Strategic Document
Revision 1.06, July 6, 2001
The purpose of this document is to provide a summary of the current activities and future directions of the DICOM standard. The content of the document is largely based on information submitted by individual working group chairs. WG10 will update this document on an annual basis and submit it for review and approval by the DICOM Committee at the committee’s annual RSNA meeting.
A Brief Background of the DICOM Standard
The introduction of digital medical image sources in the 1970’s and the use of computers in processing these images after their acquisition led the American College of Radiology (ACR) and the National Electrical Manufacturers Association (NEMA) to form a joint committee in order to create a standard method for the transmission of medical images and their associated information. This committee, formed in 1983, published in 1985 the ACR-NEMA Standards Publication No. 300-1985. Prior to this, most devices stored images in a proprietary format and transferred files of these proprietary formats over a network or on removable media in order to perform image communication. While the initial versions of the ACR-NEMA effort (version 2.0 was published in 1988) created standardized terminology, an information structure, and unsanctioned file encoding, most of the promise of a standard method of communicating digital image information was not realized until the release of version 3.0 of the standard in 1993. The release of version 3.0 saw a name change, to Digital Imaging and Communications in Medicine (DICOM), and numerous enhancements that delivered on the promise of standardized communications.
The DICOM standard now specified a network protocol utilizing TCP/IP, defined the operation of Service Classes beyond the simple transfer of data, and created a mechanism for uniquely identifying Information Objects as they are acted upon across the network. DICOM was also structured as a multi-part document in order to facilitate extension of the Standard. Additionally, DICOM defined Information Objects not only for images but also for patients, studies, reports, and other data groupings. With the enhancements made in DICOM (Version 3.0), the Standard was now ready to deliver on its promise not only of permitting the transfer of medical images in a multi-vendor environment, but also facilitating the development and expansion of picture archiving and communication systems (PACS) and interfacing with medical information systems.
Scope of DICOM
The DICOM Standards Committee exists to create and maintain international standards for communication of biomedical diagnostic and therapeutic information in disciplines that use digital images and associated data. The goals of DICOM are to achieve compatibility and to improve workflow efficiency between imaging systems and other information systems in healthcare environments worldwide. DICOM is a cooperative standard. Therefore, connectivity works because vendors cooperate in testing via scheduled public demonstration, over the Internet, and during private test sessions. Every major diagnostic medical imaging vendor in the world has incorporated the standard into their product design and most are actively participating in the enhancement of the standard. Most of the professional societies throughout the world have supported and are participating in the enhancement of the standard as well.
DICOM is used or will soon be used by virtually every
medical profession that utilizes images within the healthcare industry. These include cardiology, dentistry,
endoscopy, mammography, opthamology, orthopedics, pathology, pediatrics,
radiation therapy, radiology, surgery, etc.
DICOM is even used in veterinary medical imaging applications.
Technology Overview
The DICOM Standard addresses multiple levels of the ISO OSI network model and provides support for the exchange of information on interchange media. DICOM currently defines an upper layer protocol (ULP) that is used over TCP/IP (independent of the physical network), messages, services, information objects and an association negotiation mechanism. These definitions ensure that any two implementations of a compatible set of services and information objects can effectively communicate.
Independence from the underlying network technology allows DICOM to be deployed in many functional areas of application, including but not limited to communication within a single site (often using various forms of Ethernet), between sites over leased lines or virtual private networks (VPNs), within a metropolitan area (often using ATM), across dial-up or other remote access connections (such as by modem, ISDN or DSL), and via satellite (with optimized protocol stacks to account for increased latency).
At the application layer, the services and information objects address five primary areas of functionality:
· Transmission and persistence of complete objects (such as images, waveforms and documents),
· Query and retrieval of such objects,
· Performance of specific actions (such as printing images on film),
· Workflow management (support of worklists and status information) and
· Quality and consistency of image appearance (both for display and print).
DICOM does not define an architecture for an entire system; nor does it specify functional requirements, beyond the behavior defined for specific services. For example, storage of image objects is defined only in terms of what information must be transmitted and retained, not how images are displayed or annotated. DICOM can be considered as a standard for communication across the “boundaries” between heterogeneous or disparate applications, devices and systems.
