SIIM: How to take the archive out of PACS

2016 06 30 11 32 39 405 Portland Oregon 200 V2

PORTLAND, OR - Are you interested in deconstructed PACS, but not ready to take the plunge? Deploying a vendor-neutral archive (VNA) that takes over archive management from the PACS could be a prudent stepping-stone, according to a Friday talk at the Society for Imaging Informatics in Medicine (SIIM) conference.

Using this approach, the VNA would be able to fully manage the imaging records within the archive, while radiologists would not experience any change to the traditional front-end PACS functionality that they're used to, according to Bob Coleman, senior director for imaging informatics at MaineHealth in Portland, ME.

"This idea of using your PACS in a slightly different way than you traditionally have could be viewed as a stepping-stone to more of a fully deconstructed or best-of-breed model where you take what your PACS has done and break it up," he said.

MaineHealth has had a traditional PACS for about 15 years and is in the process of deploying a VNA, adding an enterprise image viewer, consolidating radiology and cardiology PACS archives, and implementing an enterprise imaging strategy.

"As we looked at what we were doing, we really started kicking around this idea of an 'archiveless' PACS, a name that we made up," Coleman said.

Deployment models

Currently, the most common approach to PACS and VNA deployments configures modalities to store new images to the PACS, which then archives the image objects to the VNA. In this approach, the PACS still manages the archive as well as the quality control (QC) workflow and determining which prior exams are relevant.

"You're using a [VNA] to store your data, but the PACS is still managing the archive," Coleman said. "The PACS is still the system that's controlling what's in your archive. For example, you can't just go into your VNA to delete a study. You have to go into the PACS and tell your PACS to delete the study, and the PACS will actually trigger the VNA to delete the study."

In the "archiveless" PACS model, the PACS no longer manages the archive and the VNA takes over more of the functionality traditionally provided by the PACS. The PACS is still used for radiologist tasks such as accessing the worklist, opening studies, performing diagnostic review, and launching 3D capability and speech recognition software.

"All of those things do not change," he said. "But your back-end processing -- your DICOM routing, your [quality control], the things that happen in the back office, so to speak -- all of that could actually be moved to your VNA," he said.

In contrast to the traditional PACS/VNA model, imaging modalities would be configured to send newly acquired studies directly to the VNA. Based on configurable rules, the VNA would then forward the images and relevant prior studies to the appropriate PACS, Coleman said. The VNA would also manage the QC workflow and identify relevant prior exams.

As a result, PACS would now act essentially as a cached workstation, receiving everything it needs "just in time" from the VNA. Also, the PACS would delete exams a short time -- approximately 30 days -- after they are needed. Users could also use PACS to pull data directly from the VNA if it had not previously been routed, Coleman said.

A mixed or hybrid model that combines both methods is also possible. Some PACS could operate in an archiveless mode, while others -- such as a cardiology PACS -- could operate in more of a traditional mode.

"You may do one approach for radiology and one approach for cardiology," he said. "This is probably what we are going to be implementing at MaineHealth, sort of a straight into the VNA from the radiology side and more of a traditional approach on the cardiology side. The [cardiology] data is going to flow from the modality into the cardiology PACS, which will archive to the VNA and pull [it] back when needed."

Key benefits, challenges

The archiveless PACS model eliminates shared ownership of the archive with the PACS, enabling the VNA to fully manage the imaging records within the archive, Coleman said. In addition, it simplifies the invocation of a backup PACS if the primary PACS fails.

"If the VNA is smart enough to push everything to a PACS or a reading system just in time, you could just as easily push to a backup system," Coleman said.

It also supports the routing of data to different remote PACS to support nighttime reading and lays the foundation for potential replacement of PACS in the future. Other benefits include the freedom to implement the VNA's data life-cycle management capabilities and the ability to route local and external exams to the PACS in an identical fashion. It also may avoid the need to license software from your traditional PACS vendor to integrate with the VNA. What's more, the model offers shared quality control tools and methodologies that can be employed across different PACS, according to Coleman.

The archiveless PACS approach is not without challenges, however. For example, it can be difficult to get people inside and outside of the organization -- including vendors -- to understand the vision, benefits, and necessity of changing the current model, he said.

Other challenges include having to manage the acquisition, QC, and reading status of each study, coordinating the reading task status with the availability of study image data in the reading system. All copies of imaging records need to be in sync between the VNA and the PACS, and display performance requirements also have to be met.

The archiveless model will probably not save money initially; savings will likely come in the future, however, if you decide to replace your PACS or move to more of a deconstructed PACS model, Coleman said.

How does the archiveless model look in action? When providing subspecialty reading services for a local imaging practice, images are acquired and archived remotely at its location but read locally, he said. On the other hand, off-hours reading service for MaineHealth involves remote interpretation but local acquisition and archiving of the imaging studies.

Some issues that haven't been solved yet are image markups, electronic medical record (EMR) integration, 3D integration, and easy access to prior studies, Coleman said.

"Anything that is not pushed very directly is not accessible in a remote system," he said.

A path to deconstructed PACS?

Just like the deconstructed PACS model, the archiveless PACS concept is based on the VNA taking over handling of order processing, identifying and prefetching prior studies, modality worklist, image QC, DICOM routing, and exam archiving. However, the PACS is still responsible for reading worklists, diagnostic image display, dictation/speech recognition integration, 3D integration, inter- and intracommunication, and nondiagnostic image display. It shares responsibility for image markups/key images with the VNA, according to Coleman.

In the deconstructed PACS model, those PACS functions are completely taken away. Reading worklists and dictation/speech recognition integration are handled by the electronic medical record, while the VNA handles markups/key images. The diagnostic viewer handles diagnostic image display and 3D integration and shares responsibility for inter- and intracommunication with the enterprise viewer. The enterprise viewer also handles nondiagnostic image display.

"We're kind of preparing for moving down this road to deconstructed PACS in the future," he said.

Final advice

If you're happy with your traditional PACS and are unlikely to change it in the future, Coleman recommends a traditional VNA model for those who are interested in consolidating archives but are risk-averse.

An archiveless option could be a stepping-stone "if you may replace your PACS at some point, if you may move more to a deconstructed model, if you like the idea of being able to point your [VNA] to a backup system or perhaps to point your VNA to a nighttime reading service, or if you perhaps are looking for more of an evolutionary change as opposed to ripping everything out and replacing everything," he said.

"Remember you're not impacting radiologists with this archiveless model," Coleman said. "From their perspective they're still using the same system that they've been using forever, so there's no real significant change there as long as you can get all of the data into the PACS just in time."

The final step would be deconstructed PACS, "if you truly want to change your system and replace it with something completely different, or if your EMR is encroaching on the things that have traditionally been managed by your PACS," Coleman said.

Page 1 of 775
Next Page