Breast screening is back in the news.
At the request of the Government and Cancer Research UK, an independent expert panel has scrutinised all the available evidence on the pros and cons of mammography. And this morning they’ve published their findings in The Lancet , putting these into context for the first time.
A key finding is that breast screening has significant downsides, as well as undoubted benefits: in addition to the 1,300 lives saved each year, the panel acknowledged that around 4,000 women are ‘overdiagnosed’ by the programme.
Overdiagnosis is a complicated concept – these are genuine cancers spotted by the screening programme (and which wouldn’t have been detected otherwise), which would have grown so slowly they would never caused problems during a woman’s life.
And since there’s absolutely no way to tell the difference between these slow-growing cancers and those that are life-threatening, virtually all women diagnosed through screening are treated – usually with surgery , and potentially with hormone treatment , radiotherapy and – occasionally – chemotherapy .
The concept of ‘harmless’ cancers is likely to be new to many. And the fact that one in five cancers diagnosed by the screening programme wouldn’t have caused harm if left alone is certainly likely to cause concern.
As well as overdiagnosis, screening has other downsides. About four out of five women called back for more tests, turn out not to have cancer. These false alarms cause anxiety, while some women need a biopsy to rule out cancer, and this can be painful.
And the X-rays used to carry out breast screening are linked to a very small number of cancers every year.
But on balance, because of the fact that it saves so many lives, we think breast screening should continue, and we recommend that women to go for screening when invited.
So, now that we better understand the harms breast screening can cause, what are researchers doing to minimise them?
Three key questions, if properly answered, could make a big difference: firstly, who to screen, secondly, what to screen them with, and finally, what to do with cancers when they’re found.
Thankfully the answers to these crucial questions are already being sketched out in labs and hospitals around the world. So let’s tentatively peer into the future and see what breast screening may look like in a decade’s time.
Everyone’s risk of cancer is different, and this depends – to a degree – on the genes we inherit from our parents (although how we lead our lives also affects our chances of cancer).
Women who inherit faulty versions of certain genes – BRCA1, BRCA2 or CHEK2 for example – have a high chance of developing the disease, and can be offered much more regular monitoring with techniques like MRI.
But in recent years, researchers – including those funded by Cancer Research UK – have discovered scores of more subtle genetic variations between people, which can also affect breast cancer risk (albeit to a much lesser degree).
The effects of these variants, known as ‘SNPs’ , add up – the more a woman carries, the higher her risk.
Armed with this knowledge, could we improve screening by offering it only to women at above-average risk, based on a simple DNA blood test?
Professor Paul Pharoah , a Cambridge-based Cancer Research UK epidemiologist, thinks that focusing screening on women with a higher-than-average risk could significantly alter the balance between risks and benefits.
“Every woman has a different risk of breast cancer,” he says. “Some women are at very low risk, some women are at very high risk, and most women are somewhere in the middle.
“One way of thinking about this is to imagine a group of people who are dealt a hand of cards, and scored for each card from an ace to a king. A few would get only high-scoring cards (and be at a high risk), others would get only low-cards (and be at low risk) but most would get a combination of low and high”.
In recent years, researchers have tried to calculate the size of the difference between the lowest risk and the highest risk women, and how much of this is down to genetics. This ‘genetic component’ of risk turns out to be responsible for about a third of the difference between highest and lowest (with lifestyle accounting for the rest). And, Pharoah says, researchers have already tracked the genetic variations responsible for about four-tenths of the genetic component.
And that, with a bit more number crunching, has some fairly startling implications.
“Looking at the way these known subtle genetic variants affect cancer risk, the half of the population at highest risk actually accounts for just over two-thirds of all breast cancer cases,” says Pharoah.
“Or, put another way, the low-risk population probably won’t get much benefit from screening, compared to the potential harms.”
If these low-risk women could be identified, they could possibly be offered less – or even no – screening.
Professor Pharoah thinks that before too long we’ll see trials in which low-risk women are offered the chance to have no screening, while high-risk women are offered regular mammography. If this works this could, in theory, lead to fewer mammograms, fewer false positives, and potentially lower levels of overdiagnosis.
UCL’s Dr Nora Pashayan is another of Cancer Research UK’s genetics experts, and works with Professor Pharoah as part of a Europe-wide initiative called COGS , to try to draw up a road-map to bring this forward.
According to Dr Pashayan, there are four key hurdles to overcome – effectiveness, cost-effectiveness, acceptability and feasibility.
“First, you need to prove this is effective. In other words, does it actually work, in terms of detecting cancers more effectively and minimising the harms? That’s what we need to establish first.
“Second, you need to prove it’s cost-effective – it’s no good if health systems can’t afford it and that’s another big question to be answered. Third, it needs to be acceptable – to the public, who will have their genes analysed, but also to healthcare professionals and policymakers who have to run the programme.”
The issue of genetic data isn’t trivial. Our genes can predict a lot about ourselves and our health, so how this information is stored, and who has access to it, could become a thorny ethical problem, says Pashayan.
“And finally – and this is why you need proper pilot testing – you need to show that it all works in practice, in the context of a real-life healthcare system,” she says.
So genetically stratified screening is beginning to loom large on the horizon, and could shift the balance further towards the benefits. But alongside this, other Cancer Research UK researchers are trying to refine the cornerstone of breast screening itself – X-ray mammography.
“X-raying a woman’s breasts to detect cancer is one of the most demanding tasks in radiology”, says Professor Ken Young , of the Royal Surrey County Hospital. “Taking X-rays of the breast is particularly demanding, as we need to be able to detect very small calcifications, and small differences between the soft tissues”.
Professor Young also runs the National Coordinating Centre for the Physics of Mammography in Surrey, and is one of Europe’s leading experts in this field.
