Workstation purchasing hinges on practice preferences

Look, listen, and test drive. That's the advice users give tips on how to select a 3D workstation. Any number of practice-specific variables can influence which system and software are most suitable.

Three-D workstations can be divided into thick and thin clients. In thick clients, 3D software resides on a high-end workstation that permits fast access to individual users but can be costly to purchase and update. Thin-client systems rely on software loaded on a central workstation server, accessible via multiple small computers or a PACS. While costs and maintenance are efficient and scalable, these systems can be slow if many users are online at the same time.

Both types of workstations have their fans. Thin-client users appreciate access to workstations throughout the hospital, during conference presentations, and from home using a virtual private network. Those who prefer thick-client setups cite increased processing power and speed.

How the software will be used, and by how many users, will determine which is best. Will 3D be the primary mode of interpretation for most CT and MR studies? Or will it be a supplemental tool only occasionally used to generate images by referring physicians? Will workstations be installed in the 3D lab and used by technologists only, or will 3D software be available for use in the reading room as well?

While key differences remain among vendor systems, 3D workstations are moving closer together in terms of functionality. Image display quality, user-friendliness of the interfaces, and availability of automated tools are a few ways radiologists can evaluate workstations.

"No one system is the best," said Dr. Shawn Teague, an assistant professor of radiology at Indiana University School of Medicine in Indianapolis. "Every workstation has its strengths and weaknesses. Within the same application there may be features of one workstation that are better than another's."

Computer speed and memory is another big issue, he said. Because of the large data files generated by multislice CT, speed and the size of random access memory are crucial. Data must be available when a radiologist sits at the workstation. If a physician must spend time pushing or pulling data to the system, it is less likely to be used.

The ability to load and manipulate a case with as few clicks as possible is the highest ranked 3D workstation feature, according to an informal survey of audience members at the 2005 Society of Computed Body Tomography and MR meeting. Participants who attended the society's workstation face-off event ranked features that are useful in manipulating and displaying clinical cases. Other desirable attributes included intuitive user interface, automation, and clarity of screen layout.

Early adopters of 3D systems note an increasing trend toward 3D features merged with PACS workstations. PACS are known for image navigation and communications capabilities, while dedicated 3D workstations are considered more powerful and versatile for image processing.

It's not yet clear which form 3D will ultimately take: full 3D image processing capability on PACS or basic functions only, with advanced features available only on more sophisticated systems. But, clearly, the convergence is happening.

"PACS and stand-alone workstations are going to have to merge," Teague said. "There will be one computer where you can sit and look at cases dynamically on the fly."

Such a move would eliminate the rap that volumetric viewing takes too long. Radiologists find it time-consuming to ensure that the case is on the 3D workstation and to process, save, and send images to the PACS.

"Every two minutes extra that we spend on a case means there are 10 or 20 cases fewer that we will read that day," Teague said.

Similarly, a key limitation of today's 3D workstations is the lack of workflow integration with PACS, said Dr. Matthew Barish, director of the 3D and image processing lab at Brigham and Women's Hospital.

"We need to have the 3D available on the PACS as part of the hanging protocol, so it anticipates the type of images you are likely to want to generate and then begins constructing those images for you," he said.