When President Bill Clinton proclaimed the first National Mammography Day in 1993, no one would have believed what we are facing today in 2016 – the near-constant bashing of this important screening tool. In 1993, there was only one way to detect breast cancer, and mammography was on a roll. Ultrasound was emerging as a diagnostic tool, but few considered its potential in screening. As for breast MRI, it was still 10 years away from introduction to routine clinical practice.
And now, today, with multi-modality imaging offering a near-guarantee of early detection, we have an ever-increasing mob of anti-screening activists crying: “Foul!” “You’re overdiagnosing breast cancer.” “You’re doing permanent psychological harm with your unnecessary biopsies.” “You’re only going to make the problem of overdiagnosis worse if you start adding ultrasound and MRI in your screening strategies.”
My contention is this: mammograms have been oversold from the earliest days with regard to sensitivity. Why? Not out of intent to deceive, but through the simple fact that there was no way to know how many cancers were missed. Think about it? How can you possibly know the miss rate when there’s no back-up method of imaging to cross-check your ability to detect cancer. With only one form of imaging, the only way to guess at sensitivity was to count cancers as they emerged after a negative mammogram. But what should be the interval that translates to a “miss?” 6 months? 1 year? 2 years? This totally arbitrary approach to identifying missed cancers has now been replaced with multi-modality imaging where women undergo ultrasound and/or MRI on the same day as the mammogram.
The results have been sobering, as we come to an inescapable conclusion, not with sophisticated statistics, but with grade school mathematics. If 10 cancers are discovered by mammograms, then an additional 10 are found by a second form of imaging, then mammograms only detected 50%. As it turned out, mammograms can detect 90-95% of cancers in fatty replaced breasts (only 10% of women), but the sensitivity plummets below 50% as background density increases. Cancer can hide anywhere on a mammogram where there is a white patch.
There is a powerful implication to this low sensitivity that is escaping those who will not acknowledge the painful fact of 50% sensitivity overall – mortality reductions were being demonstrated in the historic mammography trials of the 1970s and 1980s with a screening tool that missed as many cancers as it found. Now that multi-modality imaging can find the other half, imagine what we can do to lower the mortality of breast cancer! With 3-D tomosynthesis, ultrasound, contrast-enhanced mammography, MRI (and its kissing cousin Molecular Breast Imaging), we can find virtually all breast cancers at an early stage. Mammography might be marginally effective in the eyes of the critics, but “early detection” is more powerful than we ever imagined.
Amazingly, we do not need any major breakthroughs in imaging technology. We have what we need. The problem is that we are unable to use these new technologies efficiently, thus provoking the condemnation of the bean counters. Risk-based screening has been proposed as “precision medicine,” but this approach is doomed, given that you exclude the majority of eventual breast cancer patients right off the bat, and secondly, cancer yields are marginally cost-effective even when screening the patients at highest risk for breast cancer. Lifetime risks are a poor surrogate for what’s actually in the breast on screening day, and short-term risks are not much better. To me, the answer has been obvious for a long time. We need a blood test for the detection of breast cancer that tells us when to recommend adjunct imaging if mammograms are negative. Provista Diagnostics appears to have the lead in that department, and we’re in the process of confirming their test (Videssa™) in the screening setting. Stay tuned.