PET scans may improve brain injury diagnosis

Wednesday April 16 2014

“PET scans could predict extent of recovery from brain injury, trials show,” The Guardian reports. Evidence suggests that the advanced scanning devices may be able to detect faint signs of consciousness in people with severe brain injuries.

The paper reports on a study that examined how accurate two specialised brain imaging techniques were at diagnosing the conscious state and chances of recovery in 126 people with severe brain damage.

The people were scanned using Positron Emission Tomography (PET) scans, which use a radioactive tracer to highlight cell activity, and functional Magnetic Resonance Imaging (fMRI) scans, which show blood flow in the brain, to demonstrate areas of activity. The results of these scans were compared for accuracy, with assessments made using an established coma recovery scale.

The study aimed to see if the scans could accurately distinguish between a minimally conscious state (MCS) – in which there is a chance of recovery – from other disorders of consciousness.

PET scans correctly identified 93% of people with MCS and accurately predicted that 74% would make a recovery within the next year. The fMRI scans were slightly less accurate, correctly identifying only 45% with MCS and accurately predicted recovery for just 56% of them.

The brain scans also showed that a third of the 36 people who had been diagnosed as unresponsive by the coma scale actually had brain activity consistent with minimal consciousness, and just over two thirds of these people subsequently recovered consciousness.

This small study suggests that PET scanning, together with existing clinical tests, could help accurately identify people with the potential to recover consciousness.

Where did the story come from?

The study was carried out by researchers from the University and University Hospital of Liege (Belgium), University of Western Ontario (Canada) and the University of Copenhagen (Denmark). It was funded by the National Funds for Scientific Research (FNRS) in Belgium, Fonds Léon Fredericq, the European Commission, the James McDonnell Foundation, the Mind Science Foundation, the French Speaking Community Concerted Research Action, the University of Copenhagen and the University of Liège.

The study was published in the peer-reviewed medical journal The Lancet.

It was covered fairly in The Guardian and The Times, which understandably looked at the ethical implications for decisions around switching off life support or giving pain relief.

What kind of research was this?

This diagnostic study looked at how accurate two specialised brain imaging techniques – Positron Emission Tomography (PET) and functional Magnetic Resonance Imaging (fMRI) – were at correctly distinguishing between different conscious states and predicting recovery in people with severe brain damage. This included both traumatic brain damage, which is typically caused by a severe head injury, and non-traumatic brain damage, which can have many causes, such as a stroke or heart attack.

The brain imaging results were compared with an established coma recovery scale, which is used in the assessment of people with brain damage.

PET scanning involves injecting a radioactive tracer (fluorodeoxyglucose – which is why the scans are often referred to as FDG-PET), which then produces colourful 3D images that show up cell activity in the body. It is most commonly used in the diagnosis of cancer. fMRI scanning shows up blood flow in the brain, which demonstrates areas of brain activity.

The researchers point out that in people with severe brain damage and a disordered level of consciousness, judging the level of awareness is difficult. In particular, the researchers aimed to see whether the scans could accurately distinguish between “unresponsive wakefulness syndrome” and a “minimally conscious state”.

People with “unresponsive wakefulness syndrome” (previously referred to as a vegetative state) differ from people in a coma in that they have their eyes open and show a normal sleep/wake cycle, but aside from this they show no behavioural signs of awareness. Meanwhile, people in a minimally conscious state (MCS) show fluctuating awareness and response to some stimuli (such as instructions or questions). 

The distinction between them has important therapeutic and ethical implications. As the researchers say, people in MCS are more likely to suffer pain and might therefore benefit from pain-relief and other interventions to improve their quality of life. They are also more likely to recover higher levels of consciousness that those with unresponsive wakefulness syndrome. In several countries, doctors have a legal right to withdraw artificial life support from people with unresponsive wakefulness syndrome, but not those with MCS.

The researchers also say that up to 40% of such patients are misdiagnosed by traditional clinical examinations. Brain imaging methods are now being developed to complement these bedside assessments, which can assess spontaneous brain activity or specific responses to mental tasks.

