Managing the Flow in Radiology
Flex Your Data
Integrated PACS and thin clients help hospitals build a true 2D/3D infrastructure
By Keith Dreyer, MD, PhD, and Hartmut Schirmacher, PhD
High-resolution tomographic scanners and other 3D and image-fusion technologies provide a number of compelling advantages for diagnostic medical imaging. However, 3D data acquisition modalities such as CT and MR create ever-larger volumes of data, increasing the need for faster and larger servers, higher network bandwidth, workstations with large memory and fast graphics, and advanced diagnostic software.
Furthermore, the gap between the amount of information in the original data and the 2D report images sent to clinicians and referring physicians is growing steadily. Radiology departments and 3D technologists expend significant effort generating multiple series of reformatted images, cine loops, and standardized 3D views to document and communicate the diagnostic findings. Still, key physicians outside the radiology department increasingly want the flexibility to review the original data in addition to the diagnostic report.
Implementing thin-client PACS—PACS with tightly integrated 3D thin-client technology—helps hospital enterprises address these challenges.
Advantages of the Thin-Client Paradigm
The current reality in many hospitals is a centralized PACS along with a number of loosely integrated 3D review and postprocessing workstations. Data is sent from the modalities to the PACS, and then forwarded to (or prefetched by) selected workstations. Radiologists and cardiologists review the images on these workstations, and a multitude of key views or processed images are generated and sent back as snapshots and cine loops to the PACS as well as to selected recipients.
There are many problems with this “isolated workstation” paradigm. Original data is not always available where it is needed, and a significant amount of time is spent sending original and processed data between different workstations and servers. Also, additional quality control is needed to make sure that all generated diagnostic images are correctly archived and transferred to all recipients. Workstation hardware is often too slow or does not provide enough memory for efficient review of large 3D studies. Departmental workflow becomes manually intensive, software versions and optional application packages are not consistently available across workstations, and referring physicians and clinicians can review only snapshot images, but cannot use these views as bookmarks into the original data.
|Analysis of abdominal arteries and leg vessels using the advanced vessel tool of Mercury’s thin-client server solution.|
The key differentiator of a thin-client PACS (PACS with tightly integrated 3D thin-client technology) is a true central 3D processing paradigm, along with efficient streaming technology to enable thin clients to act as fully capable front ends to all viewing and processing functions of the PACS. All DICOM data remains on the server (no data transfer prior to launching the 3D viewer), all operations are performed directly on the server, and all functions can be accessed from anywhere in the hospital enterprise via thin clients.
Since data are no longer sent back and forth explicitly, the time required for data transfer, quality control, and workflow management is reduced. Network resources are used more evenly; peak bandwidth problems are effectively eliminated.
To allow for efficient diagnosis and patient treatment, 2D, 3D, and 4D data as well as large studies and reports should be accessible immediately throughout the hospital enterprise. But in addition to image distribution, clinicians increasingly rely on specialized clinical applications that have to be tightly integrated into the diagnostic workflow. A thin client providing applications such as CT cardiac analysis and surgical planning in addition to state-of-the-art image quality and manipulation improves the workflow across the borders of the different modalities, specialties, and groups.
CT cardiac analysis is an excellent example in which a thin client-based approach allows the radiology and cardiology departments to share image data and work together effectively. Both can review the 3D and 4D cardiac data in a similar fashion and share key images for easy navigation to the relevant views. The radiologist can use the 3D navigational capabilities and tools of volumetric cardiac CT to identify suspected abnormalities. The same application provides the cardiologist with quantitative analysis tools for the left ventricle, including wall motion, wall thickening, and ejection fraction. Thus, radiologists and cardiologists can perform their assessment of the key functional parameters directly within the PACS workflow, and recall the results from this analysis anytime and anywhere in the hospital.
For instance, one cardiac scan easily can produce 500 images with current scanners. Assuming 10% RR interval scans for functional analysis, one study is composed of up to 5,000 images. Clearly, the loading times of such data sets are unacceptable on workstations. With a client-server approach, diagnoses can be accelerated significantly through preprocessing the data on the server machine. The result is a time- and cost-efficient workflow for diagnosis and therapy among an interdisciplinary team of radiologists, cardiologists, and cardiac surgeons.
With a 3D-enabled thin client, physicians now are able to review primary tomographic data throughout the hospital with all of the advantages of MPR and 3D navigation. Interactive 3D images are available hospital-wide immediately after the scan is finished and can be reviewed interactively on existing PC hardware.
From the hospital’s point of view, a thin-client PACS with clinical applications removes technical barriers between different modalities and departments, creates a much more homogeneous and manageable IT infrastructure, and helps to exploit existing modality and workstation equipment and other IT resources. For radiologists, clinicians, and referring physicians, the thin-client PACS provides a significant increase in productivity and flexibility through instant access to all diagnostic image data anywhere, anytime.
