Mammography Recall Guideline Needs Second Look

Skilled radiologists with high recall rates more likely to detect breast cancer

MONDAY, March 8, 2004 (HealthDayNews) -- Radiologists who exceed the recommended 10 percent recall rate for screening mammograms may be catching more cancers, says a new study, suggesting the guideline may need to be reconsidered.

The study, which appears in the April 15 issue of Cancer, found a group of highly skilled radiologists had recall rates that varied from 7.7 percent up to 17.2 percent. However, tumor detection rates also went up as recall rates increased.

A mammography recall occurs when a radiologist sees something suspicious on a mammogram and recalls the patient for further testing and evaluation. Recall rates in the United States have traditionally been higher than in other parts of the world, and some studies have suggested higher recall rates don't always translate into higher detection.

Since every recall involves additional expense and probably additional anxiety for the woman involved, radiologists want to keep their recall rates as low as they can without compromising detection rates. A less than 10 percent recall rate, while not mandated, has been commonly accepted in the United States as a reasonable goal.

"In this field we have a recommended practice guideline to not recall more than 10 percent of the screening mammograms reviewed," says study author David Gur, a professor of radiology at Magee-Women's Hospital of the University of Pittsburgh Medical Center.

"The question is, is this practice guideline appropriate in groups of very well-trained mammographers? We found in a group of well-qualified radiologists [that] those who recall more than the recommended guidelines detected more cancers," he says.

According to the American Cancer Society, the reason it can be difficult to interpret mammography data is the mammogram image is much like a fingerprint -- no two mammograms are exactly alike.

In the current study, Gur and his colleagues collected data from three years of screening mammograms. Nearly 100,000 mammograms were interpreted by 10 different radiologists at Magee-Women's Hospital or their breast-imaging centers. Each radiologist performed more than 3,500 interpretations during that period. Some did as many as 16,000.

The study researchers found a wide variability in the recall and detection rates. The recall rates were as low as 7.7 percent and as high as 17.2 percent, with an average recall of about 11 percent. Breast cancer detection rates varied from 2.6 to 5.4 per 1,000 mammograms read.

"In our view, once you are well-trained, you should not artificially pressure radiologists to reduce recall rates," Gur says.

In an editorial in the same issue of the journal, Dr. Rachel Brem, director of the Breast Imaging and Interventional Center at George Washington University Medical Center, agrees with Gur's assessment. "Without this unnecessary pressure, the diagnosis of breast carcinoma can be optimized and patients therefore can be afforded the best chance for survival," she says.

"I don't think there should be a target number," Brem adds. "There are too many variables."

She says many factors can influence recall rates, such as the density of a woman's breast, whether a woman has had breast surgery, or if a woman brings her previous mammogram for comparison.

Dr. Julia Smith is an oncologist at the New York University Cancer Institute and the Bellevue Prevention Clinic in New York City. "It would be great if we had a 100 percent sensitive and specific test, but that doesn't exist," she says.

She adds that previous studies have shown most women would rather undergo additional testing that turns out negative than take a chance on missing detecting cancer in its early, and often most treatable, stage.

"Trying to second-guess the art of medicine is dangerous," says Smith. "The minute you put pressure on radiologists to adhere to a certain number of callbacks, you're asking them to eliminate the art of interpretation."

More information

To learn more about mammography, visit the National Library of Medicine or the American Cancer Society.

Related Stories

No stories found.
logo
www.healthday.com