This is why – nearly 15 years ago – we began co-funding a huge trial to find out whether regularly screening healthy women could spot ovarian cancers before they cause symptoms, when treatment is more likely to be successful.
The UK Collaborative Trial of Ovarian Cancer Screening (UKCTOCS) is run by Professor Ian Jacobs, who’s now based at the University of New South Wales Australia, and Professor Usha Menon from University College London, both of whom are gynaecologists who have worked on the problem of early diagnosis of ovarian cancer for many years.
And since it began, the trial has uncovered a whole range of important information about the potential of ovarian cancer screening, including how effective different tests can be at detecting cancers, and how women are affected emotionally by participating in ovarian cancer screening.
Today, the trial team have published one of their most long-awaited findings: their first analysis of whether ovarian cancer screening could actually prevent deaths from the disease.
But there’s a frustrating catch – while the data seems to show a possible benefit, there’s still a lot of statistical uncertainty in the results, and it’s far from clear how many lives it would save in practice.
So the researchers want to follow the women on the trial for a few more years to see for sure whether screening does actually save lives, and to show how that stacks up against the harms (which include unnecessary surgery and potential complications).
And while today’s result is important, there are several things that follow from it:
- We don’t think there’s enough evidence for the NHS to introduce a national screening programme at this stage
- We don’t recommend women use privately available ovarian cancer blood tests (including the test used in the trial, known as the ROCA test)
To explain why we think this, let’s look at the data – and the trial – in more detail.
What did the trial do?
In 2001, the UKCTOCS team set out to compare two different methods of screening to see whether either of them could spot ovarian cancer early, and whether this could save lives.
Between 2001 and 2005 they recruited more than 200,000 women across the UK aged between 50 and 74, and randomly assigned them to one of three groups.
The largest group, 100,000 women, formed the ‘control’ group – they weren’t given any screening, but acted as a comparison for the two screening methods to be measured against.
A second group, of about 50,000 women, were screened every year with an ultrasound scan, to look for abnormalities in their ovaries that could be a sign of cancer. If this showed something unusual, the woman had further investigations.
The third group, also about 50,000 women, were given an annual blood test to measure the levels of a protein called CA125, which is often found at high levels in women who have ovarian cancer (but which can also be raised for other reasons).
Previous screening studies had used a cut-off for CA125 levels (i.e. with measurements over a certain level triggering further tests) – but one big problem with using a cut-off is CA125 levels can vary widely from woman to woman.
So rather than using an arbitrary cut-off, the UKCTOCS researchers developed a more sophisticated method – the ‘Risk of Ovarian Cancer Algorithm’ (or ROCA, for short). This looks at how an individual woman’s CA125 levels change over time, in the hope that this can give a better idea about whether a woman might have an early-stage ovarian cancer.
After each annual blood test, the results were fed into ROCA, and depending on each woman’s score, one of three things happened: women with ‘normal’ ROCA scores continued with annual screening; those whose results suggested they were at ‘intermediate risk’ had their blood test repeated after 3 months; and those at ‘elevated risk’ had repeat blood tests, and an ultrasound, within 6 weeks.
It’s worth mentioning here that the ROCA test is now available privately, and several of those running the trial – along with Cancer Research UK (via our technology transfer company, Cancer Research Technology) – will receive a proportion of proceeds from its sale.
What did UKCTOCS show?
As we said earlier, the trial has already shown that the annual blood test using ROCA can detect ovarian cancers early. But what was missing was the crucial piece of information: does this actually make a difference? Can it detect ovarian cancers early enough to save lives?
In their analysis – published in The Lancet – the researchers initially looked at all fourteen years of data, to see if the number of women who’d died from ovarian cancer was lower in either of the groups that were being screened.
Using this analysis – which they’d proposed to do when the trial was first designed – the researchers didn’t see any clear cut effect from screening, using either method. As they write:
The mortality reduction did not reach conventional levels of significance in either [screening] arm in the primary analysis
But while the trial has been running, other studies have suggested that screening trials can take longer to show an impact than many had suspected. And so other trials, like the large US PLCO study, have subsequently taken this into consideration when analysing their results.
