Skip to main content

Together we are beating cancer

Donate now
  • Science & Technology
  • Health & Medicine

New breast cancer ‘blood test’ is still work in progress

by Henry Scowcroft | Analysis

3 May 2012

0 comments 0 comments

A mammogram

Mammography is currently used to detect breast cancer

Earlier this week, the news was full of stories about a ‘blood test’ that can apparently predict a woman’s risk of breast cancer, years before she develops the disease.

This sounds exciting. Being able to work out someone’s risk of cancer well in advance of developing the disease could be incredibly powerful.

People at higher risk could take extra steps to change their lifestyle, be offered extra screening or other monitoring, or even take drugs to prevent the disease.

But how valid are these claims? How would such a test work? What’s the science behind it? And how does it fit into the bigger picture of breast cancer prevention and detection?

We’ve been poring over the research paper that led to the headlines, and spoke to Imperial College’s Dr James Flanagan, the researcher who led the study together with Cancer Research UK’s Professor Robert Brown – to get his take on it.

The bottom line is the ‘test’ reported here is more of a proof-of-concept than a fully developed ‘cancer blood test’.

The principles that underpin the research are definitely promising, and suggest that in future the way we understand a person’s cancer risk (and what we do with this info) will be much more sophisticated.

But there’s a long way to go before a clinically useful test can be built from these findings.

Let’s take a look at how it all fits together.

Cancer risk, gene faults, and genetic variation

Cancer is ultimately caused by faults in the genes in our DNA that cause cells to multiply out of control.

For much of the last half-century, cancer researchers have been trying to understand how these faulty genes cause cancer, and how to use this knowledge to prevent and detect the disease.

Because of this research, people with multiple cases of cancer in their family can now be tested for rare faults in genes like BRCA1, BRCA2 and APC – faults that are known to substantially increase a person’s chances of cancer. Carriers can be offered extra screening or other treatments, and this has undoubtedly saved many lives.

But, crucially, only a small proportion of people who develop cancer do so because of an inherited gene fault. So researchers have been looking for common, more subtle genetic differences that explain why some people are more likely to develop cancer.

These efforts have borne fruit, and researchers are now beginning to paint a detailed picture of scores of tiny, common, genetic variations between individuals, each of which seems to cause a small change in risk, but which can add up to something more significant.

All of us carry some of these variations – precisely which ones we carry, and how many, plays a role in our chances of developing cancer later in life.

But this is very much work in progress, and this picture isn’t yet complete. And we don’t fully understand precisely how these variations influence a person’s risk, how they interact with lifestyle factors like smoking or obesity, or why some variations are linked to certain types of cancer.

And unlike testing for faulty genes, tests that can accurately predict a person’s cancer risk based on genetic variation are still some years away.

On top of this, researchers are still trying to work out whether offering people more or less screening, monitoring, or other treatments, based on these subtle variations in their DNA, actually saves lives.

Nevertheless, both these types of test look for changes in our DNA sequence – the code that contains information about how to make proteins.

But over the last few decades, it’s become apparent that on top of our DNA lies a whole extra layer of information – so-called ‘epigenetic’ information. And it’s this that Dr Flanagan’s team – in particular, PhD student Kevin Brennan – has been studying

The DNA double helix

DNA molecules are like a twisted ladder

Beyond our DNA

DNA molecules are long, twisted ladder-like molecules that live in the nucleus of each of our cells. Before a gene in our DNA can be activated, (i.e. the information on the ‘rungs’ of the ladder are ‘read’), one or more proteins first needs to stick to the outside of the DNA ladder to physically untwist it.

Cells carefully control which genes are active (i.e. which bits of the ladder are ‘untwisted’) at any given point in time by adding tiny chemical ‘tags’ to the DNA, known as epigenetic markers.

Some tags – called methyl groups – attract clusters of proteins that keep the DNA ladder tightly closed, meaning that genes can’t be switched on. The process of adding these methyl groups is known as methylation.

Scientists studying DNA methylation have discovered that it tends to occur in clumps along the length our DNA, and that these patterns are copied from cell to cell as they divide.

Researchers have also found evidence that exposure to various things over the course of our lives can influence our epigenetic patterns – things like tobacco smoke, radiation, alcohol and diet.

And they’ve also discovered that these patterns go completely, catastrophically awry in cancer. This allows cancer cells to turn on genes that should be off (and vice versa), and grow and divide out of control.

