HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Appendix All Versions of this Article: 2441060930v1 244/1/249    most recent Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Yu, C. S. Articles by Chan, P. Search for Related Content PubMed PubMed Citation Articles by Yu, C. S. Articles by Chan, P.

Relapsing Neuromyelitis Optica and Relapsing-Remitting Multiple Sclerosis: Differentiation at Diffusion-Tensor MR Imaging of Corpus Callosum¹

¹ From the Departments of Radiology (C.S.Y., K.C.L., W.Q.) and Neurology (Y.X., H.S., P.C.), Xuanwu Hospital, Capital University of Medical Sciences, 45 Chang-Chun St, Xuanwu District, Beijing 100053, People's Republic of China; and State Key Laboratory of Cognitive Neuroscience and Learning, Beijing Normal University, Beijing, People's Republic of China (C.Z.Z.). Received May 29, 2006; revision requested August 1; revision received August 14; accepted September 18; final version accepted November 15. Supported in part by the Natural Science Foundation of China (no. 30670601), Beijing Scientific and Technological New Star Program (2005B21), and Beijing Natural Science Foundation (7042026). Address correspondence to K.C.L. (e-mail: kunchengli{at}yahoo.com.cn).

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCE IN KNOWLEDGE IMPLICATION FOR PATIENT CARE References

 
Purpose: To prospectively assess sensitivity and specificity of diffusion indexes of the corpus callosum (CC) for differentiating relapsing neuromyelitis optica (RNMO) from relapsing-remitting multiple sclerosis (RRMS), by using final clinical diagnosis as the reference standard.

Materials and Methods: Participants provided informed consent; the study was approved by the institutional review board. Forty-six consecutive patients with RRMS (18 men, 28 women; mean age, 37.7 years; range, 18–58 years) and 26 consecutive patients with RNMO (two men, 24 women; mean age, 38.6 years; range, 19–59 years) underwent diffusion-tensor magnetic resonance imaging. Mean diffusivity (MD) and fractional anisotropy (FA) of the region of interest (ROI) of the CC in the midsagittal plane were measured and used as discriminative indexes. Bayesian classification with leave-one-out cross-validation was used to determine diagnostic accuracy. Differences in diffusion indexes of ROIs among groups were evaluated by using the Kruskal-Wallis test, followed by the Mann-Whitney U test for multiple comparisons and Bonferroni correction.

Results: Mean MD (8.48 x 10^–4 mm²/sec) and FA (0.729) of the ROI in patients with RNMO were significantly (P < .001) different from those (MD = 10.64 x 10^–4 mm²/sec, FA = 0.599) in patients with RRMS. Sensitivity and specificity for differentiation were 92.3% (24 of 26 patients with RNMO) and 93.5% (43 of 46 patients with RRMS) for FA and 88.5% (23 of 26 patients with RNMO) and 89.1% (41 of 46 patients with RRMS) for MD, respectively.

Conclusion: Measurement of diffusion indexes of the CC may be useful for distinguishing patients with RNMO from those with RRMS.

Supplemental material: http://radiology.rsnajnls.org/cgi/content/full/2441060930/DC1

© RSNA, 2007

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCE IN KNOWLEDGE IMPLICATION FOR PATIENT CARE References

 
Both relapsing neuromyelitis optica (RNMO) and relapsing-remitting multiple sclerosis (RRMS) are recurrent demyelinating diseases. Although they are different in regard to onset age, sex preponderance, lesion distribution, attack severity, cerebrospinal fluid profile, pathologic findings, and immunopathologic characteristics (1–11), they also have overlapping clinical manifestations, magnetic resonance (MR) imaging findings, and cerebrospinal fluid findings. This brings difficulties in differentiation, especially at the early stage of the diseases. Because the optimum treatments for the two disorders are different (12–16), correct diagnosis critically affects the therapeutic outcome.

