Diffusion Magnetic Resonance Imaging - Assignment Example

The advancements in imaging modalities were able to increase the understanding about the various biological systems and how these work together in living organisms. In terms of the resolution of images, Magnetic Resonance Imaging (MRI) is considered to be one of the most advanced, mainly due to its use of magnetic field in detecting the various differences in organ and tissue densities (Jones, 2011). With the discovery of peculiar characteristics of water and how it affects the movement of molecules through diffusion, aside from relying on the magnetic field generated by various molecules in cells, the additional characteristics of particle diffusion in living cells were able to contribute to the greater advancement of the use of MRI in the medical field. Introduction to Diffusion In living systems, diffusion is a transport process wherein molecules or particles randomly move and mix without the requirement of large quantities of motion (Johansen-Berg and Behrens, 2009, p.3). This process happens slowly through time, as the result of the combination of Brownian motion among particles along with the interaction of the properties of the surrounding medium, which in living systems mostly consists of water (Moritani, Ekholm, and Westesson, 2009). Through Brownian motion the particles become dispersed outward, and the speed of the dispersion can give observers information on the properties of the medium given the knowledge of what kind of particles was dispersed in it. Also, particles will continue to diffuse freely until an area of restriction called barrier is reached, which could affect the measurement of the diffusion rates through compromising the accuracy of the distance travelled by the diffused particles. This idea of harnessing the properties of fluids in living systems became one of the basic tenets of Diffusion Magnetic Resonance Imaging (dMRI), wherein the interactions between the mobility of water protons to dispersed particles such as macromolecules or the permeability of surrounding membranes are measured to provide images which give out the size, shapes, and orientations of scanned portions of the human body (Cercignani and Horsfield, 2001). The rates of the diffusion of various particles suspended in any kind of medium in comparison to the diffusion rates of the medium’s particles can be measured using the apparent diffusion coefficient (ADC), which can then be used as data to provide detailed imaging of structures and processes in living systems such as the permeability of membranes and equilibrium that would not be otherwise detectable in other methods such as X-rays or computed tomography (CT) scans (Luypaert, et al., 2001, p.20). ADC can be measured and analysed through the use of the following steps: calibration of gradient coil and determination of b-value using a standard (e.g. water); calculation of b-value from gradient waveforms in the diffusion-weighted pulse sequence; using one image with no diffusion gradient and repeated scanning using various b-values (including registration of the same spot at different angles); and post processing for motion correction. Since the discovery of the anisotropic property of water in living systems wherein diffusion rates of particles vary in all directions, these differences in diffusion rates of particles as well as the fluids within biological systems have been observed through early studies using MRI, and through these variation in diffusion rates scientists were able to view not only normal cellular structures but also distinguish some peculiarities in cells such as oedemas or blockages in blood vessels (Mori and Barker, 1999). In combining the use of diffusion rates along with the non-ionising nature of MRI scanning, dMRI is seen as advancement to imaging technologies by providing images with higher resolution and detail in comparison to other image modalities used today. How MRI Sensitises Diffusion The process of diffusion reduces the strength of signals recovered, and this property is harnessed by using diffusion rates to differentiate highly mobile tissues from those with low mobility. In the process of dMRI, effects of diffusion are usually removed or sensitised by the use of spin-echo, wherein two large diffusion sensitising gradient pulses induce phase changes in the water signal in a single direction: the first gradient pulse reorients various spins that lie on various positions within the magnetic field gradient, which then accumulate different phase shifts; while the second gradient pulse aims to reverse the first pulse by 180°, so that the net phase shift would be zero, cancelling out the effects of the first pulse (Cercignani and Horsfield, 2001, p.S12). Using steady-state free procession (SSFP) sequences that utilise low angle rf pulses produce multiple echoes which reduce the signal decay, allowing for the reading of signals from gradient echoes, spin echoes and stimulated echoes, providing a faster way of diffusion imaging using moderately strong gradients. Static molecules will have a net phase change of zero since there were no movements during and after the pulses were released, while non-static molecules will experience non-zero phase changes which then contribute to a degree of signal loss or attenuation (Cercignani and Horsfield, 2001, p.S12; Reiser, et al., 2008). This signal loss contributes to the differences that each type of molecule exhibits, which in turn help in building the image through variances in signals that were reflected by each particular molecule the magnetic field pulses ran into (Jones, 2011, p.59). Such differences can then be used to identify whether or not tissues