First published online September 27, 2006
British Journal of Radiology (2007) 80, 307-320
© 2007 British Institute of Radiology
doi: 10.1259/bjr/52670770
3D MRI in multiple sclerosis: a study of three sequences at 3 T
R J Mills, MRCP
1
C A Young, FRCP
1 and
E T S Smith, FRCP
2
Departments of 1 Neurology and 2 Neuroradiology, The Walton Centre for Neurology and Neurosurgery, Liverpool, UK
Correspondence: Dr Roger J Mills, Neurology, Walton Centre for Neurology and Neurosurgery, Lower Lane, Fazakerley, Liverpool L9 7LJ, UK. E-mail: rjm{at}crazydiamond.co.uk
 |
Abstract
|
|---|
The objective of this study was to assess the feasibility of using 3D acquisition at 3 T for imaging patients with multiple sclerosis (MS). Feasibility was assessed by three criteria based on acquisition time, specific absorption rate (SAR) and image quality. 47 patients with clinically definite MS underwent imaging in a Siemens 3T Trio MR scanner. Patient safety data were obtained following the scan sessions. The study had local ethics approval. The following three-dimensional (3D) sequences, all acquired coronally, were used: T2 fluid attenuated inversion recovery (FLAIR) (repetition time (TR) 6000 ms, echo time (TE) 353 ms, inversion time (TI) 2200 ms), 0.5x0.5x1 mm voxels, acquisition time 10 min 38 s; T2 turbo spin echo (TSE) (TR 3000 ms, TE 354 ms), 1x1x1 mm voxels, acquisition time 8 min 29 s; T1 inversion recovery (IR) (TR 2040 ms, TE 5.56 ms, TI 1100 ms), matrix 512x448 (0.5x0.5 mm pixels), 0.5x0.5x1 mm voxels, acquisition time 7 min 38 s. Total acquisition time was 26 min 45 s. Example images are presented. 3D scanning at 3 T provides highly detailed, high quality images with acquisition times tolerated by MS patients, even by those with severe disability. The volumetric data are suitable for a wide variety of post-processing techniques; the authors suggest that 3D studies at 3 T should be considered as the possible brain imaging protocol for either cross-sectional or longitudinal studies in MS and that the 3D T2 FLAIR sequence should be considered for the purposes of radiological diagnosis.
|
Introduction
|
|---|
The objective of this study was to assess the feasibility of using three-dimensional (3D) volume acquisition sequences at 3 T for whole brain imaging in a cross-section of patients with multiple sclerosis (MS).
|
Background
|
|---|
The study was set in the context of a diencephalic, MS lesion distribution study whose protocol demanded high resolution, contiguous slice images of no greater than 1 mm thickness. It was surmised that a volume acquisition would be desirable since this would yield a high spatial resolution and images could be reconstructed for viewing in different planes (particularly useful for identifying diencephalic and mesencephalic anatomy). Also, by using a 3 T machine, the high field strength per se would improve the imaging of deep brain structures [1–6]. However, such an imaging protocol would have to satisfy predefined feasibility criteria (see below) and it was realized that these criteria would also be relevant for future studies, and even routine diagnostic imaging.
|
Methods and Materials
|
|---|
The feasibility criteria were defined as follows: (1) reasonable acquisition time, defined as a duration of the total protocol of less than 1 h and, from the patient's perspective, a total scan time which was easily tolerable with a low incidence of head movement; (2) specific absorption rate of each sequence less than 3.2 W kg–1; 3) a minimum image quality such that lesion conspicuity was equal to or greater than "standard" [7] diagnostic two-dimensional (2D) sequences at 1.5 T.
