Breast screening: a balancing act
For almost 25 years, women in the UK aged over 50 have been regularly invited for breast screening. However, in recent years the screening programme has come under attack, so much so the government is undertaking a national review. Abi Millar talks to Dr Daniel Kopans and Dr Robin Wilson to investigate the value of screening initiatives, and to see if the physical and psychological consequences are unjustifiably damaging.
In October 2011, Professor Sir Mike Richards announced that the NHS Breast Screening Programme (NHSBSP) would undergo an independent review. Ongoing at the time of writing, the aim is not merely to revise the programme's guidelines; rather, it will assess all the underlying evidence, asking whether the benefits of screening really do outweigh the potential harms.
The review is causing a stir, not least because it raises the disquieting prospect that breast screening, in and of itself, is open to question. Since its inception in 1986, when the Forrest report declared 'screening by mammography can lead to prolongation of life in women over 50', the NHSBSP has screened millions of women, and diagnosed thousands of cancers.
Unlike diagnostic mammography, which occurs when a woman presents with symptoms, screening mammography targets populations as a whole. The aim is to diagnose latent cancers which - not yet symptomatic - are small and responsive to treatment.
According to the NHS, around 1,400 women a year can expect to have their lives saved as a corollary of the programme. Of every 400 women that attend screening for a decade, one will be treated successfully for a cancer that otherwise would have killed her.
A misled public?
We might think these survival rates are justification enough, and public opinion is certainly on the side of the screening programme. In fact, there is evidence to suggest that most people take an overly rosy view of proceedings. A Swiss study from 2001 showed that over half of women overestimated its benefits, and of the women invited for screening, around three quarters generally attend.
This is not surprising when you consider the timbre of the NHS leaflet. "If you make the literature fully comprehensive, it turns out to be ten to 12 pages long, and full of detailed epidemiological facts and figures," says Dr Robin Wilson, consultant radiologist at the Royal Marsden in London. "So you do tend to oversimplify. There has been a tendency to say that screening is wonderful without explaining the side effects in detail."
Nonetheless, behind the certitude of the government guidelines lies a complex and controversial tale. This is a story of sparring statisticians, with wildly divergent takes on the data - a wrangle this latest review hopes to mediate.
In 2000, Danish scientist Dr Peter Gøtzsche published a paper in The Lancet, claiming there was no convincing evidence in favour of screening. It did not reduce breast cancer mortality, he said, and the randomised controlled trials (RCTs) that formed its basis were unreliable.
A year later, this paper was reprinted, along with a Cochrane review. Gøtzsche and colleagues, constituting the Nordic branch of the Cochrane Collaboration, analysed data from seven breast screening trials, leading to a damning conclusion: the number of lives saved was overstated. Screening entailed over-aggressive treatments and more mastectomies, with few benefits to women of any age.
Although the IARC, an international working group, summarily rejected Gøtzsche's analysis, the ruckus has rumbled on. Recently, Gøtzsche penned a book, Mammography Screening: Truth, Lies and Controversy, which presents his divisive findings in laypeople's terms. And laypeople, it seems, to the chagrin of the NHS, are starting to take note.
"The critics are quite vocal, and therefore they get a lot of press," says Wilson. "I think the Department of Health sees this as having a negative effect on the screening programme. This review has been set up to come to a decision about the adequacy, or not, of the programme as it is."
Of course, the NHSBSP has never been static, with its guidelines and targets perennially subject to scrutiny. Take the ages of the women screened. Initially, women between 50 and 64 were invited for triennial mammographic screening, but the age range was then extended up to 70. At present, there is a pilot scheme underway to stretch the parameters from 47-73.
Despite such flexibility, the argument set out by the Forrest report - that 'there is a convincing case on clinical grounds for the screening of symptomless women' - is only now being seriously contested.
The review panel members have been selected for their impartiality. Comprising four experts in epidemiology or oncology, with Professor Sir Michael G Marmot as the chair, none of its members have published on breast cancer screening before. There is no ostensible agenda, no conflict of interests. The charity Cancer Research UK is also involved, providing support but wielding no influence over what is said.
All sides of the argument are being considered. While the bulk of the scientific community does defend the breast screening programme - Professor Richards has cited the figure as 'well over 90%' - the review should determine if the naysayers have anything more up their sleeve than a statistical sleight of hand.
One man staunchly in the pro-screening camp is US scientist Dr Daniel Kopans, one of the heavyweights of the mammography world. A professor of radiology at Harvard Medical School, he has spoken damningly about the Nordic Cochrane Centre.
"The so-called Nordic Cochrane Group and its collaborators are pretty much the only opposition to mammography screening," he says. "Recently, in a letter to the editor of The Lancet, over 40 experts in breast healthcare were severely critical of this group, and its anti-screening agenda. Gøtzsche's analysis triggered numerous re-reviews, and all repudiated his assertions."
