News Release

Computer-aided detection system for mammography does not change recall, detection rates

Peer-Reviewed Publication

Journal of the National Cancer Institute

A study of more than 100,000 screening mammograms acquired at a large academic practice and interpreted either with or without the results of a computer-aided detection system has found that computer-aided detection does not change either the recall rate or breast cancer detection rate compared with mammograms interpreted without a computerized detection system. These results, published in the February 4 issue of the Journal of the National Cancer Institute, differ from those of earlier studies that suggested that computer-aided detection may increase the breast cancer detection rate by as much as 20% without substantially affecting the recall rate.

Several computer-aided detection systems are approved by the U.S. Food and Drug Administration to help radiologists identify regions of the breast that might be cancerous. (Computer-aided detection systems are different from digital mammography; digital mammography refers only to the manner in which the mammogram itself is acquired, and computer-aided detection systems refer to software programs designed to identify suspicious areas on either film or digital mammograms.) Computer-aided detection systems were approved based mainly on results of studies done in controlled settings; however, there is little information available on how such systems actually perform in a clinical setting.

David Gur, Sc.D., and colleagues from the University of Pittsburgh, studied the mammography recall and cancer detection rates at a clinical breast imaging practice at Magee-Womens Hospital of the University of Pittsburgh Medical Center for the 3-year period during which a computer-aided detection system was introduced. From 2000 to 2002, 24 radiologists had interpreted 56,432 mammograms without the use of the computer-aided detection system and 59,139 mammograms were interpreted with the aid of such a system. There was no statistically significant difference in recall rates (about 11% for both groups) or breast cancer detection rates (about 3.5 cases detected per 1000 screening mammograms). When the researchers looked at results from only the seven highest-volume radiologists at the center, results were very similar.

"…[A]ctually observed changes in our practice were substantially lower than expected," the authors write. "This is not to say that the use of computer-aided detection would not be beneficial or cost-effective in other practices. Rather, we suggest that, at its current level of performance, computer-aided detection may not improve mammography recall or breast cancer detection rates … in academic practices similar to ours that employ specialists for interpreting screening mammograms."

In an editorial, Joann G. Elmore, M.D., of the University of Washington, Seattle, and Patricia A. Carney, Ph.D., of Dartmouth Medical School, note that there are a few limitations to the study, including the fact that the data did not take into account the characteristics of the women in the study and that there was limited follow-up information on the women diagnosed with breast cancer. Nevertheless, they say, the study underscores the importance of ongoing studies of new technologies. "As more studies on computer-aided detection in the community setting are published and as computer-aided detection systems improve, we hope that women's lives can ultimately be saved by this new technology," they write.

###

Contacts:

  • Claire Collins, University of Pittsburgh Medical Center, 412-647-3555, CollCX@upmc.edu
  • Editorial: Susan Gregg-Hanson, Harborview Medical Center, 206-731-4097, sghanson@u.washington.edu

    Citations:

  • Gur D, Sumkin JH, Rockette HE, Ganott M, Hakim C, Hardesty L, et al. Changes in breast cancer detection and mammography recall rates after the introduction of a computer-aided detection system. J Natl Cancer Inst 2004;96:185–90.
  • Elmore JG, Carney PA. Computer-aided detection of breast cancer: Has promise outstripped performance? J Natl Cancer Inst 2004;96:162–3.

    Note: The Journal of the National Cancer Institute is published by Oxford University Press and is not affiliated with the National Cancer Institute. Attribution to the Journal of the National Cancer Institute is requested in all news coverage. Visit the Journal online at http://jncicancerspectrum.oupjournals.org/.


  • Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.