SCAR initiative challenges image reading process

A major initiative by the Society for Computer Applications in Radiology is looking for ways to shift the image interpretation paradigm.Transforming the Radiological Interpretation Process, or TRIP, will begin at the closing session of SCAR's annual

A major initiative by the Society for Computer Applications in Radiology is looking for ways to shift the image interpretation paradigm.

Transforming the Radiological Interpretation Process, or TRIP, will begin at the closing session of SCAR's annual meeting in Boston next month. The initiative is designed to find novel solutions to the problem of information and image data overload through research, education, and discovery.

"Because of the number of images being produced by new modalities and the fact that we're seeing more patients now, radiologists are looking at substantially more images than they ever did before," said TRIP committee chair Richard L. Morin, Ph.D., of the Mayo Clinic in Jacksonville, FL.

As an example, current CT modalities produce a number of images an order of magnitude higher than earlier models. Whereas older CT studies may have generated 40 images, current exams easily produce 300 to 600 or more images.

"We believe it is necessary to examine the way radiology images are interpreted," Morin said. "A transformation or shift is necessary - otherwise we're going to wear out the radiologists. There's no way we can keep doing this."

TRIP backers hope to foster interdisciplinary research on technological, environmental, and human factors to address the information explosion, focusing on large image data sets, navigation devices, staffing, and training. Organizers are putting together a joint workshop or congress, probably with a federal government agency, Morin said. The idea is to see what imaging experts can come up with to relieve radiologists of overload.

"TRIP has the potential to fundamentally change the practice of radiology, covering everything from reading room design, ergonomics, and workstation design to improved service to patients and physicians, " said Dr. David Channin, chief of imaging informatics at Northwestern University Medical School.

The question is how to facilitate radiologists to better view, analyze, understand, navigate, filter, and communicate their "value" to the referring physicians, according to Channin. The TRIP initiative will be closely tied to reporting/coding/billing and workflow activities under way in DICOM Structured Reporting, IHE reporting workflow integration profile, and the RSNA's RadLex project.

"If we agree that the job of the radiologist is to reduce uncertainty in the diagnostic effort of the clinician, then the TRIP initiative has the greatest potential to make us better at what we do best," Channin said.

He predicted we will see a host of graphical user interface techniques by which radiologists can efficiently navigate large data sets, annotate very precisely the image features found, and then group these findings into meaningful, succinct differential diagnoses.

"Big blobs of free text will become extinct," he said. "Radiologists do not have the time to generate them, and clinicians do not have the time to read them."

Notification and communication tools will deliver these structured reports not only to electronic medical records but to personal electronic devices such as wireless, Bluetooth, two-way messaging, smart phones, and so on.

"As TRIP initiative technologies mature, we will see the ability to mine patient-specific and aggregate structured data collections to improve the efficiency of first-generation tools and to make new tools available," Channin said.

The June 10 closing session, "Medical image interpretation: the collision between humans and data," will be TRIP's first public exposure.

For the session, Morin has assembled a number of imaging community leaders, including experts from NASA, USC Computer Animation Laboratory, and the National Imagery and Mapping Agency, to offer their insights. Moderators from radiology will assess the likely effects of large volumes of data on image interpretation and management.