“The big change happening at the moment is digital mammography – using computers to capture, manipulate and store the image, rather than film. This is being rolled out nationally, and the NHS Breast Screening Programme is now 75 per cent digital,” he says.
Digital mammograms are better quality, need a slightly lower X-ray dose, take up less storage space (saving money) and – importantly – can be re-analysed using a range of sophisticated computer techniques.
“Although digital mammography is itself an improvement in quality, perhaps its greatest benefit is that it sets the scene for big improvements in the future,” he added. One of these improvements is known as ‘digital breast tomosynthesis’, or DBT.
“Conventional mammography, known as 2D mammography, usually involves taking two x-rays of each breast,” says Young. “In DBT, between 10 and 25 shots are taken, from slightly different angles, but each at a far lower radiation dose.”
The result: a similar overall radiation dose, and – thanks to the power of computer processing – a 3D image of the inside of a woman’s breast. Studies have already shown the technique’s considerable promise.
“We need to wait for the results of ongoing trials to be certain, but this has the potential to improve both the sensitivity and specificity [reducing the rate of false alarms and missed cancers] of breast screening, and to detect cancers at an even earlier stage,” says Young. This, in turn, would result in more lives being saved by screening.
So how far away is routine DBT?
Several large trials comparing the technique to regular 2D mammography are underway. One of the largest is based in Norway , and the NHS-funded TOMMY trial is still recruiting women. Results are expected over the next two years.
Another key project Professor Young is running is the Cancer Research UK-funded OPTIMAM study . This project involves a number of ‘virtual’ trials, aimed at discovering the best and most reliable ways of carrying out both 2D digital mammography and DBT.
After experimentally modelling a variety of different types of 2D and DBT mammograms, Young’s team are testing how effectively radiologists can spot cancers in them. This will help further refine how 2D digital mammography and DBT is eventually used in the NHS.
“Within the next year or so, tomosynthesis will start being routinely used in the NHS to re-test women recalled after their first mammogram,” predicts Professor Young, “and, depending on the trial results, I’d expect it to begin being used as a front line screening tool within five years”.
Techniques like DBT should result in a more effective, efficient screening programme, which can spot cancers earlier, leading to fewer false alarms and – in all likelihood – increase the programme’s benefits. But will this tackle the issues of overdiagnosis?
Here, Professor Young is keeping his feet firmly on the ground. “In the end, there’s a limit to what you can do with imaging,” he says. “Although improved imaging can boost screening’s benefits and minimise harms, a small amount of overdiagnosis is probably inevitable”.
So do we need a better test than X-raying a woman’s breasts? This is much further off, but there are rays of hope here too. For example, we’re jointly funding a trial of a blood test that, according to lead investigator Professor Charles Coombes, could one day replace mammography if everything goes according to plan.
But for the mean time, to truly tackle the issue of overdiagnosis, we urgently need to work out which breast cancers are the potential killers, and which ones doctors could safely leave alone.
And to do this, we need to go back to the laboratories, and the extraordinary recent advances in our understanding of breast cancer biology.
In March this year, a Cancer Research UK-funded study called METABRIC made huge waves in the scientific world. Researchers in Cambridge and Vancouver completely redefined the way we think about breast cancer, plotting out ten different types, each with different molecular fingerprints.
But METABRIC was just one of a whole series of discoveries about breast cancer’s inner workings that have been published this year, with no fewer than five key papers appearing in a single issue of Nature in June.
Thanks to research like this, we now know that breast cancers are phenomenally diverse , yet can be grouped into at least ten different types. And researchers are already building on this knowledge to try to turn it into ways to improve diagnosis and treatment .
On top of this, detailed analysis of the make-up of normal breast tissue is yielding clues to how different types of breast cancer develop, and how they behave. Again, our scientists are leading the way .
Put this all together, and you realise we’re standing on the edge of a complete transformation in how breast cancers are diagnosed. We’re just a few years away from being able to take a sample of tissue from a woman’s breast tumour, and run tests that yield clues as to where it came from, what drives it and – hopefully – how to kill it.
And that, ultimately, will be how we get a handle on overdiagnosis, and firmly shift the balance towards the screening’s benefits.
Gazing into a crystal ball is always dangerous in medical research, but in this case, we feel justified in doing so. The research we’ve discussed above outlines a tentative sketch of what breast screening may look like in the next decade – call it 2020s vision, if you will…
In her late 20s, a woman will be offered a free blood test on the NHS. The DNA in it is analysed, and the results combined with information about her lifestyle – her weight, her family history, whether she has children and how many, and when she started her periods – to produce a tailored breast screening schedule.
Perhaps she’ll only need screening every six years, starting from 50. Or maybe, because of tell-tale genetic markers, she’ll be advised to go annually from 35. Or maybe she won’t need screening at all.
When and if she does go for screening, 3D images captured by state-of-the-art kit, using minimal amounts of radiation, will be analysed by powerful computers, and any signs of cancer marked out and verified by trained experts. The chances of missing a lump, or mistaking a benign cyst for a cancer, will be extremely low.
If a suspicious lump is spotted, a nurse will take a biopsy, which will be run through a battery of hi-tech tests, analysing the genes and other molecules inside. This will give her doctor information on what sort of cancer it is and, crucially, whether and how to treat it, or whether to leave it and, instead, keep an eye on it.
And if she does need treatment, the molecules inside the tumour will give the doctor a clear idea of the best treatments to use.
This is where we want to get to. As we’ve said, there’s a lot of hard, painstaking work to get there – and no-one can truly claim to be able to predict what the results of scientific research will be, nor how long it will take.
But with your help and support, and with the dedication of our passionate researchers, sooner or later, we’ll get there.
We hope it’s sooner.