T. ScARAbino, F. Di Saile, F. Esposito, M. Tosetti, M. ARmillottA, R. AgAti, U. SAlvolini
Diffusion magnetic resonance imaging (DWI) studies the diffusivity of water, i.e. the random microscopic movement of water molecules, or Brownian motion, induced by thermal energy.
MR is the only imaging technique enabling non-invasive observation of diffusion in vivo, and provides high-resolution images of deep-seated organs without interfering with the diffusion process itself.
Although intriguing, this description still does not however explain the strong and growing interest raised by DWI in the scientific community and among clinicians in the last decade. The reason for its success lies in the astonishing detail of the data provided. In fact, based on the diffusion-driven incoherent motion detected with DWI, the water molecules probe tissue structures on a microscopic scale that is well beyond the resolution commonly achieved with other techniques, providing information on the microscopic and even the molecular structure of brain tissue that is virtually unobtainable using conventional MRI. In the diffusion times typically used in DWI acquisitions (in the order of tens of milliseconds), water molecules diffuse in the brain over distances of microns or tens of microns, interacting with microscopic tissue components such as cell membranes, myelin and macromolecules.
Since the typical voxel size of a DWI image is a few cubic millimetres, DWI analyses an overall diffusion effect representing statistically the distribution of the Brownian motion of water molecules, and provides unique data about the structural organization of tissues at the microscopic level .
Furthermore, mounting evidence indicates that accurate selection of sequence parameters will allow DWI to become sensitive to specific subcompartments of brain tissue, providing hitherto inaccessible tissue dynamics information.
Unlike most tissue parameters that can be studied with MRI, such as T1 and T2, the physical process of DWI is totally independent of the MR effect or the magnetic field. For this reason DWI acquisitions entail no specific requirements in terms of magnetic field strength. Nonetheless, since the diffusion-weighting effect basically consists of an attenuation of the MR signal, DWI poses peculiar demands in terms of signal to noise ratio (SNR), thus warranting the use of high-field MR units. High-field magnets are essential for advanced applications of DWI, such as tensor imaging (DTI) and tractography (TI).
Although most soluble molecules exhibit Brownian diffusion, water is by far the most convenient molecular species to be studied with DWI, due to its concentration in biological tissue, the highest natural abundance of its MRI visible nucleus ('H), its small dimension permitting easy access to most biological compartments, and the water-based organization of biological tissues, where most molecules show some degree of hydrophi-licity.
Other metabolites may also be investigated through the study of diffusion of'H and other nuclei. Despite their much lower intrinsic SNR, non-water and non-'H studies can provide valuable and interesting information about specific biological compartments governed by different principles to those regulating water diffusion . However, such unfavourable SNR mandates the use of high-field MR units for non-water diffusion studies.
The basic principles of diffusion MRI were introduced in the mid-1980s , combining MR strategies with earlier concepts applied to encode molecular diffusion effects in the NMR signal by using bipolar magnetic field gradient pulses . The movement of water molecules is affected by various tissue components (cell walls, membranes, intracellular organs, macro-molecules). Diffusion may be restricted in all directions (isotropic diffusion) or in one particular direction within voxels (anisotropic diffusion), as in structured tissues (e.g. cerebral white matter).
Molecular diffusion can be evaluated using MR techniques with ultrafast sequences sensitized to movement. In particular, diffusion weighting is often combined with spin-echo echoplanar sequences. The sensitization to molecular diffusion is implemented using two strong and fast bipolar diffusion gradients applied symmetrically before and after the 180° radiofre-quency (RF) refocusing pulse, resulting in proton de-phasing and subsequent rephasing. The water molecules diffusing freely during and after application of the first dephasing gradient are not completely repha-
sed by the second gradient, unlike what is seen in stationary tissues. This process results in a signal intensity attenuation that is proportional to the amount of molecular displacement.
Was this article helpful?