The services and objects that are defined in DICOM are designed to address specific, real-world applications (such as the performance of an imaging study on an acquisition device). As such, DICOM is not a general-purpose tool for distributed object management. In general, information is transferred “in bulk” according to a “document” paradigm.
By contrast, general-purpose standards for distributed object or database management generally provide lower level, more atomic access to individual attributes. Though the DICOM standard does provide the so-called “normalized” services for patient and study management, these have not proven popular, and the “composite”, document-oriented, services have prevailed. This is mostly likely a consequence of the natural division of functionality between different vendors, devices and applications. For example, the ability to “set” or “change” a patient’s name is generally implemented in a proprietary and centralized manner. To safely distribute responsibility for such a change across boundaries between different applications requires more underlying support than DICOM currently possesses (such as support for transactions and two-phase commitment).
At the present time, the pressing needs in DICOM (as indicated by the priorities of the various working groups) are to address issues relating to security, performance, new modality technology, and workflow management. These needs are being successfully addressed using the conventional “underlying” DICOM technology. Where there are interfaces to standards based on other technologies (such as HL7 V2.x and 3), the focus for harmonization is on a shared “information model”. It may be the case that the nature of the underlying technology needs to be revisited in the future, whether it is to make use of more sophisticated off-the-shelf distributed object management tools such as CORBA, or less sophisticated encoding tools such as XML. However, the current priority is to address improvements in functionality to better meet the needs of the end-user, rather than to adopt an alternative technology for the sake of it. This priority is continually reinforced by a desire to remain compatible with the installed base of equipment.
When specific new technology is required, such as in support of new features such as security and compression, the strategy is to adopt proven international, industry or de facto standards. Accordingly, network confidentiality and peer authentication in DICOM are provided by the use of either TLS (an Internet standard) or ISCL (an ISO-based standard). Similarly, rather then develop medical-image-specific compression schemes, DICOM adopts standards developed by ISO/IEC JTC 1/SC 29/WG 1 such as JPEG and JPEG 2000. For interchange media, standard file systems compatible with conventional software (such as ISO 9660 and UDF) are used.
DICOM’s Relationship to Other Standards
Throughout the development of DICOM, much attention was devoted to establishing working relationships with other related standard initiatives throughout the world. The initial version of the standard leveraged prior work by ASTM. The Internet protocol TCP/IP was adopted in 1993. In the nineties, solid cooperation with CEN, the European Committee for Standardization, resulted in a number of jointly developed supplements. In parallel, the convergence of a Japanese interchange media format (IS&C) with DICOM required much joint work where JIRA, the Japan Industries Association of Radiological Systems played a major role. In the USA, DICOM participated in the early coordination efforts for healthcare standards with the ANSI-HISBB from which DICOM adopted a harmonized patient name structure, and started progressively to define links with HL7. This cooperation has now entered in a very active phase with the creation, in 1999, of a joint DICOM-HL7 working group. Finally, it was a logical step for DICOM to establish a Type A liaison with the ISO Technical Committee 215 at its creation in 1999. ISO TC 215 has decided not to create an imaging working group, but to rely on DICOM for bio-medical imaging standards.
DICOM is also focusing its attention to the evolution of standards linked to the Internet. DICOM strategy is to integrate Internet Recommendations as soon as they are stable and largely disseminated in consumer commercial products. In this evolution, much care is taken to ensure that the consistency of the DICOM standard is maintained with its large installed base. DICOM already uses standard healthcare enterprise intranets and soon the e-mail exchange of DICOM objects (Standard MIME type) should be possible. It is clear that the use of DICOM objects and services in commonly used information technology applications will grow in the future.
DICOM’s Organizational Structure
DICOM is a standards organization administered by the NEMA Diagnostic Imaging and Therapy Systems Division. The complete bylaws of the DICOM Standards Committee are available on the NEMA web site at www.nema.org. Working groups of the DICOM Committee perform the majority of work on the extension of and corrections to the Standard. Working groups are formed by the DICOM Committee to work on a particular classification of tasks. Once formed, working groups petition the DICOM Committee to approve work items for which the working group will execute the plan delineated in the work item. Once the output of a work item (generally a supplement or correction proposal) has been completed, it is submitted to Base Standards Working Group (WG6), for their review. Supplements to the standard then go through a public comment period, after which the DICOM Committee authorizes the supplement for letter ballot by DICOM members. Letter ballots require approval by two-thirds of those voting affirmative or negative and return of more than one-half of the ballots sent to members in good standing relative to letter ballots. Since the working groups perform the majority of work on the extension of and corrections to the Standard, the current status and future directions of the DICOM standard are best represented by review of each working group.
