Mammograms No Better With Computer's Help

July 27, 2011 -- Commonly used computer-assisted detection (CAD) makes mammograms more costly -- but not better at finding cancers, a large-scale study finds.

CAD is now used to help interpret three-fourths of mammograms in the U.S. It adds 9% to 15% to the cost of a mammogram.

Despite the extra cost, CAD doesn't improve breast cancer detection or help find cancers at a more favorable stage for treatment, according to an analysis of some 685,000 women who underwent more than 1.6 million mammograms.

"In real-world practice, our study suggests CAD has little impact on the outcomes of screening mammography," study leader Joshua J. Fenton, MD, MPH, of the University of California, Davis, tells WebMD.

Fenton's team compared mammograms read with and without computer-assisted detection. They found that:

• CAD slightly increased the number of false-positive mammograms -- that is, it increased the number of women called back for further testing who turned out not to have breast cancer.
• CAD did not increase the detection rate of more dangerous or invasive breast cancers.

"The way it is currently used in practice, CAD will slightly increase a woman's chance of unnecessary recall for further testing, but probably does not affect the chance it will detect early breast cancer," Fenton says.

The problem isn't so much with CAD as with the people using it, says Robert A. Smith, PhD, director of cancer screening at the American Cancer Society. Smith was not involved in the Fenton study.

"CAD is not a substitute for competence in reading mammograms," Smith tells WebMD. "CAD is not an autopilot. It is an aid. It can be very effective if you are good at reading mammograms. But if you are not a capable reader, it will result in what Fenton found: more false positives and no improvement in finding more subtle cancers."

Another problem has less to do with CAD than with the limits of what mammograms can detect, suggests Donald A. Berry, PhD, head of the division of quantitative sciences at the University of Texas M.D. Anderson Cancer Center in Houston. Berry's editorial accompanies the Fenton study in the Aug. 3 issue of the Journal of the National Cancer Institute.

"Mammograms find cancer early. The question is, which cancers are they finding? They disproportionately find cancers that are slowly growing, which results in overdiagnosis and overtreatment," Berry says.

Even if CAD improved mammography, Berry suggests, the difference would be negligible.

"From the perspective of a woman getting a mammogram, the incremental benefit from CAD is essentially zero," he says.

Fenton essentially agrees.

"Most dangerous breast cancers are very hard to detect with mammography to begin with," he says. "By adding more frequent mammograms or a tool like CAD, we don't capture those very dangerous cancers."

Smith, however, points to a 2008 study showing that in expert hands, CAD truly does improve breast cancer detection. In that study, experts who used CAD improved their breast cancer detection rate from 81.4% to 90.4%, with only a slight increase in the number of women unnecessarily recalled for further testing.