MR imaging examinations are important, because MR findings have been included in the diagnostic criteria for multiple sclerosis and neuromyelitis optica (5,17,18). Results of many studies (19–21) revealed that damage in brain tissue that appears normal, which is confirmed by using pathologic examination results but is not seen on conventional MR images, may be detected by using diffusion-tensor MR imaging in patients with multiple sclerosis and patients with neuromyelitis optica.

Diffusion-tensor MR imaging helps to measure the random motion of water molecules and provides information about the size, orientation, and geometry of brain tissue (22–24). Within a coherently arranged white matter tract, water molecules diffuse faster in the direction parallel to the tract than in perpendicular directions. Pathologic processes, which change the microstructural environment, result in altered diffusion (25,26). Molecular diffusion can be measured by using some quantitative indexes derived from diffusion-tensor MR imaging data. Mean diffusivity (MD) and fractional anisotropy (FA) are two of the most used indexes. MD is a measure of the mean magnitude of molecular motion, and FA reflects the directionality of molecular motion (22,27,28).

Results of a recent diffusion-tensor MR imaging study (29) in patients with RNMO revealed that abnormal diffusion in brain tissue was present in the regions continuous with the spinal tracts or optic nerves but not in the corpus callosum (CC). In contrast, abnormal diffusion in the CC is commonly seen in patients with RRMS, whether visible lesions in this structure are present or not (30,31). Considering the foregoing, we hypothesized that measurements of MD and FA in the CC would help differentiate RNMO from RRMS. Thus, the purpose of our study was to prospectively assess the sensitivity and specificity of diffusion indexes of the CC for differentiating RNMO from RRMS by using the final clinical diagnosis as the reference standard.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCE IN KNOWLEDGE IMPLICATION FOR PATIENT CARE References

 
Patients and Reference Standard
The institutional review board at Xuanwu Hospital approved the study, and written informed consent was obtained from each participant. Twenty-six patients with RNMO (two men, 24 women; mean age, 38.6 years; age range, 19–59 years) and 46 patients with RRMS (18 men, 28 women; mean age, 37.7 years; age range, 18–58 years) were included. Patients were prospectively recruited from among 90 consecutive patients with RNMO or RRMS who were all examined at Xuanwu Hospital from April 2003 to July 2005 and who were recruited collectively from Xuanwu Hospital; Peking Union Medical College Hospital, Beijing, China; General Hospital of People's Liberation Army, Beijing, China; and Peking University Third Hospital, Beijing, China. The diagnosis of RNMO was based on the criteria proposed by Wingerchuk et al (5). The diagnosis of RRMS was made according to the modified McDonald criteria (18,32).

Patients were excluded if (a) their conditions satisfied the diagnostic criteria for both RNMO and RRMS, (b) they had brain imaging artifacts on images, (c) they had received corticosteroids or immunosuppressants within 3 months, (d) they had complications in the brain, and/or (e) they had undergone surgical procedures in the brain. Two neuroradiologists (K.C.L. and C.S.Y., with 13 and 10 years of experience in MR image evaluation, respectively) and two neurologists (H.S. and P.C., with 11 and 18 years of experience in neurology, respectively) assigned final clinical diagnoses to all patients in consensus by reviewing all case histories, physical examination results, MR images, and laboratory data. We obtained demographic and clinical information for all patients. Eighteen control subjects (one man, 17 women) with no history of neurologic disorders and a normal finding at neurologic examination were also recruited. Their mean age was 37.9 years ± 10.7 (standard deviation) (age range, 19–55 years).

