CAD gains ground in climb to routine clinical application

July 1, 2008

Now that computer-aided detection has become part of routine clinical work for cancer screening in mammograms and is being applied in the differential diagnosis of cancer in the lung and colon, it's only a matter of time before it rates as the standard of care for diagnostic examinations in daily clinical work.

Now that computer-aided detection has become part of routine clinical work for cancer screening in mammograms and is being applied in the differential diagnosis of cancer in the lung and colon, it's only a matter of time before it rates as the standard of care for diagnostic examinations in daily clinical work.

Given today's powerful computer workstations, radiologists are clamoring for tools that will allow them to read smarter and faster. Manufacturers, in turn, are scrambling to not only meet high-tech demands but integrate products effectively into the workflow.

Nodule detection on CT scans, though not a new idea, is an example of a technique that's matured to the point that implementing it in a user-friendly way is a priority, said Dr. Heber McMahon, a professor of radiology at the University of Chicago. He added that new releases not yet commercially available have drastically reduced false positives.

Though CAD is not always reimbursed, and is therefore hard to justify in practice, McMahon's colleagues are looking at the bigger picture.

"If it can increase production and reduce the medicolegal liability for false positives, I think people will pay more attention," he said.

Also on the horizon are analysis of interstitial lung disease on CT scans, dual-imagery chest units, and programs that will track pathologies over time. An automated tool capable of detecting, measuring, and making comparisons could be a big productivity boost, McMahon said.

"A lot of the scans that we do are for follow-up on known cancer, and we're manually measuring these. Even using electronic calipers is very labor-intensive," he said.

Though some attendees at the European Congress of Radiology in March suggested that computer-aided detection and diagnosis tools could become so advanced as to make radiologists obsolete, McMahon doubts it.

"It really is a gradual progression over years and turns out to be a challenge to produce a system that can compete with a skilled radiologist,'' he said.

Meanwhile, reports from the 2008 American Roentgen Ray Society meeting in Washington, DC, suggest that adding CAD to CT colonography screening would make detection even more effective. Recent trial results showing the effectiveness of CTC screening plus news of possible reimbursement for the procedure may increase CTC utilization and the number of annual colon screening exams.

CAD in mulimodality breast image interpretation-mammography, ultrasound, and MRI-is being introduced slowly into clinical use.

"Such methods have the potential to be used as another diagnostic test,'' said Maryellen Giger, Ph.D., a professor of radiology at the University of Chicago.

Two recent studies suggest CAD for breast MRI may be closing in on true cancer detection. Researchers from Penn Diagnostics and George Washington University found that CAD could identify up to 96% of true malignancies on contrast-enhanced MRI. Investigators said, however, they must next develop ways of reducing false positives to remove obstacles to CAD's clinical acceptance.