51 patients with clinically definite MS [8] were chosen at random from a patient database. All subjects gave written, informed consent. Scanning was performed on a Siemens 3T Trio machine (Siemens, Erlangen, Germany) using a body coil as the transmitting coil and an 8 channel head coil as the receiving coil. T2 weighted, 3D sequences, largely based on the manufacturer's recommended settings, were trialled on the first four patients, adjustments being made in order to overcome such problems as wrap. For the purposes of the study, their data were discarded. Both a 3D T2 turbo spin echo (TSE) and a 3D T2 fluid attenuated inversion recovery (FLAIR) sequence were applied to the remaining 47 patients. In order to optimize anatomical understanding of the diencephalon, a T1 weighted, 3D sequence was added to the protocol and was applied to 40 patients. Two patients had additional sequences with contrast (0.1 ml kg–1, i.e. half the manufacturer's recommended dose [9], dimeglumine gadopentetate 469.01 mg ml–1); 11 also had a standard 2D T2 TSE, for comparison, on the 3 T machine and then went on immediately to have another 2D T2 TSE and a 3D T2 TSE on a Siemens 1.5 T Symphony scanner which was located in the same department. Images from nine normal subjects were also obtained using the 3 T system. Post-scan safety data (heart rate, respiratory rate and blood pressure) were obtained on all subjects and any adverse events during the imaging recorded. For 2 of the 47 patients this was their first experience of any type of neuroimaging. The study had both local ethics committee (04/Q1501/154) and research governance committee (Walton Centre for Neurology) approval.
Viewing and post-processing of images were performed on OsiriX (v1.7.1; http://homepage.mac.com/rossetantoine/osirix/) on a Macintosh computer (http://www.apple.com/) by both a neurologist (RJM) and an experienced neuroradiologist (ETSS). Selected images were also read from hard-copy. Image quality was assessed quantitatively by calculation of signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) on a random subset of the patients and all patients who had undergone the 1.5 T imaging. Identical, circular regions of interest (ROIs), of 0.05 cm2 area, were placed over the brightest (or darkest on T1 inversion recovery (IR) images) MS lesion, over cerebrospinal fluid (CSF) in the lateral ventricle, over normal appearing white matter (NAWM), over normal appearing grey matter (NAGM) and outside the body to measure noise. The SNR was calculated as mean signal of tissue divided by standard deviation (SD) of noise. The CNR was calculated as mean signal of lesion minus mean signal of tissue divided by SD of noise [10]. Values were taken from one slice and one lesion in each patient.
|
Results
|
|---|
Subject characteristics
The MS patients had the following characteristics: 34 females, 13 males; mean age 47.0 years (SD 8.1, range 34–63 years); 5 had primary progressive, 26 had relapsing remitting and 16 had secondary progressive disease; mean duration of MS was 16.9 years (SD 9.4, range 2–40 years); range of disability by Expanded Disability Status Scale (EDSS) [11] was 1.5–8.5 (modal values 4.0 and 6.0). The normal subjects had the following characteristics: 5 females, 4 males; mean age 41.1 years (SD 6.8, range 28–52 years).
Scan protocols
Three 3D sequences were used. All were acquired in the coronal plane (aligned to the posterior surface of brain stem): T2 FLAIR (TR 6000 ms, TE 353 ms, TI 2200 ms), matrix 512x440 (0.5x0.5 mm pixels), 160 contiguous slices of 1 mm thickness, acquisition time 10 min 38 s; T2 TSE (TR 3000 ms, TE 354 ms), matrix 256x192 (1x1 mm pixels), 160 contiguous slices of 1 mm thickness, acquisition time 8 min 29 s; T1 IR (TR 2040 ms, TE 5.56 ms, TI 1100 ms), matrix 512x448 (0.5x0.5 mm pixels), 176 contiguous slices of 1 mm thickness, acquisition time 7 min 38 s. Total acquisition time for the three sequences was 26 min 45 s.
Table 1
shows the setup parameters in detail; included for comparison are those for the "standard" 2D axial T2 TSE on both the 3 T and 1.5 T machines and an abbreviated 3D sequence (only 96 slices) on the 1.5 T machine.