MR Imaging Examination
All participants were imaged with a 1.5-T MR unit (Sonata; Siemens Medical Systems, Erlangen, Germany). The brain was imaged by using the following sequences with an identical field of view (240 x 210 mm), number of sections (30), section thickness (4 mm), and intersection gap (0.4 mm): (a) T2-weighted turbo spin echo (repetition time msec/echo time msec, 5500/94; number of signals acquired, three; echo train length, 11; matrix, 256 x 224), (b) T1-weighted spin echo (650/6; number of signals acquired, three; matrix, 256 x 224), (c) fluid-attenuated inversion recovery (8500/150; inversion time, 2200 msec; number of signals acquired, three; echo train length, eight; matrix, 256 x 224), and (d) spin-echo single-shot echo planar (5000/100; number of signals acquired, 10; matrix, 128 x 112). A total of seven image sets were acquired: six with noncollinear diffusion-weighting gradients and a b value of 1000 sec/mm² and one without diffusion weighting.

The following pulse sequences were performed to image the cervical and thoracic segments of the spinal cord, respectively: (a) turbo spin-echo T2 weighted (3000/120; number of signals acquired, three; echo train length, 20; matrix, 512 x 256; field of view, 280 x 280 mm) and (b) spin-echo T1 weighted (500/12; number of signals acquired, three; matrix, 256 x 256; field of view, 280 x 280 mm). T2- and T1-weighted imaging were performed to obtain seven sagittal sections, with a 3-mm section thickness and an intersection gap of 0.3 mm. T2-weighted imaging was also performed to obtain 20 transverse sections with a 5-mm section thickness and an intersection gap of 5 mm.

Calculation of Diffusion Indexes
The diffusion tensor of each voxel was calculated by using a linear least-squares fitting algorithm (22). After diagonalization of the diffusion tensor was performed, diffusion-tensor eigenvalues were obtained. MD and FA were derived for each pixel according to the following equations:

and

Here, 1, 2, 3 are eigenvalues describing the magnitude of diffusivity in the directions of maximum, median, and minimum diffusion, respectively. These calculations were performed by one author (C.S.Y.).

Region of Interest Definition
The 30 transverse sections were first interpolated into 136 sections to obtain a voxel size of 0.94 x 0.94 x 0.94 mm. A region of interest (ROI) of the CC was defined as all pixels of the structure seen in the midsagittal plane of the reconstructed FA images, because the outline of the CC in this plane was easily identified and this ROI was expected to be reproducible across participants. ROIs were manually outlined by one neuroradiologist (C.S.Y., with 10 years of experience in MR image evaluation) who was blinded to patient identity. Each of the marginal pixels of the ROI was identified when it was located at the margin of the CC and showed high signal intensity on FA images (Fig 1). MD and FA of each ROI were then measured. Intrarater reliability of demarcation of ROIs was also assessed. The same rater outlined the ROIs in the 72 patients at least 6 months later. Mean MD and FA were measured.

View larger version (122K): [in this window] [in a new window] [Download PPT slide]   Figure 1: Midsagittal FA MR image shows placement of ROI in CC.

Intraclass correlation coefficients calculated by using a one-way random model were used to test the intrarater reproducibility. We calculated mean values, standard deviations, and 95% confidence intervals of MD and FA in the ROIs of all participants. Because the variances of the three groups were not equal, we used nonparametric (Kruskal-Wallis) analysis of variance and the Mann-Whitney U test to test for differences between groups. Bonferroni correction was used to correct these multiple comparisons, and P .01 was considered to indicate a statistically significant difference.

We compared the demographic and clinical features between patients with RNMO and patients with RRMS with the ² or t test, and P .05 was considered to indicate a statistically significant difference. We tested the relationships of diffusion indexes with disease duration and age by using nonparametric Spearman correlation coefficients, and P .05 was considered to indicate a statistically significant difference. All statistical evaluations were performed with software (SPSS for Windows, version 11.5; SPSS, Chicago, Ill).

Bayesian Classification and Cross Validation
Bayesian classification is a fundamental statistical approach for pattern recognition (33). A Bayesian classifier combines training data with a priori knowledge to obtain, given a new sample, the a posteriori probability of each class. The new sample is then assigned to the class with the maximal a posteriori probability. Details can be found in Appendix E1 (http://radiology.rsnajnls.org/cgi/content/full/2441060930/DC1).