Such methods may help distinguish between people in an MCS and those with unresponsive wakefulness syndrome.

What did the research involve?

The researchers included 126 people with severe brain damage that were diagnosed at the University Hospital of Liège, in Belgium, between January 2008 and June 2012. They included people with both traumatic and non-traumatic causes for their brain damage. The results were:

  • 41 had been diagnosed with unresponsive wakefulness syndrome
  • 81 had been diagnosed as being in a minimally conscious state (MCS)
  • 4 patients had been diagnosed with locked-in syndrome, (a state where the person is fully conscious but behaviourally unresponsive). These people acted as a control group

The researchers carried out repeated clinical assessment of the patients using a behavioural test called the Coma Recovery Scale-Revised (CRS-R). This is thought to be the most validated and sensitive method for diagnosing disorders of consciousness. The scale has 23 items and is used by specialist staff to assess hearing, vision, motor function, verbal function, communication and level of arousal.

The researchers then carried out imaging using PET and fMRI scans, though not all patients were assessed with each technique (if the person moved too much to obtain a reliable scan, the procedure was left out).

  • For the PET, the person was injected with the imaging agent fluorodeoxyglucose before undergoing a scan. The scan from each person was contrasted against 39 healthy adult controls 
  • For the fMRI scan, patients were asked to do various motor and visuospatial tasks during the imaging session – including imagining playing tennis or walking into a house. The patterns of activity in the brain were also compared to those obtained in 16 healthy volunteers

12 months after the initial assessment, the researchers assessed the patients using a validated recovery scale (the Glasgow Outcome Scale – Extended). This assesses their level of recovery and disability and places the person into one of 8 categories ranging from 1 (death) to 8 (having made a good recovery). They also obtained an assessment of each patient’s outcome from medical reports.

The researchers then calculated the diagnostic accuracy of both imaging techniques, using the CRS-R diagnoses as the reference “gold standard”.

What were the basic results?

The main results:

  • PET scanning accurately identified 93% of people in a minimally conscious state (95% confidence interval (CI) 85-98) and had a high level of agreement with behavioural CRS-R scores
  • fMRI was less accurate at diagnosing a minimally conscious state (MCS), correctly identifying 45% of patients (95% CI 30-61) and had lower overall agreement with behavioural CRS-R scores than PET imaging
  • PET correctly predicted outcome after 12 months in 74% of patients (95% CI 64-81), and fMRI in 56% of patients (95% CI 43-67)
  • 13 of 42 (32%) of patients who had been diagnosed as unresponsive with CRS-R showed brain activity compatible with minimal consciousness on at least one of the brain scans; 69% of these (9 of 13) people subsequently recovered consciousness
  • The tests correctly identified all patients with locked-in syndrome as conscious

How did the researchers interpret the results?

They say the results show that, used together with the Coma Recovery Scale, PET scanning might be a useful diagnostic tool in disorders of consciousness. They also say it would be helpful in predicting which people with MCS might make a long-term recovery.


This is a valuable diagnostic study that tested how accurate PET and fMRI imaging are at distinguishing between different levels of conscious state and helping to predict recovery. 

Diagnostic assessments are traditionally made using bedside clinical tests – but as the researchers say, judging the level of awareness in people with severe brain damage can be difficult.

In particular, the researchers wanted to see whether the scans could accurately distinguish between people with “unresponsive wakefulness syndrome” and “minimally conscious state”, as distinguishing between these two states can have important therapeutic and ethical implications. The study found that PET scanning in particular had a high accuracy for diagnosing MCS and for predicting recovery time.

It’s particularly noteworthy that PET scans detected brain activity in some people who had been diagnosed as unresponsive by the standard Coma Recovery Scale test, and two-thirds of these people subsequently recovered consciousness.

However, the study has some limitations, including its small size, some missing data and possible differences between people who were and were not lost to follow-up. As the researchers acknowledge, their study used a complex method of statistical analysis, so there is a risk of false results.

At a practical level, these specialist types of imaging techniques are expensive and complicated to set up, so could have resource implications.

Analysis by Bazian
Edited by NHS Website