MR Spectroscopy - Report Example

Question Surface coil localization Localization in MR spectroscopy can be achieved through the application of physical localization of the coil. This technique can be achieved through the application of simple surface coils. An example of such coils could be small, ring-shaped and made from a variety of wires and shapes. These surface coils normally have a very high sensitivity and are limited to penetration. They are applied in examining the anatomy of the surface on the patients’ body. (Lambert & Mazzola, 2004). This sensitivity of these coils is determined by the filling factor and also their size i.e. smaller coils are normally more sensitive than larger coils. Therefore, for surface coil localization, multi-ring surface coils are preferred since they provide a uniform field and also sensitivity and power similar to standard surface coils. An MR spectroscopy carried out using a conversion of two or three rings will lead to a spatial localization as that produced by a volume coil but with an improved RF power and SNR characteristics. Figure 1 below show surface coil localization in MRS. (McRobbie, 2007) Figure 1(Source: McRobbie, 2007) This type of localization has its own advantages and disadvantages. Advantages A simple surface coil produces a homogenous field with an in-depth penetration of the coil. This therefore means that a good signal to noise ratio (SNR) is achieved by the exclusion of noise coming from outside the region of interest. (Styles, Scott, & Radda, 1985) This is achieved in three ways: a) The use of a simple pulse sequence i.e. the application of a 90 degrees pulse in order to flip the net magnetization vector on a transverse plane. In this case, no multiple gradients are applied and the length of the time of acquiring the signal is close to zero TE which leads to a minimization of the dephasing coherence (T2 decay) and incoherent effect, therefore reducing the signal to noise ratio (SNR). (McRobbie, 2007) b) A surface coil is characterized by a high filling factor hence the area of the acquired sample is similar to that of the surface coil hence reduction in SNR. This is because the sample fills the sensitive area of the coil and a smaller coil is used for a smaller sample. (McRobbie, 2007) c) Normally, the signal detected fades away from the coil. However, with the coil attached to the region of interest, this may results to an increase in SNR. The use of multi-ring coils in this type of localization leads to an improved water suppression and localization and also reduced outer voxel contamination coupled with a minute loss in terms of SNR and also an increase in SAR. This is because surface coils are placed at the region of interest hence are very close to the source signal. (Cowin, 2011) This technique can be used for both excitation of the signal and receiving of the signal in the area of interest. This therefore leads to a reduced effect of pulse radiation since only a small area is exposed to RF magnetic field radiation hence reduction in specific absorption rate (SAR) level. (Hecht, 1997, p. 223-330) Normally, the surface coil localization technique has a very short TE hence will metabolite with a very short TE i.e. as illustrated by Phosphorous and 31P examined with the surface coil. (Cowin, 2013) Surface coil resolution also leads to a high resolution spectroscopy without the use of strong gradients. This therefore reduces safety considerations such as the effect of peripheral nerve stimulation (PNS) and acoustic noise. Disadvantages The use of strong gradients leads to the production of eddy currents hence may affect the quality of the spectrum produced. (Cowin, 2011) The surface coil localization leads to inhomogeneous RF field. This is achieved when the surface coil is used as a transmitter leading to the power of the RF field dropping as it moves away from the plane of the coil leading to inaccurate distribution of the signal intensity. (Cowin, 2011) The intensity of the signal produced is dependent on the spins in the radiofrequency (RF) field and the coil geometry. This is disadvantageous is due to the fact that the surface coil does not transport the flip angle constantly across a slice. An area from which the spectrum is obtained from is difficult to ear-mark and do not have any clear borders since the intensity of the signal in that area is non-uniform. Figure 2: Radiofrequency (RF) field profile applied from surface coil. (Source: Lambert & Mazzola, 2004) Application of surface coil. The surface coil has several limitations associated with its use. However, there are some metabolites changes