| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1
Department of Radiology, Chang Gung Memorial Hospital, 199 Tun-Hwa North Rd.,
Taipei, Taiwan 105, R.O.C.
2
Department of Public Health, Chang Gung University, 259 Wen-Hwa First Rd.,
Kweishan, Taoyuan, Taiwan 333, R.O.C.
3
Institute of Biomedical Science, Academia Sinica, 128 Yen-Jo-Yuan Rd.,
Nangung, Taipei, Taiwan 115, R.O.C.
Received March 21, 2000;
accepted after revision August 10, 2000.
Presented in part at the annual meeting of International Society for
Magnetic Resonance in Medicine, Philadelphia, May 1999.
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. We examined 30 healthy adults with a 1.5-T scanner four times on 3 days using single-voxel spectroscopy over the left hippocampus, chemical shift imaging over the left hippocampus, and chemical shift imaging over the bilateral hippocampi. Metabolite ratios were derived from the integral values of three major peaks: N-acetylaspartate (NAA), choline-containing compounds (Cho), and creatine plus phosphocreatine (Cr). The random-effects model of one-way analysis of variance was used to evaluate the reproducibility in terms of coefficient of variation; the mixed-effects model was used to compare the results of different hippocampal regions and spectroscopic techniques.
RESULTS. Most coefficients of variation for the NAA/(Cho+Cr) ratio were less than 20%. All the coefficients of variation for the posterior hippocampus (15-25%) were less than those for the anterior hippocampus (20-44%). The posterior hippocampal NAA/(Cho+Cr) ratio of unilateral chemical shift imaging had the lowest coefficient of variation (<16%). Single-voxel spectroscopy and unilateral chemical shift imaging had similar coefficients of variation for the anterior hippocampal NAA/(Cho+Cr) ratios (17-20%). There was a significant difference in metabolite ratios measured in different hippocampal regions (p < 0.01) and in those acquired with different spectroscopic techniques (p < 0.001).
CONCLUSION. The NAA/(Cho+Cr) ratio is the most reproducible parameter for hippocampal MR spectroscopy on a 1.5-T scanner. Regional variation and technical differences in metabolite ratios must be considered when interpreting proton spectra of the hippocampus.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The hippocampus has been studied with proton MR spectroscopy in several cerebral diseases, such as epilepsy [4,5,6,7,8,9,10,11,12,13], Alzheimer's disease [14], and schizophrenia [15]. Both single-voxel spectroscopy [16] and chemical shift imaging [17, 18] have been used to evaluate the metabolite profile of the hippocampus. Although promising clinical applications have been suggested, diverse results due to different measurement techniques, acquisition parameters, and data presentations have been noted [19]. These methodologic inconsistencies make it difficult to compare data from different research centers and preclude widespread applications in clinical practice. Furthermore, reproducibility of hippocampal proton MR spectroscopy has not been specifically evaluated. Only after the precision and repeatability of this technique has been determined can hippocampal proton MR spectroscopy be confidently applied for the detection of metabolite changes caused by disease or treatment.
The goal of this study was to assess the reproducibility of hippocampal metabolite ratios obtained on an automated MR spectroscopy system. Our research plan was designed to mimic the condition in which MR spectroscopy is performed for daily clinical practice and not to be a reproducibility study with a phantom in an MR laboratory. Both single-voxel spectroscopic and chemical shift imaging techniques were evaluated and performed in the same group of healthy adults. Specifically, between-acquisition reproducibility, within-acquisition reproducibility, intertechnique differences, and interregional variations were the major items investigated.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
All the proton MR spectroscopic studies were performed on a 1.5-T whole-body MR unit (Magnetom Vision; Siemens, Erlangen, Germany) with a standard circularly polarized head coil. The subjects were examined while they were supine with foam pads under their shoulders and neck for support. The subjects were instructed to extend their necks, approximately 20-30° backward, to make the long axis of the hippocampus approximately parallel to the transverse plane. A regular head-holder and tape were then used to stabilize the subject's head. Transverse T2-weighted images of the whole brain were obtained using a fast spin-echo sequence (TR/TE, 3787/99; matrix, 256 x 256; field of view, 20 x 20 cm; echo train length, 11; number of acquisitions, 1; scan time, 1 min 27 sec). Any intracranial abnormality excluded the subject from further examination. A series of orthogonal T2-weighted images for localization were then obtained using a fast imaging technique with a steady-state precession sequence (6.32/3.00; flip angle, 70°; matrix, 200 x 256; field of view, 23 x 23 cm; slice thickness, 5-7 mm; interslice gap, 5-7 mm; number of acquisitions, 1; scan time, 15 sec).
