1300 North 17th Street Suite 1752 Arlington, VA 22209 (703) 841-3281 http://dicom.nema.org MINUTES DICOM WG-27 (Web Technology for DICOM) June 6, 2011 MITA / NEMA Headquarters Arlington, VA USA Members Present Represented by ACR/Johns Hopkins University James Philbin Agfa Healthcare Imaging Rob Horn Aspera, Inc. Diego Dugatkin Corelab Partners David Clunie GE Healthcare Chris Lindop (P) Harris Healthcare Solutions Timothy Culp Merge Healthcare William Lingley Newport Instruments Robert Leif (P) Philips Healthcare Cor Loef Tarbox, Lawrence Lawrence Tarbox (P) Toshiba Med. Research Institute, USA Kevin O’Donnell Members Absent Voting Representative AAOMR Allan Farman Canada Health Systems Carl Zeiss Medical Software GmbH Compressus DeJarnette Research Systems Ben Macerola Tobias Kurzke Laszlo Gasztonyi Fred Ganong 1 ________________________________ DICOM Working Group Twenty-seven June 6, 2011 ETIAM Medical Connections, Ltd Siemens AG – Healthcare Sector Emmanuel Cordonnier David Harvey Nikolaus (Niki) Wirsz Alternate Representatives, Observers and Others Present: Coelho Prado, Thiago Cowan, Ron Csipo, Dezso Dawson, Tim Fauquex, Jaques Kumi. Daniel Laconti, Michael Lacroix, Marie-Jose Lee, Andy McNamara, David Medema, Jeroen Nolte, Bjorn Revet, Bas Schafer De Martini, Richard Augusto Solomon, Harry Stockham, Charlie Tsannes, Alexis Vastagh, Stephen Presiding Officer: Brazil, Fed. Univ. of Santa Catarina, Dept. of Informatics Merge Healthcare-WG-06 Object Forge Vital Images Open DICOM (P) Aspera Konica Minolta-WG-06 GE Healthcare – WG-06 Harris Healthcare Solutions Virtual Radiologic Corp. Philips Healthcare Siemens Healthcare-WG-06 Philips Healthcare-WG-06 Brazil, Fed. Univ. of Santa Catarina, Dept. of Comp. Sciences GE Healthcare – WG-06 DeJarnette – WG-06 Aware, Inc. MITA, Secretariat (P) =By phone and web meeting Cor Loef, Co-Chair James Philbin, Co-Chair 1. Preliminary Events Participants identified themselves and their employers. The meeting was called to order at 9:00 a.m. USA Eastern time. S. Vastagh reviewed the antitrust rules for the meeting. 2. The “MINT” Proposal Note: WI or WIP means a DICOM Work Item/Work Item Proposal (a project to make substantial changes/additions to the DICOM Standard.) The group met in joint session with WG-06 in the morning, continued in a breakout session after lunch until 3 pm and rejoined WG-06 thereafter. 2.1 Jim Philbin’s Presentation: 2 ________________________________ DICOM Working Group Twenty-seven June 6, 2011 Disclosures: CTO of H2MI, Hopkins/Harris startup focused on cross enterprise medical imaging and Med. Advisory Boards: Siemens, Merge, Vital Imaging (Toshiba Med. Sys. Group) Jim Philbin made a presentation titled: Problems with DICOM (posted in the meeting folder) to argue for the MINT proposal (to transfer of images with a web-based protocol versus the current TCP/IP. Lengthy discussion followed that mentioned: DICOM Push model is complicated and involves lots of storage overhead. Core of discussion: is it a problem in the protocol, or in a bad implementation. One member spoke for DICOM in the meeting: 200 destinations, 1.2 M studies a year transmitted with no problems. Issues with unique identifiers, and accession numbers. For some, or even most of the issues, IHE is mentioned by David as the place to resolve these. Viewing: it was argued that DICOM is too slow. The group concluded that it must be clear what is the Handshake referenced and need to know about parsing Meta data as well as about the transfer of the images to the processing servers. Will be clarified that it is a transfer between servers. Problem: why cannot DICOM saturate the wire? The transfer today will not fill-up the wire. This is still controversial without apparently sufficient evidence. Evidence should be presented. The proposal needs to list the DICOM issues, and the use cases. With that information the Group can decide what to address. Keeping the Meta data up-to-date: this has always been outside of DICOM, the internal database representation. However, the outside behavior has to be specified. IHE work includes separating the Image Manager and Image Archive. Normalizing data is a huge issue. Archives detached from database are out-of-date. Enhanced objects on the Study level: what is the need for this? Breast MR example of hundreds of series. A specific implementation could be one multi-frame object. It could be an IHE profile for Breast MR. Separate out the issue: end of transmission. Gap in WADO: push/post C-STORE. Performance over the network: Memory mapped protocols seem to work well; although, at certain moment the latency is an issue. A proposal from Aspera (posted in 3 ________________________________ DICOM Working Group Twenty-seven June 6, 2011 the meeting folder) was presented: to negotiate a network transport method. If not supported, fall back to existing DICOM. What to use for the handshake? There was a long discussion on application layer vs. socket layer solutions. WG-27 proposed the following Action Items to WG-06: ACTION 1: Create a new SOP class for PACS to take classic CT/MR/PET objects and create an enhanced object for it what the workstation can use for their applications. In the meantime use Multi-Frame SC: stop gap, and learning experience for writing new SOP class. WG-06 will propose a WI item for this for the September DSC committee, and will set up AdHoc Group to define the SOP Class. ACTION 2: WG27 can define the web based equivalents for C-STORE, C-GET, CMOVE and Query methods for this SOP Class. Unanimous approval in WG-27. Already being done in different environments, and we cannot and don’t want break this. C-STORE does not pass easily thru firewalls. Availability and reliability, you want to replicate as soon as possible. Use HTTP or any other protocol that could transport over available ports. Adding new objects to an existing study is possible. ACTION 3: Restful approach equivalent to Web Services in supplement 148. We want to support the full functionality in both methods ACTION 4: Continue with our WG27 actions for QIDO, both Restful and Web Services based, and later address NADO. Create a paper along the way listing and explaining the characteristics of both methods. Unanimous approval in WG-27. Already being done in different environments, and we cannot and don’t want to break this. ACTION 5: Create a web based solution for UPS. (This action was not presented to WG-06.) WG-06 concurred with the ACTION items proposed by WG-27. 3. Future Meetings of WG-27 Several t-cons were be scheduled to discuss and follow-up on the action items: June 29, July 13, Aug. 8, 2011. 4. Adjournment The meeting was adjourned at 4:30 p.m. USA Eastern Time 4 ________________________________ DICOM Working Group Twenty-seven June 6, 2011 Submitted by: Reviewed by Legal Counsel: Stephen Vastagh, Secretary, June 29, 2011 5 ________________________________ DICOM Working Group Twenty-seven June 6, 2011