Example images
Figure 1 shows an example image, in the plane of acquisition, from each of the three study sequences (subject is a 57-year-old female, duration of MS 10 years, EDSS 7.5, primary progressive disease).
View larger version (84K):
[in this window]
[in a new window]
|
Figure 1. Coronal three-dimensional 3 T (a) T2 fluid attenuated inversion recovery (FLAIR), (b) T2 turbo spin echo (TSE) and (c) T1 inversion recovery (IR) images.
|
|
There was clear demonstration of both white matter disease and high anatomical detail across all sequences. On the T2 weighted images, the white matter not containing obvious lesions, which might ordinarily be classed as NAWM, appeared to have a coarse texture and was not, in fact, thought to be normal (see below).
Figures 2–4 show both acquisition and reconstructed plane images for each sequence (subject is a 35-year-old female, duration of MS 12 years, EDSS 3.5, relapsing remitting disease).
View larger version (86K):
[in this window]
[in a new window]
|
Figure 2. Three-dimensional 3 T T2 fluid attenuated inversion recovery (FLAIR) orthogonal multiplanar reconstruction (MPR).
|
|
View larger version (89K):
[in this window]
[in a new window]
|
Figure 4. Three-dimensional 3 T T1 inversion recovery (IR) orthogonal multiplanar reconstruction (MPR).
|
|
The reconstructed images retain high anatomical detail and excellent depiction of lesions at the corticomedullary junction, in the posterior fossa, in the corpus callosum and adjacent to CSF spaces.
The post-contrast images (Figure 5) clearly demonstrate the major vasculature as well as small cortical vessels and intraparenchymal vessels in addition to the meningeal enhancement normally seen at 3 T. There were no examples of an enhancing lesion in the patients scanned.
View larger version (115K):
[in this window]
[in a new window]
|
Figure 5. Three-dimensional 3 T T1 inversion recovery (IR), post-contrast, orthogonal multiplanar reconstruction (MPR) (63-year-old female, duration of MS 27 years, EDSS 5.0, relapsing remitting disease).
|
|
Image quality
The SNR and CNR values for all the sequences used are shown in Table 2. Comparisons were made, by t-test, across the T2 TSE sequences only (grey columns).
There were no differences in either the SNR or CNR values between the 3 T 3D and the 1.5 T 2D T2 TSE sequences. The 3 T 2D sequence gave the highest signal and greatest CNRs across all tissues; however, the increase in lesion to NAWM CNR was not significantly better on 3 T 2D than the 3 T 3D sequence (p = 0.11). The 1.5 T 3D sequence was poor with SNRs significantly worse than both the 1.5 T 2D and 3 T 3D for all tissues; the lesion to NAWM CNR was approximately 56% less on 1.5 T 3D than the 3 T 3D sequence.
Image quality examples: 2D vs 3D
Figure 6 shows a comparison between 2D axial T2 TSE images at 1.5 T (Figure 6a) and 3 T (Figure 6b) and a reconstructed image in the same plane from the 3D 3 T T2 TSE sequence, all are from the same subject (a 35-year-old female, duration of MS 13 years, EDSS 4.5, relapsing remitting disease). Both the 2D images are 5 mm thick, the reconstructed slice is 1 mm thick, hence the boundaries of the lateral ventricles are much more distinct on the thinner slice. Lesions are most conspicuous and demarcated on the 3D reconstructed image; simply comparing these three images alone, there are 7 plaques on the 1.5 T 2D slice, 5 plaques on the 3 T 2D slice and 13 plaques on the 3 T 3D reconstructed slice. Again, the coarse texture of the white matter can be seen on the 3 T 3D image, especially when compared with either of the 2D images.
Image quality examples: 1.5 T vs 3 T 3D
Figure 7 shows a comparison between a 3D volume acquisition at 1.5 T (Figure 7a) and 3 T (Figure 7b) (subject is a 50-year-old male, duration of MS 20 years, EDSS 6.5, secondary progressive disease). The quality of the 3 T image greatly exceeds that of the 1.5 T image, largely due to superior CNR (148 vs 65) and SNR (252 vs 91) (Table 2
). There are 6 plaques visible on the 1.5 T image and 13 plaques on the 3 T image.
