Epigenetics for breast cancer screening

A few weeks ago I was at a cancer research seminar, and was chatting to a consultant oncologist over coffee. He explained to me his view that the important advances in cancer over the next decade or so are likely to be in better diagnosis, rather than better treatment.

I don't know if he's right, of course, but what he said certainly makes sense. Cancer is incredibly difficult to treat once it's at an advanced stage, and the sad fact is that despite all the progress we've made in cancer treatment over recent decades, patients with metastatic cancer almost invariably die of their disease, despite the best efforts of the medical profession. However, if caught early, many cancers are treatable. It's therefore logical to think that if we can diagnose cancer more reliably at earlier stages, it's going to result in more people spared a horrible death from cancer than any advances in treatment that we might plausibly make in the next decade or so could do.

So it was interesting to hear in the news today that researchers from Imperial College have published results of a study investigating epigenetic markers as predictors of breast cancer. They found that methylation of peripheral blood DNA can predict breast cancer risk.

This looks, on the face of it, to be just the sort of thing we need: a means of predicting who might develop breast cancer before symptoms become apparent.

It's interesting research, of course, and certainly advances our knowledge. However, I think there are some really important caveats which don't seem to have made it into the media.

The problem is that there is a huge difference between a known risk factor and a useful screening test. If you have a risk factor, such as one of the epigenetic markers studied in the recent study, then you are more likely to get breast cancer. But it doesn't mean that you will. You might get it, or you might not. All it tells you is that your risk of getting it is higher than it would have been without the risk factor.

That's not what we need as a screening test. For a screening test, you need to know that if you test positive, then you have a high chance of developing the disease, and that if you don't, then you have a low chance. Simply knowing that you're at increased risk isn't much help by itself, unless you're at such hugely increased risk that you can be pretty sure of getting the disease.

Treating cancer is a major undertaking. The commonly used treatments – surgery, chemotherapy, and radiotherapy – are all unpleasant and have significant risks, so we really don't want to be using them on someone who doesn't actually have cancer. Obviously no-one should start treatment purely on the basis of a screening test alone without further diagnostic tests, but as we know from experience with mammography, things are not perfect in that regard. Many women have cancers diagnosed on mammography which would never have harmed them if left untreated.

While the latest research increases our knowledge, and is therefore welcome, it is nowhere near good enough to be used as a screening test at this stage. Obviously this is a new area of research, and further research into epigenetic markers of cancer may refine the techniques to the point where they could be used as screening tests, which would be an exciting development.

I do hope, however, that if we get that far, the lessons of mammography screening will have been learned and we won't end up with another screening test that has the potential to do as much harm as it does good.