being scanned are still functional or not, since any kind of swelling or oedema can cause restrictions in the diffusion rates and areas, causing significant differences between the ADC of normal and abnormal tissues (Reiser, et al., 2008, p.318). Sensitisation is often used in dMRI in order to label the spins within the diffusion system based on whether the particles are static or not, and in doing so enables diffusion-weighted imaging (DWI) much more sensitive to density and diffusion rates of particles within particular scan sites (Jones, 2011). For instance, attenuation of the signal strength can result from low-density and high diffusion rates within tissues surrounded by higher amounts of water, while densely-packed tissues give off considerably stronger signals and in a way delineates the portions where membranes or other lattices are present (Hagmann, et al., 2006). The degree of diffusion sensitisation is directly proportional to the weakness of the, wherein higher values of gradients will have higher degrees of sensitisation due to the application of either strong or longer diffusion gradients, leading to images with more detail and information due to extended echo times that decrease signal to noise ratio. B-values also have an effect on the weightings of the images generated through dMRI. Achieving Different Weightings In order to provide contrasts within the tissue images, signals are changed or weighted, in such a way that T2-weighted images show sharp contrasts between different types of tissues, depending on the b-value (s/mm2) or degree of diffusion weighting used (strong or weak magnetic diffusion gradient) (Jones, 2011, p.84). This makes the b-value highly dependent on both time and diffusion gradient amplitude, wherein long times in between the generation of pulses as well as having very strong diffusion gradients can greatly attenuate detected signals. But by increasing the b-value, adequate signals that have a low noise, as well as short spacing between repeated echoes are created, which could then characterise the level of induced sensitivity on the diffusion within the target organs being scanned (Johansen-Berg and Behrens, 2009; Reiser, et al., 2008). B-values can be adjusted by either applying the diffusion gradients for a long time or applying them strongly, making the images T2-weighted as well. However, there must also be sufficient reduction of diffusion effects using the CarrPurcell-Meiboom-Gill sequence (CMPG) since the varied outward movement of particles could affect the strength of the signals recovered arising from build-ups in phase accumulation. Since the b-values can affect the outcome of the clarity of images being scanned, aside from using just a single intensity, at least two are used to vary the weighted images and show the contrasts between the different diffusion gradients being used (Reiser, et al., 2008, p. 136). In using at least two different b-values during a scan, greater image clarity can be achieved since in the case where either one of the two b-values were unable to detect something, the additional scans using other b-values can compensate for this effect. This is particularly relevant in DWI since the images are mostly run under programs that utilise statistical analyses in order to generate the computerised images according to the data obtained (Mori and Barker, 1999). By choosing the most appropriate b-values for the type of tissue with a particular ADC value and pulse direction, for example in tissues with higher density and hence would have lesser diffusion rates (e.g. liver tissue) b-values used would be lower in comparison with tissues having higher rates of diffusion due to higher water content (e.g. brain and spinal tissue) which in turn have higher diffusion rates, this decreases the chance that the images would have low signal-to-noise ratio (Koh and Thoeny, 2010). The use of an inappropriate b-value could affect the DWI by increasing artefacts which, aside from the nascent effects of proton density leads to a hazy or distorted image (Luypaert, et al., 2001). Additional processes such as repetition of scans with varying b-values, applying tractography to define diffusion direction, applying registration (aligning different images to same coordinate space) as necessary, as well as reducing image noise through post-processing are also employed to improve the quality and resolution of the final image. Measurement of Diffusion Diffusing particles usually spread outward in a circular fashion, but normally this only happens in isotropic media outside living systems. In the case of biological and living samples, the diffusion of particles and water molecules are rather ellipsoid due to the surrounding noise that prevents further outward movement and lagging of some particles (Basser, 1995, p.337; Koh and Thoeny, 2010, p.5). Also, due to the presence of noise and other particles that prevent the uniform diffusion of particles, instead of measuring the whole diffusion ellipsoid only once, it is more applicable to take various sample measurements in intervals by varying the gradient strength and allowing it to move only in one direction, in addition to the use of six separate measurements of diffusion via six orthogonal directions along individual directions (x,y,z) and simultaneously between x and y, y and z, and z and x axes to measure the size, shape, and orientation of the diffusion ellipsoid, mainly because there are expected changes in signal intensity due to the properties of the sample’s structure, as well as the expected large loss of signal due to diffusivity. (Basser, 1995, p.338). While it is possible to use other methods of measuring diffusion such as