View larger version (89K):
[in this window]
[in a new window]
|
Figure 7. (a) 1.5 T three-dimensional T2 turbo spin echo (TSE) (best window level shown) (1 mm thick, 6 lesions) and (b) 3 T three-dimensional T2 TSE, right (1 mm thick, 13 lesions) images of the same patient.
|
|
The resolving power of the 3 T 3D examination was such that small structures, e.g. Virchov–Robin spaces or the intraparenchymal vessels seen in the post-contrast images, in the order of 600–700 µm in width, could be distinguished. Similarly, some of the smallest MS lesions detectable were between 0.9 mm and 1 mm in diameter.
The coarseness of the texture on the T2 weighted images of apparently unaffected white matter was either due to artefact from the high field strength, or was a real phenomenon representing abnormality in the white matter in MS. If it was purely a result of high field strength, then no difference in the texture of white matter would be expected between the MS patients and the normal subjects. In order to investigate this, identical circular ROIs of 0.45 cm2 area were placed over lesion free white matter for each patient and over white matter for each normal subject and the SD of the signal, on the T2 FLAIR images, recorded. There was a small but significant increase, by t-test, in the SD of the signal from the white matter of the MS patients compared with the controls: mean difference 0.81 (95%CI 0.39–1.23, p<0.01), suggesting that the visible heterogeneity of signal was possibly related to the MS disease process rather than artefact. This altered texture was not thought to be caused by normal vascular structures since the signal changes were more subtle and less well defined than the discrete punctate appearances typical of Virchow–Robin spaces (see below). It was also thought to be different from what is sometimes known as "dirty-appearing" white matter (DAWM) [12], i.e. larger confluent areas of increased signal not discrete enough to be described as an MS lesion. Figure 8 shows a T2 TSE image of a large area of DAWM off the left occipital horn of the lateral ventricle. Also seen within this are bright, punctate Virchow–Robin spaces.
View larger version (91K):
[in this window]
[in a new window]
|
Figure 8. 3 T reconstructed axialT2 turbo spin echo (TSE) showing a confluent area of diffuse increased signal typical of dirty appearing white matter (DAWM). Small punctate areas of higher signal can be seen within this, representing Virchow–Robin spaces.
|
|
Figure 9 shows an example of white matter, magnified from the left frontal lobe, of a normal subject (Figure 9a) and an MS subject (Figure 9b,c) of the same sex and a similar age. There are subtle irregular changes in the brightness of the signal in the white matter of the MS patient, which are different from DAWM and which do not appear to be in keeping with vascular structures since there are no corresponding low signal changes on the T1 image.
View larger version (105K):
[in this window]
[in a new window]
|
Figure 9. 3 T reconstructed axialT2 turbo spin echo (TSE) images showing white matter in (a) a normal subject and (b) a multiple sclerosis (MS) patient. (c) Also shown is the corresponding T1 image for the same MS patient. There appears to be a coarser signal texture, on the T2 image, in the MS patient which is not thought to be due simply to Virchow–Robin spaces.
|
|
Normal subjects at 3 T
As part of the lesion distribution study, nine normal control subjects underwent the same imaging protocol. As mentioned above, the resolution of the images allowed the highly vascular nature of the brain parenchyma to become apparent with innumerable Virchow–Robin spaces visible on the T2 TSE with corresponding low signal on the T1 IR sequence (Figure 10). This perivascular high signal was generally suppressed with FLAIR. Such small perivascular spaces are not generally visible at 1.5 T, certainly with slice thickness greater than 1 mm.
