According to a paper presented at the December 2013 annual meeting of the Radiological Society of North America, breast tomosynthesis (a limited CT scan of the breast) is a better test than the digital mammogram because it is a more sensitive test that finds breast cancers (22% more), is it associated with a lower recall rate because of false positive tests (15%fewer), and has a higher positive predictive value (46% better).
Sounds impressive, right? However, are the facts being presented honestly, and what about the downside of doubling the dosage of radiation compared to the digital mammogram? The truth of the matter is that the benefits have been exaggerated by playing a game with statistics. Want to know about the real numbers? Here you are!
Yes, there's a 15% lower recall rate, but the recall rate drops from 10.4% to 8.8% for a whopping 1.6% gain...not nearly as impressive as a 15% improved recall rate. What about the increase of cancer detection ob 22%. Well, that turns out to mean that the cancer detection rate increased from 4.28% to 5.25%. That's not even a full 1%! Oh, and what about the positive predictive rate increase of 46%? That turns out to mean an increase from 4.1% to 6.0%, which is a 1.9% improvement.
How impressed are you now?
They also managed to forget to let us know that there's twice as much radiation from tomosynthesis than from a digital mammogram. That means there's a doubling of the number of cases of breast cancer caused by the radiation of the test and twice as many deaths. Now, I'm playing the statistical game the authors of this study are. The real numbers are there's one breast cancer caused by every 1000 mammograms, so there'd be two caused by tomosynthesis. It takes 9000 mammograms for one person to die of a radiation induced mammogram, so there'd be two for every 9000 tomosynthesis tests and 8 more cases of breast cancer that would be treated for it. How are we doing so far?
Lastly, how many lives are saved by digital mammograms and from tomosynthesis. Let's stick with the 9000 number to make comparing easy. We save about one life for every 2000 digital mammograms, so we'd save a grand total of 4.5 lives. And, how about for tomosynthesis? They'd do at the best 22% better, considering that all the cancers they detected that digital mammograms missed would be saved...that'd be the massive number of one more saved life!
