Research initiative explores image management issues

The urgent need to more efficiently manage the volume and flow of images in the radiology and medical fields has grown dramatically as scanning technology has evolved. Initially, soft-copy display, especially for multislice MR and CT studies, often simply reproduced the view radiologists would expect to see if they were reading formatted film. Images were tiled on the monitor using some predefined format, such as a display of four or 12 images. The images were frequently transferred using multiple preset window and level settings, such as bone or soft tissue.

As soft-copy reading became more sophisticated, radiologists could quickly and easily adjust their workflow to use stack mode instead of tile and to perform interactive windowing and leveling instead of sending duplicate slices from the scanner. Advances in acquisition modalities are driving rapid growth in the number of images per study, making the interpretation process more complicated and time-consuming.

The number of images available for review and interpretation has grown dramatically with the transition to the filmless healthcare enterprise. An informal study in the department of radiology at the Mayo Clinic in Jacksonville, FL, found that the number of images generated and stored per day grew from 1500 in 1994 to 16,000 in 2002. Extrapolating these volumes to the year 2006 presumes that approximately 80,000 images will be acquired every day. At a viewing rate of one image per second, the radiologist of the future, using today's practice strategies, will require 22.2 hours a day to interpret imaging studies.

These numbers make it obvious that changes are necessary. The number of images produced will not decrease, so the interpretation process must adapt to become efficient and effective. That transformation is under way.

DEFINING THE PROBLEM

By identifying the problem, the Society for Computer Applications in Radiology took the first step toward addressing it. In 2003, SCAR created the Transforming the Radiological Interpretation Process (TRIP) initiative to study the problems of image overload. The society took the second step earlier this year, when a successful TRIP conference, "Transforming Medical Imaging," helped to further define the problem.

The TRIP initiative is designed to focus academic, government, private-practice, and industry expertise on the impending crisis in imaging volume. Its ultimate goal is to improve the quality and safety of patient care.

Representatives from all areas of medical and scientific imaging, research, industry, and government agencies discussed the future of medical imaging and measures for managing the explosion in number and complexity of images and imaging technologies. The conference, held in Bethesda, MD, was partially funded by the National Institute of Biomedical Imaging and Bioengineering of the National Institutes of Health. Attendees from NIBIB, the FDA, and the National Library of Medicine expressed their interest in the TRIP initiative and the ongoing discussion of issues affecting the interpretation process in clinical care.

A key objective of the initiative will be obtaining government support for multidisciplinary research efforts and focusing attention on the need for better quality patient care.

The TRIP initiative has three basic objectives:

- improving the efficiency of interpretation of large data sets;

- improving the timeliness and effectiveness of communication; and

- decreasing medical errors.

The conference helped define both the initiative and the image management problems radiology faces. A consensus emerged on the urgent need to provide temporary and long-term solutions for a clinical problem of significant importance.

TRIP aims to foster research, education, and discovery of innovative solutions to the problems of information and image data overload. It will promote interdisciplinary research on technological, environmental, and human factors involved in managing and exploiting imaging data.

The two-day conference included plenary, scientific, and breakout sessions. Plenary sessions were presented by leaders in six major research areas related to TRIP: human perception, image processing and computer-assisted detection, data set visualization, information navigation, databases and integration, and evaluation and validation.

The lively and collegial breakout sessions allowed attendees to participate in one of four in-depth group discussions. The sessions culminated in group recommendations on issues facing the field, major impediments to progress, and the short- and long-term outlook for radiology. Future programs will continue to focus on and refine the challenges and issues that medical imaging confronts.

The conference presenters concurred that the number of images the average radiologist reads and interprets will soon outpace human capabilities.

"The radiology workflow is growing faster than the workforce," said Dr. Ramin Khorasani, vice chair of radiology at Brigham and Women's Hospital.

The outcomes of the conference will be published in the Journal of Digital Imaging.

NEXT STEPS

The first TRIP conference brought together presenters and speakers from all areas of imaging, as well as academia, government, and the corporate world. TRIP corporate partnerships provide ongoing support for annual research conferences/workshops, quality research by young scholars through the SCAR informatics grant program, and forums in which the various groups can interact and share results. Inaugural TRIP partners include SCAR corporate members Agfa Healthcare, Barco, Siemens Medical Solutions, Fujifilm Medical Systems USA, GE Healthcare, Stentor, Emageon, LumenIQ, TeraRecon, and Vital Images.

Following publication of the second TRIP white paper, SCAR will hold a forum for these corporate partners. From the outset, the aims of the TRIP initiative have reflected the concerns of SCAR's diverse and unique membership. Only in this society do the researchers, clinicians, vendors, and providers who work in imaging technologies come together on an equal footing to share insights and results and identify the direction of future work. These constituencies share in the risks of adopting new technologies, programs, services, and processes designed to improve performance and patient care. For more information about the society or the TRIP initiative, visit the SCAR Web site at www.scarnet.org.

Dr. Morin is the Brooks-Hollern professor of radiologic physics at the Mayo Clinic in Jacksonville, FL. Dr. Andriole is director of imaging informatics at Brigham and Women's Hospital in Boston.