Technical Update

Ann Scherzinger, Ph.D.

The New Image Standard­ DICOM: Why You Should Care

As hospitals aspire to create the universal patient medical record, standards have evolved to facilitate the incorporation of images into these electronic charts. Since vendors are initially reluctant to give up their proprietary hold on image storage and displays, it has been up to the medical community to demand the creation of universal standards. DICOM (Digital Imaging and Communications in Medicine) has become the standard protocol for image transfer and storage in Radiology. This standard was created by the National Electrical Manufactures Association (NEMA, http://medical.nema.org/ ), reacting to demands from the radiology user community, as a means to facilitate the electronic distribution and review of clinical images. Since its first versions, initiated by the American College of Radiology (ACR) and NEMA in the 1980s (labeled ACR-NEMA), the standard has grown to encompass most medical imaging specialties including endoscopy, ophthalmology and cardiology.

The DICOM standard is a full communication standard, giving not just image formats but also protocols for network communications, transfer, storage and display of images and structured reports. The standard is continually evolving and documents describing the current version, 3.0, as well as past and future version changes can be found at ftp://medical.nema.org/medical/dicom/.

Figure 1

DICOM is based on object oriented software principles and thus images are "information objects" with defined classes. In the standard, endoscopy is defined as an imaging modality, ES, and the endoscopy image is one of a set of images labeled as visible light (VL) images. Visible Human images, considered an external-camera photographic modality (XC) would also fall in the VL class of images. All images contain a leading header followed by the image data. Headers can vary in size depending on the information object, but all begin with a 128 byte offset followed by the characters "D","I","C","M". This is followed by tags which describe patient demographics, image acquisition parameters and the special attributes associated with a particular image type. A sample of an US Doppler image with some of its header attributes is shown in Figure 1. Special attributes associated with VL images are listed below in Table 1.

Tag identifiers are assigned to all image attributes and are further explained in the Data Dictionary (Part 6 of the standard) and in the Information Object Definitions (Part 3 of the standard). EUS images are included in the Ultrasound (US) modality where attributes used to define the image and transducer type, transducer position and image orientation are used to distinguish the various US imaging modes. The US class has an extensive range of attributes used to define gray scale, Doppler and cine image data as well as typical US measurement data.

Since DICOM is a full communication protocol, vendor conformance requires that medical devices employ a basic set of software and network communication capabilities. Compliance may be limited to only a subset of services and information objects so it is important to review a devices DICOM Conformance Statement prior to purchase. Additionally, one can test aspects of conformance by the electronic review of DICOM images from one vendor on another’s workstation. Vendors can facilitate this by presenting DICOM images on their websites for user download and review.

Although there are still many problematic implementation issues, DICOM has provided the ability to query, retrieve, store and display clinical images from multiple vendors, using a single image browser. Free browsers are available on the web for basic viewing applications, see http://www.psychology.nottingham.ac.uk/staff/cr1/dicom.html, while more sophisticated browsers are commercially available through many equipment vendors and software suppliers. As always, these browsers need to be tested for their ability to read the desired information object.

DICOM features are being developed to integrate imaging with all aspects of the patient electronic medical record through the direction of clinical scenarios developed by the Integrating the Healthcare Enterprise (IHE) committees, http://www.rsna.org/IHE/index.shtml and http://www.himss.org/asp/issuesbytopic.asp?TopicID=11 . Throughout its development, the DICOM standard has been driven by the user community. A list of user representatives on all of the DICOM working groups is given on the NEMA website and user comments are routinely invited on all standard additions. Imaging scientists and healthcare providers are urged to participate to insure the most optimal sharing of this critical medical information.

Editor:
John C. Deutsch, M.D.
Duluth, MN

Editorial Board:
Manoop S. Bhutani, M.D.
Galveston, TX
William R. Brugge, M.D.
Boston, MA
Peter R. McNally, D.O.
Denver, CO
Iqbal S. Sandhu, M.D.
Salt Lake City, UT
Thomas J. Savides, M.D.
San Diego, CA