One of the biggest problems in radiology when you consider enterprise and multi-enterprise image distribution, becomes managing workflow and data flow," says Gary Wendt, M.D., M.B.A., vice chairman of informatics for the University of Wisconsin department of radiology in Madison. Images generated on any number of modality scanners must be routed to specific workstations for post-processing procedures and to the clinicians who require that information to provide appropriate patient care.
For example, if a spinal imaging study is acquired, will it be read by an orthopedic surgeon, neurosurgeon or other specialist, and where will that clinician be located physically within the enterprise to view the post-processed image data set?
Wendt uses a McKesson Horizon system with a central common RAID (redundant array of independent disks) repository on-demand system that is integrated via their EMR (electronic medical record) system to clinical data maintained by their HIS/RIS capabilities.
Warren S. Edwards, Ph.D., vice president of engineering at McKesson Medical Imaging stresses the importance of a logically centralized architecture where all of the data that comes into the system is stored using a high availability configuration, with no single point of failure. They recommend use of a WAN (Wide Area Network) infrastructure with built-in redundancy in two or more database servers.
Coupling distribution challenges with the massive volume of image data sets produced by today's multislice scanners requires careful allocation of network resources.
"With some of the CTAs [computed tomography angiograms], you can have 1,000 to 2,000 images, and you can be pushing a Gigabyte of data. If it's sent into the PACS, and it's brought up once or twice in the workstation, you multiply your total data flow by two or three," continues Wendt. "If you have an auto-routing situation, and you push that data set to 30 workstations, you've multiplied your traffic over the network so that you're now pushing 30 Gig over your network."
John C. Weiser, Ph.D., diagnostic imaging physicist for Xtria Healthcare in Frederick, Md., explains that a standard chest x-ray file would contain about 10 megabytes per image. However, with digital mammography, the typical image file is closer to 40 megabytes. He describes one network configuration solution involves use of a VLAN (virtual local area network), to remove imaging data flow from their primary IT network, thereby providing its own dedicated bandwidth.
In addition to their VLAN network design, Weiser explains that the Lahey Clinic (Boston) utilizes three levels of image compression between the server and the workstation. Lossless compression is used for primary diagnostic functions. In that scenario, the image remains mathematically unchanged once it has been decompressed. They use lossy compressions for clinical compression that provides images to referring physician offices, and deep compression is used for archiving image data sets. Most clinicians would find it difficult to discern which images had undergone lossy or lossless compression.
Another issue involves adherence to HIPAA regulations for patient privacy. Peter McClennen, general manager of global market for GE Medical Systems Information Technology, explains that Centricity Enterprise Web used as a data and image distribution system requires password access. If the user is within the hospital, they use a workstation located on a secured network. Outside the institution, they access the network through a VPN (virtual private network) or SSL (secure socket layer).
MORE DIAGNOSTIC INFO
Current workstations leverage computing power to offer advanced applications that benefit clinical management of patients who often exhibit complex disease processes.
Silicon Graphics, Inc. (SGI) has worked for many years with OEMs to provide components that they could integrate into their various modalities. Volumetric visualization is a trend that they see in enhanced workstation capabilities.
Michael Brown, SGI marketing team manager for advanced graphics explains that they can take the entire volume of data from a CT scan, enable the clinician to view images from whatever angle they prefer, manipulate various levels of transparency to see through certain structures, or even allow them to be time varying so that they can see a heart beating or muscles contracting.
The Advantage Workstation from GEMS offers software analysis tools that permit clinicians to accomplish