A leave-one-out cross-validation approach (34) was employed to evaluate the performance of our classifiers. In each round of the leave-one-out validation, one participant was selected as a testing sample. The remaining participants were used as training samples to construct the classifier with formulas 1–4 (see Appendix E1). The testing sample was then classified with the trained classifier by using formula 5. Such procedure was repeated until each participant was tested one time. The performance of our classifier was evaluated with sensitivity and specificity. Sensitivity was determined as TP/(TP + FN), and specificity was determined as TN/(TN + FP), where TP is a true-positive finding (test results showed RNMO and the final clinical diagnosis was also RNMO); TN, a true-negative finding (test results showed RRMS and the final clinical diagnosis was also RRMS); FP, a false-positive finding (test results showed RNMO but the final clinical diagnosis was RRMS); and FN, a false-negative finding (test results showed RRMS but the final clinical diagnosis was RNMO). We also tested the predictive ability of our method in patients with disease duration of less than 2 years (eight patients with RNMO and 23 patients with RRMS) to provide evidence of the ability of the method to help differentiate between the diseases at an early stage.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCE IN KNOWLEDGE IMPLICATION FOR PATIENT CARE References

 
Participants
Patients were recruited from April 2003 to July 2005. There was a greater female preponderance and more severe disability among patients with RNMO than among patients with RRMS (P < .05) (Table 1). No significant difference was found in age at imaging, age at onset, and disease duration between the two groups (P > .05). Optic neuritis was more frequent in patients with RNMO than in patients with RRMS (P < .05). However, the presence of attack-related weakness and somatosensory symptoms was not significantly different between the two groups (P > .05). Normal findings on brain MR images and extensive spinal cord lesions were more frequent in patients with RNMO than in those with RRMS (P < .05). However, visible lesions in the CC were more common in patients with RRMS than in patients with RNMO (P < .05).

View larger version (15K): [in this window] [in a new window] [Download PPT slide]   Figure 2: Flow diagram of study design for assessment of predictive performance of diffusion-tensor MR imaging, with MD and FA of CC as discriminative variables. Right = correct prediction, Wrong = incorrect prediction.

Comparison of Diffusion Indexes
Statistical comparison revealed significant differences overall in MD and FA of the CC among groups (Kruskal-Wallis one-way analysis of variance, P < .001). When post hoc (Mann-Whitney) tests were performed, patients with RRMS had a higher MD and a lower FA in the CC than patients with RNMO and control subjects (P < .001), and no significant differences were found in these indexes between patients with RNMO and healthy control subjects (Table 2).

Performance
Distribution maps of MD and FA of the ROIs from patients with RRMS and those with RNMO are shown in Figure 3. With FA as a discriminative index, sensitivity and specificity for differentiation were 92.3% (24 of 26 patients with RNMO) and 93.5% (43 of 46 patients with RRMS), respectively. With MD as a discriminative index, sensitivity and specificity were 88.5% (23 of 26 patients with RNMO) and 89.1% (41 of 46 patients with RRMS), respectively (Fig 2). In patients with disease duration of less than 2 years, sensitivity and specificity were 87.5% (seven of eight patients with RNMO) and 82.6% (19 of 23 patients with RRMS) for MD and 62.5% (five of eight patients with RNMO) and 95.7% (22 of 23 patients with RRMS) for FA, respectively.

View larger version (13K): [in this window] [in a new window] [Download PPT slide]   Figure 3a: Distribution maps of MD (x 10–4 mm2/sec) and FA of CC in patients with RNMO and those with RRMS.