And when the team took this time lag into account, they got different results: it seemed to show that the risk of death from ovarian cancer was about 20 per cent lower among the women given an annual blood test. (The result was similar for women given an ultrasound, but more women had unnecessary surgery in this group).
Based on this, and assuming women are screened for between 7 and 11 years, the researchers estimate that for every 10,000 women screened every year with a blood test, about 15 lives could be saved.
But here’s the catch: their analysis also includes what’s known as a ‘confidence interval’ – effectively the size of the uncertainty in the result. So the ‘20 per cent reduction’ is actually in the middle of a much bigger range. And because of the low numbers of women who have so far developed and died from ovarian cancer in the trial – about 650 out of 200,000 – this range is from about 0 per cent up to about 40 per cent, and the ‘true’ figure is likely to lie anywhere inside it.
We’ve summarised this in the graphic below:
As the graphic shows, it’s not clear whether there is an actual reduction in deaths, and even if a reduction is there, how big it might be.
And this is vitally important to know, because screening also has downsides, most of which occur because no test is perfect – any test can have unintended consequences.
In ovarian cancer screening, these harms include:
- Unnecessary surgery from false alarms
- Complications due to surgery
- Anxiety caused by screening and false alarms
To put these into context, the UKCTOCS trial suggests that for every three women who had surgery to check for ovarian cancer (based on their blood test), two of them turned out not to have cancer. Scaling that up over a whole population adds up to a LOT of unnecessary surgery.
And this surgery isn’t a trivial procedure, it carries risk: about 3 out of every hundred women who had surgery also had ‘major complications’ because of it; this includes things like infections or damage to other organs.
There’s also the toll that taking part in screening can have on a person’s mental wellbeing. Previous findings from the UKCTOCS group show that, following unnecessary surgery for a false alarm, women were more likely to be worried and distressed, as were women who had more repeat blood tests as part of screening.
These types of psychological harms can have a real impact on a person’s mental health, and need to be taken into account when working out the balance of risks and benefits.
All screening programmes carry risks, including things like false positive and negative results – when people are told that they have cancer when in fact they’re cancer free, or are incorrectly given the all-clear. There’s also the risk of overdiagnosis (diagnosing and treating cancer that may never have gone on to cause any symptoms or do any harm in a person’s lifetime). This is something we’ve written about when discussing other screening programmes.
All of these potential harms can be big risks to manage. And how these harms stack up against the benefits influences whether a screening programme can be introduced, and who should be offered screening. In the UK, the National Screening Committee makes recommendations on which screening programmes should be available, based on a clear set of criteria.
But, based on the evidence we have so far on ovarian cancer screening, there isn’t enough evidence to say exactly what the benefits and harms of ovarian cancer screening are, nor how they stack up. It’s not clear that screening definitely saves lives, and though it’s clear that there are some harms, we need more research on these too.
The lack of a clear picture of harms and benefits of ovarian cancer screening is also why we would not recommend using commercially available screening kits.
This is why more follow up of the women in UKCTOCS is so important. After almost 15 years of research it can be frustrating to think we are so tantalisingly close to an answer, but as the researchers themselves acknowledge in their paper:
Further follow up to assess mortality reduction is required before firm conclusions can be reached
We’ll be waiting to see what longer follow up of the UKCTOCS trial will reveal about ovarian cancer screening, but until then we don’t know if ovarian cancer screening will tip the scales towards the benefits.
Fiona Osgun is a health information officer at Cancer Research UK
- If you have questions or concerns about ovarian cancer, we have information on our website, or you can contact our nurse information team.
Jacobs, I J., Menon, U., et al. (2015). Ovarian cancer screening and mortality in the UK Collaborative Trial of Ovarian Cancer Screening (UKCTOCS): a randomised controlled trial The Lancet : 10.1016/S0140-6736(15)01224-6