MRI-Based Measurement of Brain Stem Cross-Sectional Area in Relapsing-Remitting Multiple Sclerosis Tomos R. Chivers, Cris S. Constantinescu, Christopher R. Tench From the Division of Clinical Neurology, University Hospital NHS Trust, Queen’s Medical Centre, Nottingham, UK (TRC, CSC, CRT).
ABSTRACT PURPOSE: To determine if patients with relapsing-remitting multiple sclerosis (RRMS) have a reduced brain stem crosssectional area (CSA) compared to age- and sex-matched controls. The brain stem is a common site of involvement in MS. However, relatively few imaging studies have investigated brain stem atrophy. METHODS: Brain magnetic resonance imaging (MRI) was performed on patients and controls using a 1.5T MRI scanner with a quadrature head coil. Three-dimensional magnetization-prepared rapid acquisition gradient-echo (MPRAGE) images with 128 contiguous slices, covering the whole brain and brain stem and a T2-weighted image with 3 mm transverse contiguous images were acquired. We measured the brain stem CSA at three sites, the midbrain, the pons, and the medulla oblongata in 35 RRMS patients and 35 controls using a semiautomated algorithm. CSA readings were normalized using the total external cranial volume to reduce normal population variance and increase statistical power. RESULTS: A significant CSA reduction was found in the midbrain (P ࣘ .001), pons (P ࣘ .001), and the medulla oblongata (P = .047) postnormalization. A CSA reduction of 9.3% was found in the midbrain, 8.7% in the pons, and 6.5% in the medulla oblongata. CONCLUSIONS: A significantly reduced, normalized brain stem CSA was detected in all areas of the brain stem of the RRMS patients, when compared to age- and gender-matched controls. Lack of detectable upper cervical cord atrophy in the same patients suggests some independence of the MS pathology in these regions.
Keywords: Multiple sclerosis, brain stem, atrophy, MRI. Acceptance: Received February 20, 2015. Accepted for publication February 27, 2015. Correspondence: Address correspondence to Tomos R. Chivers, Division of Clinical Neurology, C Floor, South Block, University Hospital NHS Trust, Queen’s Medical Centre, Nottingham, NG7 2HU, UK. E-mail: [email protected]. Funding: This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. J Neuroimaging 2015;25:1002-1006. DOI: 10.1111/jon.12244
Introduction Multiple Sclerosis (MS) is a disease that affects the central nervous system (CNS) causing chronic demyelination, inflammation, and axonal loss. Its distinguishing feature is the development of lesions throughout the CNS. Progressive atrophy of the CNS is a recognized feature of MS, and is thought to be a better correlate of permanent disability than T2 lesion load, which varies more unpredictably compared with disability.1–3 Magnetic resonance imaging (MRI) studies of atrophy in MS have shown that there is a correlation between axonal loss and CNS atrophy.4 Further studies have also shown that demyelination also plays a role in atrophy, but demyelination is potentially reversible whereas axonal loss is not.5,6 Many MRI studies have compared the brain and spinal cord of relapsing-remitting MS (RRMS) patients with those of unaffected controls. The correlation between brain atrophy and progression of RRMS is widely accepted,7 and detectable atrophy appears early in the disease.8,9 Spinal cord atrophy appears to exist in both progressive forms of MS and relapsing-remitting forms of MS.10–12 While there have been many studies focused on atrophy of the brain and spinal cord, relatively little research has focused on the brain stem. The brain stem is a common site for pathology in MS with one study of 114 clinically definite MS patients
1002
Copyright
finding brain stem lesions in 68%.13 Pathology of the brain stem due to MS can cause eye movement disorders, vertigo, tremor, urinary problems, hemiplegia, facial pain, paralysis, and hearing problems. Many of these manifestations may develop early in MS. In a recent study of MS patients, brain stem reflexes were abnormal in 87% and evoked potentials in 83%.14 Other studies have shown a strong correlation between brain stem lesion load, diffusion abnormalities, and disability.15 Early studies looking at brain stem atrophy utilized computer tomography (CT) to examine the brain stem. One study of 200 MS patients found brain stem/cerebellar atrophy in 53 of the patients and found there was a strong correlation between brain stem atrophy and brain stem dysfunction.16 CT images, however, lack the contrast and resolution necessary to accurately take measurements, so recent studies have utilized MRI.5 Of the MRI studies that have looked at brain stem atrophy in MS, there has been some inconsistency. Those that have detected atrophy in MS patients, including RRMS, when compared with controls, have found levels of volume reduction that vary from 11%17 to 20.6%.18 However, other studies, such as one looking at 49 RRMS patients, have found no noticeable atrophy compared with age-matched controls.19 Modern techniques such as voxel-based morphometry (VBM) have been used to show atrophy in brain structures.20
◦ 2015 by the American Society of Neuroimaging C
VBM has been used to show white matter atrophy in the brain stem and that this correlates with lesion volume.21 VBM, however, is better suited to studies with large groups of subjects where it is not practical to take manual measurements.22 In this study we measure the cross-sectional area (CSA) of three brain stem structures: the midbrain, pons, and medulla oblongata. We use a semiautomated algorithm with good reproducibility in 35 RRMS patients, and 35 age- and gendermatched control subjects; these subjects were previously used to study atrophy in the upper cervical cord.23 Intersubject variance is reduced by normalizing results to the total external cranial volume (TECV), a dimension measure we assume to be unaffected by multiple sclerosis.