View larger version (13K): [in this window] [in a new window] [Download PPT slide]   Figure 3b: Distribution maps of MD (x 10–4 mm2/sec) and FA of CC in patients with RNMO and those with RRMS.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCE IN KNOWLEDGE IMPLICATION FOR PATIENT CARE References

 
It is sometimes difficult to discriminate RNMO from RRMS because both diseases usually manifest as having a relapsing-remitting course and have similar clinical features. Characteristics that have been used to distinguish RNMO from RRMS include the following: (a) Patients with RNMO are more likely to be female and have higher age of onset and more severe attacks (3,4). (However, overlaps in clinical manifestations are common in these two disorders [35,36].) (b) Lesions in patients with RNMO are typically restricted to the spinal cord and optic nerves (5), but seven of 26 patients with RNMO in our study had brain lesions, and some patients with RRMS have manifestations involving mainly the spinal cord and optic nerves. (c) Signal intensity abnormality extending over three or more vertebral segments on spinal cord MR images is supportive of RNMO (5), but this feature was also present in 15 of 46 patients with RRMS in our study. (d) Cerebrospinal fluid examination results show pleocytosis (>50 cells per cubic millimeter), neutrophils, and absence of oligoclonal banding in most, but not all, patients with RNMO (3,5,6).

Therefore, the above features, though typically characteristics of RNMO, cannot be confidently relied on to discriminate RNMO from RRMS. Even with the proposed diagnostic criteria (5,18), these two diseases cannot be absolutely separated, because the condition of some patients may simultaneously satisfy criteria for both RRMS and RNMO. Without a clear diagnosis, it is difficult to choose an optimal treatment, because ß-interferon (13,14) and glatiramer acetate (12) are indicated for RRMS, whereas prednisone and azathioprine (37), rituximab (15,16), and plasma exchange (16) are proposed for RNMO. Thus, to optimize therapeutic effects, it is important to develop a method to differentiate RNMO from RRMS.

In patients with RRMS, damage to the CC commonly manifests as visible lesions at MR imaging, discrete lesions not seen on conventional MR images, and wallerian degeneration (19–22). However, results of a previous diffusion-tensor MR imaging study (29) showed that patients with RNMO did not have abnormality in the CC, which indicated that either the CC was intact or that it had occult damage that was not detectable by using diffusion-tensor MR imaging. Nevertheless, these findings led us to hypothesize that diffusion indexes of the CC might help differentiate RNMO from RRMS.

We found that both MD and FA of the CC were significantly different between patients with RRMS and patients with RNMO, which suggested that the severity of damage to the CC differed in these two diseases. The possible causes for the difference were as follows: (a) 20 of 46 patients with RRMS had visible lesions in the CC that directly caused abnormal diffusion in this structure, (b) 26 of 46 patients with RRMS without visible lesions in the CC would have already had abnormal diffusion because of the presence of unseen discrete lesions and wallerian degeneration, and, (c) in contrast, patients with RNMO did not have or had only minor damage in the CC.

The results of our study suggest that both MD and FA of the CC have the potential to differentiate patients with RNMO from those with RRMS, and FA seemed to have a slightly better performance. We speculate that the better performance of FA may be caused by sampling error or the distinction between the sensitivities of the two indexes. When all patients were regarded as a training set, the cutoff values were 0.688 for FA and 8.90 x 10^–4 mm²/sec for MD. In clinical practice, we may use these two cutoff values to differentiate RNMO from RRMS. Some authors have reported that a serum autoantibody marker of neuromyelitis optica could help differentiate neuromyelitis optica from multiple sclerosis (38). We assume that our method, used together with immunoglobulin G detection of neuromyelitis optica, may improve diagnostic accuracy.

Generally speaking, the classification performance will be improved if two or more independent variables are used. Although MD and FA of the CC were not correlated in control subjects, they were significantly correlated in patients with RRMS and those with RNMO. This indicates that MD and FA of the CC are not independent variables in these patients. Therefore, we did not consider examining the classification performance of these two indexes in combination.

Recognizing the importance of differentiating RNMO from RRMS as early as possible, we investigated the correlation between diffusion indexes and disease duration but did not find any significant correlations. These findings may indicate that a difference in diffusion indexes might exist at the early stage of the diseases. As a preliminary exploration, we tested the predictive ability of our method in patients with disease duration of less than 2 years and found that the accuracy for differentiating RNMO from RRMS was more than 80%. Of course, we recognize that a follow-up study of patients with clinically isolated syndromes should be performed to check if our method could help reliably diagnose RRMS and RNMO at the early stage.

