Computer-Aided Mammograms Fail to Catch More Breast Cancersby and
Technology adds $400 million to annual U.S. health care bill
Study finds women get no benefit from Hologic, Icad equipment
Breast cancer screening performed with computer-aided technology from companies like Hologic Inc. and Icad Inc., designed to boost the accuracy of mammograms, failed to find more tumors, according to a study that suggests its widespread use should be curtailed.
The technology, approved in the U.S. in 1998, is used in 90 percent of mammograms performed each year at a cost of more than $400 million, according to lead researcher Constance Lehman, director of breast imaging and co-director of the Avon Foundation Comprehensive Breast Evaluation Center at Massachusetts General Hospital in Boston. That should end, she said.
“There is absolutely no question from our research that computer-aided detection offers no benefit to catching cancers that otherwise would have been missed or improving performance,” Lehman said. “We should certainly stop charging for it. We want to use our health-care dollars on technology that will improve the health of our patients.”
Mammograms have been the focus of controversy since 2009, when the U.S. Preventive Services Task Force recommended that most women under 50 forgo breast cancer screening and for women 50 and older to do it every other year instead of annually. The guidelines were intended to reduce the risk of false positive findings and unnecessary biopsies. The American Cancer Society, however, says the tests should be performed each year starting at 40. The current study shows that regardless of how often mammograms are performed, computer-aided detection doesn’t improve their accuracy.
The study examined mammograms from 323,973 women who underwent digital screening with or without the computerized assistance, which highlights potential problems for radiologists to review after an initial analysis. Cancer was detected in the same proportion of women regardless of which method was used, with similar accuracy in finding actual tumors and ruling out false positives.
Representatives of Hologic and Icad had no immediate comment on the study. Hologic gets about 37 percent of its sales from breast health products, while Icad relies entirely on revenue from medical imaging systems.
Insurance companies pay more than $20 per mammogram for use of the technology, and the U.S. government’s Medicare program pays about $7, the study found. Health plans constantly examine the safety and effectiveness of medical care when developing their coverage plans, said Clare Krusing, a spokeswoman for America’s Health Insurance Plans, an industry group.
Aetna Inc. will review the study and take it into consideration when deciding future coverage plans, spokeswoman Cynthia Michener said. Cigna Corp. spokesman Matt Asensio declined to comment.
The findings back up two previous studies that generated similar results but were criticized because they involved only older women or older technology that was read by less experienced radiologists. The present study, published in JAMA Internal Medicine and funded in part by the National Cancer Institute, is one of the largest of the technology and was designed to address those shortcomings, Lehman said.
The findings won’t change recommendations from the American Congress of Obstetricians and Gynecologists, which establishes mammography guidelines that are followed by many insurance companies. While doctors recommend women get regular mammograms, the type of imaging used for the screening and any computerized assistance is done at the discretion of the radiologists, said Christopher Zahn, vice president for practice at the organization.
Joshua Fenton, a family doctor who conducts research on the effectiveness of cancer screening at the Center for Healthcare Policy and Research of the University of California at Davis, said the U.S. should learn from the experience with computer-aided detection in mammography. Investment in new technology should be done only after its benefits are proven and its cost is justifiable, he said.
“Payments for ineffective services like CAD combine to bloat our health-care economy,” he wrote in an editorial that accompanied the study. “If we could curtail use of many similarly ineffective tests and interventions, we could significantly reduce U.S. health-care expenditures while augmenting resources for effective care or well-designed studies of promising innovations.”