that are important to acquire with MR spectroscopy and only can be acquired in the signal when applying very short TE which can be achieved by using surface coil localization. For instance, in 31P MRS (Phosphorus magnetic resonance spectroscopy), a short sequence is required on the surface coil due to its short TE and low signal to noise ratio. (Cowin, 2013). Therefore, the use of the use of same coil for simultaneous excitation and receiving RF pulses is the simplest way of applying the use of a surface coil. This is done by limiting the area of interest to only the region under the surface. The flip angles of the spins are determined by the amplitude of RF field and the RF pulse is expected to decrease dramatically as moving further away from the surface due to the flat shape of the coil. Normally, spins at the surface will experience 90o RF pulse with production of a signal while those away from the surface will experience 1800 RF pulse with no signal generated. The spins that are experiencing a 90o RF pulse are located further deeper in the sample as the power of the pulse increases. However, this method has some disadvantages. This include; It produces poor boundaries and size and shape of the localized region cannot be controlled hence a signal with a different pulse from outside a localized signal will be detected. Also, there is no way of avoiding the excitation of the area near to the surface coil hence will contribute to the final signal. This method can also be achieved by the application of selective gradients that lead to provision of a spatial encoding of the larmor processional frequency hence allowing one to be able to determine the magnetic response in 3D. (Styles, Scott, & Radda, 1985). A slice selective 180 degree pulse applied after time tau and then the signal acquired as shown in figure 2 below. At exactly a 900 magnetization vector flip on a transverse plane, a maximum signal is achieved. However, this maximum flip angles cannot be achieved in surface coils since the position of the slice is dependent on the radio frequency magnetic field i.e. B1 and also the maximum degree of intensity. This therefore means that, only a single spin will experience an exact 900 RF pulse as the others will experience flip angles other than 90 degrees. Hence, the final signal intensity will be low since spins are rotated at different angles and TE will have increased in a way that T2 decay plays role in in the final signal. Figure 3 (Source: Bottomley, 2005) Question 2: PRESS and STEAM sequences Point Resolved Spectroscopy (PRESS) and Stimulated Echo Acquisition Mode (STEAM) are techniques used for localization in MRS. Acquiring a signal from PRESS and STEAM is due to several factors based on the sequence frame. An application of slice selective of 90 degree RF together with two 180 degree RF pulses in order to obtain a signal intensity with PRESS and STEAM techniques. The PRESS sequence is same as the spin echo sequence but a perpendicular selective gradient with an RF pulse is applied to the other slice gradients acquired hence an echo will only be effected from only a voxel. (Bottomley, 2005) The resultant echo that is normally the transverse plane of the whole time and is acquired from the net magnetization vector. The application of the 180 degrees pulses is possible with the refocus on effect of coherent dephasing from magnetic field inhomogeneity (T2) coupled with the rephasing gradients due to the application of slice selective gradients. (Cowin, 2011) However, incoherent dephasing caused by molecules interaction i.e. T2 decay is based on effective time and is irrecoverable. The effective time in this case is the time between the application of RF at 90 degrees excitation and the acquiring of the signal. Therefore, TE period considered to be long which result in increasing the T2 decay which effect in reducing the acquiring signal intensity and also the signal of some molecule structures with short TE would be lost in transverse plane. So only tissue structures that have long T2 relaxation would be presented in the signal that obtained from PRESS localization. (Castillo, Kwock, & Mukherji, 2009) STEAM Localization This kind of sequence is applicable so that it can store transverse magnetization for a certain period of time in a longitudinal plane so as to avoid long effective time and hence reduce the effect of T2 decay on the signal. This is normally achieved through the application of three 90 degrees of RF pulses and then the stimulated echo signal is acquired. This therefore means that the spins that had been regularly spread are influenced by the