The DICOM Standards Committee and its Working Groups
|
|
|
The DICOM Standards Committee
Secretariat NEMA
Secretary Howard Clark, NEMA
Vendor Co-Chairman David Clunie, M.D. ComView Corporation
User Co-Chairman Lloyd Hildebrand, M.D., American Academy of Ophthalmology
Date of Last Update: July 6, 2001
Scope:
The DICOM Standards Committee exists to create and maintain international standards for communication of bio-medical diagnostic and therapeutic information in disciplines that use digital images and associated data. The goals of DICOM are to achieve compatibility and to improve workflow efficiency between imaging systems and other information systems in healthcare environments worldwide.
Working Group 1 (Cardiac and Vascular
Information)
Secretariat NEMA
Secretary Howard Clark, NEMA
Co-Chairman Jon Elion, MD, FACC, Heartlab
Co-Chairman Harry Solomon, GE Medical Systems
Date of Last Update: July 6, 2001
Scope:
To develop standards for the digital interchange of the cardiovascular images, physiologic waveforms and core clinical information, which characterize a patient undergoing a catheterization procedure.
Roadmap:
The WG1 roadmap is a long-term strategy to specify information objects and services to fully digitize and integrate data flow within and beyond the catheterization laboratory. The ultimate goal is a comprehensive digital cardiovascular record for the patient, of which the cath lab record is a significant part.
The cath lab is a multi-modality mix of many types of equipment from many different manufacturers. Moreover, cardiovascular medicine requires consultation and interaction between many medical disciplines. Standardized data interchange is critical in this environment.
The initial WG1 work effort was directed toward the definition of X-ray Angiography (XA) image exchange, including both a Basic Cardiac profile and a 1024 profile for media-based exchange. The next step was the development of DICOM cardiovascular waveform exchange, with specializations for hemodynamics, cardiac electrophysiology, and electrocardiography. Completing the data acquisition aspect is definition of intra-vascular ultrasound (IVUS) image exchange.
The next level of integration involves definition of various structured data interchanges in the cath lab, to support the interpretation and effective use of the basic images and waveforms. This structured data includes procedure logs, quantitative and qualitative measurements, and clinical reports, and requires specialization of the DICOM Structured Report services.
Finally, the overall workflow management of the cath lab within the cardiology department and within the larger institution will be evaluated. This will be done with regard to adapting the DICOM Study Management services to the particular needs of cardiology practice (as distinct from radiology).
Short Term Goals:
Completion of DICOM standards for IVUS image exchange, currently in Public Comment.
Specification of IODs, Templates, and SOP Classes for cath lab structured reports:
· Hemodynamics
· Quantitative Coronary and Ventricular Analysis
· Cath Lab Procedure Log
· Electrocardiography
· Cath Lab Clinical Report
Drafts templates are available for prototyping and demonstration.
Technical support for public demonstrations of DICOM for the cardiovascular community.
Current Status:
WG1 plans to meet approximately four times per year for 1-2 days each, plus interim discussions by telephone conference.
Current Work Items:
IVUS (Supplement 48)
Cath Lab Structured Reports (pending approval)
Risks:
The integrated data flow in the cath lab requires the implementation of multiple DICOM Services in the equipment, such as DICOM image and waveform interchange, and various types of structured reports. It will be important to establish a forum (not necessarily within WG1) that will promote the adoption of comprehensive interoperation profiles for the cath lab, similar to the effort of IHE in radiology.
Further, the various types of objects must be synchronized to provide clinically meaningful and accurate data acquisition. For example, IVUS images must be accurately synchronized to ECG and audio waveforms, and to the procedure log, to post-process for 3-D reconstruction and to be able to identify the sites imaged and the acquisition context. These multiple objects of different SOP classes must be effectively managed as a coordinated whole, and effective technical mechanisms for inter-equipment synchronization need to be developed.