For single-voxel spectroscopic studies, the volume of interest (20 x 20 x 15 mm) was placed over the left anterior hippocampal region and included part of the head and body (Fig. 1A). The anterior margin of the volume of interest was within 3 mm anterior to the peduncles, the lateral margin within 3 mm lateral to the lateral border of left temporal horn, and the inferior margin within 1 mm below the inferior border of the hippocampal head or subiculum. The position of the volume of interest was meticulously adjusted in orthogonal images to ensure standard placement. The procedure was performed by an MR spectroscopy technician and was then checked by a neuroradiologist to ensure technique consistency. The method of point-resolved spectroscopy [21] was applied for spectrum acquisition. After the transmitter and receiver were automatically adjusted, water signal was automatically shimmed to within a line width of 3-5 Hz. After preirradiation of water resonance by applying three chemical shift selective pulses [22], water-suppressed single-voxel spectroscopy was then performed (1500/135; field of view, 24 x 24 cm; number of acquisitions, 256; data points, 1024; scan time, 6 min 31 sec). For continuous single-voxel spectroscopic measurements, signal from the same volume of interest was acquired without changing the size and location.
|
For the unilateral chemical shift imaging studies, the volume of interest (20 x 40 x 15 mm) included as much as possible of the left hippocampus (Fig. 1B). In bilateral chemical shift imaging studies, a large volume of interest (from 60 x 60 x 15 mm to 80 x 60 x 15 mm) was applied to encompass both hippocampi (Fig. 1C). The anterior margin of the volume of interest was 5-10 mm anterior to the peduncles, the lateral margin within 3 mm lateral to the lateral border of the ipsilateral temporal horn, and the inferior margin within 1 mm below the inferior border of the left hippocampal head or subiculum. A hybrid spectroscopic technique, point-resolved spectroscopy-chemical shift imaging, was used to avoid contamination of lipid signal from the skull. After the transmitter and receiver were adjusted automatically, water signal was automatically shimmed to within a line width of 8-12 Hz for unilateral and 10-15 Hz for bilateral chemical shift imaging studies. After preirradiation of water resonance by applying three chemical shift selective pulses [22], water-suppressed chemical shift imaging was then performed (1500/135; field of view, 16 x 16 cm; number of acquisitions, 2; phase encoding, 16 x 16; voxel, 10 x 10 x 15 mm; data points, 1024; scan time, 12 min 55 sec). For continuous chemical shift imaging measurements, signal from the same volume of interest was acquired without changing the size and location.
|
|
Automated processing of the raw data was accomplished using a commercially available spectral analysis software package (version B23A, Magnetom Vision; Siemens). The automated spectral analysis included the following procedures: Fourier transformation with a K-space Hamming filter (50%); apodization with a gaussian filter (center, 0.0 msec; width, 256 msec); subtraction of residual water signal; correction of frequency shift, baseline, and phase; and integration of peak areas by gaussian curve fitting. The NAA peak was identified at 2.00 parts per million (ppm) (range, 1.80-2.20 ppm), the Cr peak at 3.02 ppm (range, 2.92-3.12 ppm), and the Cho peak at 3.22 ppm (range, 3.12-3.32 ppm) (Fig. 1D). Spectra and integral values were then displayed on the base image with an overlaid grid indicating the anatomic location of each voxel (Figs. 1E and 1F). The integral value of each peak was dimensionless and represented relative measurement of the amount of each metabolite. Metabolite ratios were calculated in terms of NAA/(Cho+Cr), NAA/Cho, NAA/Cr, and Cho/Cr. There were eight spectra, arranged in a 2 x 4 matrix (Fig. 1E), in unilateral chemical shift imaging. Ratios from voxels of the anterior two rows were averaged and assigned as the value of the anterior hippocampus, and those of the posterior two rows were averaged and assigned as the value of the posterior hippocampus. For chemical shift imaging that encompassed bilateral hippocampi, the eight voxels in the left hippocampus that corresponded to those in unilateral chemical shift imaging were selected for analysis. A spectrum was excluded if the integration of any peak could not be accomplished using the automated analysis software. The time required for data processing was less than 3 min for a single-voxel spectroscopic study and less than 10 min for a chemical shift imaging study.