T2 relaxation time. (Charatcharoenwitthaya & Lindor, 2007) Figure 1 below shoe the STEAM TE of 20ms from animal fat and in vivo from a human liver. Figure 4 (Source: Charatcharoenwitthaya & Lindor, 2007) At the application of the third RF pulse, the transverse magnetization goes back to the transverse plane, hence transvers magnetization is achieved at T1 in a process known as stimulated echo as illustrated in Figure 2. However, this process has its own drawback in that 50 percent of the magnetization is used to provide the final signal. This is because the whole spins are spread regularly along the positive Y. However, this is advantageous is a way that during the first and the final T-periods, the signal intensity is reduced because of the T2 relaxation hence the sequence of the stimulated echo in part compensates for losing 50% of the signal in the period of time T among the second and the third pulses of a stimulated echo sequence. However, the final signal magnitude is not affected by the dephasing of the spin around the z-axis. (Cheong et al, 2004) Figure 5: STEAM pulse sequence (Source: Cheong et al, 2004) In the first 90 degree, the selected pins are flipped in a transverse plane and after a period of time, they are totally dephased due to T2 hence forming a disc shape along the x-y plane. In the second spin, the RF flips the disc into a longitudinal plane and the spins start to recess around the z plane and parallel to transverse plane creating a sphere of dephased spins. Lastly, after the third 90 degree pulse, the cone is rotated onto the longitudinal axis and the spins distribute themselves evenly in a transverse position with each of the halves moving in opposite directions. (Kim et al., 2008) The three rotation cases are shown in the figures below. Figure 3: Spins in the X-Y plane Figure 4: Second RF 90 spin Figure5: The cone shape (Source: Kim et al., 2008) The signal Intensity The net magnetization is generated along the positive y axis and the stimulated echo can be acquired after the third 90 degree RF pulse. However, it is accumulated from the spins that are distributed evenly in the positive y side of the sphere, as seen in figure 6. This implys that the stimulated echo is formed from spins that has net magnetization component in the z, x and y axis. Therefore the signal intensity can only be acquired from the averaging of the sum of net magnetization of y components which is only 50% of the signal intensity that would be obtained if all spins are aligning in the y plane (Cowin, 2013). Figure 6: The stimulated echo (Source: Bottomley, 1999) The TE in STEAM is shorter when compared with that in PRESS. This can be attributed by dephasing of the spins between the second and the third 90 degree RF pulse. However, it does not effect in the final stimulated echo signal as the net magnetization was in longitudinal plane. In STEAM, T2 decay effects are only presented in the first and the final Tau periods while in PRESS, the effects are along the echo sequence. Therefore for metabolites with shorter T2 relaxation, STEAM sequence is preferred while for those with long TE, PRESS is used since it enables one to acquire a larger signal. (Cowin, 2013). Question 3: Peak area metabolite concentration Magnetic resonance spectroscopy (MRS) can be used to provide information about the concentrations of particular metabolites and changes in their levels. The concentration of metabolites is normally referred as the amount of chemical distribution per unit volume and this concentration can be measured in moles. Therefore using single-voxel magnetic resonance spectroscopy (MRS), the concentration ratio over a selected volume of tissue can be estimated with the consideration of various factors in determining the concentrations from the peak area. (Bottomley, 1999) From the information available in various literature, it is clear that the concentration of a metabolite can be measured using the peak area but not in a direct way. Therefore, the argument that peak area is the direct measurement of metabolite concentration is not valid. This argument can be supported by three factors. First is that, during the measuring of the metabolite concentration, the concentration of the volume selected keeps on varying and the metabolites bound are not MR observable. This variation occurs in different magnitudes in different tissues i.e. the concentration sodium (NA) in intra-cellar is 5 millimolar, while in the blood plasma, the concentration is about 145 millimolar. The second factor to be considered is the recognition of the number of the equivalent protons that contribute to the acquired signal in MR spectrum. (Barker et al., 1993) Before any effort to determine the