View larger version (135K):
[in this window]
[in a new window]
|
Figure 10. Coronal, axial and sagittal(a–c) T2 turbo spin echo (TSE) and (d–f) T1 inversion recovery (IR) images of a normal subject showing innumerable Virchow–Robin spaces in the white matter.
|
|
Tolerability, safety and adverse events
The procedure was well tolerated by almost all subjects, even those with severe disability; the most disabled patients participating were of EDSS 8.5, i.e. wheelchair bound with limited upper limb use. Table 3 shows the number of sequences degraded by movement artefact, as determined by subjective visual assessment of the images. The sequences were applied in the same order to each patient, starting with the T2 FLAIR. The overall proportion of sequences which displayed movement artefact was about 7% and in each case, the degradation was classed as slight. Only one patient had artefact in all three sequences.
View this table:
[in this window]
[in a new window]
|
Table 3. Number of 3 T 3D sequences with movement artefact. 10 out of 134 individual sequences displayed movement artefact
|
|
There were three minor adverse events: one patient became anxious and aborted the scan just prior to the termination of the T2 FLAIR sequence, she was subsequently rescanned without complication; one patient complained of transient facial flushing; and one patient developed vertiginous symptoms lasting 5 min following completion of the scan. Post-scan observations in all patients were within acceptable limits, with none of the subjects exhibiting tachycardia or tachypnoea: mean heart rate (HR) 73 beats per min (bpm) (range 61–90 bpm), mean respiratory rate (RR) 13 min–1 (range 10–20 min–1), mean blood pressure (BP) 129/80 mmHg (systolic range 100–169 mmHg, diastolic range 60–106 mmHg).
|
Discussion
|
|---|
It is important in patient based studies that imaging should be tolerated easily by the subjects. Sequences for research protocols are often developed on normal volunteers who may be able to undergo prolonged image acquisition without difficulty. This may be unsuitable for patients with severe neurological disability, a consideration which is likely to become more important as the attention of clinical trials in MS turns to patients with progressive disease and higher disability. In the present study, selection and adjustment of the sequences was deliberately performed on patients from the outset. This provided practical guidance on which scan times might be reasonable for use on subsequent patients. The chosen protocol could be completed in under half an hour; some institutions recommend safety guidelines that no patient should be in the bore of a 3 T magnet for more than 1 h [13] and so adequate time would have been available for further sequences, if required. The protocol was tolerated by all the subjects in the study with only a small risk of movement artefact; it was also notable that the noise from the scanner did not cause distress. The first feasibility criterion of reasonable acquisition time was satisfied.
At high field strength and with intensive acquisition, it is important that the specific absorption rates of the examinations are within safe limits. All sequences were within the Medicines and Healthcare products Regulatory Agency (MHRA) guideline limits of 4 W kg–1 [14] and also fell below the manufacturer's guideline limit of 3.2 W kg–1, thus the second feasibility criterion of acceptable SAR was satisfied.
It was found that, when using volume acquisition with voxels of 1.0x1.0x1.0 mm or less, both acquisition time and SNR were largely dependent on field strength, as expected. At 1.5 T, the 3D sequence image quality was relatively poor and the acquisition time for even an abbreviated sequence exceeded that on the 3 T by about 50%. It was shown that the SNR and CNR of the volume acquisition at 3 T were comparable with the thicker slice, 1.5 T 2D axial images, at least for the T2 TSE sequences. However, many more lesions were visible in the 3 T 3D images. By these objective measures, the third feasibility criterion of image quality being equal to or better than 1.5 T 2D images was satisfied; certainly, the opinion of both the investigators and other neuroradiologists [15] was that the images, subjectively, were of high quality. Very small structures, i.e. submillimetre, of both anatomical and pathological nature could be discerned. Subtle visual abnormalities in lesion free white matter were noticeable on the T2 weighted images, something which is only normally made apparent by diffusion tensor imaging or calculation of magnetization transfer [16–18]. To the authors' knowledge, such visually detectable texture abnormality in the brain of MS patients has not been reported previously. Others have shown texture abnormality in the spinal cord of MS patients [19], but not in the brain [20]. Further work is necessary to determine whether this texture abnormality has any clinical or other radiological correlation.