Our study possibly contained certain limitations. Ascertainment and referral bias might have influenced the results, because the specificity of the proposed criteria for RNMO and RRMS (5,18) was not 100% and could engender misclassifications. The limited spatial resolution of diffusion-tensor MR imaging data might have also affected the reliability of our results. If imaging had been performed with a 3.0-T MR unit, a set of images with higher spatial resolution and more reliable results could have been produced.

From the results of our study, we can conclude that measurement of diffusion indexes of the CC may be helpful for differentiating patients with RNMO from those with RRMS.

	ADVANCE IN KNOWLEDGE

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCE IN KNOWLEDGE IMPLICATION FOR PATIENT CARE References

 

• Measurement of fractional anisotropy and mean diffusivity of the corpus callosum may be helpful for differentiating patients with relapsing neuromyelitis optica from those with relapsing-remitting multiple sclerosis; sensitivity and specificity were 92.3% and 93.5% for fractional anisotropy and 88.5% and 89.1% for mean diffusivity, respectively.
  

	IMPLICATION FOR PATIENT CARE

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCE IN KNOWLEDGE IMPLICATION FOR PATIENT CARE References

 

• Diffusion-tensor MR images can help in evaluating a patient suspected of having relapsing neuromyelitis optica or relapsing-remitting multiple sclerosis.
  

	FOOTNOTES

 

Abbreviations: CC = corpus callosum • FA = fractional anisotropy • MD = mean diffusivity • RNMO = relapsing neuromyelitis optica • ROI = region of interest • RRMS = relapsing-remitting multiple sclerosis

Authors stated no financial relationship to disclose.

Author contributions: Guarantors of integrity of entire study, C.S.Y., C.Z.Z., K.C.L.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript final version approval, all authors; literature research, C.S.Y., C.Z.Z., K.C.L., Y.X., W.Q.; clinical studies, C.S.Y., H.S., P.C., W.Q.; statistical analysis, C.S.Y., C.Z.Z., Y.X., H.S.; and manuscript editing, all authors

	References

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCE IN KNOWLEDGE IMPLICATION FOR PATIENT CARE References

 