number
of
signals
that contribute in
NMR spectrum, the structure of the metabolite has to be known and also the number of non-equivalent protons that contribute to the different peaks i.e. singlet, doublet and triplet. The third factor is concerned with T1 and T2 weighting. This is based on the fact that different metabolites and protons in single volume may have different T1 and T2 relaxation. (McRobbie, 2007) The T1 weighting image occurs when TR is very short and the when the tissue has not fully recovered. In longitudinal plane, the signal from 90 degree pulses will be affected by a process referred to as partial saturation. For example, inositol (T1= 1 sec) and taurine (T1= 1.7sec) both acquired from solution with same concentration but different T1 weighting time, and the TR that use is about 2 second. Therefore as inositol will recover faster, the peak intensity that obtained from inositol is higher than the peak intensity of taurine (Kim et al., 2008) Short T2 elements can be used to achieve broad peaks with different baselines. If a spectrum is achieved with a long TE time, this leads to the flat baseline and narrower peaks will be acquired from the metabolites with a long T2 time. If the time echo is decreased, the signals from the whole elements will rise with also the metabolites that have a very short T2, such as in Lipids (CH2). (Lambert & Mazzola, 2004) However, the long TE will decrease in the signal intensity and T2 decay increase and therefore lower signal to noise ratio. For instance from Table 1 above, using 100 msec TE value for inositol with a T2=110 and chlorine with a T2=350, it is found out that the inositol has a lower signal intensity to that of chlorine. As illustrated in Figure 7 below, the width of the peak can be affected by the T2 decay as the components with short T2 decay give a broad peak and component of long T2 decay will give a narrow and sharp peak. (Kim et al., 2008) The quality of the localization makes the calculation of the correct volume a little difficult. Therefore, a reference signal must be used for calibration. The post-processing stages i.e. the baseline correction and area quantification are subtracted, so as to obtain an accurate peak quantification. This baseline correction is produced by a mathematical mixture with the baseline offset and higher and linear order polynomials. (Hecht, 1997) The baseline errors would result in uncorrected peak high and also an increasing peak height due to overlapping peaks. Therefore, even though the same parameters maybe used for the same patient in different places, the total amount of signal that is acquired from MR spectroscopy may not be the same. The amount of the signal is affected by several factors including, hardware, degree of accuracy of RF pulses, patient position in relation to the coil, and the accuracy of shimming. (Cheong et al, 2003) Figure 7: Overlapping peaks (Source: Cheong et al, 2003) However, the baseline correction can be done using three types of automation. These include automation without any operator effort, selecting a function so as to simulate a curve that perfectly simulates the baseline and selecting points and areas of the spectrum such that the operator and then the program does fit a curve to the selected points. As it can be seen in Figure 7, the peak quantification spectroscopy provides a special chemical of hydrogen atoms to be different inside a chemical. (Castillo, Kwock, & Mukherji, 2004) This therefore helps in identifying the quantification of the metabolites. Lastly, spin-spin coupling of magnetic moments result in splitting of resonance frequency into multiplets of intensities and the phase direction will be relay in the TE and the mixing time (MT) period. The spin-spin coupling constant J (e.g. lactate) can effected by the function of TE period can change the peak phase of two spins coupled group from positive to negative mode (Figure 8). (Charatcharoenwitthaya & Lindor, 2007) Inverted doublet at TE=144 (negative peak At TE= 288 positive peak Figure 7: the effect of TE on evolution of the lactate doublet (Source: (Charatcharoenwitthaya & Lindor, 2007) Question 4: Applications of MRS Investigate Stroke Atherosclerotic diseases like stroke normally affect the supply of blood to the brain hence resulting in cerebrovascular accidents. Stroke can be classifies as ether ischemic or hemorrhagic. A large of number of strokes i.e. 80% are due to cerebral ischemia, which is a condition caused by inadequate supply of blood to the brain. On the other