|
|
|
One-way analysis of variance with replications taken into account was used to evaluate the reproducibility of the hippocampal spectroscopic studies in terms of the coefficient of variation [23]. The coefficient of variation is the variation in a measurement relative to its mean, which is defined as the within-subject standard deviation divided by the within-subject mean [23]. However, in our study, the standard deviation of replicates appeared to change proportionately with the mean, which is not consistent with the assumption of constant variance in analysis-of-variance model. It was necessary to justify the statistical model according to Chinn [24] by performing logarithmic transformation of the original value, calculating variance components using the random-effects one-way analysis of variance, and deriving the coefficients of variation of the original scale from the square root of the within-mean square. The logarithm transformation makes the within-subject standard deviation approximately independent of the mean and also leads to a better estimation of the coefficient of variation [24]. In general, coefficients of variation that are less than 20% are desirable and those greater than 30% are undesirable [23].
For metabolite ratios acquired with different MR spectroscopic studies, between- and within-acquisition reproducibility were evaluated using the modified random-effects one-way analysis-of-variance model:
yij = µ + 
i +
eij (i = 1, 2,..., 30, and j = 1, 2,...,
ni), where yij is the jth replicate of
logarithm transformation of the metabolite ratio for the ith subject,
i is a random variable representing between-subject
variability, eij is a random variable representing
within-subject variability, and ni is two or three for the
within- or between-acquisition study, respectively. Both
i and eij are assumed to follow
a normal distribution with a mean of zero and with variances of
A2 and 2,
respectively. Here, eij is independent of
i and other eij.
In clinical practice and longitudinal studies, a change can be confidently
detected only if the difference between two measurements exceeds the upper
limit of the confidence interval of the coefficient of variation. The upper
limit of the 95% confidence interval for the mean square error is calculated
as follows: (degree of freedom x mean square error) /
2 0.975, df.
The lower limit of the 95% confidence interval for the mean square error is
calculated as follows: (degree of freedom x mean square error) /
2 0.025, df.
The 95% confidence interval for the coefficient of variation can then be estimated by taking the square root of the upper and lower limits of the 95% confidence interval for the mean square error.
The mixed-effects model of one-way analysis of variance was used to determine whether there is significant difference among various MR spectroscopic studies and between different regions of the hippocampus by treating subjects as random variables and studies and metabolite ratios as fixed variables. The analysis was accomplished by adding one fixed-effect, either MR spectroscopic studies or hippocampal regions, to the modified random-effects model described earlier.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
|
For the unilateral hippocampal chemical shift imaging studies, there were eight voxels in the left hippocampus. A total of 960 spectra were present in unilateral chemical shift imaging studies. Metabolite ratios could not be obtained from 41 (4.3%) of the 960 spectra because of poor spectral quality. Most of the spectra excluded from this study were located in the anterior hippocampal region (31/41 [76%]), especially the first row. Compared with the other metabolite ratios, the NAA/(Cho+Cr) ratio had the lowest between- and within-acquisition coefficients of variation in both the anterior and posterior hippocampi. All the coefficients of variation for the anterior hippocampus were greater than 20%, except those for the NAA/(Cho+Cr) ratio. All the coefficients of variation for the NAA/(Cho+Cr) ratio and the NAA/Cho ratio of the posterior hippocampus were less than 20% (Table 1). All the coefficients of variation for the posterior hippocampus were less than the corresponding ones for the anterior hippocampus. The NAA/(Cho+Cr) ratios of the posterior hippocampus had the lowest between- and within-acquisition coefficients of variation (15% and 16%, respectively). The upper limits of the 95% confidence interval for the between-acquisition coefficients of variation of the anterior and posterior hippocampal NAA/(Cho+Cr) ratios were 25% and 18%, respectively. Figure 2B shows the distribution of the NAA/(Cho+Cr) ratios derived from unilateral chemical shift imaging studies of the anterior and posterior hippocampi. This distribution showed that the range of the means ± standard deviation was much smaller in the posterior hippocampus than in the anterior hippocampus. There were significant differences in the NAA/(Cho+Cr), NAA/Cho, and Cho/Cr ratios for the anterior versus the posterior hippocampus (p < 0.001), but not in the NAA/Cr ratio (p = 0.12).
