■ site map
June 2010 - December 2013
II - Mammography
1. Mammography story
What most people know about standard X-ray mammography is that it was trusted and widely embraced for decades as the lifesaving early breast cancer detection test. But, times change. What we learned during the past decade casts serious doubts on this image. In all likelihood, the benefits of mammography were exaggerated, and its risks downplayed or ignored.
The ball started rolling after a small 1999 Swedish study has found that, contrary to the expectations and beliefs, breast cancer mortality in large population of Swedish women was reduced only by a small fraction of what conventionally projected - and officially endorsed - figures were implying should have happened.
It prompted the Danish National Board of Health to ask The Nordic Cochrane Centre (NCC) in Denmark for an independent review. So the NCC researchers took on the task of thoroughly analyzing major breast cancer trials up to that date, with special attention paid to those from Sweden, where most of the large random controlled trials of mammography screening took place.
They found that most of the trials were methodologically flawed, thus unreliable. The two that passed the scrutiny, agreed with the small Swedish study in that they
did not show any appreciable mortality reduction from mammography screening.
The Danes' conclusion was mercilessly direct: the available reliable evidence
does not justify public screenings with X-ray mammography.
The news sent tsunami waves through the world of mammography: all the medical, health and governmental organizations recommending and practicing it, as well as many millions of women who were told exactly the opposite - time and again.
How is it possible that so many people religiously followed such a shaky concept for so long?
Here's a quick look at how it all unfolded, mainly based on the background paper for The Institute of Medicine report by B.H. Lerner.
Screening mammography: making of a myth
Not that there was no objections to the use of X-ray mammography for public screening programs before. As far back as 1970s, when it was at its beginnings, the opponents were pointing to the lack of evidence for both, its benefits and its possible negative long-term effects.
Bailar wrote that the possible benefits of X-ray mammographic screening "have received much more emphasis ... than its defects" (Mammography: a contrary view, 1976).
The advocates of X-ray screening, led by U.S. radiologists like Gershon-Cohen, Egan, Zuckerman and Strax, kept their course. After the first trial designed to evaluate X-ray mammography as a screening tool - Health Insurance Plan of Greater New York (HIP, 1971), proposed and implemented by Strax - showed earlier detection and produced significantly reduced breast cancer mortality numbers in the screened population, public X-ray screening for breast cancer was well on its way.
What wasn't known at the time is that the quality of mammography in the HIP trial was so poor (39% sensitivity), that not only more tumors were detected without it, by physical breast examination, but their average size was also smaller than of those detected by mammography. In addition, randomization of participating women had gross, never fully addressed or explained biases favoring positive numbers in the screened population (e.g. far more women with prior breast cancer excluded from the screened than from the control group).
So, as early results of the American Cancer Society (ACS) and National Cancer Institute's (NCI) large-scale trial (Breast Cancer Detection Demonstration Project, 1973-80) with some 270,000 women aged 35-74 seemingly confirmed the benefits, the two organizations went on with recommending X-ray mammography screening of asymptomatic women.
The critics, like NCI's own Bailar, were pointing out that such uncontrolled demonstration project
does not allow for meaningful conclusions
(in addition, the project also used termography and clinical breast examination in the screened population, further clouding the exact benefit due to X-ray mammography).
Bailar also contended that it shouldn't be assumed that every early detected, small abnormal breast tissue growth - or what appear as such on the mammographic image - will become breast cancer. It will take nearly three decades from then to this quite logical warning for it to be definitely documented as one of mammography's biggest negatives: overiagnosis. He also wondered why the effect of exposure to ionizing radiation - a known carcinogen - has been ignored.
After Bailar's views were published, they were given press coverage, and became publicly known. That prompted a series of investigations on all sides, from NCI establishing that HIP trial does not show benefit of screening for women below age 40-49 (Breslow et al. 1976; screened women in this age group actually had higher mortality in the trial, although not statistically significant), to examining the actual radiation dose delivered by mammography X-ray machines - which found that some machines delivered significantly higher radiation dose than assumed - and to publicly exposing initial reservations within the NCI with respect to the demonstration project.
In the bad need of winning public support for the screening, leaderships of ACS and NCI responded by correcting some project's aspects that could compromise it in the public eye, including modifying consent form for participating women to acknowledge potential risk, standardizing official radiation doses, and deciding to offer screening only to those women under 50 who were at high risk.
But the desire to confront breast cancer was already morphing into aggressive marketing. According to the ACS criteria, some 80% of all women would classify as "high risk" -
nearly eight times
higher a rate
NCI, who seemed to be responding to the legitimate concerns about screening more responsibly than the ACS, investigated the nature of 506 small abnormal growths detected during the project. Turned out, 66 of these 506 pathological specimens were benign growths; nevertheless, 53 of them resulted in some form of mastectomy (McDivitt, 1978).
That was indicating a very serious risk from screening: unnecessary treatment, including invasive treatment and disfigurment. Working group outside the NCI was appointed to re-examine these specimens, and found that only three of the 53 mastectomies, at the most, were unnecessary.
McDevitt stud by his results. Surprisingly - or not - the issue, which could decide the fate of public X-ray screening,
was brushed aside unresolved, and pretty much forgotten.
The 66 women were not even informed by the project about possible misdiagnosis and mistreatment; that "formality" was left up to the physicians who treated them. There were more important things to do.
Within the next few years, based in good part on the data from the demonstration project - as unreliable as it was by design - ACS went on to recommend baseline screening for women aged 35-39 (1980) and for women aged 40-49 every one or two years (1983), in addition to those over 50y of age.
In 1988, the NCI, along with the American College of Radiology and 11 other medical organizations, joined the recommendation for the 40-49 group. However, NCI withdrew its support for screening this age group in 1993, based on results from ongoing randomized clinical trials generally implying no sufficient benefit.
In 1997, a 13-member advisory panel within the NCI - which, ironically, was assembled because then-NCI-director Klausner believed in enacting screening recommendation for women in their forties - based on their view of the available data and 32 expert-testimonies, concluded that existing evidence does not support such screening recommendation.
The panel was furiously attacked by the pro-screening camp. Even the U.S. Senate got involved; with female voters on their mind, and with mammography establishment lobbying at their doorstep, the Senate unanimously passed suggestion to the NCI advisory board to reject the panel's decision. It worked. NCI opted to adopt the ACS policy recommending screening for women in their 40s at least once every two years.
But its old recommendation wasn't aggressive enough for ASC anymore; the same month, March 1997, it switched to a new, yearly screening recommendation for women aged 40-49; it was quickly followed by the American College of Radiology.
Some other organizations remained unconvinced as to the benefit of screening women below 50; one of them, USPSTF (United States Preventive Services Task Force), is similar to the denounced NCI panel in that it consists of experts whose particular expertise is outside (in this case) of the breast cancer field. The main purpose of such a panel is promoting objective, unbiased assessment. Obviously, there was still plenty of room left for biases elsewhere.
Oddly enough, these further victories of the pro-screening camp were taking place despite seemingly well documented ongoing research
seriously questioning the rosy picture of screening efficacy.
Let's take a closer look at the actual evidence of the efficacy of mammography screening, before the small Swedish study finally blew the whistle.