Keith Dreyer, MD, PhD, is the vice-chairman of radiology at Massachusetts General Hospital, Boston, and assistant professor of radiology at the Harvard Medical School. Hartmut Schirmacher, PhD, is the senior product manager in the Life Sciences Group at Mercury Computers, Berlin.
Four Rural Hospitals to Share PACS Platform
To improve efficiency for radiologists, four competing New England-area hospitals using the same radiology group will centralize their PACS. All four hospitals—Alice Peck Day Memorial Hospital, Lebanon, NH; Mount Ascutney Hospital and Health Center, Windsor, Vt; New London Hospital, New London, NH; and Valley Regional Hospital, Claremont, NH—operate in rural areas, and depend on the services of Valley Radiologists Professional Associates, Claremont, a single radiology group with just six members, for their imaging needs. The hospitals will consolidate their imaging needs using the AMICAS Vision Series PACS from AMICAS Inc, Boston, as a central platform.
Originally, the hospitals planned to each purchase a separate PACS, but the possibility of integration gave way to the enterprise-platform plan. The four hospitals came together in the name of patient care, said Peter McClennen, president and COO of AMICAS. “It’s a very competitive marketplace,” he said. “These are businesses, and they compete. With this group of people, they felt that if they could create an integrated imaging environment, it really would be a much better patient care solution.”
The platform allows each hospital to maintain their existing IT systems while taking advantage of a centralized PACS. Alice Peck Day runs on a MEDITECH HIS; Mount Ascutney has a CPSI HIS; New London has a McKesson HIS; and Valley Regional operates on a Siemens HIS. “Not only do these four hospitals have different HIS systems, but they also have different RIS systems and different EMRs,” McClennen explained. “But they’re going to have one single integrated imaging system. The key is having a system that’s built for integration. We believe our ability to smoothly integrate with all of them simultaneously is unique.”
Implementation will be hospital by hospital, McClennen noted, with a target duration of around 90 days. “First, we put the main infrastructure in, and we have an enterprise project manager who’s responsible for the whole program,” he said. “Then we go hospital by hospital, with Alice Peck Day being the first.”
A crucial feature of the AMICAS PACS is the ability to restrict access to data when necessary. “It has to have the ability to quarantine data per hospital, because some medical specialists are not approved to see data from other facilities,” McClennen said. “So, having technologies that allow you to restrict and approve access, especially with HIPAA and the political and business concerns of multiple ownership entities, was really important.
“But for radiologists or people who work at all of the facilities, we can give them enterprise access, and then they have the ability to see everything. The main benefit of that is quality of care, because then the right radiologist reads the right study regardless of where it’s rendered.”
More daunting than meeting the technological challenges, however, was balancing the political and business concerns, McClennen noted. “This is the first time I’ve ever seen this done,” he said. “That’s why I think it’s so exciting. Quality of care trumped business and political issues.”
IBM GMAS Improves Archiving for Orlando Regional
For large hospital systems like Orlando Regional Healthcare, Orlando, Fla, archiving data is no small feat. To help address the demand generated by its eight hospitals and two data centers, Orlando Regional recently implemented the IBM Grid Medical Archive Solution (GMAS) from IBM Corp, Armonk, NY. Incorporating servers, storage, software, and services, the GMAS provides hospitals, clinics, research institutions, and pharmaceutical companies with a multitier, multiapplication, and multisite approach to archiving.
Orlando Regional uses the system to address the exponential growth of fixed content data that the health care system has been experiencing.
“Like most big hospital systems, we have to replicate our digital images for radiology PACS, cardiology, and our medical records,” explained Alex Veletsos, CTO at Orlando Regional. “The Grid actually instantaneously creates a second copy of it and places it wherever you want it to be.”
This saves many trips between the hospital system’s two data centers, which are located 8 miles apart. “The advantage of GMAS is it does that automatically for us,” Veletsos said. “We don’t have to manually go in and create that copy; we don’t have to buy replication software to do that because it’s included with GMAS. Most importantly, it has the intelligence to prohibit replication of corruption.”
GMAS is based on IBM’s System Storage EXP3000, System Storage Grid Access Manager, System x 3650 servers, and IBM Global Technology Storage Servers. Delivered in prepriced bundles, the system is designed to eliminate many of the IT department’s manual administration tasks.
“It’s an all-in-one solution,” Veletsos said. “It doesn’t come in bits and pieces. If you tried to do this on your own as the CTO of a hospital, you would have to gather your own software, your own administrators. When you get this, you get everything. It makes the implementation easier, and it makes the administration easier.”
Orlando Regional’s GMAS solution has a capacity of 50 terabytes and was installed in less than 5 weeks, though data migration took a little longer. It now stores images from Orlando Regional’s radiology PACS, HIM medical records, and cardiology PACS applications.
“It changes our outlook from the traditional storage and archival strategy that we had in the past to a more automated way, and avoids replication and corruption,” Veletsos said.
IBM Corp introduced the GMAS at the annual meeting of the American Telemedicine Association, Washington, DC, held in mid May in Nashville, Tenn.