|
A total of 960 spectra were selected for evaluation of reproducibility of chemical shift imaging studies that encompassed bilateral hippocampi. Poor spectral quality made evaluation impossible in 81 voxels (8.4%), most of which were in the anterior hippocampus (67/81 [83%]). Similarly, the NAA/(Cho+Cr) ratio had the lowest between- and within-acquisition coefficients of variation in both the anterior and posterior hippocampi (Table 1). All the coefficients of variation of the posterior hippocampus were less than the corresponding ones of the anterior hippocampus. However, only between- and within-acquisition coefficients of variation for posterior hippocampal NAA/(Cho+Cr) were less than 20%. All the other coefficients of variation were greater than 20%, and most were more than 30% for the anterior hippocampus. The upper limits of the 95% confidence interval for the between-acquisition coefficients of variation for the anterior and posterior hippocampal NAA/(Cho+Cr) ratios were 35% and 22%, respectively. Figure 2C shows the distribution of the NAA/(Cho+Cr) ratios derived from the anterior and the posterior hippocampus in chemical shift imaging studies that encompassed bilateral hippocampi. Similarly, the range of the means ± standard deviation was smaller in the posterior hippocampus than that in the anterior hippocampus. There was also significant regional difference in the NAA/(Cho+Cr), NAA/Cho, and Cho/Cr ratios (p < 0.01) but not in the NAA/Cr ratio (p = 0.49).
|
There was a significant difference among all the metabolite ratios obtained
on single-voxel spectroscopy, unilateral chemical shift imaging, and chemical
shift imaging encompassing bilateral hippocampi (p < 0.001),
except the Cho/Cr ratio of the posterior hippocampus (p =
0.21).
,
|
|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Different metabolite ratios, including NAA/(Cho+Cr) [4, 7, 12], NAA/Cho [9, 11, 14], and NAA/Cr [10, 11, 13, 14], have been applied for evaluation of cerebral diseases. Most researchers agree that NAA is located primarily within the neurons and regard NAA as a specific neuronal marker [1, 4, 6, 10, 27]. Cho is a constituent of phospholipid metabolism in the cell membrane [1, 3]; increased Cho intensity may reflect increased membrane turnover. Cr intensity has been regarded as an internal reference because it is fairly stable in different cerebral regions and abnormal conditions [1,2,3]. However, because of the anatomic complexity and the resultant field inhomogeneity, the spectral resolution of the hippocampal region is always worse than that of the upper cerebral structures [8, 27]. Field inhomogeneity, combined with the long T2 of cerebrospinal fluid, causes suboptimal water suppression and complicates baseline assignment in the left side of a proton spectrum, reducing the precision of peak integration for Cr and Cho [28, 29]. In the present study, the impacts were reflected by large standard deviations and a wide range of NAA/Cho, NAA/Cr, and Cho/Cr ratios, with most of their coefficients of variation greater than 20%. On the contrary, the NAA/(Cho+Cr) ratio had the smallest standard deviation and the lowest between- and within-acquisition coefficients of variation, most likely because it is easier and more accurate to define the Cho+Cr peak complex than to delineate them separately.
The change in the hippocampal NAA/(Cho+Cr) ratio of an individual subject over time must be greater than 20-25% (the upper limits of the 95% confidence interval for between-acquisition coefficients of variation) to confidently show an effect due to disease progression or therapeutic intervention, rather than to biologic or technique variation. On the other hand, the upper limits of the 95% confidence interval for between-acquisition coefficients of variation of the other metabolite ratios were much higher, even in the range of between 30% and 40%. These findings suggest that the NAA/(Cho+Cr) ratio is most precise because it does not need to distinguish the close Cho and Cr peaks, and when two results of a single subject over time are compared, the ratio NAA/(Cho+Cr) will detect a smaller change than the other ratios. Which ratio to use in a given patient depends on these findings and other considerations such as whether, and in what direction, Cho and Cr intensities are affected by the disease process. For example, the NAA/(Cho+Cr) ratio is more sensitive for detecting abnormalities caused by mesial temporal sclerosis, which decreases NAA intensity and increases Cho intensity [4, 7, 12]. However, even though it is more reproducible than other ratios, the NAA/(Cho+Cr) ratio may be not as sensitive for the evaluation of metabolite changes caused by dementia, for which Cho and Cr intensities remain unchanged [14].
