Accommodating Non-DICOM Images: Is Your Enterprise Imaging Strategy Diverse Enough?

Twitter icon
Facebook icon
LinkedIn icon
e-mail icon
Google icon
 - Michael J. Gray
Michael J. Gray, Principle, Gray Consulting, San Francisco

Healthcare organizations are beginning to understand that they need a strategy for how they will handle medical imaging across the enterprise.  This concept of enterprise imaging must consider more than the replacement of an existing radiology or cardiology PACS, and it means more than image enabling the EMR with interfaces to those departmental PACS. 

The implied scope of enterprise imaging goes beyond consideration of the traditional, DICOM-oriented departmental PACS to include images produced during such procedures as surgery, endoscopy, ophthalmology, and numerous other “oscopies” and “ologies” from bronchoscopy to urology.  This broader concept of enterprise imaging must consider non-DICOM as well as DICOM images, still-frame images as well as video clips, image sets that are the result of ordered procedures, as well as image sets that are simply captured on mobile devices during office visits or encounters in departments such as dermatology and emergency.

Achieving the broader concept of enterprise imaging requires the development of a strategic plan that must encompass more than scheduling PACS replacements and insuring EMR access to radiology and cardiology images. The broader plan would determine how these other “non-traditional” images will be acquired, identified, described, managed, and displayed. 

Perhaps the first order of business for the organization is to determine the nature of the problem that needs to be solved.  Is there a need to build another dozen or so PACS solutions for the ’ology and ’oscopy departments, or is the problem image enabling the EMR with all of these non-DICOM images produced outside of radiology and cardiology? 

Standing up another handful of standalone PACS runs contrary to the current populist approach to data consolidation.  If any of the ’ology and ’oscopy departments really need a PACS solution, perhaps a single PACS can be built to address this collective.  After all, there are more similarities than differences in the image data objects and workflow in these departments. 

In most healthcare organizations, the real problem appears to be how to provide access to these images from the EMR. Therefore, it seems likely that the best approach to solving the proliferation and management of “other images” is to consider this a part of the strategy to image-enable the EMR.

Image enabling the EMR: History

Image enabling the EMR used to be all about accessing and displaying DICOM images being managed by an existing PACS like radiology and cardiology, maybe with maternal fetal medicine and endoscopy.  Access was achieved by building an interface between the EMR and the PACS.

Display was achieved through the use of the clinical display application associated with the PACS.  The most common alternative to establishing individual EMR interfaces to each PACS and using multiple display applications, is to push studies from the smaller PACS to the radiology PACS and use the radiology viewer to display all types of images. 

This choice is understandable. In most healthcare organizations, the radiology PACS is the largest existing image data management system. The infrastructure is already in place. Presumably a favorable ingestion and management fee can be negotiated with the PACS vendor.

Radiology PACS probably is a better image manager than an enterprise content management system that was designed to manage and display documents. The organization has long-term experience with radiology PACS, especially its associated acquisition technology and workflow. The features and functions of the clinical viewer are good enough for most basic image display use cases. Significantly, the radiologists can use their experience to provide guidance to the clinicians and caregivers.

Unfortunately, there are complications when extending this radiology-centric solution beyond DICOM studies.  The majority of the ’oscopy and ’ology imaging procedures produce MPEG video clips, not still frame images. The files are very large. There are interface issues. The workflow is labor intensive, because video editing is required. Each of these issues is problematic to most radiology PACS.

Recent approach

Arguably the better approach to image enabling the EMR is to deploy a more universal image viewing solution, one that is capable of displaying both DICOM and non-DICOM image objects, and still-frame images as well as video clips. A key characteristic of the “univiewer”—my term for the universal viewer—is that it requires a single interface to the EMR. 

Ideally, a univiewer would present a single-pane view of all image types associated with a patient’s episode of care and offer a broader range of display features and functions than the typical radiology-oriented viewer. The univiewer’s zero-/near-zero footprint client paired with server-side rendering and pixel streaming technology offers a superior level of performance.