1. Bergamaschi R, Ghezzi A. Devic's neuromyelitis optica: clinical features and prognostic factors. Neurol Sci 2004;25(suppl 4):S364–S367.[CrossRef][Medline]
2. Ghezzi A, Bergamaschi R, Martinelli V, et al. Clinical characteristics, course and prognosis of relapsing Devic's neuromyelitis optica. J Neurol 2004;251:47–52.[CrossRef][Medline]
3. Wingerchuk DM, Weinshenker BG. Neuromyelitis optica: clinical predictors of a relapsing course and survival. Neurology 2003;60:848–853.[Abstract/Free Full Text]
4. O'Riordan JI, Gallagher HL, Thompson AJ, et al. Clinical CSF, and MRI findings in Devic's neuromyelitis optica. J Neurol Neurosurg Psychiatry 1996;60:382–387.[Abstract/Free Full Text]
5. Wingerchuk DM, Hogancamp WF, O'Brien PC, Weinshenker BG. The clinical course of neuromyelitis optica (Devic's syndrome). Neurology 1999;53:1107–1114.[Abstract/Free Full Text]
6. Nakashima I, Fujihara K, Fujimori J, Narikawa K, Misu T, Itoyama Y. Absence of IgG1 response in the cerebrospinal fluid of relapsing neuromyelitis optica. Neurology 2004;62:144–146.[Abstract/Free Full Text]
7. Correale J, Fiol M. Activation of humoral immunity and eosinophils in neuromyelitis optica. Neurology 2004;63:2363–2370.[Abstract/Free Full Text]
8. Mandler RN, Davis LE, Jeffery DR, Kornfeld MK. Devic's neuromyelitis optica: a clinicopathological study of 8 patients. Ann Neurol 1993;34:162–168.[CrossRef][Medline]
9. Lefkowitz D, Angelo JN. Neuromyelitis optica with unusual vascular changes. Arch Neurol 1984;41:1103–1105.[Abstract/Free Full Text]
10. Lucchinetti CF, Mandler RN, McGavern D, et al. A role for humoral mechanisms in the pathogenesis of Devic's neuromyelitis optica. Brain 2002;125:1450–1461.[Abstract/Free Full Text]
11. Mandler RN, Dencoff JD, Midani F, Ford CC, Ahmed W, Rosenberg GA. Matrix metalloproteinase and tissue inhibitors of metalloproteinases in cerebrospinal fluid differ in multiple sclerosis and Devic's neuromyelitis optica. Brain 2001;124:493–498.[Abstract/Free Full Text]
12. Flechter S, Vardi J, Pollak L, Rabey JM. Comparison of glatiramer acetate (Copaxone) and interferon beta-1b (Betaferon) in multiple sclerosis patients: an open-label 2-year follow-up. J Neurol Sci 2002;197:51–55.[CrossRef][Medline]
13. Frank JA, Richert N, Bash C, et al. Interferon-beta-1b slows progression of atrophy in RRMS: three-year follow-up in NAb- and NAb+ patients. Neurology 2004;62:719–725.[Abstract/Free Full Text]
14. Saida T, Tashiro K, Itoyama Y, et al. Interferon beta-1b is effective in Japanese RRMS patients: a randomized, multicenter study. Neurology 2005;64:621–630.[Abstract/Free Full Text]
15. Cree BA, Lamb S, Morgan K, Chen A, Waubant E, Genain C. An open label study of the effects of rituximab in neuromyelitis optica. Neurology 2005;64:1270–1272.[Abstract/Free Full Text]
16. Keegan M, Pineda AA, McClelland RL, Darby CH, Rodriguez M, Weinshenker BG. Plasma exchange for severe attacks of CNS demyelination: predictors of response. Neurology 2002;58:143–146.[Abstract/Free Full Text]
17. McDonald WI, Compston A, Edan G, et al. Recommended diagnostic criteria for multiple sclerosis: guidelines from the International Panel on the diagnosis of multiple sclerosis. Ann Neurol 2001;50:121–127.[CrossRef][Medline]
18. Polman CH, Reingold SC, Edan G, et al. Diagnostic criteria for multiple sclerosis: 2005 revisions to the "McDonald Criteria." Ann Neurol 2005;58:840–846.
19. Werring DJ, Clark CA, Barker GJ, Thompson AJ, Miller DH. Diffusion tensor imaging of lesions and normal-appearing white matter in multiple sclerosis. Neurology 1999;52:1626–1631.[Abstract/Free Full Text]
20. Evangelou N, Esiri MM, Smith S, Palace J, Matthews PM. Quantitative pathological evidence for axonal loss in normal appearing white matter in multiple sclerosis. Ann Neurol 2000;47:391–395.[CrossRef][Medline]
