Verification of SR Documents

All other questions regarding DCMTK

Moderator: Moderator Team

Post Reply
Message
Author
Martin Henriksen
Posts: 3
Joined: Tue, 2008-02-19, 12:59

Verification of SR Documents

#1 Post by Martin Henriksen » Mon, 2008-03-10, 12:55

The company, I work for, is signed up as "Report Creator" for this years Connectahon.

But, using DCMTK we run into a problem concerning MESA test 605.
In this test, we have to create a Structured Report, which is PARTIAL (ie. not completed) and VERIFIED.

The functions in DCMTK (ver. 3.5.4) does not allow for verifying PARTIAL documents. (And it is not possible to "un-complete" a completed document)
(As far as I understand the DICOM standard, this is actually allowed...although it may not seem very logical :) )

Is this a known problem with DCMTK or are we doing something wrong ?

brgds
Martin

Marco Eichelberg
OFFIS DICOM Team
OFFIS DICOM Team
Posts: 1293
Joined: Tue, 2004-11-02, 17:22
Location: Oldenburg, Germany
Contact:

#2 Post by Marco Eichelberg » Mon, 2008-03-10, 13:09

Unfortunately the DICOM standard does not describe the exact meaning and relationship of the "completion" and "verification" flags. The DCMSR module indeed implements a policy where a report must be completed in order to be verified. A completed document cannot be "un-completed", you can only create a revised version using DSRDocument::createRevisedVersion(), which is essentially the same except that the document receives new UIDs and inserts a reference to the old document into the predecessor document list. This is also a deliberate design choice. The whole point of calling a document "complete" is that the only thing that should happen to the document after completion is verification. Modifications of a completed document are a revision, and this is what the revision mechanism in DICOM SR and DCMTK is good for. From this perspective I would say that MESA test 605 if flawed if it requires the creation of a partial but verified document. I suggest you contact Steve Moore (or whoever developed the SINR test cases) concerning this issue.

akaraivanov
Posts: 6
Joined: Fri, 2004-12-03, 11:34
Location: Denmark

#3 Post by akaraivanov » Mon, 2008-03-10, 13:31

Marco Eichelberg wrote:Unfortunately the DICOM standard does not describe the exact meaning and relationship of the "completion" and "verification" flags. The DCMSR module indeed implements a policy where a report must be completed in order to be verified.
...
The way I read the standard, completion-flag and verification-flag are semantically quite ortogonal. Looking into PS 3.3 C.17.2 (SR Document General Module) I fail to find the requirement you mention about. Would it be possible to give more precise reference to the dicom standard?

Thank you very much.
Alexander Karaivanov

Jörg Riesmeier
ICSMED DICOM Services
ICSMED DICOM Services
Posts: 2217
Joined: Fri, 2004-10-29, 21:38
Location: Oldenburg, Germany

#4 Post by Jörg Riesmeier » Tue, 2008-03-11, 12:06

I agree with Marco that this was a deliberate design decision. Of course, you could manipulate the DICOM dataset directly after calling DSRDocument::write() or even introduce a new API method for this purpose. However, I don't think that this would be a good idea ...

akaraivanov
Posts: 6
Joined: Fri, 2004-12-03, 11:34
Location: Denmark

#5 Post by akaraivanov » Tue, 2008-03-11, 12:19

Jörg Riesmeier wrote:I agree with Marco that this was a deliberate design decision. Of course, you could manipulate the DICOM dataset directly after calling DSRDocument::write() or even introduce a new API method for this purpose. However, I don't think that this would be a good idea ...
Thank you for you answer, Jörg.
For me it is still not clear if PARTIAL & VERIFIED is legal combination according to DICOM & IHE. I put that question on comp.protocols.dicom hoping for someone to throw some more light on that. Yes, if it is, me and my colleague Martin we'll have to workaround it as you say and possibly come back with feature request for dcmtk.

Just thinking for a second about it - would not it make sense one to use PARTIAL & VERIFIED for objects that contain preliminary results?

Alexander.

Jörg Riesmeier
ICSMED DICOM Services
ICSMED DICOM Services
Posts: 2217
Joined: Fri, 2004-10-29, 21:38
Location: Oldenburg, Germany

#6 Post by Jörg Riesmeier » Tue, 2008-03-11, 13:00

Reading the relevant sections in David Clunie's SR book in addition to the standard text, I would agree that the DICOM standard is pretty silent on the use of the Completion and Verification Flag.
Just thinking for a second about it - would not it make sense one to use PARTIAL & VERIFIED for objects that contain preliminary results?
This is discussed in David's book and he describes a couple of real world examples. However, he does not explicitly say that a PARTIAL report can be VERIFIED nor does he say that this is not possible -- this is at least what I understand from the text :-)
I put that question on comp.protocols.dicom hoping for someone to throw some more light on that.
I am also curious what the "official" reading is ...

Jörg Riesmeier
ICSMED DICOM Services
ICSMED DICOM Services
Posts: 2217
Joined: Fri, 2004-10-29, 21:38
Location: Oldenburg, Germany

#7 Post by Jörg Riesmeier » Tue, 2008-03-11, 13:28

Here is an excerpt from the book:
a “complete” but unverified document may be an “interim,” “preliminary” or “draft” of some kind, distinguishable perhaps from a “partial” report which is just a piece of a document, not an entire document

akaraivanov
Posts: 6
Joined: Fri, 2004-12-03, 11:34
Location: Denmark

#8 Post by akaraivanov » Tue, 2008-03-11, 13:47

Jörg Riesmeier wrote:Here is an excerpt from the book:
a “complete” but unverified document may be an “interim,” “preliminary” or “draft” of some kind, distinguishable perhaps from a “partial” report which is just a piece of a document, not an entire document
And here is another quote from the same book :D
Finally, the use of the word “state” is probably inappropriate in this context, though it has been used here for clarity. The verification and completion flags are really only descriptors. In particular, there is no express or implied “state machine” in the standard for these attributes, and no transitions between states defined.

Jörg Riesmeier
ICSMED DICOM Services
ICSMED DICOM Services
Posts: 2217
Joined: Fri, 2004-10-29, 21:38
Location: Oldenburg, Germany

#9 Post by Jörg Riesmeier » Mon, 2008-09-29, 15:20

As a follow up: It seems that this open issue will be solved with a new CP (proposed by David Clunie). Excerpt from this draft CP: "A value of 'VERIFIED' shall only be used when the value of Completion Flag (0040,A491) is 'COMPLETE'."

Here's the relevant discussion in comp.protocols.dicom.

akaraivanov
Posts: 6
Joined: Fri, 2004-12-03, 11:34
Location: Denmark

#10 Post by akaraivanov » Mon, 2008-09-29, 15:27

Jörg Riesmeier wrote:As a follow up: It seems that this open issue will be solved with a new CP (proposed by David Clunie). Excerpt from this draft CP: "A value of 'VERIFIED' shall only be used when the value of Completion Flag (0040,A491) is 'COMPLETE'."

Here's the relevant discussion in comp.protocols.dicom.
Great! Thank you for the hint Jörg!

Post Reply

Who is online

Users browsing this forum: Bing [Bot] and 1 guest