Early detection from mammography screening: two new studies provide more information to evaluate harms vs. benefits

June 18, 2010

Women are told that mammography screening saves lives, but evidence that it decreases breast cancer deaths is conflicting and continues to be questioned by some scientists, policy-makers and members of the public. One recent study highlights that mammography increases the detection of low-risk cancers that may never become life-threatening, while another highlights the potential harms of rapid adoption of new detection technologies that have little proven effectiveness.

A study presented at the annual meeting of the American Society of Clinical Oncology (ASCO) in Chicago found that screening mammography is much more likely to detect low-risk breast cancers with a good prognosis that otherwise would likely not have become clinically significant. The study compared the proportion of low-risk cancers, as categorized by a prognostic 70-gene signature, among a cohort of 222 women aged 49-69 years diagnosed with breast cancer in the pre-screening era (prior to 1992) with that of 205 women aged 49-60 years diagnosed in the screening era.

Among the cohort of women diagnosed with breast cancer in the prescreening era, 46% had low-risk cancers with good prognosis. Among the women diagnosed in the screening era, 58% of the cancers were low risk. And among the 94 women with screen-detected breast cancer, 67% had a low-risk cancer. Interestingly, there was no statistical difference seen in the proportion of biologically low-risk cancers diagnosed among women too young for routine screening, those younger than 40 years between the pre-screening era (17 of 68 cancers) and the screening era (17 of 56 cancers). Researchers said these findings confirm that screening preferentially detects low-risk tumors, highlighting the need for better tools to predict which breast cancers will become life-threatening, and which will not.

In a separate study, published in the Archives of Internal Medicine, researchers at UC Davis Cancer Center found that the use of computer-aided detection (CAD) systems to interpret mammograms increased from 5% in 2001 to 27% in 2003 after Congress required Medicare to cover CAD, despite the fact that prior studies showed use of CAD does not necessarily help radiologists find breast tumors more accurately. The use of CAD has been found to result in 32% more women recalled for additional testing and 20% more women undergoing a breast biopsy with no clear improvement in the detection of breast cancer. The authors highlight this as an example of “expensive technology widely adopted in clinical practice before proven effective.” They go on to state that although most consumers and doctors presume that newer technologies are better than existing ones, we as a society must demand scientific evidence that illustrates the clinical benefits, important associated harms, and whether the benefits outweigh those harms. Health care coverage should guarantee access to care that is proven effective. Care for which the effectiveness is still uncertain could still be covered, but as part of a quality clinical trial or otherwise appropriately contributing to the evidence base through clinical research.