It is interesting to find that the posterior hippocampal metabolite ratios in the unilateral chemical shift imaging studies had better reproducibility than those in the single-voxel spectroscopic studies. In general, single-voxel spectroscopy has more accurate spatial localization, better water suppression, and shorter acquisition times compared with the chemical shift imaging technique [16,17,18, 30]. In current single-voxel spectroscopic studies, the volume of interest was positioned in the anteromedial part of the left mesial temporal lobe where suboptimal shimming due to the susceptibility effect would hinder the advantages of single-voxel spectroscopy. Furthermore, it is impossible to select the same exact volume of interest for different studies, resulting in various degrees of extraneous tissue contamination. This problem was more dominant in single-voxel spectroscopic studies for which a large volume of interest was used and only one spectrum was obtained for the reproducibility study. On the contrary, unilateral chemical shift imaging studies used a small voxel size, excluded bad spectra, and averaged metabolite ratios for the reproducibility evaluation. Better field homogeneity, less tissue contamination, and more data included for analysis all contributed to better reproducibility of the chemical shift imaging of the posterior hippocampus. Anatomic complexity and field inhomogeneity over the anterior hippocampal region also caused many anterior voxels to be excluded from the study because of poor spectral resolution. The combination effect of worse curve fitting and fewer voxels led to a wider distribution of the metabolite ratios in the anterior than in the posterior hippocampus.
Most between- and within-acquisition co-efficients of variation in bilateral hippocampal chemical shift imaging studies were greater than the corresponding ones in single-voxel spectroscopy and unilateral chemical shift imaging. This finding was most likely because of the fact that a large volume of interest that encompassed not only bilateral hippocampi but also a huge amount of extraneous tissues—including the adjacent temporal lobes, midbrain, cerebrospinal fluid spaces, vermis cerebelli, and vessels—made suboptimal shimming in chemical shift imaging studies even worse. Consequently, bilateral hippocampal chemical shift imaging studies had suboptimal spectral resolution, excluded many spectra from analysis (81 versus 41 in unilateral chemical shift imaging studies), and had a wide distribution of metabolite ratios.
For evaluation of within-acquisition repeatability, two measurements were performed continuously without changing the volume of interest to minimize the variation due to head position, volume selection, physiologic condition, and magnet fluctuation. Accordingly, within-acquisition reproducibility was expected to be much better than between-acquisition reproducibility. However, in the present study, only 11 of the 20 within-acquisition coefficients of variation were less than the corresponding between-acquisition coefficients of variation, eight of which were only smaller by less than 1%. It was interesting to find that measurements obtained during the same session did not have obviously better reproducibility than those obtained separately on different days. It is unlikely that biologic or machine variability influenced this finding because continuous measurements were performed within a short period of time (<30 min) and between-acquisition reproducibility was similar to that of within-acquisition for all the spectroscopic studies. Motion artifacts caused by movement of the subject during the prolonged examination time, which was necessary for continuous acquisitions, could be the key factor; these artifacts could result in poorly resolved spectra that were susceptible to being integrated inaccurately or excluded from the study. The expected high repeatability of within-acquisition measurements was thus hampered but might be improved with better training or positioning and fixation of the head that is more comfortable for the subject.
Several studies concerning the reproducibility of cerebral metabolite ratios measured by in vivo proton MR spectroscopy have been published [28, 29, 31,32,33]. Marshall et al. [31] reported the coefficients of variation for the NAA/Cho and NAA/Cr ratios were 10-26%, using single-voxel spectroscopy with a TE of 135 msec and a volume of interest of 8 cm3 in the parietal white matter. Narayana et al. [32] and Jackson et al. [33] suggested that the overall coefficient of variation was 36% for the NAA/Cr ratio and 39% for the Cho/Cr ratio, using single-voxel spectroscopy with a TE of 100 msec and a volume of interest of 27 cm3 in the frontal lobe. Despite different acquisition parameters and volumes of interest, coefficients of variation for corresponding metabolite ratios in our hippocampal single-voxel spectroscopic studies (Table 1) were comparable with those of Marshall et al. and much better than those of Narayana et al. and Jackson et al. This finding implies that even though the hippocampal region has worse field homogeneity than the upper cerebral region, reasonable reproducibility of hippocampal single-voxel spectroscopy can be obtained with an automated spectroscopic technique. Using proton chemical shift imaging with different TEs (272, 20, and 30 msec) and volumes of interest (>85 x 85 x 15 mm) in the high cerebral region, Tedeschi et al. [28], Jackson et al., and Charles et al. [29] reported between-acquisition coefficients of variation were 18.0-22.2% for the NAA/Cr ratio, 12.8-25.8% for the NAA/Cho ratio, and 16.0-21.0% for the Cho/Cr ratio. In