Enhanced DICOM Objects - Detailed Proposal
1. Proposed Workitem: Enhanced DICOM Image Profile
- Proposal Editor: Dick Donker
- Editor: Dick Donker
- Date: 29 september 2008
- Domain: Radiology
The original DICOM cross-sectional and multi-frame projectional objects needed a significant revision because of technology advancements, and the introduction of new acquisition protocols for which there was no standard description.
DICOM has created a new "enhanced family" of multi-frame image objects for MR, CT, PET, and most recently XA, XRF and 3D X-Ray (cone beam CT and tomosynthesis), and there is a new 3D Ultrasound object in development. In addition, new types of data such as raw data and MR spectroscopy can now be stored in a standard manner. Products have now entered the market that are capable of creating some of these new objects. There is a need to promote their adoption for storage in and retrieval from Image Managers and Archives as well as their display by Image Displays.
It is desirable to profile the multitude of possible clinical acquisition protocols supported by these objects in order to constrain alternative means of encoding the same information and to make feasible and interoperable the consistent display on workstations.
Additional profiles for improved query requirements for these objects, as well as per-frame retrieval of image contents may also be warranted.
3. Key Use Case
Profiles for a few clinical use cases have already be defined by WG16 and the committee for the advancement of DICOM, and have been demonstrated at the SCAR and RSNA conference by a number of vendors. DICOM WG16 and the NEMA (MITA) Advanced CT/MR Taskforce has requested that these profiles be elaborated and promoted to IHE profiles:
3.1 CT/MR Cardiac Imaging
New CT/MR cardiac images containing standard attributes for location and timing information.
3.2 MR Diffusion
MR diffusion images containing standard attributes for diffusion B values, direction and diffusion image type.
3.3 CT/MR Contrast Perfusion
Intra-venous contrast enhanced images containing standard attributes for timing and phase.
3.4 Multi-stack Spine Imaging
Images are acquired at multiple disk space levels of the spine such that multiple slices at each level are parallel to the disk space, each in a separate stack from slices at different angles through other disks. These images are exchanged in an interoperable way with standard attributes describing the stacks and their geometries.
3.5 MR Spectroscopy
Single-voxel, multi-voxel or multi-slice 1D MR spectra are exchanged with reference and metabolite images that contain information to allow correct display, post-processing and analysis.
Consider adding PET-CT to this growing list (Comment added by J. Wallis)
4. Standards & Systems
For the use-cases proposed, the Enhanced CT, MR and MR Spectroscopy SOP Classes provide the necessary functionality to be profiled.
For additional query and retrieval, DICOM Sup 119, Instance and Frame Level Retrieve SOP Classes, (which has now reached Letter Ballot status) may be useful.
The use cases may constitute the basis for a new IHE enhanced DICOM objects profile, which specifies for each all that are applicable of the following:
- The required attributes (mostly already defined as mandatory in the standard),
- The relevant functional grouping
- The content of the dimension module
- The use of a Real World Values LUT
- The use of a Supplemental Color LUT
- The relevant attributes for display on a workstation
6. Technical Approach
No new actors are needed. A new set of requirements will be added to the Acquisition Modality, Evidence Creator, Image Manager/Archive and Image Display actors.
None, see above
No new transactions are needed. A new set of requirements will be added to the Image Stored, Query Images and Retrieve Images transactions
New transactions (standards used)
None, see above
Impact on existing integration profiles
The embedded image part in Scheduled Workflow will be impacted
New integration profiles needed
A new content profile should be included for each modality with specific acquisition and display requirements.
Breakdown of tasks that need to be accomplished
Use as input the work that has been done for the enhanced CT/MR demonstration, and create new Modality Content profiles based on this existing work.
- Reference Input - [Enhanced DICOM Objects Clinical Scenarios.doc]
7. Support & Resources
DICOM WG16 has requested that IHE creates this profile, and is available for support. Philips is prepared to provide the author for the new Profile.
If some enhanced modalities are not addressed this year (e.g. US), adding them later might disrupt or expose holes in whatever profiles we do release.
We can't address all possible clinical scenarios this year. Adding them later may disrupt or expose holes in whatever profiles we do release.
We may get no response from the clinical society or the vendors, because the large-scale implementation of the enhanced MR and CT objects still has to happen. And if PET/XR don't get involved, how do we know our spec is "modality unbiased". Need to actively recruit displays and target "value" for them. Need to recruit clinicians to get a value-statement and useful display reqs. If we get neither, the profile is in trouble.
- If we are unable to get Clinical Volunteers to help with the definition of the Enhanced Objects this does not bode well for the Profiles. With regard to the vendors it is a chicken/egg problem: some vendors are waiting for this profile.
The profile might be addressing a non-existent problem
- It is certainly possible for the same clinical study to be encoded differently between different vendors. It is possible the different encodings might not cause problems but DICOM felt it necessary to spend time profiling them for their demonstration. The chicken-egg problem certainly seems to need addressing.
Vendors might not implement the profile.
- There are currently only one or two Display Vendors, and getting them involved is critical to success. Some modalities and PACS have already demonstrated prototypes and some have released in product.
9. Open Issues
The NEMA Committee for the Advancement of DICOM would like to organize a first informal testing of this profile with systems that envisage or already give support for enhanced CT and MR objects at the January 2007 connectathon. Answ: we assume this NEMA desire is still valid for the 2009 connectathon.
We need to consider how this profile must be packaged, as a separate profile per modality, or per clinical scenario?
- Consider a "General Enhanced" profile with the general guidelines (if any), and specific profiles for certain clinical exam types.
- The General profile would provide general guidance/fallback for when you're not doing a specific exam type.
- It would also provide behaviors that are typically common to all exam types.
- Support for the General profile would be a pre-requisite for the Specific profiles.
- Specific profiles would address a certain exam type and can include requirements for each modality that does that type of exam like what is currently included in the Acquisition Modality today for specific Modalities
- When Specific behaviors apply, they override the General.
- Consider 2-3 specific Clinical Profiles to go with the "General" Profile in the first year
- The 5 clinical specialities are too much to cover.
- Need to do at least a couple to better understand what is "General"
- Need multiple modalities so we can figure out how that should be documented.
- 2 of the use cases are MR specific
- "Cardiac Imaging" is very open ended and harder to address without dealing with US, XR and PET
- Cardiac should maybe be done collaboratively with IHE Cardiology (are they active this year?)
- Proposal, do either:
- 3 Profiles = General Enhanced Object Profile + CT/MR Contrast Perfusion + CT/MR Spinal
- 4 Profiles = the above + Cardiac Imaging
Should we address enhanced XR at also at this moment in time” If not, how will an enhanced XR profile be added later?
- Answ: Add it later as a seperate use case. For now, lets stick to the ones we have experience with from the NEMA Demo.
Should we address PET/CT now?
- Answer: Same as XR.
Downgrading Enhanced to Standard
- Should consider taking a stand on how this should happen and who should do it
- How much display behavior capabilities do we establish as baseline
- Would need to keep it basic, would need clinician input
How much clinician input is needed for defining the display requirements?
- Answ: estimated required input is somewhat low.
- Use cases are pretty straight forward.
- Should contact specialty societies directly or work through RSNA, or work through vendor luminaries, etc.
10. Tech Cmte Evaluation
Effort Evaluation (as a % of Tech Cmte Bandwidth):
- 45% for Addressing CT/MR Cardiac, MR Diffusion, CT/MR Contrast Perfusion
- 30% for Addressing MR Diffusion, CT/MR Contrast Perfusion
Responses to Issues:
- See italics in Risk and Open Issue sections
- Dick Donker