The principal advantage of the 3 T 3D sequences was that a true volumetric data set, of high spatial resolution, was acquired providing the opportunity of applying many post-processing techniques. For example, the ability to both reconstruct and simultaneously view images in three orthogonal planes greatly enhances their diagnostic power, not only in localizing lesions in three dimensions but also by resolving questions of partial voluming. For longitudinal studies in MS, great precision is required in scan alignment at each examination if using 2D sequences, usually prolonging the time spent in the scanner for the patient, in order to allow co-registration of lesion sites; 3D data can be aligned by post-processing techniques thus reducing the precision required for scan alignment [21]. A typical MRI outcome measure in trials for MS therapy is rate of cerebral atrophy [22]; a 3D data set would be an ideal substrate for whole brain volume estimation. Another advantage of using high field strength MRI is that smaller doses of contrast medium are required in order to enhance images [23].
Many studies have failed to correlate MS lesion load with various clinical or biological markers of disease [17, 24, 25]. This may be due to underestimation of lesions at lower field strengths; there is emerging evidence that 2D imaging at 3 T, compared with 1.5 T, increases significantly the detection of MS lesions [26–30], or it may be due to suboptimal scan geometry [18, 31, 32]. Thin, contiguous slice imaging at 3 T may therefore have considerable impact on studies of lesion load or distribution.
Some disadvantages of the 3 T 3D sequences were, first, the data required large amounts of storage space (typically 190 Mb, including localizer, per patient). Second, the 3 T 3D acquisition times were about three times longer than individual 2D sequences, of standard slice thickness, at 1.5 T, which may be used for diagnostic purposes. Although not problematic for the patient, this may have an implication for departmental throughput. However, in a research setting, it should be considered that the apparent time penalty may be offset by the time of performing extra 2D sequences for different planes of acquisition and the time required for scan alignment and possible re-scan time following subject movement. Notwithstanding this, some 2D sequences used in MS research are long, e.g. 12 min 18 s for 1.5 T axial T2 FLAIR of 3 mm slice thickness [27].
Each of the three sequences employed in the study provided different pathological and structural information which may all have been useful in the assessment of a patient with MS. However, the most common reason for imaging in MS is to reveal the extent of white matter disease and, to this end, the authors felt that the T2 FLAIR sequence gave the greatest information. Not only were lesions well demonstrated, but also other abnormality in apparently lesion free white matter was detectable. It was easier to distinguish lesions from normal vascular appearances using this sequence, compared with the T2 TSE sequence. Therefore, it would be the sequence of choice if necessity dictated that only one sequence could be used, perhaps for example in a time constrained, diagnostic setting.
It was not possible to comment on the sensitivity of the 3D 3 T sequences to lesion enhancement. This would require a separate study and a patient group purposively chosen for more active disease. However, there is little reason to assume that it would be significantly worse than standard 2D imaging [27, 28].
|
Conclusion
|
|---|
Whole brain volume acquisition at 3 T can produce very high quality images in a relatively short time and is entirely feasible for application to MS patients, even those with severe disability, something which was not readily achievable at 1.5 T because of excessive acquisition time and poor image quality. The high spatial resolution reveals visual abnormalities in the texture of lesion free white matter on T2 weighted images. 3D acquisition carries the advantage of allowing all the post-processing techniques available for volumetric data and the authors recommend such sequences for research protocols, and suggest that the 3D T2 FLAIR alone is suitable for the purposes of radiological diagnosis.
|
Acknowledgments
|
|---|
The authors would like to thank the patient and control subjects for their willing participation in this study. Also, thanks go to the staff at the Magnetic Resonance and Image Analysis Research Centre, Liverpool University, for their assistance with this work.
Received for publication July 21, 2006.