Putting all this together raises an interesting question. Can looking at these epigenetic patterns before cancer develops – and at differences in these patterns between individuals – yield any clues about a person’s chances of subsequently developing the disease?

Epigenetics and cancer risk

Several studies in recent years have suggested that the answer could be yes. For example, in 2008, blood samples from bladder cancer patients in a Spanish study were found to have different methylation patterns from people without the disease. And last year, a US study, also of bladder cancer, found a similar thing .

One important point about all these studies is that they looked at DNA from white blood cells – the cells of our immune systems. But although the immune system is heavily involved in cancer (as we’ve blogged about before), this wasn’t why researchers looked here. It was because they had no choice – the red cells in our blood don’t contain any DNA. In Dr Flanagan’s words, “they’re all we can get our hands on for this type of research”.

In 2009, Dr Flanagan’s team published results of a study of methylation patterns in the DNA from breast cancer patients’ white blood cells. They looked at epigenetic ‘tags’ on a whole range of genes involved in cancer, and found that one gene in particular seemed to be methylated in patients but not in women without the disease – a gene called ATM, which is known to be involved in several types of cancer.

But despite these tantalising results, a glaring weakness in all of the studies to date has been that they looked at the blood of people who already had cancer.

Cancer – and its treatment – can cause all sorts of changes to our bodies, so no-one could be sure that these epigenetic differences were caused by the cancer itself, or whether they’d existed before. So it wasn’t clear whether epigenetic markers could help predict a person’s risk.

Prospective studies

To try to address these flaws, with funding from Breast Cancer Campaign and Cancer Research UK, Dr Flanagan’s team turned to three large forward-looking (or ‘prospective’) studies – the European Prospective Investigation into Cancer (EPIC), the UK’s Breakthrough Generations Study, and the Australian KConFab study. These studies have been running for several years, collecting blood samples from healthy people, and following what subsequently happened to them.

This allowed Dr Flanagan’s team to identify women on these studies who had developed breast cancer, and then to go back and analyse blood samples taken many years before they were diagnosed.

The researchers identified 640 women who developed breast cancer in the three studies, and Brennan painstakingly measured whether the ATM gene in each of their blood samples was methylated.

They then repeated the process on a similar number of women on the studies who didn’t have breast cancer. And then, with the assistance of The Institute of Cancer Research’s genetic epidemiologist, Professor Montse Garcias-Closas, they put all their results into a computer to crunch the numbers.

They found that women who had the highest levels of methylation in the ATM gene in their white blood cells were nearly twice as likely to have subsequently developed breast cancer. And the effect was even more apparent in younger women. But it’s important to point out: not every women who had a methylated ATM gene developed breast cancer, and not every women who developed breast cancer had a methylated ATM gene.

What does all this mean?

According to Dr Flanagan, this is the first study with sufficient size and rigour to be able to find a definite link between an epigenetic marker in the blood and cancer risk.

“To be honest, most of the previous studies to try to look at this have been too small – we haven’t been able to measure this with any certainty,” he told us, adding “but this isn’t a ‘blood test’ yet. We need to look at how this fits in with other factors”.

How could this finding be used to help patients in future? Currently, people are offered cancer screening based on their age. But many researchers think that this ‘one-size-fits-all’ approach won’t last forever, and one day genetic testing will be used to offer people more tailored screening.

Dr Flanagan thinks that, in future, epigenetics could also help determine from what age, and how often, women should be offered screening. “That’s absolutely where we’re heading with this research,” he said, “but we’re not there yet.”

Next, Flanagan plans to look at how these epigenetic markers affect risk in people with different DNA variations, and he plans to collaborate with researchers at the Institute of Cancer Research to do so. Another plan is to look at the type of breast cancer women with methylated ATM genes develop, and to look in detail at the type of white blood cells that contain these modifications, to find clues as to how this marker actually increases risk.

So some of the media reports of an imminent ‘blood test’ that can “predict a woman’s breast cancer risk” were wide of the mark. Although it’s early days, this research suggests that epigenetic markers that can be measured in our blood – along with variations in our DNA – are a significant piece of the complex and fascinating jigsaw puzzle that is cancer risk.



  • Brennan, K. et al (2012). Intragenic ATM Methylation in Peripheral Blood DNA as a Biomarker of Breast Cancer Risk Cancer Research, 72 (9), 2304-2313 DOI: 10.1158/0008-5472.CAN-11-3157