Technology has revolutionized the way doctors diagnose and treat patients. But according to a new study by Seattle researchers, it can’t best the professionals yet—at least when it comes to breast cancer detection. On Monday, a study published in the journal JAMA Internal Medicine showed that computer-assisted detection (or CAD) for mammography does not help radiologists identify breast cancer. And because the technology adds at least $400 million to annual health care costs, that’s a big problem.
The makers of the imaging system, Hologic Inc. and Icad Inc., originally marketed their product as a radiologist’s “second set of eyes” and a “spellchecker.” But the study found that radiologists were more likely to miss cancers when using CAD. Their accuracy was actually 7 percent better when they did not use the computer-assisted review.
“The bad-news story from society’s perspective is we’re spending a lot of money on something that sounded like a great idea—and just isn’t,” Diana S.M. Buist, an investigator with Seattle’s Group Health Research Institute who helped conduct the study, told the Seattle Times.
More than 90 percent of mammograms in the U.S. are now computer-assisted. Since its 1998 approval by the Food and Drug Administration, there has never been strong evidence that it works. In fact, there was previous evidence to the contrary. In 2006, the same Seattle research group behind Monday’s findings published a study showing CAD’s shortcomings. In 2011, yet another study from the University of California, Davis, came out with similar results: CAD does not improve breast cancer detection.
According to Bloomberg Business, these studies were criticized “because they involved only older women or older technology that was read by less experienced radiologists.” The most recent study was designed to fix that.
As Joshua Fenton of the University of California, Davis’ Health System argues in his editorial accompanying the study, it’s probably time we stop spending health care dollars on unnecessary technology.