Countercurrents: Is now the right time to pull the plug on mammography?

My views on this are well known, and I have much more information on my web-site under Breast screening, mammography, overdiagnosis.

Curr Oncol. 2019 Jun; 26(3): 162–163. Published online 2019 Jun 1. doi: 10.3747/co.26.4835PMCID: PMC6588072PMID: 31285658

S.A. Narod, MD^*Author informationCopyright and License informationDisclaimer

In their report of 10 December 2018, the Canadian Task Force on Preventive Health Care gave a lukewarm endorsement of screening mammography¹. After a review and meta-analysis of the literature, the Task Force estimated a 0.85 relative risk of dying of breast cancer for women who attended mammography compared with those who never did, with no differences by age group.

That is to say, regular mammography attenders faced nearly the same risk of dying of breast cancer as women who did not attend (85% to be precise), and in no subgroup did the level of protection exceeded 15%. Further, the evidence in favour of any benefit was judged to be very low (ages 50–59 years) or low (ages 60–69 years). The downgrading of the evidence for mammography came largely from a re-analysis of the Swedish quasi-randomized trials. An updated analysis of the Canadian National Breast Cancer Study showed that, although cancers detected at screening had a better prognosis than cancers detected in the control arm, no benefit was associated with screening in terms of breast cancer mortality at 30 years post-randomization².

A relative risk of 0.85 is derisory and in any other setting would be dismissed as nugatory. We would never think to adopt a new screening test that performed that badly. And it can hardly be claimed that such a small effect justifies early detection as the “key to survival,” as is often bruited about. The Task Force also had no appetite for either breast self-examination or clinical breast examination.

We have heard the meretricious arguments of the most vocal calumnists. They say that the Canadian National Breast Cancer Screening trial was biased—claiming widespread false randomization—but that was impossible³. Of greater interest are the several claims that mammography in “the real world” performs even better than reported in the Swedish trials.

An observation trial from Sweden pegged the benefit at a 60% reduction in deaths⁴, and a trial from Canada, at 40%⁵. The difference between the Task Force estimates and those estimates is not attributable to the “real world,” but to real bias. The Canadian and Swedish studies compared, at a population level, breast cancer mortality in women who attended mammography screening with those who did not attend. But women who attend mammography clinics are selected to have no cancer at study onset, whereas women in the comparison group might well have been diagnosed with breast cancer in the past⁶.

Consider a women who has breast cancer and who receives an invitation to attend screening in the mail. I am pretty sure she doesn’t go. But her death would be considered an event in the comparison group (non-attendees). Between 2004 and 2014, the U.S. Surveillance, Epidemiology, and End Results database captured 126,511 deaths from breast cancer. Of those 126,511 deaths, 88,322 (70%) were diagnosed before 2004.

If the Task Force believes the screening mammography benefit is 0.85, why bother to recommend it? Surely the reasonable thing is to abandon mammography and spend the money elsewhere. Instead, they offer the avuncular opinion that each woman should discuss the risks and benefits with her physician. That sounds like a compromise, but diligent physicians might be perplexed upon consulting the Canadian Medical Association Journal recommendations and finding the ball fired back into their court.

Part of the mission of the Task Force is to consider all the prevalent opinions and to separate those of disinterested parties from those of parties with something to gain. The fact that they are not practitioners of the art of mammography is what makes them credible. Unless a more forceful recommendation against mammography emerges, the taxpayer will continue to pay. I propose that, as is the case for other screening tests such as colonoscopy and prostate-specific antigen, it is reasonable that the woman who wishes to have a mammogram should pay for the test out-of-pocket. Neither of the former two screening tests for men are paid by ohip (the Ontario Health Insurance Program), but they are available at the patient’s choice (fecal occult blood testing is covered by the province).

The Task Force cites overdiagnosis and issues of sensitivity, specificity, and cost as harms and the reasons contributing to mammography failure. I don’t think that that case holds. If the screening paradigm were true, then finding some cancers early would surely do some good. I have come to a different conclusion about why mammography fails: that if a cancer is going to metastasize, it will do so before it is clinically (or radiologically) apparent, and that breast cancer does not metastasize in the interval between the mammogram and the mass^7–9. I propose that the cancer will have metastasized by the time of diagnosis—whether by imaging or palpation—but will not metastasize thereafter. Under the parallel model, the size of the cancer is a marker of aggressivity, and the bigger the cancer, the more likely it is that metastases are present at diagnosis. But the latter model does not imply that the chance of metastases can be lessened by advancing the date of detection. This is an area that we hope to explore further.