Diffusion Tensor Imaging MRI technique helps produce satisfactory images of nerve bundles

TU DelftNov 15 2010

Standard MRI scans have so far been unable to produce satisfactory images of nerve bundles. However, this is now possible by an MRI technique called Diffusion Tensor Imaging (DTI). Matthan Caan succeeded in improving the DTI method during his PhD research at TU Delft, enabling him to produce more accurate images of the damage that radiotherapy and chemotherapy cause in young leukaemia patients.

Alzheimer's disease
With MRI, images can be obtained of various parts of the body. Unfortunately, these scans are not sufficiently refined for the imaging of nerve bundles. Matthan Caan concentrated on adjusting the conventional MRI technique via diffusion tensor imaging (DTI), so that better images of the nerve bundles could be obtained.
Via DTI, a method that is under considerable development, changes in the so-called white matter of the brain can be mapped. These changes can be caused by the ageing process or by Alzheimer's disease, among other things.

White matter
The white matter contains the connections between the nerve cells, and the principle of DTI concerns the movement of water molecules within the white matter. The molecules can move easily along the length of the nerve bundle, but not so easily on the perpendicular. This is because they are obstructed by the wall of the nerve cell and a protective layer (of myelin). Using DTI, the difference in the water molecules' freedom of movement allows conclusions to be drawn on the health of the nerve bundles in the white matter.

AMC
Caan succeeded in improving the DTI method during his research in collaboration with his supervisor, Dr Frans Vos. He demonstrated this at the Academic Medical Centre in Amsterdam, through research into the side effects of chemotherapy on the brains of young cancer patients. 'Usually, the scans are examined point by point. We combine measurements over a much bigger range, examining the whole nerve bundle.'
This allowed Caan to show that a lower dose of chemotherapy treatment causes much less damage to the brain. 

Self-learning
'In a conventional DTI analysis, all the voxels (3D pixels) are considered independently', says Caan. 'However, we expect that, because of the many connections in the brain, more areas are involved in complex brain diseases, such as schizophrenia. We have therefore introduced a self-learning system that can perform a comparative study of these areas.'

Rotterdam Scan Study
DTI images could eventually be used for early diagnosis of Alzheimer's disease, for example. This would be possible by analysing changes in the white matter among large groups. Extensive population studies are necessary in order to allow differences between groups of patients to be shown statistically.

In the coming years, the Delft research group, Quantitative Imaging, led by Professor Lucas van Vliet and in which Caan carried out his research, will benefit from the Rotterdam Scan Study, in which 5,000 people are participating. It is expected that a substantial statistical basis will be available in five years