radioactive or fluorescent tracers, as well as infrared spectroscopy, Rayleigh scattering and laser and neutron scattering, these methods are deemed unsuitable for in vivo systems due to a great deal of invasiveness in comparison with the use of magnetic resonance, making MRI a better option for use in vivo systems. A two-pulse gradient is employed as the default diffusion sensitive sequence, known as the Stejskal-Tanner pulse field gradient, which exposes the medium to a magnetic field with differing strengths that affect spins depending on their orientation along the gradient’s axis (Hagmann, et al., 2006, p. S210). This in turn labels the spins and can be monitored as tracers, because spins that have moved during the first of the two gradient pulses will not return to their initial state and by undergoing a total phase shift will decrease the signal intensity of the measured spectroscopic signal (Hagmann, et al., 2006, p. S212). The addition of magnetic field gradients in MRI was significant since without it the amplitude and duration could not create significant changes in the signals detected, and hence would not be able to show clear contrasts within the images (Jones, 2011, p.58). This is because aside from the possibility of attenuation of signal strength, the diffusion effects of magnetic susceptibility-induced gradients may actually contribute to the loss of clear signals through compounding of the small and undetected variances during diffusion, and can only be mitigated by increasing the gradient pulses during the course of the scan (Jones, 2011, p.59). If longer diffusion times are required or if spin echo cannot be used, stimulated echo (STE) may be used to lengthen echo times to study slow diffusion rates. STE considerably increases echo times when compared with spin echo, but the signal is greatly reduced to half. Thus the strength of the reflected signals not only contain information with regards to the kind of molecules present within the systems, but also the diffusion rates of the fluids within these tissues since the areas with high diffusion rates have tendencies of highly-attenuating the pulses, leading to signal decay which translates to a hollow portion in the image generated. Likewise, while densely-packed molecules prevent high diffusion rates due to the presence of barriers, these allow a strong reflection of signals, making the generated images brighter and much more solid. Effects of Diffusion on Stationary and Moving Spins It has been previously mentioned that the diffusion rates of stationary and moving particles’ nuclear spins will create different responses to magnetic fields generated by the MRI scanner. This is due to how bipolar pulses (positive/negative pulses) would move the various particles depending on whether these are static or not, as well as to the orientation of the spins with respect to the origin of the magnetic gradient field (Jones, 2011, p.45). However, even if the directions of the gradient fields are changed, any particle that exhibits a non-zero displacement value will be affected by the field in any way due to being reversed by a negative gradient during the second pulse, thus accounting for signal attenuation during the course of the pulse coming from the . Likewise, for molecules that were non-moving will not be thoroughly affected by the pulse due to the phase shift being proportional to the gradient amplitude during the time when the pulses were generated, and will be able to generate strong contrast against particles that were displaced. However, it must also be remembered that other sources of magnetic fields such as eddy-currents produced by the scanner itself could affect the spins of the molecules, which in turn could either add noise or remove some signals altogether, and thus must be corrected through the use of techniques such as double-spin echo or diffusion gradient application schemes to prevent further distortions to the final images (Koh and Thoeny, 2010). Tissues with high fluid content or have many mobile elements or particles also have high rates of diffusion, which means that the particles can be expected to have non-zero phase shifts at any given moment due to a phase offset, and when a gradient field passes through these tissues by virtue of DWI, the images would be dark due to the rapid decay rates of the signals (Luypaert, et al., 2001). Likewise, in tissues with less mobility and hence would have lower rates of diffusion, since the signals would be expected to have more areas to bounce off to, there would be lesser rates of signal attenuation, resulting to these dense areas appearing bright. These also allow scanning of the connections between various tissues such as blood vessels since these tissues have uniform characteristics such as being relatively more static that the fluid that are transported through them, and are dense enough to reflect the signals back due to being impermeable in terms of the measurement of diffusion rates (Mori and Barker, 1999). Effects of Physiological Motion on Diffusion-Weighted Images The high sensitivity of dMRI up to the point of detecting Brownian movement in various tissues and surrounding media may have strong points such as the detection of diffusion rates, but this also poses strong negative effects such as the rapid decay of signals due to rapid particle outward movement, which could in turn affect image resolution. Due to the strong effects of diffusion and motion in relation to dMRI techniques, it is important that patients undergoing such scans must make as little movement as possible. This is because any kind of movement could alter the orientation of the nuclear spins in relation to the pulse gradient, since