Cancer screening is never straightforward. At first glance, it seems like an obviously good idea. By searching for early signs of cancer in apparently healthy people, doctors can spot cases of disease when they can be most easily treated, or even spot hallmarks of the disease’s early stages. Lives could be saved.
But the reality is more complicated, involving a fine balance of benefits and risks. No screening test is perfect, and there’s always a chance that a cancer will be missed. Alternatively, there’s a chance that the test will raise false alarms, by suggesting that someone might have cancer when they actually don’t. These “false-positive” results can worry people unnecessarily, as well as pushing them towards unneeded tests or treatments.
“False-positive” results are generally regarded as a ‘bad thing’. But even that’s too simple an interpretation, as a new study from Finland demonstrates. The study, a clinical trial of the controversial PSA test for prostate cancer, tells us that false-positives are common. It also shows that men who get a false alarm:
- are likely to get another one the next time they go for a PSA test
- are likely to refuse future invitations to screening, and
- are likely to actually be diagnosed with prostate cancer the next time round
The third result, in particular, is a fascinating one. It suggests that men who get a false-positive result through PSA testing, in the words of the researchers, “constitute a special group”. They could well go through unwarranted tests, but they could also harbour missed cancers that only turn up later.
PSA testing revisited
In brief, there’s no national prostate screening programme in the UK but men can ask their GP to have a PSA test. The idea is that men with prostate cancer have higher levels of the PSA protein. However, men with high levels don’t always have prostate cancer, men with low levels sometimes do, and PSA tells us nothing about whether a tumour will grow quickly enough to cause problems.
There’s still a lot of disagreement about whether PSA testing saves lives. You may remember that two large trials published conflicting results on this subject last year. A European trial (the ERSPC) found that screening does save lives, but a US trial (the PLCO) said that it doesn’t.
But saving lives is just part of the story. Even if PSA testing prevents people from dying from prostate cancer, we’d still need evidence that it does so without causing too much harm.
Consider this – admittedly slightly flippant – scenario. Every prostate cancer patient has a head, so you could set up a screening programme (a bizarre one, admittedly) where you tested every man in the country to see if they had a head. If they did, you’d send them for treatment. This test would be guaranteed to pick up 100 per cent of prostate cancers but it would also lead to a huge number of false-positives since, clearly, having a head doesn’t automatically mean you have cancer.
This is an extreme example, but the same principle applies to more sensible screening tests. In order to judge whether a particular test is worthwhile, you need to see how many people are falsely suspected of cancer because of the test, and whether this is justified by the benefits. That’s where the new study comes in.
As we mentioned above, there’s a large prostate screening trial running across Europe, called ESPRC. The new results, published in the British Journal of Cancer, (which Cancer Research UK owns) come from the Finnish part of this trial – its largest component.
It involves more than 80,000 men, some of whom were randomly invited to three rounds of PSA testing, with four-year gaps between each round. Roughly 30,000 men attended their first round of screening and more than 10,000 of these men went on to attend all three rounds.
The study showed that false-positives are a common part of PSA testing. In any individual round of testing, the majority of positive results are false alarms (between 60 and 70 per cent), while just over a quarter lead to an actual cancer diagnosis. Among the men who attended at least one round of screening, 1 in 8 had at least one false-positive result.
It’s worth noting that the researchers were using a fairly high cut-off level of PSA (4 ng/ml) – i.e. the level above which they were thought to have suspected prostate cancer. This sets a pretty high bar for a positive result and should minimise the number of false positives. Nonetheless, many still crept through.
Among the men who get a false alarm in one round, more than half will get another false alarm in the next one. Many men without tumours have persistently high PSA levels for some other reason, so they keep on testing positive. That’s a lot of extra worry and more potential for unneeded tests.
Indeed, in this trial, every third man who got a false alarm went through two biopsies within 4 years of their result. That’s probably an underestimate too, as it doesn’t account for any visits to private doctors.
However, the study also shows that false-positives aren’t entirely meaningless. If men had a false alarm during one round of screening, they were 3-9 times more likely to be diagnosed with prostate cancer during the next round, compared to those who had a straightforward negative result.
In absolute terms, the odds of being diagnosed with prostate cancer are 12-20 per cent (between 1 in 5 and 1 in 8) if the men had a false alarm during the previous PSA test, but just 1-4 per cent (between 1 in 25 and 1in 100) if they had a negative result.
All in all, a quarter of men who get a false-positive result are eventually diagnosed with prostate cancer. Thankfully, the majority of these cancers (around three quarters) are still in their early stages. But the worry is that some are missed because false alarms put people off from further tests. Indeed, men are 1.5-2 times more likely to give screening a miss the next time round if they get a false positive.
This paradoxical problem of false-positives shows that we urgently need better ways of working out a man’s risk of prostate cancer, particularly among those who get a false-positive result in a PSA test.
The challenge for researchers is to work out which of these men are most likely to harbour a problematic prostate tumour; which have a slow-growing one; and which have no tumour at all.
Kilpeläinen, T., Tammela, T., Määttänen, L., Kujala, P., Stenman, U., Ala-Opas, M., Murtola, T., & Auvinen, A. (2010). False-positive screening results in the Finnish prostate cancer screening trial British Journal of Cancer DOI: 10.1038/sj.bjc.6605512