News Release

Radiologists’ interpretation of mammograms varies widely

Peer-Reviewed Publication

Journal of the National Cancer Institute

The interpretation of mammograms varies widely among radiologists practicing in a community setting, according to a new study in the September 18 issue of the Journal of the National Cancer Institute. Younger, more recently trained radiologists had two to four times more false-positive interpretations than older radiologists, the study found.

These findings, culled from the real-world setting, bolster evidence from earlier studies that examined radiologist variability using test sets of mammograms. In this study, Joann G. Elmore, M.D., of the University of Washington School of Medicine, Harborview Medical Center, and her coworkers examined results from 24 community radiologists' interpretations of 8,734 screening mammograms from 2,169 women over an eight-and-a-half year period.

The authors found wide variation in how frequently different radiologists noted masses, calcifications, and other suspicious lesions. For example, one radiologist did not observe any calcifications, while another radiologist noted calcifications in more than 20% of the films read. The radiologists also varied widely in their diagnostic interpretations and recommendations for additional screens and biopsies.

The rate of false-positive readings among the radiologists ranged from 2.6% to 15.9%. However, after adjustment for differences in patient, radiologist, and testing characteristics, the range of false-positive rates narrowed to 3.5% to 7.9%. Women who were younger, were premenopausal, were using hormone replacement therapy at the time of the mammogram, had a family history of breast cancer, or had had a previous biopsy were more likely to have a false-positive result.

A false-positive result was also more likely for women who had mammograms in the 1990s than women who had mammograms in the 1980s. "This increase in false-positive rates may be related to the fear of malpractice litigation, given the prominence in North America of malpractice litigation for delayed detection of breast cancer," the authors explain.

The variability in false-positive rates could also be accounted for by the radiologist's age, gender, and number of years since graduating from medical school, the authors found. Younger, more recently trained radiologists had two to four times the number of false-positive readings than radiologists who graduated more than 15 years prior to the study.

In a related editorial, M. Robyn Andersen, Ph.D., of the Fred Hutchinson Cancer Research Center in Seattle, and her colleagues say that although the current study suggests that radiologist training is an important factor in determining the rate of false positive readings, they note that the quality of mammographic interpretation cannot be addressed without also examining the number of cancers missed. "Knowing where to trade off an increase in sensitivity for a decrease in specificity is not a matter of formula," they say. "It needs to be carefully considered in a policy context."

Andersen and her colleagues point out that double readings have improved both sensitivity and specificity of mammographic interpretations in many countries. They say that public policy in the United States needs both a focus on training and an immediate re-examination of double reading as a national policy.

###

Contact: Susan Gregg-Hanson, Harborview Medical Center, 206-731-4097; fax: 206-731-8605, sghanson@u.washington.edu

Editorial: Kristen Woodward, Fred Hutchinson Cancer Research Center, 206-667-5095; fax: 206-667-7005, kwoodwar@fhcrc.org

Elmore J, Miglioretti D, Reisch L, Barton M, Kreuter W, Christiansen C, et al. Screening mammograms by community radiologists: variability in false-positive rates. J Natl Cancer Inst 2002;94:1373–80.

Editorial: Kessler L, Andersen R, Etzioni R. Much ado about mammography. J Natl Cancer Inst 2002;94:1346–7.

Attribution to the Journal of the National Cancer Institute is requested in all news coverage.


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.