21 years of MRI
Professor John Duncan explains the difference that MRI has made to the diagnosis and treatment of epilepsy over 21 years, and how it has helped to change lives. Watch the video of his interview
When did you first realise the potential for MRI (magnetic resonance imaging) in the diagnosis and treatment of epilepsy?
Clinical MRI was really invented in the early 1980s and it was evident right away that it had the potential to show us the causes for epilepsy in the brain. We realised at end of the 1980s that this was potentially of huge importance for epilepsy surgery.
One in 3 people who develop epilepsy are not controlled with medication. If one can identify abnormalities in one part of the brain causing the seizures, one can target that area and plan to remove the piece you think is causing the epilepsy and so effectively cure the person of their epilepsy.
Did Epilepsy Society invest in a scanner straight away?
No, initially we bought time by using the MRI scanners at St Mary's Hospital and Great Ormond Street Hospital in London. In the early '90s we saw huge potential in the scans. We were finding abnormalities in people's brains that led them to having curative surgery.
Twenty-one years ago there were very few MRI scanners in the country. There was gross under provision compared to today's standards.
So at Epilepsy Society we had a fundraising campaign to create the building and raise money for the first 1.5 Tesla magnet which opened at our Chalfont Centre in Buckinghamshire in May 1995. We were given a charitable grant to offer MR scans to people from around the country. This was of huge benefit.
Has the scanner been upgraded since it was first installed at Epilepsy Society?
Since 1995 technology has moved on and our scanner has very much kept pace. In 2003 it was completely replaced with a more powerful 3Tesla scanner. In 2010 the magnet remained the same but the computing hardware was replaced.
As technology has got better, scanners have become more sensitive and we can identify abnormalities that we may not have previously seen. For this reason, people who we have seen in clinic 5-10 years ago, have been invited back and re-scanned with latest technology. This has given us the possibility of finding something subtle that wasn't picked up before.
Are MRIs only useful for people who are being considered for epilepsy surgery?
MRI scanning can also help to provide useful information for people who aren't candidates for surgery. For example we may be able to identify a malformation of the brain in someone. If we are able to show this to the person and their family, it can take a lot of the mystique out of the problem. It's reassuring for them to know their epilepsy is nothing that they've done or brought on themselves.
It is good to be able to explain that though surgery may not be a possibility because the abnormality is too wide or affects both sides of the brain, this is why medication is necessary. It does not take problem away but it does help to understand it better.
How exciting has it been to see deeper and deeper into the brain?
It has been fascinating. What gets me up in the morning is the ability to make new scientific advances and to apply them to people who one sees in the clinics. -real people with real problems. It is exciting to apply the latest scientific advances which allow us to be more precise and more specific. We can identify problems that we could not see before.
What exactly do the latest scientific advances allow you to see?
As well as looking at brain structure we can also see white matter networks in the brain. These form the critical connections or 'wiring' between different areas of the brain. If these pathways are unable to communicate, the brain may not be able to function properly even though separate areas are working. This helps us to plan our surgical approach so we don't cause harm.
We use functional MRI (fMRI) to see where different functions are occurring. We have developed this here over last ten years. For example we can see where in the brain a person thinks of words, names, pictures and describes objects verbally. We can pick out different parts of the brain responsible for different language functions.
We can see how some medications may affect these functions . One side effect from some epilepsy medication may be a difficulty in thinking of words. The medication may affect parts of the brain that deal with language function.
I have tried this myself. I have done a language fMRI test then taken epilepsy medication and done the test again. The difference the medications can make has been very striking.
The effect that a medication may have can vary enormously from person to person. It is very similar to the different response people have to alcohol. Some people have a huge tolerance of alcohol while others are not able hold their drink. If we are worried about side effects, we can give a person a test dose of a drug we are considering and then carry out a language fMRI. If we see a bad effect we know to steer away from that medication or to use it very cautiously. We can also ensure that people are on the look-out for any side effects themselves so they can report back and we can adjust their medication accordingly.
Many people with epilepsy report memory problems. Are you able to use MRI scanning to look at memory function in the brain?
Memory is an area of ongoing work here at the Chalfont Centre. In the last few years we've developed tests for memory for words and memory for pictures. We know how that works in the temporal lobes of the brain which are a common source of the type of epilepsy that doesn't respond well to medication. The temporal lobes are also the most common site for surgery. These tests are an important factor in deciding if that surgery is the right thing for a person.
This is very important work that has been taken up around the globe. Whether we can use this to determine affect of medication on memory function is not yet clear and that is one of the avenues of research to follow in the next couple of years.
Does knowledge of the risks involved make it harder for someone to decide whether to have surgery themselves?
The decision to undergo surgery is obviously a huge one. We can't say for definite ” this is what will happen to you, this will be the outcome. “ The best we can do is to give someone the odds. We have carried out brain surgery on a thousand people and we can use their scans and outcomes to work out the risks for other people.