Besides a small working cache, there is no need for long-term management of the data outside of the PACS. The univiewer can federate searches for both new studies and relevant priors across multiple, discoverable data repositories on the organization’s network.

Driving demand for access

A number of issues are driving this expanded scope of the image-enabled EMR. Most organizations will soon discover that once physicians become accustomed to actually using their EMR, they will quickly recognize that access to PACS-based, DICOM image procedures is no longer sufficient. Today, enterprise imaging is expected to go beyond radiology, cardiology and any other department that happens to have a PACS.

There is wide recognition that the other ’oscopy and ’ology imaging procedures have clinical value in teaching and documentation, as well as treatment planning and follow-up. Informal (mobile) imaging is already happening, as there are numerous clinical uses for predominantly still frame images captured by smart phones and tablets in use cases found in dermatology, emergency, anatomical pathology and wound care.

The recent approval by many healthcare organizations of “bring your own device” (BYOD) policies has exponentially increased technology availability and capture opportunity, but there is poor governance over the use of mobile imaging. An enterprise imaging strategy must establish who exactly should be allowed to perform such imaging, what methodologies should be used and, most importantly, what security measures should be in place. Needless to say, failing to immediately establish effective governance policy for this new imaging modality potentially exposes the organization to a high level of risk.

An expanded view of enterprise imaging means that the organization’s strategy for image enabling the EMR must now include all images with clinical value.  This includes all DICOM studies with associated orders, all ’oscopy and ’ology studies that are primarily non-DICOM video clips with associated orders and all informal (mobile imaging) image sets primarily comprised of still frame images without associated orders. 

In addition, the strategic plan must address the fact that ordered procedures have an associated reimbursement, while informal image sets collected during an encounter with no formal order have no associated reimbursement. This fact makes informal imaging somewhat cost-sensitive.

Technology issues

The enterprise imaging strategic plan—more specifically, the organization’s strategy for image enabling the EMR—must recognize and address a number of technology issues.  DICOM compatibility means relatively simple image acquisition, management and display, so any deployment strategy should address these imaging sources first.

Likewise, DICOM provides a relatively simple pathway for creating a formal order for the imaging procedure by generating patient and study identification (ID) numbers via processes like Modality Worklist.  The order also provides a relatively simple methodology for associating a standardized description for the image study. Therefore, an enterprise imaging deployment strategy should accommodate all of the ordered procedures before tackling those without an order.

The strategic plan then must consider the following technology issues related to how the other “non-traditional” image studies will be acquired, identified, described, managed and displayed. 

  • Video imaging systems like those used in the ’oscopies and ’ologies will require a specific acquisition interface solution.
  • Lengthy video clips will require an editing feature to produce key clips or still frames from the original video runs to reduce bandwidth and storage demands.
  • Video requires a decision on the nature of the data object format: DICOM wrapping or encapsulation, or native MPEG.  This latter decision subsequently affects the data management and display choices.
  • The absence of an order significantly complicates the process of mobile image acquisition as it requires a methodology for creating/associating the patient ID without the help of Modality worklist.
  • Mobile image acquisition requires a methodology for creating a unique Image/study Identifier like the accession number assigned by a PACS.
  • Mobile image acquisition requires an editing feature to delete unwanted images from the set.
  • Mobile imaging requires a methodology for appending a study descriptor without the benefit of an existing department-oriented lexicon. This process has to be largely automated and eliminate as much variability as possible. This is the classic “head CT vs CT head” issue.  Subsequent access to mobile image sets and any possibility for hanging protocols demands a standard nomenclature for the study descriptors.
  • Mobile imaging requires a methodology for removing all of the images and patient identifying information from the capture device once the images have been forwarded to a collection server, or other designated target, to protect patient privacy.
  • Mobile imaging also requires a decision on the nature of the data object format: DICOM wrapping or encapsulation, or native JPEG and MPEG.  This decision also affects the data management and display choices.

