Cancer Research UK on Google+ Cancer Research UK on Facebook Cancer Research UK on Twitter

Let's beat cancer sooner

An MRI scanner and brain tumour images

Advances in imaging techniques are helping to analyse childhood brain tumours

Despite the huge progress that’s been made in treating many types of childhood cancer, brain tumours are lagging behind. Although these diseases are relatively rare – and the chances of surviving have increased over the past few decades – brain and spinal cord tumours are the leading cause of cancer death in children, claiming more than one hundred young lives every year.

One of the major challenges in treating these tumours in children is working out what’s going on deep inside their growing brains. There are many different types of brain tumour, each requiring different treatment approaches and with varying chances of survival. But despite advances in imaging, such as MRI scanning, that help doctors see inside the ‘black box’ of the skull, it’s still hard to figure out exactly what sort of tumour a child has without resorting to invasive surgery.

In two related papers, published earlier this year in the European Journal of Cancer, Professor Andrew Peet and his team at the University of Birmingham have taken a step forward in developing a non-invasive technique that could help doctors to diagnose childhood brain tumours more accurately before surgery, as well as helping them choose the best way to treat them. And Cancer Research UK’s support, through our Imaging Programme, was essential for making it happen.

Let’s take a closer look at what they found.

What’s the problem?

Childhood brain tumours are a complex and diverse group of more than a dozen diseases affecting different tissues in the brain in various ways. And while an MRI scan can provide a lot of information about where a tumour is lurking, it often can’t help doctors pin down the exact type, nor how aggressive it might be.

The next step towards getting a definitive diagnosis is usually surgery to remove the tumour, which can then be analysed by a pathologist to give a more detailed answer. But in many cases, it would be incredibly helpful to have more information about the tumour before heading into the operating theatre. This would enable doctors to plan the best possible approach and decide whether other treatments, such as chemotherapy, are likely to be needed too.

And there are also implications for the extent of the surgery itself. For example, some types of brain tumour need to be completely removed to give the best chances of surviving, while in other cases it’s OK to leave a small amount of tumour as long as it’s mopped up later with chemotherapy and radiotherapy. The more tumour a surgeon has to remove, the greater the chances that healthy brain tissue will end up getting damaged. But it’s hard to know how far to cut if accurate information about the type of cancer is only available in hindsight, once the pathology reports come back from the lab days later.

There have been some recent advances to help surgeons make these kinds of decisions in the operating theatre, such as making tumours ‘glow-in-the-dark’, but knowing more in advance would be a huge help.

Wobbling molecules to analyse tumours

To try and shine light inside this black box, Professor Peet and his team are using a technique called Magnetic Resonance Spectroscopy (MRS), which allows them to analyse the levels of certain molecules inside tumours. It relies on the fact that molecules ‘wobble’ in particular ways when they’re in a strong magnetic field. This wobbliness is measured and analysed using complex computer programmes, building up a characteristic ‘fingerprint’ called a spectrum, which looks a bit like the ridges and valleys of a mountain range:

An MRS spectrum

A typical MRS spectrum from a brain tumour, taken from Peet et al (2012) Nature Reviews Clinical Oncology

Researchers have already used this technique to analyse brain tumours in adults, but it isn’t usually used in children’s cancers, which are quite different from those in grown-ups.

Because childhood brain tumours are rare, the researchers had to work together with doctors in 10 treatment centres in Europe and South America to gather enough suitable patients. In total, they managed to analyse 97 children with a range of different types of cancer, including 11 with ependymoma, 29 with medulloblastoma, and 38 with pilocytic astrocytoma, as well as a handful of children with other less common tumour types.

The scientists found that each of these three major tumour types had a distinctive spectrum, which could be easily identified. What’s more, the fingerprints of tumours in a particular part of the brain known as the posterior fossa looked slightly different from those growing elsewhere.

But it’s not just in diagnosing the particular type of tumour in a child where this technique could be useful. Patients who appear to have the same type of cancer can respond very differently to the same treatment. However, at the moment it’s difficult for doctors to predict which children are likely to do better and benefit from milder treatment, and which might do worse, needing a more intensive approach.

Peet team

Professor Peet and his team in Birmingham

Studying 115 children who had come to Birmingham Children’s Hospital with a brain tumour, Professor Peet and his team found that their MRS technique can also be used to predict how individual patients would fare after treatment. They discovered that children whose tumours have certain chemical fingerprints had a relatively good outlook, while those with different signatures were likely to do worse – in particular those whose tumours contain lots of fatty molecules.

What next?

These results tell us that MRS is promising as a non-invasive way of analysing childhood brain tumours, providing information that could help surgeons and doctors make better decisions about how best to treat their patients. And the team’s work also reveals more about the molecular changes that are happening deep within brain tumours, helping to explain how they develop and grow.

This is the largest study ever undertaken using MRS for analysing childhood brain tumours. And although there still needs to be a bit of tweaking on the technical side of things, there’s a good argument for incorporating the technique in further large international clinical trials to bring benefits to more children in the future.

MRS is available already in many major hospitals and can be used in addition to MRI scanning in helping to diagnose and treat children with brain tumours. However, the sophisticated analysis developed by Professor Peet and his team is not yet ready to be used routinely in the clinic, so it won’t change how patients are diagnosed and treated right now.

Childhood brain tumours are still a huge challenge for doctors and scientists, but research is our best way of beating them. Cancer Research UK funding played a vital role in this project, through our Imaging Programme in Birmingham.

In turn, that funding simply wouldn’t be there without the generosity of our supporters who –  like us – want to see the day when no child loses their life to cancer.  We’ve got a long way to go, but every step takes us in the right direction.


Wilson M. Et al (2013). Magnetic resonance spectroscopy metabolite profiles predict survival in paediatric brain tumours, European Journal of Cancer, 49 (2) 457-464. DOI:

Vicente J. et al (2013). Accurate classification of childhood brain tumours by in vivo 1H MRS – A multi-centre study, European Journal of Cancer, 49 (3) 658-667. DOI: