The physicist behind the MRI scanner...

Under the supervision of Evelyne Balteau, MR physicist at the Cyclotron Research Centre of the University of Liege, Elodie André has just completed a study showing that MRI technology can supply an in-depth knowledge of the human brain, the white matter it contains and the diseases that can develop within this matter. She focused on diffusion phenomena and diffusion imaging to examine the way water molecules diffuse in the brain. In particular, she focused on the diffusion tensor imaging (DTI) model, a simplified representation of diffusion in the brain, and its extension, the diffusion kurtosis imaging (DKI) model,  a promising technology.

In its own way, MRI has revolutionised the world of medical imaging by offering hitherto unknown possibilities for the exploration of the human body. This technique has become an integral part of the examinations recommended by doctors to diagnose many pathologies, as a complementary source of information to other techniques such as CT (X-ray Computed Tomography) or PET (Positron Emission Tomography). Yet none of us, when undergoing an MR scan, ever imagines that physicists, thanks to their hard work, constantly keep improving the interpretation of the data acquired by these machines!

As an engineer specialized in physics, Elodie André, under the supervision of Evelyne Balteau, MR physicist at the Cyclotron Research Centre of the University of Liege, has just carried out a study (1) – and completed her doctoral thesis (2) - which shows that MRI technology can supply in-depth information about the human brain, its white matter and the diseases that can develop within it... as long as one can (even) better decode the information resulting from the technology. As Evelyne Balteau explains, “this study is part of a research program intended to push the boundaries for the interpretation of MR images, while at the same time automating rigorous data processing for more reliable and reproducible results". It is an ambitious and daunting task...

Water, water everywhere!

It is worth taking a small trip back in time for those who have not been following the development of this technology or who cannot remember what MRI means. This powerful medical diagnostic tool provides three-dimensional images with high anatomical precision. A relatively recent technology, it has developed rapidly since its design in 1973 by Paul Lauterbur and Peter Mansfield (Nobel Prize winner for medicine in 2003). The first images of the human body were acquired in 1977. 

MRI is non-invasive and has no known side-effects. The technology is based on the physical phenomenon of nuclear magnetic resonance discovered by Felix Bloch and Edward Purcell in 1946 (Nobel Prize for physics in 1952). It consists of observing the nuclear magnetic resonance (NMR) of protons of water contained in our body which, as everyone knows, is made up 70 to 80 % water. In practice, MRI examines the response of nuclei subjected to an external magnetic field and electromagnetic excitation. The excited atom is the proton (H⁺), the principal component of water (H₂O). The energy absorbed by the proton during excitation is restored during the relaxation process, taking the system back to the equilibrium and generating the signal recorded by the MRI equipment. The recorded signals are analyzed by computer in order to reconstruct a 3D image of any area chosen beforehand.  

The intensity of the signal emitted by an element of volume (known as a voxel) depends on the concentration of water and Nuclear Magnetic Resonance (NMR) parameters (especially the relaxation time, indicating the return to magnetic equilibrium after excitation) of tissues encountered in the voxel considered and the method of acquisition applied (MRI sequence). The result is a three-dimensional image of the distribution of water in a patient’s body. The higher (lower) the intensity of a signal from a given point of the body, the whiter (darker) the point corresponding to the image is (respectively). The MRI sequence modifies the contrast between tissues. It is chosen according to the type of tissue that needs to be highlighted or the type of pathology that is being detected (such as tumors). 

Nonetheless, the contrast obtained can be insufficient to suitably differentiate the healthy parts of the body from those affected by a disease. A very simple way of influencing a signal in MRI is to increase the contrast, either by increasing the examination time to allow for more acquisitions, or by using a contrast agent (like gadolinium). This agent makes it possible to show the presence of tumors or other such pathologies more clearly. 

Everything moves...

As a major modern medical imaging technique, MRI is constantly under development. "These developments lead to a higher quality control and improved processing of the acquired data", says Evelyne Balteau. The work of Elodie André is part of this process. She focused on the quality of data obtained during the study of diffusion phenomena and diffusion imaging, which indicate the way in which water molecules diffuse in the brain producing a contrast in the image. As mentioned above, the results obtained from an MRI brain scan make it possible to observe the white and grey matter of the brain as a whole in a non-invasive manner. There is no need for surgery or risky biopsies, and no need to wait for post-mortem dissection results in order to have a perfect three-dimensional image of the brain! “In diffusion”, explains Elodie André, “we are more interested in the white matter”.  Another essential point: these structural studies of the brain indicate whether the white matter is healthy or not. This element, which completes the information collected by means of cognitive methods, is important for the diagnosis of many pathologies.

 "In order to clearly understand what diffusion involves on a macroscopic level, you might imagine a drop of ink added to a glass of water. The drop of ink diffuses into the water. The concentration of ink molecules which is dense in the ink drop, becomes homogenous throughout the entire glass of water and balances out without any extra stirring, explains Elodie André

What is examined here is therefore the diffusion of the water molecules themselves in the human brain. These molecules are not static. The direction, or directions they take are important. "In fact the molecules diffuse more along the fibres, because there are fewer barriers than perpendicularly (among others, the membranes and myelin). By observing this movement, it becomes possible to extrapolate the direction of nerve fibres. Thanks to repeated measurements in different directions and carried out over the entire brain, it is possible to characterize the diffusion in 3D. For each voxel (each volume element of the image), the diffusion tensor can be represented by an ellipsoid which can vary in size and is more or less stretched along the main direction of diffusion of the water molecules in this voxel",  the physicist continues. 

If the diffusion is equally probable in all directions, it is said to be isotropic and the tensor will be represented by a sphere that can vary in size, depending on the ease of diffusion (presence or absence of obstacles). In the brain, many obstacles can constrain the diffusion which will therefore be reduced (as compared to diffusion in a glass of water) and possibly not to the same extent in all spatial directions. "Thus, in the white matter of the brain, we can find ellipsoids along the direction of the fibres: in this case the diffusion is said to be anisotropic. One of our diffusion parameter of interest, which is called fractional anisotropy,  has a quantitative value. It varies from 0 (isotropic diffusion) to 1 (very anisotropic). By using this parameter to compare two groups of individuals, we can detect the area where the white matter is affected by disease. The result of these calculations and comparisons is often difficult to interpret when we are looking at a damaged brain area", explains Elodie André.

Removing complexity

The most standard model used to describe diffusion in the brain is the diffusion tensor imaging (DTI) model. Elodie André has been working on an extension of this model, the DKI or Diffusion Kurtosis Imaging model. "The diffusion can be described by various models of increasing complexity. The diffusion tensor is a very simple model yet diffusion in living tissues is complex. This complexity is related to the presence of blood vessels, fibres, cell membranes and variations in the permeability of the latter… And one of the parameters accounting for this complexity is the kurtosis. In the more basic DTI model, we assume that the molecule diffuses following a Gaussian distribution (bell-shaped curve). This is indeed the case for diffusion in a glass of water but is no longer the case in the presence of obstacles. The difference with the Gaussian curve is the kurtosis”, explains Evelyne Balteau.

As described by Elodie André, "the kurtosis technique makes it possible to quantify this non-Gaussianity, and therefore to indicate whether the diffusion at a given point follows a Gaussian distribution ( considered as a reference) or not. If the measured kurtosis is zero, it is because we are dealing with a Gaussian case of homogenous liquid. Kurtosis can be positive (with a curve that is narrower than the Gaussian curve) or negative (with a curve that is wider than a Gaussian curve)."

The mean kurtosis value reflects the underlying micro-structures. "To date there is no well-defined physiological model which makes it possible to simulate the results of kurtosis measurements due to the complexity of the problem and the high number of factors intervening.  There is no clear description or simple link between kurtosis and the micro-elements that form the structure of tissues (obstacles). However, measurements carried out in different situations, and among different populations, make it possible to see that the kurtosis varies locally in some areas of the brain", states Evelyne Balteau. The idea consists in using the mean kurtosis as a biomarker which makes it possible to more reliably diagnose diseases such as Alzheimer’s or Parkinson’s and to assess more accurately the stage the disease has reached thanks to the more sensitive and potentially more relevant information that this technique offers with regard to the integrity of white matter. 

"Kurtosis measurements among different populations make it possible to create a dictionary of kurtosis identity patterns for each population. Kurtosis could therefore be used as a diagnostic tool in conjunction with and complementary to other existing tools. Yet kurtosis is not the only biomarker of interest in diffusion MRI, other models are focusing on e.g. the 3D tracking of myelin fibers which connect the different parts of the brain to each other and to the rest of the body. Kurtosis is one aspect of the pattern, complementary to other features and parameters collected during cognitive assessments, and helping building a more general picture of a pathological situation and its evolution", states Evelyne Balteau.

Quiet, no noise please...

The major interest of Elodie André’s study lies in her contribution to image quality which makes it possible to go further, acquiring more valuable data in a higher number of diffusion directions. "For each diffusion direction, an image of the brain is obtained. The signal drops as a function of the diffusion weighting we impose: in DKI, a more important weighting is applied. The signal received is therefore weaker, more ‘noisy’. The objective of my research is to find a method for improving the quality of the signal and, therefore, to increase the quality of the result" 

Her idea consisted in extracting the signal from this noise by using what are called the moments of the distribution. “On one hand, we measure the noise, on the other, we remove it from the signal”, she clarifies. To succeed, she designed a quite rapid method. "It is not yet perfect” she says, “but it already gives a more precise signal”. The two experiments she carried out in her study confirmed this.

Eliminating dependency

The first experiment was carried out by positioning the head of the subject in different ways, more or less close to the MRI coil (the radio-frequency (RF) antenna collecting the signal). "This made it possible to produce a distinct spatial distribution of the signal-to-noise ratio (SNR) for each acquisition, because the elements of the RF coil have a sensitivity that decreases with distance. The signal collected is therefore higher in the areas of the brain closer to one of these elements and weaker for a more distant region. In principle, for a given individual, the results obtained in two different positions should be identical. But in practice, due to the difference in SNR , this is not the case. However, with my correction, we get the same parameters. This indicates that with this method, we are no longer dependent on the SNR and we see identical kurtosis maps after correction".

The second experiment was aimed at the variability of results between subjects, in order to verify whether the correction reduced the variability due to the low signal-to-noise ratio. Twenty-five individuals took part in the experiment. As figure 9 presented in the study shows, without correction, kurtosis is artificially higher and variability is more important (see the first column) in particular at the front of the brain (top of the slice), where the signal-to-noise ratio was lower. On the other hand, when high-performance correction methods were used (columns 2 and 3), the contrast was better (the white matter is represented in yellow) and the variability due to the low signal-to-noise ratio is reduced. "We can conclude that using the correction made it possible to discard differences due to noise, while demonstrating the finer and more interesting differences between groups of subjects. It is then possible to use fewer subjects to draw solid conclusions and avoid false positives (with erroneous results due to noise)", explains Elodie André.

Listen to the silence...

Elodie André has shown by means of her study that it was important to pay attention to the noise that influences final results, but she plans to go further. "It is easy to implement our correction method which is now used routinely at the Cyclotron Research Centre. The objective is to share it more widely so that everyone can use it." To succeed in this objective, in collaboration with other laboratories (in London and Lausanne), a toolbox is under development.  Its implementation should allow every department using MRI technology to make use of the method, and therefore to benefit from an optimal and shared noise correction technique without the need to be a physicist. 

Sharing optimal resources is the key solution to one of the current difficulties: in the absence of a shared state-of-the-art method, each group produces “local” results, depending on local data acquisition procedures and local data processing techniques, making it impossible to compare results obtained by the different group. "By sharing optimal methods, it would be possible to compare and even combine results from different laboratories more easily", argues Elodie André.

“We have on the one hand, engineers and physicists implementing and developing computer tools to get the best results from the technologies available for medical applications, and on the other hand, medical doctors with high practical expertise, basing their diagnosis on the visual analysis of images. Often one side seems completely opaque to the other. How many doctors do computer programming? How many physicists can recognize the corpus callosum or the substantia nigra? And yet, their interaction is indispensable. For example, the tool developed by the former to control image quality, will be based on the expertise of the latter to avoid any confusion between artefacts and real pathology. Interactions in multidisciplinary fields are essential!

Unfortunately, in hospitals, due to lack of time and means, little time is devoted to the adjustment of procedures and the training of medical teams for the use of tools developed by research teams. As a result, complex equipment can be underused, while outdated procedures keep being used routinely. It is true that these procedures and the expertise of medical doctors make it possible to rapidly produce a correct diagnosis. However, physicists, thanks to the support they can provide to medical teams, still have a bright future, reducing misinterpretation related to human errors, improving the reproducibility and the reliability of results, fully taking advantage of available techniques and using new ones as soon as they become available” 

(1) "Influence of noise correction on intra -and inter- subject variability of quantative metrics in diffusion Kurtosis imaging", Plos One, http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0094531
(2) Improvement of data quality for Diffusion Kurtosis Imaging and application to clinical neurological research, Elodie André, thèse de doctorat, Université de Liège.