Ripple effect

Expanding the scope of the image-enabled EMR beyond radiology and cardiology is like the proverbial opening of Pandora’s box.  It seems like every decision has a ripple effect on the next decision.  In addition to how the video clips or still-frame images  will be acquired and identified (presumably using a suite of software solutions), there are two other key decisions that the strategic plan must identify and resolve, the interdependencies of which are both complex and sometimes frustrating.

Decision No. 1. The first decision that must be made is where and how these video clips and sets of images will be stored and managed.  The following are the more obvious options.

Existing radiology PACS. This option is probably not the best choice for this class of imaging for several reasons. First and foremost, the cost of ingesting and managing the image data levied by the PACS vendor is probably going to restrict rather than encourage this type of imaging. Secondly, this probably forces every image set to be converted to DICOM. Thirdly, this obviously adds volume to the inevitable future data migration project when the incumbent PACS is replaced.

Existing enterprise content management solution. This is a good option, assuming the ECM can handle image data, but it is not the best option. First, ECM solutions are ideally suited to document not image management. Secondly, the ECM is generally ill equipped to handle video. Thirdly, this is another potential future data migration project when the VNA is deployed.

Vendor neutral archive (VNA).  This is the best option, assuming the organization has already deployed the VNA.  The VNA vendor is accustomed to negotiating an ingestion and management cost in line with the relative value of the image set, diagnostic studies being more valuable than informal image sets with no associated reimbursement rate. Secondly, most VNAs can handle either DICOM or native object formats, adding flexibility to the acquisition solution. Thirdly, the VNA can assure cross-system compatibility, because this is what VNAs do.  Fourthly, and perhaps most significantly, the VNA virtually eliminates the need for a future data migration. 

Unfortunately, if the VNA is not already in place, the VNA option is expensive and adds the time-consuming selection and deployment process to the original task of image-enabling the EMR. If mobile imaging is already underway in the enterprise, especially if the methodologies in use are not uniform and the data is not completely secure, there really isn’t time to investigate and deploy a VNA.  The organization needs to address the mobile imaging issues immediately.

Decision No. 2. The second decision organizations need to make is how these video clips and sets of images will be displayed. The following are the more obvious options.

Existing radiology PACS clinical viewer. The radiology-oriented features and functions of this display application are not ideally suited for this broader range of images and video clips.

The univiewer. The more generic, yet feature rich univiewer is ideally suited to this broader range of images and video clips, as this is exactly what this display application was designed to deliver.

Risky business

There should be a sense of urgency in developing the broader enterprise imaging strategic plan described in this article.  Physicians are asking for this expanded scope. More significantly, the organization should recognize its exposure to the high level of risk associated with allowing mobile imaging to continue without proper governance and a well-conceived solution in place to guarantee data security.

The key question, then, is how does the organization expand the scope of image enabling the EMR—including access to all of the image sets described herein—within a reasonably short time frame and with a modest budget.  In the interest of being more inclusive, we should determine how to achieve this scope of image enabling the EMR without having to deploy a VNA.

I believe the most expedient solution, one that gets the organization on a fast track for a broader enterprise imaging strategy, is to deploy a univiewer whose associated image cache is capable of managing a reasonable volume of image data. If the univiewer cache is based on an enterprise SQL server, the organization should be able to manage a year or two of mobile images before performance begins to degrade. Those two years should be sufficient time for the organization to select and deploy a VNA to take over data management from the univiewer cache.

Radiology and radiologists specifically can play a key role in this enterprise imaging strategic plan. Radiology is where imaging began.  Radiology directors and their PACS support staff have years of experience vetting imaging solutions and supervising change management programs.

Their participation in defining the end state of the enterprise imaging solution, the associated deployment strategy and the design of the various workflows involved in image acquisition should guarantee the overall success of the project. Radiologists also can use their PACS experience to provide guidance to the steering committee that will be charged with developing the requirements document that will lead to the selection of the univiewer best suited to meeting the needs of clinicians and caregivers who will routinely use it to access images.