Since new consensus guidelines on screening mammography have just been promulgated by the Preventive Services Task Force (a panel of experts convened by the Agency for Healthcare Research and Quality, AHRQ, of HHS), there has been a firestorm of controversy. The new guidelines, as you probably know, no longer recommend routine screening of women in their 40s, scale back the recommendation for women 50 and older to once every two years, and also do not endorse the use of breast self-examinations. As news coverage tells us, many physicians and advocates are strongly in disagreement, and even portray these recommendations as showing contempt for women's lives.
People have an intuition that screening must be good, because it allows cancer to be found early, and that "saving lives" is worth even a very high cost. Here I will try to offer people a deeper understanding of these issues.
The first thing you need to know is that while it was once believed that all cancerous lesions would ultimately progress to clinically important disease, it has now become clear that this is not the case. Many of the tumors found through screening mammography are what are called Ductal Carcinoma in Situ, DCIS -- abnormal cells that are not metastatic and have not invaded beyond the inner walls of the milk duct. In fact, not all of these go on to produce disease, and new evidence shows that apparently, some of them spontaneously regress -- disappear without treatment. Yet, if they are found, the woman inevitably undergoes surgery, and often radiation and chemotherapy as well. You also need to know that some women who do develop breast cancer that might eventually kill them will die of something else first, in which case treating their breast cancer will only do harm. Detecting cancers that will never actually harm a woman, and treating her for them, is called overdiagnosis -- yes, the diagnosis is accurate, but it is not beneficial to make it. This is a potential harm of mass screening.
Another kind of harm is from false positives -- lesions which appear on the mammogram but are found not to be cancerous on biopsy. These also cause anxiety and pain. And, of course, mass screening of the population is very expensive -- it's money that could be used for other beneficial purposes, and money is not infinite.
So, you have to do the math. Just saying that somebody's life might be saved by screening women in their forties isn't enough reason to do it -- the costs may well outweigh the benefits. Furthermore, we don't actually know, in most cases, that any particular woman's life has been saved. Just because a purportedly cancerous lesion was found and removed doesn't actually lead to that conclusion. And remember, nobody's life can actually be saved -- we are all going to die, it is only a question of when and how. Since cancer surgery and treatment can have a strong negative impact on quality of life, it is not necessarily worth it even if it does extend life by a relatively small amount.
So how do the numbers work out? One place to start is with a discussion by Norton Hadler.
Let me tell you about the women of Malmö, Sweden. In 1976, half of the women between the ages of 44 and 68 were randomly selected and invited to participate in mammographic screening every 18 to 24 months for a decade. All the women -- both the screened and unscreened -- were followed for another 15 years after that decade. Over these 25 years, 9,279 of the 21,000 in the invited group, and 9,514 of the 21,000 in the control group, died of various causes.
In total, 1,320 women in the invited group were diagnosed with breast cancer, and 212 died from the disease. Breast cancer was detected in 1,205 women in the study's control group who did not get routine screening, and 274 of those women died from their breast cancer. The difference between the two groups is not considered statistically significant. So whether or not women received screening mammography did not seem to affect the chances of their doctor detecting a breast cancer, regardless of their age.
Other trials have had similar results. Note that this means that 21,000 women had to undergo routine screening to result in 62 fewer deaths from breast cancer, even you don't want to accept that this difference might be due to chance.
Other trials have had similar results. The bottom line is that if 2,000 women are regularly screened for breast cancer for 10 years, precisely 1 will benefit by avoiding dying of breast cancer. (Of course she will still die ultimately, of something else.) On the other hand, 10 women who would not have died from breast cancer will undergo breast cancer surgery, and possibly radiotherapy or chemotherapy, according to calculations by Gotzsche, et al. BMJ, February 21, 2009.
Unfortunately, most peer reviewed literature is off-limits to non-subscribers (a situation I decry) but here's the abstract of an important JAMA article. Let me summarize it for you.
When you start widespread screening, you will, obviously, have a sudden jump in the incidence rate of cancer, because you're detecting more of it. What you want to happen is that after time, the incidence of serious disease and death goes down, because early detection has prevented many cancers from becoming serious.
What has actually happened since widespread mammographic screening was intorduced is that the incidence of breast cancer has doubled; but the incidence of serious disease and death has declined very little, and perhaps not at all due to screening. Periodic screening may be good at detecting early stage lesions that will never develop into meaningful disease before the person dies of something else; but tend not to find more dangerous cancers early enough to make much of a difference. And that appears to be what is happening. A large burden of morbidity and treatment for innocuous lesions that would never have caused a problem, in exchange for a very small reduction in the rate of serious disease and death, if any. In fact, new calculations, as I said earlier, seem to imply that some lesions actually disappear on their own.
Mass screening would be much more useful if we could distinguish the dangerous lesions from ones that just bear watching, but as of now, we really can't. The shameful truth is, however, that without having good evidence that screening really does save lives or that the benefits outweigh the costs, the ACS and disease-specific organizations have undertaken mass campaigns over decades to try to get everyone to be screened. Of course there is money to be made at every stage, from radiography to lab to surgery and chemotherapy. Is that why they jumped the gun on the evidence? Who would think such a thing.
So I ask people not to be offended by these new guidelines. In fact, even every two years for women 50 and older may be too much. There is a strong case to be made.
A final word, however, which is very important! We are talking here about mass screening of the general population. If you have a sister or a mother who was diagnosed with breast cancer, particularly at a relatively young age, your risk is higher and mammography makes a lot more sense for you. Never having had children, or only having a child late in life, is also a risk factor. You should consider your own circumstances and your own personal feelings about this, and make an informed decision. I am not giving anyone medical advice. I am just presenting some facts for your consideration.
But it's not as simple as many people seem to think.