Making breast screening more accurate

Over the last few years it’s been hard to miss the ongoing debate about pros and cons of breast screening.

On the one hand, it saves lives by detecting the disease earlier.

On the other, there are downsides – most importantly, it can detect slow growing cancers that don’t need treating – so-called ‘overdiagnosed’ cancers.

But how do these balance out? Experts still disagree, but the best estimate comes from the 2012 Independent Breast Screening Review (also known as the Marmot Review), which estimated that for every one life saved by screening, three women are overdiagnosed.

Nevertheless the debate rumbles on, with reports appearing almost weekly looking at screening’s pros and cons.

So what can be done to clarify the situation? Can the balance be shifted towards the benefits?

One thing that might help would be to identify women at higher (or lower) risk of the disease, and offer them more (or less) screening (hence the excitement over a potential ‘blood test’ for breast cancer risk we blogged about this morning).

But as well as ‘overdiagnosis’ – something that’s inherently difficult to estimate – screening can have other downsides. One of these is that, for every 1000 women invited for screening, about 40 are called back for further tests, but only eight of these come back positive.

That means around thirty women in every 1000 go through the short-term anxiety of worrying they might have cancer, only to be told they don’t.

Can anything be done?

Thanks to research, it can. This week we spotted two eye-catching new research papers – one from the US, one from Germany – that point to some substantial improvements on the horizon for women being screened for breast cancer.

The breast in 3D – Digital Breast Tomography

Mammography involves x-raying a woman’s breasts, but exactly how this is done has changed over the years. The big change in recent years has been the advent of digital mammography, which is now widely used around the NHS. As we wrote when we covered the Marmot review:

“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.”

This in turn can allow radiologists to adopt an even newer technique called digital breast tomography, or DBT, which uses a moving x-ray detector and sophisticated computer software to create a 3D image of the inside of the breast. But this isn’t yet used routinely to screen women on the NHS (although it is used in some centres to examine women whose initial mammogram has shown a positive-looking result).

Several studies have shown that DBT could bring real improvements in routine use, and we’re currently funding a study called OPTIMAM2, which is working out whether and how it could be used in the NHS programme. So it was really interesting to see the results of a US study appear this week, published in the journal JAMA, looking at what happened when DBT was introduced in 13 different screening centres across the US.

“None of the studies to date have demonstrated what the impact would be in term of saving lives or overdiagnosis” – Professor Ken Young, Cancer Research UK

As well as reducing the rate women were recalled for things that didn’t turn out to be cancer, DBT also detected more cancers – cancers that conventional screening wouldn’t have picked up. In other words, fewer women were recalled, and more cancers were detected – a win-win situation.

So what needs to happen for this to be used on the NHS? We spoke to breast screening expert Professor Ken Young, who leads our OPTIMAM2 study, to find out.

“From research so far we can see that screening with DBT is showing benefits in terms of recall and detection. However none of the studies to date have demonstrated what the impact would be in terms of saving lives or overdiagnosis”.

And this is a big limitation of the new study as well. To properly understand these benefits, says Young, you need randomised trials. “But these can take a long time,” he says.

Things could be sped up by using computer software and mathematics to predict clinical benefits from early trial data , “and this is an approach that my team are beginning to develop,” says Young.

But the new study’s strength, says Young, is its large size and that it was carried out ‘in the wild’ as US centres switched to DBT, and still found a similar size of improvement in cancer detection and recall rates as the few organised screening studies.

“This means we can be pretty confident that these improvements are real, and not down to the way the studies were designed,” Young says.

The NHS has itself been studying DBT, in a long-running trial called TOMMY and the NHS Screening programme is reviewing the wider evidence about DBT. In particular, there are some significant practical issues to be overcome – particularly the much longer time required to read DBT images.

So we’re not there yet, but Young predicts the advent of routine DBT is not too far off.

Screening using MRI

X-rays aren’t the only way to detect breast cancers

So it looks like tomography will have a place in the screening programmes of the not-too-distant future, and hopefully reduce the number of women being recalled unnecessarily, and increase the number of lives saved by screening.

But x-rays aren’t the only way to detect breast cancers – you can also use a technique called magnetic resonance imaging (MRI) too. This uses intense magnetic fields to ‘wobble’ the molecules of a dye that’s been injected into the area being screened. These wobbles can be reconstructed in a computer to form an exquisitely detailed picture.

MRI can be much more sensitive than mammography, and avoids the need to use x-ray radiation.

And it’s already used by the NHS to screen women at particularly high risk of breast cancer, such as those carrying an inherited gene fault.

But it has three serious drawbacks that prevent it being used more widely. First, you need to inject a dye into the person being screened. Secondly, it takes a lot of time – up to 40 minutes – for a full MRI study. And, as a result of the time taken, and because the equipment is more sophisticated, thirdly it costs a LOT of money – far more than conventional x-ray screening.

But the second paper we spotted this week, from a research team in Germany and published in the Journal of Clinical Oncology, shows that at least two of these problems – time and money – seems to have been addressed. The German team has developed a way of using MRI for breast screening that takes about 3 minutes, rather than 40 minutes – with no apparent loss of accuracy. This is potentially a game-changer for MRI, as a single unit could screen far more women in a given day, vastly reducing the cost of the process.

However, says Young, there’s still one hurdle to be overcome – you still need to inject a dye, and this can be an uncomfortable experience compared with going for a mammogram (although many women say this is no walk in the park either).

Nevertheless, it’s a “valuable development”, he says and definitely something to keep an eye on.

Where next?

Whether you think the balance of risks and harms is acceptable (and so decide to go for screening), or not, is very much an individual decision.

Our position on all this is clear – we think that, on balance, the NHS Screening Programme is a good thing that prevents avoidable deaths from breast cancer, with the important caveat that everyone invited should be given clear, understandable information about the pros and cons.

But we also know that, over time, research will make screening more effective, quicker, less worrying and less time consuming. That’s why we come to work every day, and why we passionately believe that, with research, we will beat cancer sooner.

Henry

• Video: What it’s like to have a mammogram

References

• Friedewald S.M, et al. (2014). Breast Cancer Screening Using Tomosynthesis in Combination With Digital Mammography, JAMA, 311 (24) 2499. DOI: http://dx.doi.org/10.1001/jama.2014.6095
• Kuhl C.K, et al. (2014). Abbreviated Breast Magnetic Resonance Imaging (MRI): First Postcontrast Subtracted Images and Maximum-Intensity Projection–A Novel Approach to Breast Cancer Screening With MRI, Journal of Clinical Oncology, DOI: http://dx.doi.org/10.1200/jco.2013.52.5386

Mammogram image from Flickr