the latter is usually from one direction only (Cercignani and Horsfield, 2001). By moving the tissue this could potentially cause distortions in the signals due to the highly-sensitive nature of the scan in terms of detecting microscopic movement. There can also be signal losses and spin dephasing which can lead to image blurring, ghosting, increased artefacts and more noise as the result of movement (Koh and Thoeny, 2010). Thus, the use of dMRI among claustrophobic patients is not recommended, not is the sole usage of it in generating images that require patients to stay still for at least half an hour (Reiser, et al., 2008). However, it is possible to provide additional comfort for the patient such as increasing paddings inside the scanner so as to prevent excess movement while at the same time preventing the patient from getting fatigued in the process. Aside from the actual movement of patients, other kinds of physiological motion such as the scanning of pulsating parts of the body like the heart, pulses, or the pulsation of the brain itself could also have effects on the outcome of the scan (Johansen-Berg and Behrens, 2009). This can be corrected by using an additional six degrees of freedom through three rotations and three translations, as well as increasing the control in the process of post-acquisition of images. If these pulsating regions are not accounted for, aside from the increase in the distortion of the images, there is also a recorded increase in the diffusivity of the particles (Johansen-Berg and Behrens, 2009, p.53). Thus, aside from taking into consideration the types of tissues to be scanned, the ADC of the tissue as well as the most appropriate b-value for the gradient scan, physiological motion from the test subject itself must also be accounted for when doing the scan, so as to increase control over the final image through the removal of potential sources of distortion and blurring and in turn ensuring the high-quality of the images (Koh and Thoeny, 2010, p. 20). Conclusion The use of diffusion-weighted imaging (DWI) in MRI scans contributed greatly to the advancement of its usage. This is because of the higher resolution of images that can be achieved through the harnessing of the variances in the diffusion rates of particles commonly found within tissues. Since there is a greater dependence on the movement of molecules through a medium, it enables the delineation of barriers such as membranes and fluids upon scanning. This enables the images to be diffusion-weighted, wherein particles that exude high diffusion rates would have a sharp contrast against particles with low diffusion rates. These differences are much more captured in detail using dMRI, which increases the sensitivity of the scan to certain abnormalities within tissues or cells. In addition to the MRI scanner not using any ionising radiation, this not only make this image modality safe but also have higher resolution in presenting scanned images of the internal anatomy. Since dMRI is highly-dependent on the micro-orientation of particles within a living system, it is important for the test subject to keep still during the scan, or for the technician to use methods that would prevent the test subject from making excess movement, otherwise the slight movement could alter the direction to where the magnetic pulse field would land, thereby distorting the signals from the particles, and in a way could also distort how the final image would appear. While initially this presents to be a problem, this same sensitivity with regards to the movement of particles and the density of the medium which makes dMRI much more sensitive to certain abnormalities within tissues or cells that would otherwise be undetectable using other image modalities. Due to the characteristics of water in the presence of various particles, differences in the diffusion rates of particles and water molecules in living systems can be detected through DWI in a way that varying rates and concentrations of particles can provide strong contrasts to areas with higher densities and lower diffusion rates, thus giving the image a clearer picture of which areas are barriers or dense, the connections between the much more solid areas, and which areas allow fluids to flow normally. Bibliography Basser, P. J., 1995. Inferring microstructural features and the physiological state of tissues from diffusion?weighted images. NMR in Biomedicine, 8(7), pp. 333-344. Cercignani, M. & Horsfield, M., 2001. The physical basis of diffusion-weighted MRI. Journal of the Neurological Sciences, Volume 186, pp. S11-S14. Hagmann, P. et al., 2006. Understanding diffusion MR imaging techniques: from scalar diffusion-weighted imaging to diffusion tensor imaging and beyond. Radiographics, Volume 26, pp. S205-S223. Johansen-Berg, H. & Behrens, T. E., 2009. Diffusion MRI: From quantitative measurement to in-vivo neuroanatomy. London: Academic Press. Jones, D., 2011. Diffusion MRI : Theory, Methods, and Applications: Theory, Methods, and Applications. New York: Oxford University Press, Inc.. Koh, D. & Thoeny, H., 2010. Diffusion-Weighted MR Imaging: Applications in the Body. New York: Springer. Luypaert, R., Boujraf, S., Sourbron, S. & Osteaux, M., 2001. Diffusion and perfusion MRI: basic physics. European Journal of Radiology, Volume 38, pp. 19-27. Mori, S. & Barker, P. B., 1999. Diffusion magnetic resonance imaging: its principle and applications. The Anatomical Record, Volume 257, pp. 102-109. Moritani, T., Ekholm, S. & Westesson, P., 2009. Diffusion-Weighted MR Imaging of the Brain. New York: Springer. Reiser, M., Semmler, W. & Hricak, H., 2008. Magnetic Resonance Tomography. New York: Springer. Read More