Computer-Aided Mammography May Not Increase Cancer Detection Rates
After ]]>skin cancer]]> , ]]>breast cancer]]> is the most common cancer diagnosed in women. And after lung cancer, it is the second highest cause of cancer deaths in women. By detecting cancer at an earlier stage, ]]>mammography screening]]> has the potential to improve prognosis and increase the chance of successful treatment.
Traditional mammograms require a radiologist to interpret an x-ray of the breast to determine whether diagnostic testing is necessary. In recent years, the National Cancer Institute has supported the development of technologies to improve the accuracy of mammography. One such technology, computer-aided mammography, uses computers to bring suspicious areas to the radiologist’s attention.
The Food and Drug Administration (FDA) has approved several of these computer-aided systems for clinical use. The FDA’s approval was based on evidence that suggested computer-assisted mammography could significantly increase cancer detection rates. But these studies were conducted in a laboratory setting, and clinical studies are a better indicator of the real-world benefits of computer-aided mammography.
A new study in the February 4, 2004 issue of the Journal of the National Cancer Institute found that computer-aided mammography was no more accurate than traditional mammography in detecting breast cancer in a clinical environment.
About the Study
This study included results from 115,571 mammograms, evaluated by 24 certified radiologists in between 2000 and 2002 at Magee-Womens Hospital in Pittsburgh, Pennsylvania. Computer-assisted mammography was introduced in the Magee-Womens Hospital in 2001, so about half (51.2%) of the mammograms were computer-assisted and the other half (48.8%) were traditional.
Mammograms were considered positive if the patient was recalled (instructed to come back for diagnostic testing) and negative if a standard one-year followup was recommended. In the women who were recalled, the researchers kept track of breast cancer diagnoses.
The researchers compared recall rate and cancer detection rate between women who had computer-aided versus traditional mammograms. They repeated this analysis for the seven highest volume radiologists, who read an excess of 8,000 mammograms each over the three-year study period.
As a result of mammogram screening, 13,171 (11%) of the women were recalled and 407 (0.3%) were diagnosed with breast cancer. There were no differences in recall or cancer detection rates between traditional and computer-assisted mammograms.
Similarly, in the mammograms performed by the seven high-volume radiologists, both techniques resulted in a recall rate of approximately 11% and a cancer detection rate of 0.3%.
While these findings are intriguing, they do have certain limitations. First, a small percentage of women recommended for recall never returned for diagnostic testing, which may mean the number of breast cancer cases was underestimated. Furthermore, the researchers found that more women were returning for subsequent routine examinations in the later years of the study. Subsequent examinations typically have a lower recall rate than initial examinations, so it’s possible that recall rate increased after the implementation of computer-aided mammography.
How Does This Affect You?
These findings suggest that computer-aided mammography may be no more beneficial than traditional mammography in some clinical settings. Unlike many screening tests, mammography involves a great deal of human interpretation. Obtaining a ]]>prostate specific antigen]]> (PSA) level, for example, requires no human interpretation; it is a simple blood test that measures the amount of a protein in the blood (an possible indicator for prostate cancer). Mammography, on the other hand, requires a radiologist to interpret the x-rays.
Failure to detect breast cancer is the leading cause of medical malpractice claims in the United States. Perhaps this is one of the reasons computer-aided detection was so quickly adopted by the medical community; computerized mammography could theoretically take some of the burden of breast cancer diagnosis out of the hands of doctors.
But, according to this study, computer-aided mammography may be no more beneficial than traditional mammography. Americans love technology, and often assume it is superior to human performance. Of course, this is not always true, and the added expense of new technology is not necessarily worth the price.
Still, you can bet that researches have no plans to give up. In the future, computers may totally replace humans in the detection of breast cancer and many other diseases. For now, however, relying on a well-trained, experienced human being to interpret your mammogram seems to be your best bet.
Breast Cancer Home Page
National Cancer Institute
Learn About Breast Cancer
American Cancer Society
Elmore JG, Carney PA. Computer-aided detection of breast cancer: has promise outstripped performance? Journal of the National Cancer Institute . 2004;96:162-163.
Gur D, Sumkin JH, Rockette HE, et al. Changes in breast cancer detection and mammography recall rates after the introduction of a computer-aided detection system. Journal of the National Cancer Institute . 2004;96:185-190.
Improving methods for breast cancer detection and diagnosis. National Cancer Institute website. Available at: http://cis.nci.nih.gov/fact/5_14.htm . Accessed February 3, 2004.
Screening mammograms: questions and answers. National Cancer Institute website. Available at: http://cis.nci.nih.gov/fact/5_28.htm . Accessed February 3, 2004.
Last reviewed Feb 5, 2004 by ]]>Richard Glickman-Simon, MD]]>
Please be aware that this information is provided to supplement the care provided by your physician. It is neither intended nor implied to be a substitute for professional medical advice. CALL YOUR HEALTHCARE PROVIDER IMMEDIATELY IF YOU THINK YOU MAY HAVE A MEDICAL EMERGENCY. Always seek the advice of your physician or other qualified health provider prior to starting any new treatment or with any questions you may have regarding a medical condition.
Copyright © 2007 EBSCO Publishing All rights reserved.