Computer-aided detection (CAD) does not improve the accuracy of mammograms, a new study suggests. Most mammograms in the United States use CAD. Researchers analyzed 1.6 million mammograms performed on over 680,000 women in 7 states from 1998 to 2006. CAD did not improve the detection of invasive breast cancer. It also did not increase the chance of detecting breast cancer at an earlier stage. However, the false-positive rate increased from 8.1% to 8.6% with CAD, meaning women had a greater chance of being called back for further testing. The study was published in the Journal of the National Cancer Institute. HealthDay News wrote about it July 27.
By Robert H. Shmerling, M.D.
Harvard Medical School
What Is the Doctor's Reaction?
Technology is responsible for some of the most dramatic advances in modern medicine. Examples include CT scans and MRIs that allow doctors to safely see inside a person to detect disease (or to rule it out).
But, sometimes technology just gets in the way. That may be the case for a technique that is commonly used during mammography: "computer-aided detection."
Most mammograms use computer-aided detection. With this technique, a computer analyzes an image and alerts the radiologist to suspicious areas. The computer is supposed to reduce the chances that the radiologist will overlook an abnormality. Since abnormal areas on mammograms may indicate the presence of breast cancer, computer-aided detection should save lives.
However, a study just published in the Journal of the National Cancer Institute suggests that computer-aided detection may not be delivering on its promise.
In this new study, researchers analyzed mammograms performed on more than 680,000 women in 7 states. The accuracy of the mammograms was compared before and after the introduction of computer-aided detection. The study found that:
The higher false-positive rate may not sound like a big deal. But, considering the number of mammograms performed each year, thousands of women may undergo additional, unnecessary testing (including biopsies) due to the inaccuracy of computer-aided detection. In addition, computer-aided detection adds about $30 million to the cost of mammograms each year in this country.
Should we eliminate the routine use of computer-aided detection? It sure sounds that way.
What Changes Can I Make Now?
If you're a woman who is not having mammograms regularly, talk to your doctor about whether you should be. There is controversy about whether women at average risk should begin at age 40 or 50. The ideal frequency is also unclear. Just last week, the American College of Obstetricians and Gynecologists announced new guidelines recommending that mammography screening be offered annually to women beginning at age 40.
If you have risk factors for breast cancer, it may be even more important to have mammograms regularly. Risk factors include:
It's a good idea to have mammograms at the same reputable facility rather than at different sites. Having them at one site makes it easier for your doctors to compare results.
Know what to expect when you're having a mammogram. Up to 20% of women without breast cancer will require a biopsy for an abnormal mammogram if the test is performed as recommended each year for 10 years. Knowing this in advance can be helpful if you are ever told your mammogram is abnormal "but probably nothing to worry about."
You may be able to reduce the chances you'll have a false positive test result by not using lotions, talcum powder or deodorant under your arms or on your breasts. They may cause abnormalities on a mammogram that mimic breast cancer.
If you don't hear about the results of your mammogram, don't assume it was normal. Call your doctor to confirm that he or she received the results and that they are not abnormal.
If your mammogram is not normal, get the recommended follow-up tests. Your doctor may recommend another mammogram in a few months, an ultrasound or a biopsy. These additional tests can help sort out whether an abnormality is a false positive result or whether breast cancer may be present.
Review other screening strategies with your doctor, including self-breast examination and examinations by your doctor during office visits.
Genetic testing may be appropriate if you have a strong family history of breast cancer. Some women who carry high risk genes choose to have treatment to prevent breast cancer.
What Can I Expect Looking to the Future?
I think you'll see continued debate about the value of mammograms. This will include questions about when to start having them and how often they should be repeated. One recent analysis suggested changing the recommendations for mammography from being age-based to being risk-based. That is, it may be reasonable to have a mammogram every three years if you're at low risk and have had normal mammograms in the past. But, yearly mammograms starting at age 40 may make sense if you have a strong family history of breast cancer.
Look for researchers to study alternative screening schedules and different imaging techniques, not just on the detection of cancer but on quality of life and longevity as well.
Finally, if the results of this new study are confirmed, you can expect the popularity of computer-aided detection to wane over time.
© 2013 Aetna Inc.