21. Rocca MA, Agosta F, Mezzapesa DM, et al. Magnetization transfer and diffusion tensor MRI show gray matter damage in neuromyelitis optica. Neurology 2004;62:476–478.[Abstract/Free Full Text]
22. Basser PJ, Mattiello J, Le Bihan D. Estimation of the effective self-diffusion tensor from the NMR spin-echo. J Magn Reson B 1994;103:247–254.[CrossRef][Medline]
23. Eriksson SH, Rugg-Gunn FJ, Symms MR, Barker GJ, Duncan JS. Diffusion tensor imaging in patients with epilepsy and malformations of cortical development. Brain 2001;124:617–626.[Abstract/Free Full Text]
24. Le Bihan D. Looking into the functional architecture of the brain with diffusion MRI. Nat Rev Neurosci 2003;4:469–480.[CrossRef][Medline]
25. Anderson AW, Zhong J, Petroff OA, et al. Effects of osmotically driven cell volume changes on diffusion-weighted imaging of the rat optic nerve. Magn Reson Med 1996;35:162–167.[Medline]
26. Sevick RJ, Kanda F, Mintorovitch J, et al. Cytotoxic brain edema: assessment with diffusion-weighted MR imaging. Radiology 1992;185:687–690.[Abstract/Free Full Text]
27. Basser PJ, Pierpaoli C. Microstructural and physiological features of tissues elucidated by quantitative-diffusion-tensor MRI. J Magn Reson B 1996;111:209–219.[CrossRef][Medline]
28. Pierpaoli C, Jezzard P, Basser PJ, Blarnett A, Di Chiro G. Diffusion tensor MR imaging of the human brain. Radiology 1996;201:637–648.[Abstract/Free Full Text]
29. Yu CS, Lin FC, Li KC, et al. Diffusion tensor imaging in the assessment of normal-appearing brain tissue damage in relapsing neuromyelitis optica. AJNR Am J Neuroradiol 2006;27:1009–1015.[Abstract/Free Full Text]
30. Hasan KM, Gupta RK, Santos RM, Wolinsky JS, Narayana PA. Diffusion tensor fractional anisotropy of the normal-appearing seven segments of the corpus callosum in healthy adults and relapsing-remitting multiple sclerosis patients. J Magn Reson Imaging 2005;21:735–743.[CrossRef][Medline]
31. Coombs BD, Best A, Brown MS, et al. Multiple sclerosis pathology in the normal and abnormal appearing white matter of the corpus callosum by diffusion tensor imaging. Mult Scler 2004;10:392–397.[Abstract/Free Full Text]
32. Lublin FD, Reingold SC. Defining the clinical course of multiple sclerosis: results of an international survey. National Multiple Sclerosis Society (USA) Advisory Committee on Clinical Trials of New Agents in Multiple Sclerosis. Neurology 1996;46:907–911.
33. Huelsenbeck JP, Ronquist F, Nielsen R, Bollback JP. Bayesian inference of phylogeny and its impact on evolutionary biology. Science 2001;294:2310–2314.[CrossRef][Medline]
34. Man MZ, Dyson G, Johnson K, Liao B. Evaluating methods for classifying expression data. J Biopharm Stat 2004;14:1065–1084.[CrossRef][Medline]
35. Misu T, Fujihara K, Nakashima I, et al. Pure optic-spinal form of multiple sclerosis in Japan. Brain 2002;125:2460–2468.[Abstract/Free Full Text]
36. Weinshenker BG. Neuromyelitis optica: what it is and what it might be. Lancet 2003;361:889–890.[CrossRef][Medline]
37. Mandler RN, Ahmed W, Dencoff JE. Devic's neuromyelitis optica: a prospective study of seven patients treated with prednisone and azathioprine. Neurology 1998;51:1219–1220.[Abstract/Free Full Text]
38. Lennon VA, Wingerchuk DM, Kryzer TJ, et al. A serum autoantibody marker of neuromyelitis optica: distinction from multiple sclerosis. Lancet 2004;364:2106–2112.[CrossRef][Medline]

This article has been cited by other articles:

				D. Wheeler, V. V. R. Bandaru, P. A. Calabresi, A. Nath, and N. J. Haughey A defect of sphingolipid metabolism modifies the properties of normal appearing white matter in multiple sclerosis Brain, November 1, 2008; 131(11): 3092 - 3102. [Abstract] [Full Text] [PDF]

This Article Abstract Figures Only Full Text (PDF) Appendix All Versions of this Article: 2441060930v1 244/1/249    most recent Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Yu, C. S. Articles by Chan, P. Search for Related Content PubMed PubMed Citation Articles by Yu, C. S. Articles by Chan, P.