Accepted for publication August 17, 2006.
|
References
|
|---|
- Schwindt W, Kugel H, Bachmann R, Kloska S, Allkemper T, Maintz D, et al. Magnetic resonance imaging protocols for examination of the neurocranium at 3 T. Eur Radiol 2003;13:2170–9.[CrossRef][Medline]
- Pinker K, Ba-Ssalamah A, Wolfsberger S, Mlynarik V, Knosp E, Trattnig S. The value of high-field MRI (3T) in the assessment of sellar lesions. Eur J Radiol 2005;54:327–34.[CrossRef][Medline]
- Vinogradov E, Degenhardt A, Smith D, Marquis R, Vartanian TK, Kinkel P, et al. High-resolution anatomic, diffusion tensor, and magnetization transfer magnetic resonance imaging of the optic chiasm at 3T. J Magn Reson Imaging 2005;22:302–6.[CrossRef][Medline]
- Bink A, Schmitt M, Gaa J, Mugler JP 3rd, Lanfermann H, Zanella FE. Detection of lesions in multiple sclerosis by 2D FLAIR and single-slab 3D FLAIR sequences at 3.0 T: initial results. Eur Radiol 2006;16:1104–10.[CrossRef][Medline]
- Mugler JP 3rd, Bao S, Mulkern RV, Guttmann CR, Robertson RL, Jolesz FA, et al. Optimized single-slab three-dimensional spin-echo MR imaging of the brain. Radiology 2000;216:891–9.[Abstract/Free Full Text]
- Tubridy N, Barker GJ, Macmanus DG, Moseley IF, Miller DH. Three-dimensional fast fluid attenuated inversion recovery (3D fast FLAIR): a new MRI sequence which increases the detectable cerebral lesion load in multiple sclerosis. Br J Radiol 1998;71:840–5.[Abstract]
- Simon JH, Li D, Traboulsee A, Coyle PK, Arnold DL, Barkhof F, et al. Standardized MR imaging protocol for multiple sclerosis: consortium of MS Centers consensus guidelines. AJNR Am J Neuroradiol 2006;27:455–61.[Free Full Text]
- Poser CM, Paty DW, Scheinberg L, McDonald WI, Davis FA, Ebers GC, et al. New diagnostic criteria for multiple sclerosis: guidelines for research protocols. Ann Neurol 1983;13:227–31.[CrossRef][Medline]
- Data sheet for Magnevist®. Schering, 2006
- Kaufman L, Kramer DM, Crooks LE, Ortendahl DA. Measuring signal-to-noise ratios in MR imaging. Radiology 1989;173:265–7.[Abstract/Free Full Text]
- Kurtzke JF. Rating neurologic impairment in multiple sclerosis: an expanded disability status scale (EDSS). Neurology 1983;33:1444–52.[Abstract/Free Full Text]
- Ge Y, Grossman RI, Babb JS, He J, Mannon LJ. Dirty-appearing white matter in multiple sclerosis: volumetric MR imaging and magnetization transfer ratio histogram analysis. AJNR Am J Neuroradiol 2003;24:1935–40.[Abstract/Free Full Text]
- Smith ETS. Guidelines for use of MRI at The Walton Centre for Neurology and Neurosurgery. 2005
- Medical Devices Agency. Guidelines for magnetic resonance equipment in clinical use. 2002
- Mills RJ, Young CA, Smith ET. 3D 3T MRI in multiple sclerosis. British Society of Neuroradiology, 2005
- Laule C, Vavasour IM, Moore GR, Oger J, Li DK, Paty DW, et al. Water content and myelin water fraction in multiple sclerosis. A T2 relaxation study. J Neurol 2004;251:284–93.[CrossRef][Medline]
- Miller DH, Thompson AJ, Filippi M. Magnetic resonance studies of abnormalities in the normal appearing white matter and grey matter in multiple sclerosis. J Neurol 2003;250:1407–19.[CrossRef][Medline]