Methods Patients and Controls Thirty-five clinically definite RRMS patients24 (21 females and 14 males) and 35 control subjects (21 females and 14 males) were recruited to the study. The median age (interquartile range) for patients was 39 (32-44) years and the median age of control subjects was 34 (29-42) years. Patients were excluded if they had received immunosuppressive or immunomodulatory drugs in the last year or corticosteroids in the last 3 months so as to avoid any effect treatment may have upon the processes of inflammation or atrophy. The median disease duration was 9 (5-13) years and the median expanded disability status scale score (EDSS)25 was 3 (2-4). Approval was obtained from the local research ethics committees. All patients gave informed written consent.
MRI Acquisition
sites using a semiautomated flood fill algorithm. A seed point was selected within each of the structures and a semiautomated flood fill algorithm used to grow the seed to capture the region of interest. This was performed on all 35 patients and 35 controls. Previous studies have employed the total intracranial volume (TICV) to normalize results.5,7,18,23,28 Despite using a similar method to previous studies to normalize TICV,18,23,28 preliminary work found that using the external volume of the cranium gives a better correlation to brain stem CSA than TICV, possibly because TICV was more difficult to measure reproducibly. TECVs were recorded to allow normalization of the CSA to reduce the interindividual variations in subject size. To measure the TECV, the sagittal view of the cranium was first rotated until the frontal pole of the cerebrum was horizontal with the posterior surface of the cerebellum. The external surface of the calvaria was outlined on a number of axial slices starting with those where the cerebral tissue was first visible and ending when cerebellar tissue was no longer visible. Once these regions of interest (ROI) were established, the slice volumes were added together to calculate the TECV. TECVs were compared for males and females to ensure the method was sensitive to known differences. T2 lesions in the brain stem and cerebellum were measured on axial T2 weighted images for all of the 35 patients. All T2 hyperintense lesions were identified as ROIs and the individual volumes were added together to give T2 lesion load.
Data Analysis The brain stem CSA were normalized to the TECV using a method previously summarized by Jack et al.29 The equation used for normalization was: CSAn = CSA + α(TECVmean − TECV),
Brain MRI was performed on patients and controls using a 1.5T Siemens Vision (Siemens, Erlangen, Germany) MRI scanner with a quadrature head coil. This was used to acquire 3-dimensional magnetization-prepared rapid acquisition gradient-echo (MPRAGE) images with 128 contiguous slices, covering the whole brain and brain stem (voxel size = .98 × .98 × 1.25 mm, matrix size = 256 × 256 × 128, repetition time (TR) = 9.7 ms, echo time (TE) = 4 ms, inversion time (TI) = 300 ms, flip angle = 10°), and a T2-weighted image with 3 mm transverse contiguous images was acquired (TR 4,100 ms, TE 90 ms, matrix 256 × 256, field of view 25 cm).
where CSAn is the normalized CSA and TECVmean is the average TECV for the control subjects. The coefficient α is the gradient of the least-squares straight line fit of CSA to TECV in the controls (Fig 2). Statistical tests were performed on pre- and postnormalized data using Pearson’s or Spearman’s correlation coefficient, twosample t-tests for comparison of mean values, and a MannWhitney U test for comparison of median values. Ten subjects had repeat scans to assess scan-rescan reproducibility.
MRI Analysis
Results
Image analysis was performed using NeuRoi.26 Three slices of the brain stem, one in each of the midbrain, pons and medulla oblongata were chosen as sites to measure CSA on the T1 images (see Fig 1). The midbrain was identified as the axial slice where the superior colliculi were at their largest. The pons was identified as the most inferior of the axial slices where the superior cerebellar peduncles have yet to attach to the cerebellum. The medulla oblongata was identified as the axial slice at the superior origin of the pyramids, just inferior to the floor of the fourth ventricle. These were chosen due to the ease of consistently identifying them and they were largely surrounded by sufficient cerebrospinal fluid (CSF) to give high contrast between brain stem and surroundings. All three sites are known areas of MS pathology.27 The brain stem was rotated in the sagittal view until vertical. It was then delineated on the axial view at the three chosen
There was no significant difference in the median age of the controls and the patients (P = .2). The coefficient of variance was calculated for nonnormalized CSA using scan-rescan images and was found to be 1.7% in the midbrain, 1.9% in the pons, and 5.5% in the medulla oblongata.
