Pancreatic cancer cells - image courtesy of the London Research Institute EM Unit
Pancreatic cancer has one of the poorest survival rates of any common cancers, with only three per cent of patients still alive five years after diagnosis.
And despite our advances in understanding the disease, this figure has barely budged in the last 40 years.
One reason for the lack of improvement is pancreatic cancer often doesn’t have early-stage symptoms, and so is often very advanced by the time patients are diagnosed.
In fact, more than eight out of 10 people with pancreatic cancer are diagnosed when their cancer has already spread, which means that surgery to remove the tumour is sadly no longer an option.
But at the moment, there are no reliable ways to detect the disease early – and one is desperately needed.
So, naturally, when one catches wind of any advances in early detection in the pancreatic cancer field, there’s always a certain amount of raised hope and excitement.
Possibly a bit too much excitement, in the case of today’s story.
So what’s the actual story?
A step forward
While there is a protein in the blood, CA19-9, which some pancreatic cancers secrete, it’s not produced by all types of tumours. This is why at the moment it‘s used by doctors alongside other tests to help diagnose patients who are already suspected of having the disease.
And while some research has shown CA19-9 levels might be raised in patients up to two years before they’re diagnosed, it’s not yet accurate enough to use to diagnose patients.
So a team of scientists at Barts Cancer Institute, Queen Mary University of London, decided to look for proteins in urine that could help identify patients with early-stage pancreatic cancer.
One of the advantages of a urine-based test is that urine is, in general, much less complex to analyse and less invasive to collect.
To find out if there were any measurable signs linked to cancer, the team – led by Dr Tatjana Crnogorac-Jurcevic – looked at levels of 1,500 proteins in urine samples from 18 patients – six of whom had pancreatic cancer, six who had a benign inflammatory disease of the pancreas called chronic pancreatitis (which can increase your risk of developing pancreatic cancer) and six healthy individuals.
They noticed that three proteins – LYVE-1, REG1A and TFF1 – were higher in pancreatic cancer patients.
These proteins, they suspected, could potentially be used to help distinguish patients with early-stage pancreatic cancer from healthy individuals.
And further studies they carried out, in a larger number of patients, showed tentative signs that this might be the case.
But before we get ahead of ourselves, there’s still a long way to go before we know if this research could lead to a test that would help detect pancreatic cancer early.
The study only validated how well these three proteins predicted things using 488 urine samples, from 192 patients with pancreatic cancer, 92 patients with chronic pancreatitis and 87 healthy volunteers.
This analysis showed that urine from patients with pancreatic cancer showed elevated levels of each of the three proteins when compared to healthy individuals. So in combination, the set of proteins seems to be fairly good at differentiating healthy individuals from patients.
But – crucially – the protein levels couldn’t reliably tell the difference between chronic pancreatitis and pancreatic cancer patients. And this could be a fairly big stumbling block for a useful test, and might have implications for who it could potentially benefit.
So before we go galloping down the road to clinical tests, there are some key questions that still need answering.
The first one being: even if this test were proven reliable in future studies, who would you offer it to?
The test is unlikely be used on the general population due to the relative infrequency of pancreatic cancer, which would mean it wouldn’t be time or cost-effective. A useful test for pancreatic cancer is rather more likely to be targeted towards groups at higher risk, such as those with chronic pancreatitis, or with a family history of the disease.
Any test that aimed to spot pancreatic cancer would also need to be able to reliably tell the difference between cancer, and other non-cancerous conditions like chronic pancreatitis, or diabetes. And, as mentioned previously, this ‘three protein’ method can’t yet reliably discriminate between those with chronic pancreatitis and pancreatic cancer is much less reliable.
Another question we have is whether the protein levels are still as accurate at predicting cancer when the control group (healthy people) is as old as the pancreatic cancer patients?
The median age of healthy patients on the study, and those with pancreatic cancer, differed by over 20 years. As the risk of cancer increase with age, it is crucial to see if the results are the same when comparing apples to apples.
To address some of these questions, the team is now planning further research on urine sample from high risk groups, as well as collecting more long–term data to see if the proteins are present in urine in the time between the genetic changes that could cause the cancer to develop, and diagnosis.
But the key thing that needs to be proven, in future research, is whether the ability to detect pancreatic cancers using this test translates into earlier diagnoses and lives saved.
But the question of whom could benefit from such a test still remains.
This story touches on two hugely important areas – pancreatic cancer, and the search for ways to diagnose cancer earlier. These are both areas we’re prioritising in our new research strategy, so it’s always welcome when new findings move things forward.
But we need to be careful not to over-egg things, and raise false hope of new tests, when there’s such a long way ahead.
- Crnogorac-Jurcevic, Tatjana. ‘Identification Of A Three-Biomarker Panel In Urine For Early Detection Of Pancreatic Adenocarcinoma’. Clinical Cancer Research 21.15 (2015): 3512-3521.