- Wayne Moore GR. MRI-clinical correlations: more than inflammation alone–what can MRI contribute to improve the understanding of pathological processes in MS? J Neurol Sci 2003;206:175–9.[CrossRef][Medline]
- Mathias JM, Tofts PS, Losseff NA. Texture analysis of spinal cord pathology in multiple sclerosis. Magn Reson Med 1999;42:929–35.[CrossRef][Medline]
- Gasperini C, Horsfield MA, Thorpe JW, Kidd D, Barker GJ, Tofts PS, et al. Macroscopic and microscopic assessments of disease burden by MRI in multiple sclerosis: relationship to clinical parameters. J Magn Reson Imaging 1996;6:580–4.[Medline]
- Prima S, Ourselin S, Ayache N. Computation of the mid-sagittal plane in 3-D brain images. IEEE Trans Med Imaging 2002;21:122–38.[CrossRef][Medline]
- Bermel RA, Bakshi R. The measurement and clinical relevance of brain atrophy in multiple sclerosis. Lancet Neurol 2006;5:158–70.[CrossRef][Medline]
- Trattnig S, Pinker K, Ba-Ssalamah A, Nobauer-Huhmann IM. The optimal use of contrast agents at high field MRI. Eur Radiol 2006;16:1280–7.[CrossRef][Medline]
- Mainero C, Faroni J, Gasperini C, Filippi M, Giugni E, Ciccarelli O, et al. Fatigue and magnetic resonance imaging activity in multiple sclerosis. J Neurol 1999;246:454–8.[CrossRef][Medline]
- van der Werf SP, Jongen PJ, Lycklama a Nijeholt GJ, Barkhof F, Hommes OR, Bleijenberg G. Fatigue in multiple sclerosis: interrelations between fatigue complaints, cerebral MRI abnormalities and neurological disability. J Neurol Sci 1998;160:164–70.[CrossRef][Medline]
- Bachmann R, Reilmann R, Schwindt W, Kugel H, Heindel W, Kramer S. FLAIR imaging for multiple sclerosis: a comparative MR study at 1.5 and 3.0 Tesla. Eur Radiol 2006;16:915–21.[CrossRef][Medline]
- Nielsen K, Rostrup E, Frederiksen JL, Knudsen S, Mathiesen HK, Hanson LG, et al. Magnetic resonance imaging at 3.0 tesla detects more lesions in acute optic neuritis than at 1.5 tesla. Invest Radiol 2006;41:76–82.[CrossRef][Medline]
- Sicotte NL, Voskuhl RR, Bouvier S, Klutch R, Cohen MS, Mazziotta JC. Comparison of multiple sclerosis lesions at 1.5 and 3.0 Tesla. Invest Radiol 2003;38:423–7.[CrossRef][Medline]
- Bachmann R, Reilmann R, Schwindt W, Kugel H, Heindel W, Kramer S. FLAIR imaging for multiple sclerosis: a comparative MR study at 1.5 and 3.0 Tesla. Eur Radiol 2006;16:915–21.[CrossRef][Medline]
- Wattjes MP, Lutterbey GG, Harzheim M, Gieseke J, Traber F, Klotz L, et al. Higher sensitivity in the detection of inflammatory brain lesions in patients with clinically isolated syndromes suggestive of multiple sclerosis using high field MRI: an intraindividual comparison of 1.5 T with 3.0 T. Eur Radiol 2006;16:2067–73.[CrossRef][Medline]
- Tubridy N, Barker GJ, MacManus DG, Moseley IF, Miller DH. Optimisation of unenhanced MRI for detection of lesions in multiple sclerosis: a comparison of five pulse sequences with variable slice thickness. Neuroradiology 1998;40:293–7.[CrossRef][Medline]
- Hashemi RH, Bradley WG Jr, Chen DY, Jordan JE, Queralt JA, Cheng AE, et al. Suspected multiple sclerosis: MR imaging with a thin-section fast FLAIR pulse sequence. Radiology 1995;196:505–10.[Abstract/Free Full Text]
Top
Abstract
Introduction
7 lesions), (b) 3 T two-dimensional T2 TSE axial (5 mm thick,