Lung screening – still a need for caution, despite new blood test - Cancer Research UK

Lung cancer is often diagnosed at a late stage, when it's spread

Yesterday’s announcement that the Scottish government will trial a new lung cancer test has been widely covered in this morning’s media – both on TV and in the papers.

In case you missed the story, the NHS in Scotland is to run a trial of a proprietary blood test, called EarlyCDT-Lung, on 10,000 heavy smokers, with the aim of detecting lung cancer earlier, when it’s more treatable.

This focus on lung cancer is welcome, and comes as part of a wider initiative by the Scots government, Detect Cancer Early, which we helped campaign for and which we’re delighted has now launched.

However, the tone of some of the media coverage unfortunately overstates the stage this research is at, and readers could be forgiven for thinking a screening programme is imminent.

On top of this, there’s a bigger picture. Screening for lung cancer is fraught with difficulties and unanswered questions – chiefly the fact that chest surgery to investigate a positive result isn’t a trivial matter, and carries significant risks.

So if a test can be proven to save lives and minimise harms, it would be a significant step forward.

But with those caveats out the way, we thought it would be helpful to look at how this blood test fits in with this bigger picture. People should realise that, while this trial is helpful and welcome, a national lung screening programme is probably still some way off.

Where are we at with lung screening?

Before we look at the new test, a bit of background. Over the last year, several large trials of chest x-rays to spot lung cancer have published results. Here’s what we wrote about one of the largest, which compared two different methods of lung screening:

[There] are a few crucial reasons why doctors are not yet calling for all smokers to opt for an annual CT scan of their chest:

The number of positive screening results was considerably higher in the CT scan group than the chest x-ray group (18,000 vs 5,000). The vast majority of these results turned out to be false positives – suspicious scans that turn out not to be cancer on further investigation (in fact, around 95% of positive results in both arms of the trial were false positives).

Although there were very few negative effects of the screening procedures themselves, there were risks associated with diagnostic tests following screening. Ten people from the CT scan group and 11 from the chest x-ray group died within 60 days of having investigative procedures following screening. A further 75 people in the CT scan group had a severe complication associated with diagnostic testing (24 in the chest x-ray group).

The trial compared two different modes of screening rather than comparing screening with no testing. So although there was no increase in overall death rate in the CT scan group compared to the chest x-ray group, both arms of the trial were associated with increased risks, and it is not clear how the risks and benefits of these screening procedures compare with no screening.

There was some suggestion that lung screening results in some degree of ‘overdiagnosis’ – that is, the diagnosis of some lung cancers that would never have progressed into more serious disease or caused any symptoms. The authors are calling for more research in this area, which is a sensible way forward.

In essence, we just don’t know whether the benefits of chest x-rays outweigh the harms (this is important and we’ll come back to it later).

On top of this, CT screening is relatively expensive, and the UK probably doesn’t have enough scanners to run a population-wide screening programme.

What is the test and how does it work?

The EarlyCDT-Lung test was born in a lab in Nottingham University, and subsequently developed by Oncimmune, a ‘spin-out’ company associated with the University.

The test measures the levels of certain antibodies in a person’s blood, which often rise when they have lung cancer. This is a really exciting idea, and Oncimmune are also developing similar tests that aim to detect other cancers.

How effective is it?

There are many ways to assess how effective a screening test is. Two key measurements are a test’s ‘sensitivity’ and its ‘specificity’.

We’ve blogged about this before, in the context of ovarian screening :

When designing a test for any disease, two things are really important. The first, obvious, one is how good the test is at picking up people that have the disease. This is called ‘sensitivity’. The second important factor is how good the test is at not picking up people that don’t have the disease, which is called ‘specificity’.

It’s important to get a good balance between sensitivity and specificity. A quick thought experiment will reveal why. Imagine a test where every single person that has the disease tests positive. This sounds great, but what if the reason this happens is because everybody who takes the test at all, regardless of whether or not they have the disease, tests positive? Clearly this would be a useless test in practice. It’s highly sensitive but not in the least bit specific.

According to Oncimmune’s website, their lung test has a sensitivity of 41 per cent, and a specificity of 93 per cent.

In other words, it’ll correctly identify lung cancer in about four out of ten people who have the disease, and correctly identify about nine out of ten people who don’t have it. It will also give a ‘false positive’ result of about seven in a hundred people who don’t have the disease.

However, we spoke to Geoff Hamilton-Fairley, Oncimmune’s executive chairman, who told us that the company had managed to improve these figures, and will be publishing data about this in the next few months. So the test that will be used in the Scottish trial will be an improved version.

Another key feature of the test is that it’s able to spot some cancers that wouldn’t be visible on a chest x-ray. According to the company, this could lead to diagnosing some cancers five years before they could be spotted by other means.

How will the trial work?

The trial will involve 10,000 ‘high-risk’ smokers – people who have smoked the equivalent of 20 a day for twenty years. Half of them will be offered the blood test, the other half won’t. People who have a positive result will then be offered a chest x-ray.

Given the problems we discussed above with lung screening using just a chest x-ray, the hope is that the ‘pre-screening’ people with the blood test will ‘enrich’ the population being offered x-rays with people who are more likely to have cancer. This will hopefully lower the false positive rates from chest x-rays (which can be as high as 50 per cent), meaning the benefits might then outweigh the harms.

It could also lead to an economic benefit – advanced lung cancer is more expensive to treat than early disease, so if all goes well, the test could save the NHS money.

But here come the caveats

You’ll notice a lot of ‘ifs’ and ‘maybes’ in that last paragraph. These are questions that the trial will try to answer. There’s a long way to go before we get a full sense of how, or whether, this test should be routinely used in the NHS, despite its undoubted promise.

That said, a robust lung screening programme would be a fantastic addition to current efforts to detect cancer earlier. The disease kills nearly 35,000 people a year in the UK – anything that could reduce this terrible toll has to be worth looking at. But we need to make sure we don’t cause other problems in doing so.

And it’s worth noting that this isn’t the only horse in the race – a large lung cancer screening pilot trial using CT scanning is already underway across the UK, funded by the Department of Health.

And then there’s smoking

About eight out of ten lung cancers are thought to be directly caused by smoking. As well as trying to spot cancers early both in smokers and in non-smokers, we also need to find ways to help people quit smoking, or discourage them from starting.

Over the years, Cancer Research UK has been a world leader in campaigning for tighter controls on tobacco. We’re currently trying to persuade the UK Government to bring in laws forcing tobacco companies to sell their product in plain, brand-free packs, with prominent health messages. Although this won’t necessarily stop current smokers from smoking, there’s mounting evidence that it will discourage kids from starting.

So as well as our anticipation over this lung screening test, we’re also looking forward to campaigning for plain packaging, and protecting the next generation from the evils of tobacco marketing. You can find out how to support the campaign here. Please sign up, and help us end the packet racket.

Henry