Most of WG1’s current work is based on DICOM Structured Reporting, which is new and as yet unproven in practice. This presents several risks to the work of WG1:
· DICOM SR relies extensively on coded terminology and templates. There remain both practical and theoretical questions regarding the processes for development, distribution, and maintenance of code sets and templates for the cath lab SR information objects.
· DICOM SR does not currently support a means of controlling presentation, similar to Grayscale Presentation State for images. Rendering of SR is a major area for development, and substantial inter-product differences in presentation may affect the acceptance of SR by the clinical community.
· Tool support for DICOM SR is still rudimentary. This may be an impediment to adoption of SR information exchange.
· SR involves formalizing and standardizing much clinical knowledge, which has hitherto been informal or ad hoc. Development of SR templates thus requires much more involvement from the clinical community, as well as willingness by that community to adopt standardized formats over custom local formats.
Challenges and Opportunities:
The WG1 effort has the potential to significantly improve the operation of the catheterization laboratory, with consequent improvement in both patient care and facility administration. Major improvements in lab productivity and clinical response time to emergent cardiovascular conditions can result from integrated data flow within and between the cath department and the hospital. This integration includes both data moving into the lab (e.g., demographics, blood lab results), and data moving out (e.g., images and reports to the emergency department). The structured coding of cath lab data can facilitate analysis and improvement of patient outcomes, quality assurance, automated support for billing, as well as administrative analysis of organizational effectiveness. Beyond the institution, this data can support broad clinical research and regional or national clinical data registries.
The work effort must studiously adhere to the roadmap for progressive and incremental integration of the cath lab data flow, focusing on pragmatic next steps in the standardization process which bring real benefit to the cardiovascular user community (clinicians and administrators). A major challenge will be developing standards that facilitate the conversion of legacy cath lab systems to DICOM standard data interchange. The standards must be effective even in environments that have a mix of fully DICOM-compliant subsystems and legacy systems that may support only a limited subset of DICOM functionality.
Strong participation in WG1 from the European Society of Cardiology, as well as the American College of Cardiology, has helped ensure a broad international applicability of the work. It is critical to maintain this liaison and engagement.
Working Group 2 (Digital X-Ray) – This working group is not currently active
Secretariat
Secretary
Chairman
Date of Last Update: July 6, 2001
Scope:
To develop a DICOM standard for digital projection radiography.
Working Group 3 (Nuclear Medicine)
Secretariat NEMA
Secretary Richard Eaton, NEMA
Chairman Jeff Pohlhammer, Marconi Medical Systems
Date of Last Update: July 6, 2001
Scope:
To develop standards for the digital interchange of Nuclear Medicine and PET images.
Roadmap:
Potential work items would be generation of CPs to address the above-mentioned short-term goals.
Short Term Goals:
The chairman is currently trying to activate the committee so that several items can be addressed. These include:
· Identification and publication of the SNOMED terms currently in use for required coded attributes in the NM IOD. These were circulated among vendors at the time this IOD was created, but new implementers of this IOD have difficulty determining which SNOMED terms are applicable.
· Identification and publication of SNOMED terms that can be used for some of the optional coded attributes in the NM IOD. Of particular interest are terms that can be used to describe the image view (i.e. LAO, Vertical Long Axis, etc) and to identify imaging conditions such as stress/rest. The former would be applied to the View Code Sequence, which is already part of the NM IOD. The latter may need to be included in a new (optional) attribute.
· Clarification of several items in the NM IOD that are commonly misunderstood, such as clarification of when the Image Position and Image Orientation attributes are required.
Current Status:
This WG has been inactive for a while. The chairman has requested that the secretary contact committee members to schedule a teleconference.
Current Work Items:
None.
Risks:
None, yet.
Challenges and Opportunities:
Outside groups, such as the SNM (Society of Nuclear Medicine) have recently become more interested in DICOM, and have expressed concerns related to the issues mentioned above. This presents an opportunity to work with the SNM, and other related groups, to address these issues in a manner consistent with the architecture of the current NM IOD. The challenge is to do so in a timely manner.
Working Group 4 (Compression)
Secretariat NEMA
Secretary Howard Clark, NEMA
Chairman Alan Rowberg, MD, University of Washington (ACR)
Date of Last Update: July 6, 2001
Scope:
To provide data compression facilities for the DICOM standard and to advise on application or object-related definitions of data compression parts of the DICOM Standard created by other working groups.
