Clear clinical benefits have resulted from the widespread adoption of multidetector CT (MDCT) technology, as well as from the use of advanced, contrast-enhanced, and functional MRI applications. The resulting increase in data-handling needs, however, has been less welcome at facilities that are not equipped to deal with it. For example, a standard abdominal/pelvic study produces 60 to 70 images, but an MDCT study might yield 900 images. CT angiography can create 2,000 images, and a multiphase cardiac study might involve 8,000 images. Cardiac CT can require 4 gigabytes of memory for data storage, even before postprocessing.
Four times as much information will be generated per study as higher resolutions become standard. For facilities that are already struggling to handle today’s data, this increase will underscore the need to solve multiple problems that are already evident: network bandwidth that is too low, workstation memory that is too small, servers that are too slow, and software that is too old to handle advanced imaging data.
The Thin-client Answer
The solution to these problems, for an increasing number of imaging providers, is to shift away from PACS with 3D workstations in favor of thin-client systems that use central processing and exam databases. Streaming technology makes it possible for physicians to use their own PCs to gain access to original 3D data and to use multiplanar reformatting and 3D navigation tools. These operations involve no transmission of data from modalities to PACS to workstations and do not impede network data flow. In addition to solving problems of access and workflow, thin-client technology has considerable economic benefits to offer the enterprises that adopt it.
One such enterprise is Massachusetts General Hospital (MGH), Boston, where more than a million exams per year are generated by the hospital and the outlying sites in its health system. For primary interpretation, MGH relied on PACS, and images reached referring physicians via Web-based distribution. This system became impractical for support of the number of 3D visualizations needed, according to Keith Dreyer, DO, PhD, who is vice chair of radiology at MGH and assistant professor of radiology at Harvard Medical School. He explains that referring physicians need not the 10,000 images that an exam might generate but, instead, 3D representations of the images, along with the ability to slice into the 3D dataset to plan surgeries and measure changes. The implementation of thin-client technology brought that capability with it.
Hardware and Support
Because the use of thin-client technology allows 3D images to be viewed immediately from any PC of clinical quality, the cost savings gained by avoiding the purchase of new image-viewing hardware can be massive, particularly for larger health systems. A second large financial benefit is that there is no need to pay for maintenance and repair or for service contracts to support all that hardware. Savings are also obtained because it is not necessary to synchronize hardware upgrades throughout the system every few years in order to maintain cutting-edge capabilities. Instead, if a hardware upgrade is called for, it is limited to the central servers.
As one of the five largest US pediatric radiology departments, the pediatric imaging department of Children’s Hospital, Birmingham, Ala, has experienced the image-handling problems of high-volume providers. Its RIS/PACS could not use the 3D capabilities of the parent University of Alabama Medical Center, and it needed to give access to 3D images and image-manipulation tools to referrers. By implementing thin-client technology from Visage Imaging (Carlsbad, Calif) in 2006, the hospital was able to avoid additional hardware/support costs beyond standard PCs while integrating its RIS/PACS with the thin-client system. The entire hardware addition was limited to two radiology-department servers.
Before that change, as Stuart A. Royal, MD, clinical professor in the pediatric radiology section of the University of Alabama, Birmingham, explains, clinicians who wanted to see original data had to travel to the radiology department. Now, they have rapid access to original images and can use 3D image data efficiently, but there was no expenditure for any additional hardware.
Network and IT Resources
Image processing, when performed using a thin-client configuration, uses a central server and Web distribution. This means that the need