hand, hemorrhagic stroke is caused by a less common effect in the brain i.e. bleeding into the brain tissues. There is need for constant blood supply, glucose and oxygen to the brain’s neuron cells for them to function effectively. In an instance the supply is short-lived or cut-off to this cells, they may die hence causing permanent damage of the part of the brain tissue. Also, this may result to dysfunction f various parts of the patients which maybe depending on the affect tissue of the brain. (Castillo, Kwock & Mukherji, 2003) Several studies have shown the importance of application of proton (1H) MR spectroscopy in the detection of cerebral ischemia. This is because the 1H MR spectroscopy allows a non-invasive measurement of several brain metabolites e.g. Creatine, N-acetyl aspartate, choline and lactate (Lac). This metabolites are used to indicate the normal and pathological function of the body. (Barker et al., 1993) Several experimental studies and procedures according to Barker and his colleagues investigate the temporal variation of the brain metabolites after ischemic stroke. In this study, spectra were acquired through the application of STEAM (stimulated echo) pulse sequence over a selected region. STEAM Sequence A. Slice selection gradients Normally, three slice selective RF 90 degree pulses applied in STEAM sequences so as to acquire a signal. These slice selective gradients are applied perpendicularly with each other so as to determine a voxel shape. The first two 90 degree RF pulses occur in time tau 1, which is equal half of TE while the second and third 90 degree RF pulses are separated by time tau 2 which is also referred to as the mixing time (TM). The stimulated echo is normally acquired in time tau 1, after the third 90 degree RF pulse. (Cowin, 2013). B. Trim gradients As per the figure below, three main trim gradients are applied for a time period equal time period and in opposite direction of the slice selection gradients. This type of gradients are normally applied so as to recover the loss of spin coherence due to applied slice selection. At gradient X, the first trim occurs and this act as rephrasing gradient in the negative direction for the slice selective gradient that applied at the first RF 90o pulse. At y-durection, the second stream is applied and act as a preparation gradient for the slice selecting gradient at third 90o RF pulse. This second trim is in same direction as the slice select gradient since the phase of the spins is reversed by the 180o pulse. The third tri occurs in the z-direction and is normally used a preparation gradient for the slice selective gradient in z direction at second 90o RF pulse. Figure 8: STEAM Sequence (Source: Cowin, 2013) C. Crusher gradients In order to achieve a maximum signal intensity, five crusher gradients are applied. The first gradient is used to dephase any net magnetization in the transverse plane before the second slice selective RF pulse is applied. The next three gradients are applied to destroy any signal accumulation from outside the selected sample. Consequently, a final crusher gradient is applied after third slice selection RF pulse so as to refocus the spins dephasing of first crusher gradient. Experimental Constraints It is clear that STEAM sequence has an advantage in detecting metabolite with short TE period and will lead to a better image localization due to use of 90 degree pulse instead of 180 degree pulse and also reduce the effect of specific absorption rate (SAR). Application along TE will dramatically reduce the signal to noise ratio in STEAM. In this particular experiment, the area of interest is on the lactate doublet signal which has long TE and the phase direction of the doublet peak with constant J coupling is affected by the function of the TE period. Two advantages are associated with the use of a long TE i.e. we can obtain a positive lactate doublet peak and also the lipids and other macro molecules may mask the lactate peak and are removed from the final signal. On other hand, the use a high T will lead to a high amount of the SNR (Signal to Noise Ratio) hence PRESS sequence will be most appropriate for use. (Cowin, 2013) Results From the results obtained, there is a significant increase lactate (Lac) doublet peak at 1.33ppm and significant declines of N-acetyl aspartate at (NAA) 2.02ppm in the infarct lesion. Also, there was an increase in choline (Cho) at 3.2ppm while a Creatine (Cr) signal was decreased in infarct lesion. Discussion A. N-acetyl aspartate (NAA) In the normal brain tissue, the peak of N-acetyl aspartate (NAA) considered to be the highest singlet