In today’s world of gizmos and gadgets, we tend to trust our technology to make our lives easier and more error-free. However, a new study on computer software that is widely used to locate cancerous regions in mammograms shows that the technology may be doing more harm than good. The study shows that the software has not only failed to make breast cancer detection more accurate, but that it has also increased the risk that a patient will be erroneously told that her mammogram shows cancer cells.
The study, which may be found online in the “Journal of the National Cancer Institute,” bases its conclusions on the analysis of about 1.6 million mammograms taken at radiology facilities in several states between 1998 and 2006. The same researchers who have published this study were also the authors of a similar one conducted in 2007, when they first became skeptical of the efficacy of this mammogram technology, known as computer-aided detection (CAD).
The problem comes from the fact that mammograms themselves are not always as trustworthy as we would rather them be. According to the National Cancer Institute, the average mammogram may overlook up to 20 percent of breast cancers. However, studies show that having multiple radiologists examine a mammogram may improve the accuracy of its result. This is where CAD has been most often deployed – as a second set of eyes to look at a mammogram. Thus, CAD is being used as if it is as good as a radiologist’s eye, and it simply may not be the case that it is.