and per [4] there
is an argument this should be:
<Code code="urn:oid:1.2.840.10008.5.1.4.1.1.88.59" display="Key Object Selection Document" codingScheme="2.16.840.1.113883.4.873" system="urn:ietf:rfc:3986"/>
or to be DICOM-specific:
<Code code="urn:oid:1.2.840.10008.5.1.4.1.1.88.59" display="Key Object Selection Document" codingScheme="1.2.840.10008.2.6.1" system="urn:dicom:uid"/>
though the latter goes beyond [4] since that list does not contain an OID for
system="urn:dicom:uid" and as I said, DCMUID was never intended for transient UIDs.
I.e., codingScheme="1.2.840.10008.2.6.1" and system="urn:dicom:uid" are not really
the same thing.
So I am not sure that either of those suggestions for <Code/> is really correct.
The history of DCMUID is that it was added in DICOM Supplement 101 to handle encoding
of SOP Class (not instance) UIDs in CDA (HL7 V3 "Act") references, as a code system [5].
This usage persists in the current PS3.20 (CDA report), at least in the examples if
not the normative text [6].
There has also been a discussion in FHIR about a ValueSet for DICOM storage SOP Class
UIDs (a subset of all well-known UIDs that corresponds to [7]), which may or may not
be relevant/helpful for this discussion.
Since they may not be on the ihe-mhd-implementors list, I have cc'd a few folks involved
in the earlier FHIR imaging discussion, who may (or may not) still be interested in
this (and be aware that replies only to the list may not go to them).
Probably a subject for HL7 IISIG DICOM WG 20 to consider (again).
David
1. http://oid-info.com/get/1.2.840.10008.2.6.1
2. http://terminology.hl7.org/external_terminologies_csmr.html
3. http://dicom.nema.org/medical/dicom/current/output/chtml/part06/chapter_A.html
4. https://www.hl7.org/fhir/R4/identifier-registry.html
5. ftp://medical.nema.org/medical/dicom/final/sup101_ft.pdf
6. https://dicom.nema.org/medical/dicom/current/output/chtml/part20/sect_C.5.2.html
7. https://dicom.nema.org/medical/dicom/current/output/chtml/part04/sect_B.5.html#table_B.5-1
On 4/10/24 8:02 AM, Andreas Schnabl <a.sc...@synedra.com> wrote:
> Hi,
>
> while testing my MHD server against the FHIR Toolkit, I have noticed two code system related peculiarities in the codes.xml file delivered with it (and XDS Toolkit):
>
> 1.) In the following line
>
> <Code code="1.2.840.10008.5.1.4.1.1.88.59" display="Key Object Selection Document" codingScheme="1.2.840.10008.2.6.1" system="http://dicom.nema.org/resources/ontology/DCM"/>
>
> the code system URL "http://dicom.nema.org/resources/ontology/DCM" is given, which, according to [1], identifies the DCM coding scheme. However, in the same line, the coding scheme OID "1.2.840.10008.2.6.1" is given, which, also according to [1], identifies the DCMUID coding scheme. Looking through the actual tables for DCM [2] and DCMUID [3], I can find the definition of the given code "1.2.840.10008.5.1.4.1.1.88.59" at [3] (DCMUID), but not at [2] (DCM).
>
> So am I correct in concluding that this code lives just in the DCMUID scheme, and thus in disregarding the given URI (http://dicom.nema.org/resources/ontology/DCM)?
>
> Side question: if I am correct, then what would be the correct URI representing the DCMUID coding scheme? Just "urn:oid:1.2.840.10008.2.6.1"? Or does there exist a canonical URL for it? At least at [4], I couldn't find any entry about DCMUID, just about DCM (DICOM Controlled Terminology).
>
>
> [1] https://dicom.nema.org/medical/dicom/current/output/html/part16.html#chapter_8
> [2] https://dicom.nema.org/medical/dicom/current/output/chtml/part16/chapter_D.html
> [3] https://dicom.nema.org/medical/dicom/current/output/chtml/part06/chapter_A.html
> [4] https://terminology.hl7.org/external_terminologies_csmr.html
...
> [1] https://dicom.nema.org/medical/dicom/current/output/html/part16.html#chapter_8
> [2] https://dicom.nema.org/medical/dicom/current/output/chtml/part16/chapter_D.html
> [3] https://dicom.nema.org/medical/dicom/current/output/chtml/part06/chapter_A.html
> [4] https://terminology.hl7.org/external_terminologies_csmr.html
ell...@argentixinfo.com
Apr 14, 2024, 2:54:41 PM4/14/24
to David Clunie, ihe-mhd-im...@googlegroups.com, Harry Solomon, Boone, Keith W (GE Healthcare), Brad Genereaux, John Moehrke, Jonathan Whitby, a.sc...@synedra.com
Hi Andreas,
David summarizes the situation well, and the lack of a defined value set for SOP instance UIDs has proven a challenge in a couple of situations.
I think the only thing to add is the FHIR “clarified” their use of RFC 3986 about a year ago so that urns tied to a “schema” don’t need to use the 3986 system or prefix. (DICOM UIDs, GUIDs used as primary keys in a particular database, etc.) This means DICOM UIDs can use a “SOP class system” or “SOP instance system” or “study instance system” as appropriate which helps when you have multiple OID, GUID, and URL identifiers all claiming the 3986 system. In reality this “clarification” helps one problem but creates multiple new ones.
Elliot.
—
Elliot Silver, M.Sc., AStd.
President & Principal Consultant
Vancouver, BC
Confidentiality Notice: This e-mail message, including any attachments, is for the sole use of the intended recipient(s) and may contain confidential and privileged information. Any unauthorized review, use, disclosure or distribution is prohibited. If you are not the intended recipient, please contact the sender by reply e-mail and destroy all copies of the original message.
--
You received this message because you are subscribed to the Google Groups "IHE MHD Implementors" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ihe-mhd-implemen...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ihe-mhd-implementors/0100018ec84af563-03b281e9-2f5b-4834-8fe7-55dea74dace5-000000%40email.amazonses.com.
Reply all
Reply to author
Forward
0 new messages