For example if someone's seizures are coming from a part of the brain that is close to an area that is important for language, we might be able to say that surgery could give them a 50 per cent chance of being seizure free but they also risk having more difficulty thinking of uncommon words such as 'presumptuous' or 'unconsciously'. We would also be able to say whether their memory would be affected.
People can make up their own mind. Their decision will be determined by how hugely their life is affected by their epilepsy. I saw a gentleman in clinic who was falling to the ground every day, breaking his jaw. He could not go out of the house or get out of his chair without someone to help him.
After tests I was able to say to him that there was a 1 in 3 chance of curing his epilepsy and he jumped at the chance. On other hand these odds may not be so attractive to someone who experiences epilepsy as a funny feeling in the stomach and blank episodes.
MRI and fMRI enable us to sophisticate the odds much better in terms of benefit and risks. Different abnormalities on the brain carry different chances of a person still being free of seizures after five or ten years.
How much does computer technology guide surgery?
Further development over last five years has been to put all of the imaging data together into one 3D space. Previously we looked at all imaging separately. For example we would look at an MR scan for structure, a functional one for language, a SPECT scan for blood flow during seizures, a PET scan to show whether the brain used more or less amounts of sugar. We would then integrate all the different data mentally. We would also look at a brain scan as a series of slices.
In last five years, working with computer scientists, we are now able to look in 3D at all these scans on one computer model which we can make transparent or opaque.
This enables us to plan our surgical approach to remove an area of abnormality. It is very much like an aviation map. If you are flying from London to Bangkok you need to know where mountains are so you don't fly into them. When we are looking at the brain, we need to know where all the critical pathways are including those for language, memory, vision and movement, so that we don't harm them.
3D maps can be used in the operating theatre to precisely guide the surgery in exactly the same way as pilot use automatic pilot navigate route. This is 3D multi-modal imaging (3DMMI).
What other applications does 3DMMI have in brain surgery?
Half of the people who are considered for brain surgery won't be suitable. In a quarter of people all the information hangs together to recommend surgery, without needing further tests. In the remaining quarter we may not be exactly certain where the seizures are coming from.
In these cases we need to put recording wires into the brain to pinpoint which part of the brain is causing the trouble. We place 8-15 recording wires into the brain through very small holes. These must be precisely navigated so as not to damage any arteries or veins as this could cause haemorrhage. There is not as great a risk of damaging one of the eloquent pathways with a single wire as long as this is done in a limited way.
Using computerised 3DMMI, we can find the best trajectory to get from the skull to the target. We have developed a computer programme for doing this very swiftly and accurately. We are now developing a small robot that will execute trajectories with much greater accuracy and speed.
MRI has helped us to develop a whole pathway from initial diagnosis to working out the causes of epilepsy and whether intervention should be surgical or medical. We use imaging to help determine whether surgical treatment should be an early consideration in the treatment pathway.
We can design how electrodes should be passed into the brain safely, plan surgical resection so the surgeon knows exactly which part must be removed to give the person the maximum chance of doing well with minimum chance of causing harm.
After surgery the person comes back for a scan 3-4 months later and we can see exactly what has been taken away and what is still there. So if person is still having seizures we can discuss whether it might be possible to extend thepart that has been taken away. We We try as hard as we can get plans right from the start, .ut sometimes there is a trade off . For example, we may think seizures in a particular person are coming from the left temporal lobe. That is the part of the brain that is important for language and memory for words.. If that person undergoes a big operation, the chance of being seizure free may be higher but the chance of damage in memory and language could also be higher. We might agree to make a modest operation and hope that works. If it doesn't and the seizures continue, we might extend the operation in hope of success. Once you have taken a piece out, of course, you can't put it back.
How do you feel when you are able to reduce or cure seizures in someone whose life is highly affected by their epilepsy ?
It feels fantastic. It is a huge reward to see people come back well. If I had to pick out a small group of people in particular it would be people who have had seizures in teenage years who haven’t been able to get to college but who have then had surgery, been free of seizures and then trained as nurses or doctors and are now delivering healthcare. It is wonderful to see the transition.
What is the future for MRI?
The technology continues to move on. The main steps over the next five years will be to invest in a more powerful magnet. 7T MRI scanners are just coming in although they are not ready yet for full routine clinical use. It is a bit like a car with a very big engine. It may behard to control and more unstable.
However, over the next 2-3 years we will see 7T MRI scanners that are stable enough for reliable clinical use. This will give us more power and enable us to identify smaller abnormalities in the brain. With fMRI, it will help us develop our work further in understanding the different functions of each part of brain.
In parallel to this, we will see improvement in electronics and computing power. Scanners nowadays are hundreds times stronger than in earlier days so we are able to carry out much more sophisticated analyses. We envisage there will be continued advances in speed, precision and range. This will help us build up more and more detailed pictures of the brain.
We will also see the linking of all the MR data with new genetic data. We are just at the threshold with this development and we hope to start seeing results.
For example, certain genetic make-ups are associated with abnormality of structure in different parts of the brain and this is a field that is really going to open up in the coming years.