peak and is given as .02 ppm. NAA is presented in the nervous system and is found in grey and white matter. It is considered to be the marker of density and viability of the neuronal and axonal. Degradation of a neuron in the MRI spectroscopy map is signified by a decrease of the NAA concentration. A decrease of NAA in the sub-acute phase of stroke is combined with reduction in choline and Creatine metabolites (Barker et al., 1993). B. Lactate (Lac) Normally, the Lactate peak is a doublet and a signs at 1.33ppm which projects above or below the baseline base on the function of TE. An increase of lactate in the ischemic region is a key mark of a stroke. Anaerobic glycolysis will be induced due to oxygen reduction and result in increasing the production of the lactate that accumulate precisely in the presence of the hypo perfusion. Therefore the presence of the lactate will determine oxygen reduction (Barker et al., 1993). In the sub-acute a stroke, an increase concentration of lactate were countered in the ventricles and contralateral hemisphere. Some publications have indicated an increment of pH values and lactate in sub-acute infarcts The clinical impact of MRS on Stroke With the use of MR spectroscopy, one can obtain information of the metabolite changes in the ischemic tissue during the early stages of the stroke up to the death of the neuronal cells (Barker et al., 1993). However, a reliable interpretation is based on the following assumptions: The neural cells are dead if there is reduction in NAA Hypoxia is the first stage in the ischemia and is determined by augmented lactate. The development trend in MR spectroscopy exam would help in recognizing the accurate stage of the stroke and this may have a great impact on in the treatment decisions. Therefore, identifying the stage of stroke will increase the extreme effectiveness of some emergency medicines of stroke (e.g. clot-busting drugs) which can be achieved only in the first four hours after the stroke. (Castillo, Kwock & Mukherji, 2003) Conclusion From the above discussion, the use of surface coils is most appropriate localization even though it has some disadvantages associated with it. Peak area is not straight evaluating the metabolites but it is able with some factors that should be considered. Despite the limitations noted above, MRS studies in patients with schizophrenia are producing findings that provide tentative support for previous neuroimaging, neuropathological and neuropsychological work. The findings are compatible with current ideas about the nature and etiology of stroke and also raise further questions about the biochemical nature of the disorder and its course Bibliography Barker, P. B., Soher, B. J., Blackband, S. J., Chatham, J. C., Mathews, V. P., & Bryan, R. N. (1993). Quantitation of proton NMR spectra of the human brain using tissue water as an internal concentration reference. NMR in Biomedicine, 71, 1991-1993. Bottomley, P. A. (2005). In vivo Solvent-Suppressed Localized Hydrogen Nuclear Magnetic Resonance Spectroscopy: A Window to Metabolism? Proceedings of the National Academy of Sciences, 82(7), 2148–2152. Castillo, M., Kwock, L., & Mukherji, S. K. (2003). Clinical Applications of Proton MR Spectroscopy. Neuroimaging clinics of North America, 16(2), 269-283. Charatcharoenwitthaya, P., & Lindor, K. D. (2007). Role of Radiologic Modalities in the Management of Non-alcoholic Steatohepatitis. Clinics in Liver Disease, 11, 37-54. Cheong, J. L., Cady, E. B., Penrice, J., Wyatt, J. S., Cox, I. J., & Robertson, N. J. (2005). Proton MR Spectroscopy in Neonates with Perinatal Cerebral Hypoxic-Ischemic Injury: Metabolite Peak-Area Ratios, Relaxation Times, and Absolute Concentrations. NMR Biomed, 1(2), 32-43. Cowin, G (2011). Magnetic Resonance Spectroscopy and Applications, UQ . Hecht, H. G. (1997). Magnetic resonance spectroscopy. New York: Wiley. Kim, H., Taksali, S. E., Dufour, S., Befroy, D., Goodman, T. R., Petersen, K. F., Constable, R. T. (2008). Comparative MR Study of hepatic fat quantification using single-voxel proton spectroscopy, two-point Dixon and three-point IDEAL. Magnetic Resonance in Medicine, 59, 521–527. Lambert, J. B., & Mazzola, E. P. (2004). Nuclear magnetic resonance spectroscopy: An introduction to principles, applications, and experimental methods. Upper Saddle River, N.J: Pearson/Prentice Hall. McRobbie, D. W. (2007). MRI from picture to proton. Cambridge, UK: Cambridge University Press. Styles, P., Scott, C. A., & Radda, G. K. (1999). A method for localizing high-resolution NMR spectra from human subjects. Magnetic Resonance in Medicine, 6, 89-124. Read More