Skewed screening data
As Kopans sees it, Gøtzsche's work is palpably anti-scientific. Of the seven trials included in the Nordic Cochrane review, five showed clear benefits to screening. Coincidentally or not, Gøtzsche discarded these five because he claimed they weren't properly performed. "You cannot simply throw out trials that you don't like," fumes Kopans. "His reasons were not supported by any other reviews."
Although Kopans says he would lend his unqualified support to any scientific review, he worries that the UK study might echo its 'corrupted' stateside equivalent. In 2009, the US Preventative Services Task Force (USPSTF) conducted an assessment of US mammography screening. Kopans believes this was influenced by the findings of the Nordic Cochrane Centre. "I think the panels were trying to do their best," he concedes, "but they were guided in their deliberations by others whose agenda was to limit access to screening."
As Kopans sees it, this task force picked and chose their data, as well as eschewing epidemiological information in favour of computer modelling.
"The results of models are determined by the assumptions built into the algorithms," says Kopans. "Consciously or otherwise, models can be designed to produce whatever outcome the modeller would like."
In the event, the USPSTF came to a fairly radical conclusion: to drop the previous US advice that all women over 40 should receive a mammogram every one to two years, and instead limit breast screening to the over 50s.
While the US Health and Human Services dictated these new guidelines be ignored, there are still many physicians in the US who follow them unquestioningly - "out of ignorance," Kopans says. He adds that there is absolutely no data to support the age of 50 as a threshold for screening.
Troublingly, Kopans believes that the UK review may not be as disinterested as one would hope. "I am told that the panel members are highly qualified, but Doug Altman is a member of the Nordic Cochrane Centre," he points out. "Clearly the Centre's work will be reviewed. Why would you include someone on the panel who is directly involved with it?"
The bigger picture
While Kopans has devoted much of his career to expounding the benefits of breast screening, he is no blinkered polemicist. Even the most ardent supporters of screening are fully aware of the downsides. You don't need statistical training to see that through scanning populations indiscriminately, any screening programme will inevitably inflict some damage.
It is important to bear in mind that most of the women screened are healthy. Of the 2.1 million screened by the NHSBSP in 2009/10, fewer than 17,000 had a detectable breast tumour. The rest suffer unnecessary anxiety, with the best possible outcome being a nail-biting wait for their results.
A certain unlucky proportion are subject to false positives: while healthy, their results are ambiguous, and they are called back for a second mammogram. The psychological consequences here are obvious.
Harder by far to justify is overdiagnosis. Defined as picking up indolent cancers that would never have progressed, this is an issue that has long concerned the NHS.
"There are some low-grade in-situ tumours, with no life-threatening potential," explains Wilson. "Some would say that those types of cancer only have a 30% risk of progressing over the next 30 years. The problem is that at the moment we don't know which of these cancers will progress and which ones won't, and therefore we have to advise treatment for all of them. There are very few women who would be willing to take the risk."
The numbers affected by this issue are open to interpretation, with Gøtzsche stating that overdiagnosis accounts for 30-50% of the cancers detected, and most studies putting the figure at under 10%.
Still, it is not straying beyond the realm of orthodoxy to say that, for every woman who has her life saved by screening, another woman has to undergo an unnecessary and potentially traumatic therapy.
"Roughly for every eight breast cancers we diagnose through the screening programme, we will prevent one woman dying of breast cancer," says Dr Wilson. "Three of the seven wouldn't have died anyway because it would have been diagnosed at a treatable time. Three will die of breast cancer, because even though it was diagnosed through the screening programme, it was already too late. And one person will have treatment for breast cancer that she didn't need."
Placed in such stark terms, the ratio of benefit to harm sounds almost like a balanced equation. This, however, is to disregard what kind of value judgements we bring to the figures. Does the act of saving human lives outweigh the downsides of the programme?
Kopans and Wilson both think it does. Even in the case of overtreatment (a benign biopsy, for instance) they say that women themselves tend to see the positives in screening. "In my experience, the vast majority of women who have had the breast biopsy with benign results are thrilled that they did not have breast cancer," says Kopans. "It is a vocal minority that has raised this as a major issue."
Gøtzsche declined the opportunity to contribute to this article. He has published two titles that detail his work and clarify his position on the issue, Rational Diagnosis and Treatment. Evidence-Based Clinical Decision-Making (2007) and Mammography Screening: truth, lies and controversy (2012).
Nobody is saying that breast cancer screening is a unilateral good: the consequences, both psychological and physical, can be adverse. However, the screening programme has proven a major contributor to improvements in cancer survival rates.
In time, one hopes the whole debate will become a moot subject. "Obviously, the ultimate is for breast cancer treatment to get so good we don't have to screen for it," says Wilson. "That certainly is happening, because the mortality from breast cancer has dropped by 30%, including those who haven't been screened."
For now, screening looks set to continue, and with the results of the review due later this year, strong feelings and passionately argued views abound on either side.
"Mammography is not perfect," says Kopans. "It does not find all cancers, and when cancer is found, it's not always early enough to result in a cure, but it is the best we have at this time."
This article was first published in our sister publication Medical Imaging Technology.