The skills radiologists need to excel in filmless reading can be acquired only through practice, which may not make perfect but will improve performance. Research has shown that once radiologists adjust to reading images from workstations instead of viewboxes, their productivity soars. But the time it takes to successfully make that transition is highly dependent on several physical, workflow, and technological parameters.

Making the shift is more complex than merely changing the mode of image display, said Elizabeth Krupinski, Ph.D., an assistant professor of radiology at the University of Arizona. Physical characteristics of the environment-including the ergonomic aspects of workstations as well as sound and lighting-influence intangibles, such as radiologist comfort level, which indirectly affect performance.

Judicious use of a computer mouse, for example, allows radiologists to click quickly through images in stack mode. A day of clicking through thousands of images, however, may lead to repetitive stress syndrome.

Performance problems can also be traced to time lost navigating among multiple icons on a display. Too many icons results in gaze fixation, which is not only distracting but potentially detrimental to image interpretation.

"The less complicated an interface, the more intuitive it is," Krupinski said. "That lowers the learning curve."

TECHNOLOGICAL TOOLS

Aside from physical and workflow parameters, multiple technological tools exist to bolster soft-copy interpretation performance and productivity. The mere ability to have information and images flowing through hospital and radiology information systems can dramatically reduce the time needed to move from acquisition to interpretation and display.

Other workflow boosts can be derived from soft-copy workstations. Automatic display protocols increased radiologist efficiency by more than 10% in one, according to research conducted at the VA Medical Center in Baltimore. In addition, employing the modality work list has reduced the error rate in transmitting CT scans to the PACS from 8% to about 1.5% at the facility, said collaborators Dr. Eliot Siegel and Dr. Bruce Reiner.

In a study presented at the 2003 Society for Computer Applications in Radiology meeting, Reiner evaluated image processing techniques by readers examining normal and subtle fracture exams. The study also assessed readers' perception of image quality, fatigue, and interpretation time. The readers overwhelmingly chose multifrequency processing and gray-scale inversion as image processing tools of choice, and the two techniques also yielded better accuracy in diagnosis of subtle fractures, he said.

Digital image processing technologies are just the tip of the interpretive iceberg in soft-copy reading. Disease-specific frequency processing, 3D, multiplanar reformatting, and computer-aided detection are increasing in importance in filmless domains. The use of these technologies is undergoing increasing study, as PACS leaders hope to harness the explosion of digital data through the use of these and other tools. SCAR has launched a new initiative to spearhead research and education about best practices for interpreting and managing digital data sets.

The Transforming the Radiological Interpretation Process (TRIP) initiative aims to improve the efficiency of data set interpretation, improve the timeliness and effectiveness of communication, and decrease medical errors.

Specific areas of research include better understanding of human perception and visualization, as well as navigation, database integration, and CAD.

"If CAD can help radiologists not miss things, or not get as fatigued, by helping with some of the decisions to be made, then we may be able to avoid mistakes and actually increase search efficiency," Krupinski said.