Editorial: TRIP™ Update

Richard L. Morin¹

Published online: 27 April 2004

As you might recall, TRIP™ is the name coined by the Society for Computer Applications in Radiology (SCAR) for a concept that has been on the minds of many PACS researchers as our technology continues to evolve. TRIP^™ stands for Transforming the Radiological Interpretation Process and is trademarked by SCAR. The TRIP^™ Initiative has the potential for encouraging new research and development in new methods for handling the large image sets that can now be produced by radiology modalities.

If you look back at the history of digital image display, especially for multislice image data produced for MRI and CT studies, you would see that the original softcopy display paradigm more often than not simply reproduced the view a radiologist would expect to see if he or she were reading formatted film. Images were tiled on the monitor using some predefined format, such as four images on a monitor or twelve images on a monitor. Frequently, the images were sent using multiple preset window and level settings such as bone or soft tissue. As softcopy reading became more sophisticated, it was fairly quick and easy for the radiologist to adjust her workflow to use a stack mode instead of tile and to perform interactive window and level instead of sending duplicate slices from the scanner. With recent advances in acquisition modalities, we are seeing a rapid growth in the number of images per study which makes the interpretation process more complicated and time consuming. We did a study on the number of images per day a radiologist reading cross-sectional studies could expect to see. In 1994 at the Mayo Clinic Jacksonville, a radiologist could expect to view 1,500 cross-sectional images per day. By 2002, this number had grown to 16,000 images per day, and we estimate that by 2006 this number could grow to 80,000 images per day. If a radiologist looked at each image for 1 second, that would translate into 22.2 hours for 80,000 images. No one would argue that we have a shortage of radiologists, which may not improve, and using our current radiology interpretation methods and workflow, a day's image production could not be interpreted in a day. When you look at the numbers and the facts, it is obvious that we need to change something. We will not reduce the number of images produced, nor will we be able to lengthen a day, so we need to change the interpretation process to make it efficient and effective.

The concept for developing a program that changed the radiology interpretation process was born at a SCAR Research and Development committee meeting in July 2002. At that meeting a consensus of the committee agreed that this was a topic of major importance and concern and they wanted SCAR to take the lead in developing information about the issues and to encourage innovative approaches to solving the problems. At a strategic planning retreat in January 2003, the concept was named TRIP^™ and SCAR set out three three fundamental objectives for the initiative: first, to improve the efficiency of interpretation of large datasets; second, to improve the timeliness and effectiveness of communication; and, third, to decrease medical errors with the ultimate goal of the initiative to improve the quality and safety of patient care. We believe that with the aid of computers and processing, the interpretation process will change on three fronts: First, the workflow for image acquisition, processing, and interpretation will change. A common workflow now would include CT or MRI acquisition followed by processing by technologists that results in additional images produced in different planes and possibly including additional three-dimensional objects or screen captures of three-dimensional images. In some cases, the only images available to the radiologist may be the ones prepared by technologists with little capability of interactive processing by the radiologist other than window/level or zoom/roam. The workflow is already evolving giving the radiologist greater processing control using tools such as multiplanar reformat or maximum intensity projections. The second change might be the addition of automated preprocessing into the workflow. For example, a computer may screen images for normal or abnormal patterns and generate the normal reports. This may sound frightening and may be rejected by many as a dangerous precedent, but we should remember that many laboratory and pathology tests such as a Pap Smear are "read" by a computer. In radiology, the computer performs all the calculations for a bone density test, presenting the radiologist or physician with a series of measurements, graphs of normals, and an interpretation of the results. Finally, there is the potential of images being transformed into another form that communicates function or anatomical features without closely resembling the original image. In the future we may not have people looking at images, we may have people looking at computer output. The bull's eye graph of cardiac wall motion does not "look like" the original gated SPECT images, but it is an effective way to convey information about the study.

To promote this initiative, we decided to start with three projects. First, we started compiling a literature review regarding image interpretation, workflow, workstation design, and alternative image interpretation methodologies. This is now available at www.scarnet.org. Second, we selected the topic of "Medical Image Interpretation-The Collision between Humans and Data" for the closing session at SCAR 2003. We invited speakers from other image-intensive industries (NASA, National Imagery and Mapping Agency, the film industry) to join us in a discussion of handling large datasets. We will hold a TRIP^™ Initiative Conference in October 2004 in Washington, DC, which will include academic investigators and members from federal agencies and industry. Finally, we are working with other societies to get the message out and generate interest in the problem and potential solutions. Dr. Katherine Andriole introduced the TRIP^™ initiative to attendees at the DICOM 20th Anniversary Conference and Workshop in September 2003. The SCAR video "Medical Imaging's TRIP to the Future" was presented in the RSNA infoRAD Roundtrip Showcase at the RSNA 2003 meeting in December. In addition, "Transforming the Radiological Interpretation Process through Information Technology" was presented by George Bowers, Eliot Siegel, Bruce Reiner, and myself at HIMSS 2004 in Orlando, Florida.

This year SCAR is announcing a partner program for industry to participate in TRIP^™. Companies will be able to choose their level of commitment as Terra partners, Peta partners, or Exa partners. Partners will have access to early TRIP^™ research as well as focused forums.

We all believe that using computer resources and information technology can help the image interpretation process and that, if something is not done, the problem will become unmanageable. This is a problem waiting around the corner and the most frightening thing is that our volume estimate may be too low. We know it will take time to change the way radiologists practice and we need to start now. We are all committed to our organization (SCAR) and we want to show our membership that through our leadership, we can all benefit.

Contact SCAR|Home