Roadmap:
Develop appropriate DICOM transfer syntaxes for JPEG 2000 when it is released.
· Expected release of JPEG 2000 Part 1 is December 2000
Expectations for the next few years, after the current work items are finished:
· JPEG 2000 (ISO 15444-1 CD stage)
· lossless & lossy progressive wavelet
· MPEG-2 (WG 12)
· M-JPEG 2000?
· Retirement of unused 10918-1 processes
Use of Future Image Schemes
· Network
· any transfer syntax can be negotiated
· uncompressed default baseline
· compressed baselines?
· Media
· general purpose uncompressed
· XA/CT/MR 10918-1; US RLE & 10918-1
· future - Mammo DVD JPEG-LS ? 2000 ?
Future schemes for other data
· Single bit (scanned document)
· CCITT Fax
· JBIG
· JBIG2
· Overlays
· Waveforms (if not combined with images)
· Audio
· Future JPEG 2000 work
Beyond JPEG 2000 Part 1 (ISO 15444-1)
· Part 2 extensions (may not be royalty free)
· 3D transform (multi-slice or multi-spectral)
· Line/block based coders (major IPR issues)
· More Region-of-Interest modes
· Part 2 lags 6-9 months behind part 1
Short Term Goals:
We plan to serve a maximal number of modalities and clinical situations, with support for growing areas such as telemedicine and the use of new technologies such as DVD. We will develop the appropriate DICOM transfer syntaxes for JPEG 2000 when it is released.
Current Status:
Compression that is already available in the standard:
JPEG (ISO 10918-1) - all processes
· lossy (DCT)
· lossless
· sequential, progressive
· Huffman, arithmetic entropy coders
RLE (aka TIFF Packbits)
· ultrasound
JPEG-LS (ISO 14495-1) (DICOM CP-174)
· lossless, lossy (“near-lossless”)
Current Work Items:
Tracking the progress of the JPEG 2000 standard, evaluating its performance in medical imaging, and planning to develop appropriate DICOM transfer syntaxes for JPEG 2000 when it is released.
Risks:
The risk of JPEG 2000 not issuing as an international standard is low. We expect no problems in creating the proper transfer syntaxes. The success of other items will depend on support from the DICOM vendor community.
Challenges and Opportunities:
Multi-frame and 3D compression.
WG 12 is making progress on the incorporation of MPEG compression into the DICOM Standard. We now have a technical expert on MPEG to assist us in our work. The first phase of our work is to determine whether MPEG really offers considerably greater compression ratios and/or performance for Ultrasound cine than Lossy JPEG or JPEG 2000 compression. We have been concentrating on MPEG2 and have found that one of the key issues is the limited capability of most decoders. Most current MPEG2 decoders can only handle certain image frame sizes. We need to investigate this issue further to determine if we want to create a transfer syntax for MPEG2 with restricted allowable image frame sizes.
Relationships to other Standards:
DICOM has cross-representation to JPEG 2000.
WG 5 (Exchange Media)
Secretariat NEMA
Secretary Howard Clark, NEMA
Chairman David Clunie, ComView Corporation
Date of Last Update: June 12, 2001
Scope:
To develop DICOM standards for interchange media.
Roadmap:
Potential work items with respect to new types of media:
· Larger capacity MOD media are widely used but not actually recognized in the standard, both for 90mm and 130mm
Potential work items with respect to secure media:
· WG 14 has an assigned work item to develop media security proposals, currently at either the file or attribute level; the possibility of using a secure file-system has been raised and OSTA has begun work on Secure UDF; should this become a reality then WG 5 will need to review the implications of incorporating Secure UDF in PS 3.10, 3.11 and 3.12
Potential work items with respect to the DICOMDIR:
· Opportunities exist to improve or replace the DICOMDIR to better support:
· Large number of small files
· Longer file names (such as those based on UIDs)
· A “.dcm” extension
· Security features
Potential work items with respect to archival
· At present DICOM only defines interchange media, not archive media
· DICOM PS 3.10 files (+/- DICOMDIR) are frequently used on optical archive media, both removable and in jukeboxes
· A proposal from the VA has been received to consider defining an “archive migration format” to facilitate migrating studies from obsolete archives to new devices