Skip to main content

Together we are beating cancer

Donate now

Why don’t researchers know what causes pancreatic cancer?

by Henry Scowcroft | Analysis

6 August 2009

0 comments 0 comments

Microscope image of pancreatic cancer cells
Pancreatic cancer cells - image courtesy of the London Research Institute EM Unit

Pancreatic cancer is notoriously difficult to treat and, in many cases, therapy only extends life for a short time, rather than curing it.

We’re also a long way from properly understanding how our lifestyle and our genes interact to cause the disease – in contrast to many other forms of cancer, where the picture is less murky (although far from crystal clear).

For instance, researchers have known for some time that certain cancers are strongly linked to our lifestyle – smoking and lung cancer being the obvious example. And we’re also starting to understand how genetic variations affect people’s risk of a variety of cancers.

So why is pancreatic cancer proving so opaque to our efforts to understand it? A fascinating and thought-provoking editorial in the Journal of the National Cancer Institute explains why…

Its authors, Brain Wolpin and Meir Stampfer, both based in the US, ask the question, “why has progress been so slow in defining predisposing factors for pancreatic cancer risk?” They propose four key reasons:

1 – Relatively few people develop it

In the UK, about 7,600 people develop pancreatic cancer each year. This is comparatively low compared to, say, breast cancer (46,000 people) or bowel cancer (37,000).

In fact, the ‘age-standardised rate’ for pancreatic cancer in Europe is just 9 people per 100,000 in the general population. So it’s hard to carry out studies large enough to detect things that might affect who gets the disease. Large studies are also more expensive and time-consuming to carry out.

2 – Poor quality of data


Much of what we know about the causes of cancer comes from ‘case-control’ studies – where scientists examine the life histories or biology of people with and without disease, and look for tell-tale differences.

But this is difficult in the case of pancreatic cancer. As the report says, the outlook for many people with pancreatic cancer is quite poor, so there are limited opportunities to get detailed information about their diet and lifestyle. Also, this usually relies on memory – either that of the patient or of their next-of-kin.

But ‘remembering things’ is perhaps the least reliable way to gather hard evidence – our psychological make-up often leads us to over-emphasise certain memories, particularly under stress or bereavement. This is what epidemiologist call ‘recall bias’.

The authors make a related point when they note that, because the disease is often quite advanced,

samples … collected at the time of diagnosis are difficult to interpret, due to the large alterations in patient nutritional and functional status.

Pancreatic cancer patients can be extremely unwell, making it hard to tell if abnormalities in the sample are the cause of the cancer, or caused by it.

3 – Genetic faults are rare

The traditional way to find cancer genes is to look for families with a history of the disease, examine their DNA, find a faulty gene, and then look for faults in this gene in the wider population.

But this has so far proved a bit of a cul-de-sac in pancreatic cancer research. As the editorial notes,

…although studies offamilies with high rates of pancreatic cancer have identifiedseveral predisposing genetic variants, these variants are rareand contribute little to the overall population burden of pancreaticcancer.

In other words, the mutations found in families with pancreatic cancer don’t seem to be responsible for causing it in ‘non-inherited’ cases – i.e. most of them.

4 – No way to spot the disease early

Wolpin and Stampfer note that there’s currently no screening or early detection method for pancreatic cancer that would allow the disease to be spotted when treatments could be more effective.

So, going further, and rather pessimistically, they suspect that the final reason for the lack of progress in understanding pancreatic cancer risk is that there’s not much of an incentive yet. What’s the point of identifying risk factors for the disease, they ask, if there’s no way to detect it early?

This final point highlights the need for cancer research to proceed on all fronts – not just drug development, not just prevention research – if we’re to beat the disease.


So what do we know about pancreatic cancer risk?

Studies that look at pancreatic cancer rates over time, or in people who move from one country to another, clearly show that lifestyle has an important effect.

But what are these mysterious ‘lifestyle factors’? And why are they so tricky to pin down?

There are some things that large studies suggest can influence a person’s chance of developing the disease.

Tobacco – never far from the conversation where cancer’s concerned – counts for between a fifth and a quarter of cases, according to several studies. But if everyone gave up smoking tomorrow (and we can dream), we’d still be none the wiser about how the vast majority of cases arise.

Other things linked to the disease are generally involved with how the body generates energy – obesity, type 2 diabetes, and insulin resistance. And recently, a large study found evidence that the disease is more common in people who eat a diet rich in animal fats.

But generally, these things only have a small effect. Are we missing something?

The whole is more than the sum of the parts

A fascinating insight came earlier this year, with a paper published as part of the US NIH-AARP study, involving a whopping 450,000 people.

The study’s authors gave people scores from 0 to 5 depending on how well they stuck to five lifestyle factors – no smoking, limiting alcohol, eating a Mediterranean diet, keeping a healthy body weight and keeping active.

Those who scored the maximum of five points – i.e. whose lifestyle was generally very healthy – were nearly 60 per cent less likely to go on to develop pancreatic cancer than those who scored zero points.

This seems to be a clear indication of what Ed blogged about last year – that lifestyle is important as a whole package. A good diet on its own will only go so far – staying active, cutting down on alcohol AND keep a sensible bodyweight are all needed for the benefits to add up substantially.

So – going back to pancreatic cancer – Wolpin and Stampfer look to the future and see light at the end of the tunnel. Firstly, several large international studies, such as the US’s Pancreatic Cancer Cohort Consortium, are now in place to try to overcome the issues around patient numbers and data quality.

And secondly, greater sharing of samples – plasma, DNA and tissue – taken from consenting pancreatic cancer patients should speed up the rate of progress in understanding the disease biologically. This in turn should lead to ways to detect the disease at an earlier stage. As Wolpin and Stampfer write:

Evidenceof this potential is clearly visible in studies of other malignancies,such as [bowel] and breast cancer.

As well as continuing our current research into the disease, Cancer Research UK have, in our 5-year strategy, committed to invest in long term projects to tackle hard-to-treat cancers, specifically including pancreatic cancer.

So, although progress has certainly been slow in beating this deadly disease, the future direction of cancer research looks set to lead significant progress in the years to come.



Wolpin, B., & Stampfer, M. (2009). Defining Determinants of Pancreatic Cancer Risk: Are We Making Progress? JNCI Journal of the National Cancer Institute, 101 (14), 972-973 DOI: 10.1093/jnci/djp182

Jiao, L., Mitrou, P., Reedy, J., Graubard, B., Hollenbeck, A., Schatzkin, A., & Stolzenberg-Solomon, R. (2009). A Combined Healthy Lifestyle Score and Risk of Pancreatic Cancer in a Large Cohort Study Archives of Internal Medicine, 169 (8), 764-770 DOI: 10.1001/archinternmed.2009.46