NCRI 2015

Sponge test a cheaper and less invasive screening method for oesophageal cancer

Dr Rebecca Fitzgerald - University of Cambridge, Cambridge, UK

I was chairing a session on the earlier diagnosis of cancer and in particular talking about how molecular tests can be used, because there’s a whole new genomic era that we’re in so can we have a new generation, really, of molecular tests for earlier diagnosis of cancer. I was talking specifically about oesophageal cancer, adenocarcinoma, which is a cancer with a very poor outcome and I think we could make a real impact if we could find a way of diagnosing it earlier.

What are the current options?

At the moment the current options are endoscopy and biopsy. The main risk factor for this kind of cancer is heartburn, reflux symptoms. The people most at risk are particularly white men over the age of 50, obesity may play a role, partly because you get more reflux. So that’s a big proportion of society would fulfil that sort of risk bracket. So the problem is that you can’t really endoscope everyone with reflux symptoms and GPs have a very hard task: 10% of GP consultations are around heartburn symptoms, so who do they refer for endoscopy? The corollary of this is that as endoscopists in hospital 70% of the endoscopies we do are normal. So that’s a big burden on healthcare, it’s a fairly invasive test. So what I’ve been interested in for some time is how we really come up with a new paradigm for diagnosis; trying to find something simple, more cost effective, GP based and to couple a different cell collection device with a molecular test that could be objective, accurate and do two things with the molecular test: both diagnose Barrett’s and risk stratify because you don’t want everyone with Barrett’s to have to have an endoscopy because then you get back to the same sort of problem. So that’s really what I was presenting an update on, our work in this area. The cell collection device is called the cytosponge. This is in the primary care setting. The patient comes to see the GP with some heartburn symptoms, they swallow this pill on a string, it’s about the size of a multivitamin pill. It wriggles down to the top of the stomach, just with the warmth of the stomach it will dissolve and out pops a spherical sponge. The nurse pulls that back just a few minutes later and collects about half a million plus cells on the sponge. That just gets put in a standard preservative pot and sent to the lab. So from the patient’s point of view this is a pretty straightforward, simple test, it will make their eyes water a bit but it’s really not bad. Compared with an endoscopy it’s a lot, lot easier, it’s a lot more cost effective.

Then the lab side of it, you can be much more sophisticated. You want it easy for the patient but to really be sophisticated in terms of the molecular tools that one uses. So the first test that we’ve done quite a lot of work now on and tested over 2,000 patients in two studies is to diagnose Barrett’s and that’s using one specific biomarker called trefoil factor 3. It’s a nice simple test, it’s an immuno-based test. So standard clinical path labs are used to doing immuno-based tests and it’s scored simply positive or negative. So even if you just have one cell that stains with this antibody it’s positive. Then what we’re working on now, and I presented some new data on this today, is how you then with the same sample if it was TF3 positive so you think the patient has got Barrett’s how you then check whether they’re high risk or low risk. We’re still refining this but it seems that p53 is a critical gatekeeper mutation for patients that go on to get cancer and you can test for p53 over-expression with an immuno-marker. You can also test for mutations on the cells from the cytosponge, all from the same sample. So we’re using that. We’ve done some work to see what’s the best combination of biomarkers. p53 on its own may not be enough but the pathologist’s diagnosis in terms of whether the cells look atypical is actually also very predictive and how long the Barrett’s segment is that the patient has, if you know that information.

So using all those three things together we think we can then put patients into three buckets, a high risk group and they would go straight to endoscopy and have treatment if it’s confirmed, an intermediate risk group where we would say repeat the cytosponge in about six months, and a low risk group where really we feel pretty laid back, they’ve got Barrett’s, we think it’s low risk, swallow the cytosponge again in three years’ time. So we’re still confirming exactly how we want this risk stratification to be done but that’s the kind of way we’re thinking about it at the moment.

Is looking at risk stratification in this area a new thing or are other countries doing it too?

It’s a completely new area. So people for a long time have been thinking endoscopy, what else can you do on the biopsies to risk stratify. One of the big problems with biopsies is sampling bias. So Barrett’s is a polyclonal disease so you can miss the spots that you want with a biopsy. Typically at surveillance at endoscopy we do four biopsies every 2cm and even then we know that we sometimes miss the area. The great thing about the cytosponge is you get a nice sweep of the whole oesophagus so there’s less sampling bias. Currently the biopsies are just looked at by pathologists for dysplasia, no molecular biomarkers are typically used. People are starting to use p53 immunohistochemistry, that’s the first one that’s now gone into our British Society guidelines. Worldwide that’s still not used really. People have been looking at adding molecular tests to biopsies for quite a long time but haven’t really come up with something which is clinically applicable and accurate enough. So again the cell collection, I think, is really a pivotal part of this to get much less sampling bias. Because the cell collection is so good then I think we can do clever things with biomarkers, plus the technology has really leapt ahead so it’s much easier to look for mutations now in a sample than it was before.

What percentage of Barrett’s oesophagus turns into cancer?

It’s about one patient in 300 per year. So it’s a low absolute risk. So that’s why it’s really important that the test is simple, inexpensive and that you have the risk stratification. At the moment everyone with Barrett’s goes through repeated endoscopy every couple of years. That’s quite a big burden when most of them will never get cancer so if you can take it away from that so as much of the effort is on identifying the high risk people, as is reassuring the people actually at very low risk so that you can give them good reassurance that you’re confident that they really are low risk and then let them off the hook so they don’t have to have repeated testing.

Could this save the NHS a lot of money?

That’s the hope. The idea is we’ll do more case finding and concentrate more on the high risk people.

What about the place of a healthy lifestyle in preventing this?

Healthy lifestyle is certainly important and our understanding of diet and cancer is actually pretty primitive. We don’t really understand much about it, it’s hard to study. But no doubt obesity is a big problem in the Western world and obesity, of course, is a risk factor for all sorts of things. But obesity is one of the risk factors for this kind of cancer. So absolutely it has to be sensible to be getting people to be healthier in the first place rather than then just trying to catch the problem once it has already developed.

Do doctors give people presenting with Barrett’s oesophagus advice about diet and exercise?

You would hope so, wouldn’t you? Lifestyle factors are important to help reduce reflux and if you look at the NICE stepwise guidelines that doctors should give their patients that is part of it. It’s quite hard to do, GPs are pressed for time and I think we could all do a lot more of that. So as part of this, if you were going to introduce more systematic testing in the GP surgery then absolutely we should be talking to patients more about risk factors.