our chemical shift imaging studies of bilateral hippocampi, corresponding coefficients of variation of the posterior hippocampus (18-25%) were comparable with those reported in the literature, whereas those of the anterior hippocampus (29-41%) were much greater. These findings indicate that when automated chemical shift imaging with large volumes of interest is performed for the evaluation of the hippocampus, good reproducibility can be obtained only in the posterior hippocampal region. Because of the prominent susceptibility effect and suboptimal shimming, the reproducibility of chemical shift imaging in the anterior hippocampal region was obviously worse than in higher cerebral region. Regional variation of the coefficient of variation in different brain structures must be considered when MR spectroscopy is performed for clinical practice, especially in longitudinal studies. This study showed that there was a significant difference in the NAA/(Cho+Cr) ratio for the anterior versus that for the posterior hippocampus. Similarly, Vermathen et al. [34] studied proton MR spectroscopic imaging in the hippocampus of 14 control subjects and found NAA/(Cho+Cr) ratios were significantly higher in the posterior than the anterior hippocampus. Regional variation of the NAA/(Cho+Cr) ratios could occur as a result of a combination effect due to more neurons in the posterior than the anterior hippocampal tissue and the decreasing thickness of the hippocampus from anterior to posterior, leading to different contributions from extrahippocampal tissues [34]. These findings imply that precise voxel placement in the hippocampus is extremely important for comparison of hippocampal proton spectra from the contralateral side, between different subjects, or in longitudinal studies. In addition, our results showed that all the metabolite ratios obtained by single-voxel spectroscopy and by unilateral and bilateral chemical shift imaging techniques were different from each other, except the Cho/Cr ratio for the posterior hippocampus. This finding means that even if similar acquisition parameters are used for spectral measurement, the intertechnique difference in hippocampal metabolite ratios is significant. These findings suggest that consistent acquisition technique, precise selection of the volume of interest, and standard data presentation are important for obtaining reproducible and clinically useful hippocampal proton MR spectroscopic data.
There are several limitations to this study. First, when this study was performed, our MR scanner could apply the volume of interest only at a fixed transverse section. It would be better to align an oblique volume of interest along the long axis of the hippocampus rather than to manipulate the subject's head. The spectral quality and reproducibility can be expected to be better if the patient needs not extend the neck backward during the long acquisition time. However, further assessment is mandatory before a conclusion can be made. Second, the current study was conducted on a 1.5-T clinical MR scanner with a regular head coil, which always resulted in suboptimal separation of the Cho and Cr peaks in hippocampal spectra. Any technique that improves the signal-to-noise ratio, such as a higher field strength or a better head coil, may distinguish the Cho and Cr peaks more effectively and reduce the superiority of the NAA/(Cho+Cr) ratio over the other metabolite ratios. Third, the spatial resolution of chemical shift imaging in this study was coarse, and signal intensities from nonhippocampal tissue were always included within the voxel. Advancement in acquisition techniques and data processing that can improve local shimming and spectra fitting, may delineate more specifically the different metabolite profiles of the anterior and the posterior hippocampus.
In conclusion, the findings of our 1.5-T spectroscopic study indicate the following: when proton MR spectroscopy is applied for hippocampal evaluation, the NAA/(Cho+Cr) ratio derived from single-voxel spectroscopy or unilateral chemical shift imaging is the most reproducible parameter; when only the anterior hippocampal region is examined or patients are less cooperative, single-voxel spectroscopy is an appropriate technique because of the good reproducibility and its short acquisition time; when spectroscopic studies are performed in an individual subject over time, the difference in the hippocampal NAA/(Cho+Cr) ratio between two measurements must be greater than 20-25% for identification of a significant change; and consistent acquisition technique and volume of interest placement are mandatory for clinical application of hippocampal MR spectroscopy.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  Y. J. Choi, J. K. Kim, H. J. Kim, and K.-S. Cho Interobserver Variability of Transrectal Ultrasound for Prostate Volume Measurement According to Volume and Observer Experience Am. J. Roentgenol., February 1, 2009; 192(2): 444 - 449. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. M. Wellard, R. S. Briellmann, C. Jennings, and G. D. Jackson Physiologic Variability of Single-Voxel Proton MR Spectroscopic Measurements at 3T AJNR Am. J. Neuroradiol., March 1, 2005; 26(3): 585 - 590. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) and 1-2
([UNK]) are performed continuously without changing volume of interest to
evaluate within-acquisition reproducibility. Measurements 2 (
) and 3
(
) and the first one of the continuous measurements (i.e., measurement
1-1) are for evaluation of between-acquisition reproducibility. Single-voxel
spectroscopic studies. Line and dash lines show mean and standard deviation,
respectively.