Brain stem Atrophy After normalization to the TECV, the mean CSA of the midbrain, pons, and medulla oblongata in the patients was significantly reduced relative to the control group (P ࣘ .001, P = .001 and P = .047, respectively); see Table 1. The differences between patient and control mean normalized CSA of the midbrain, pons, and medulla oblongata were 50.2 mm2 (CI 48.0–52.4 mm2 ; 9.5%), 45.1 mm2 (CI 44.3–45.9 mm2 ; 8.9%), and 12.8 mm2 (CI 12.0–13.6 mm2 ; 6.7%), respectively.
Chivers et al: MRI-Based Measurement of Brain Stem CSA in RRMS
1003
A
B
C
Fig 1. Examples of an outlined midbrain (A), pons (B), and medulla oblongata (C).
Fig 2. Normalizing for TECV. Gradient of the least-squares straight line fit of CSA to TECV in the control = coefficient α. Table 1. CSAs in Patients and Controls Postnormalization
Midbrain Pons Medulla
Control Mean CSA (mm2 )
Patient Mean CSA (mm2 )
Mean CSA Difference (%)
530.7 ± 8.9 509.6 ± 9.8 190.8 ± 4.8
480.5 ± 7.8 464.5 ± 9.4 178.0 ± 3.8
9.5%*** 8.9%*** 6.7%*
Significantly reduced CSA in patients compared to control is indicated by asterisk. * P ࣘ .05, ** P ࣘ .01, *** P ࣘ .001.
There was no significant correlation between normalized CSA and disease duration. There was no significant correlation between EDSS and CSA in the midbrain (r = .18; P = .28) and pons (r = .07; P = .67). There was, however, a significant correlation between EDSS and CSA in the medulla oblongata (r = .36; P = .03). Mean T2 lesion load in the cerebellum and brain stem was 1.3 mL (range 0-7.9 mL). There was a significant correlation between T2 lesion load and normalized CSA in the midbrain (Pearson’s r = –.39, P = .02, CI –.67 to –.01) but there was no significant correlation in the pons (P = .22, Pearson’s r = –.21, CI –.55 to .18) and medulla oblongata (P = .52, Pearson’s r = –.11, CI –.47 to .28).
TECV The TECV in the controls was significantly smaller (P ࣘ .001) in the females (2,296 ± 51 mL) than the males (2,648 ± 42 mL). There was significant correlation between the control TECV 1004
and the CSA of the midbrain (Pearson’s r = .58; P ࣘ .001), pons (Pearson’s r = .54; P ࣘ .001), and medulla oblongata (Pearson’s r = .46; P = .005) of the controls. The gradient of the least-squares straight line fit (α) for the midbrain, pons, and medulla oblongata was .1378, .1378, and .0554, respectively. The TECV in the patient group was significantly smaller (P ࣘ .001) in the females (2,329 ± 41 mL) than the males (2,722 ± 66 mL). In the patients, there was a significant correlation between the TECV and the CSA of the midbrain (Pearson’s r = .44; P = .008), but not the pons (Pearson’s r = .23; P = .18) or medulla oblongata (Pearson’s r = .24; P = .17) of the patients. There was no significant difference (P = .46) in the mean TECV between the patients and the normal controls.
Gender In the controls, the mean CSA of midbrain (P = .02), pons (P = .001), and medulla oblongata (P = .003) was significantly smaller in females than in males. There was no difference in the mean of the normalized CSA of control males and females. In the patients, there was no significant difference between the CSA of the males and females in the midbrain (P = .06), pons (P = .72), and medulla oblongata (P = .13) (see Table 1). There was no significant difference between male and female mean normalized CSA of the midbrain and medulla oblongata; however, the female mean normalized CSA was significantly larger in the pons (P = .01).
Journal of Neuroimaging Vol 25 No 6 November/December 2015
Discussion Brain stem Atrophy This study adds to relatively sparse literature showing reduced brain stem volume17,18,28 in MS by demonstrating that three structures in the brain stem are significantly atrophied when normalized to account for natural variance and compared with matched controls. After normalization, a CSA reduction of 9.3% was found in the midbrain, 8.7% in the pons, and 6.5% in the medulla oblongata. Several longitudinal studies have suggested that brain volume decreases by .6% to 1% a year in patients with MS.5,30,31 The median disease duration of the patients in this study was 9 years, which may indicate that the hypothesized rate of atrophy in the brain could also apply to the brain stem. Further work would be necessary to confirm this. Given the disease duration and the fact the patients were free of disease modifying treatment, it might also suggest that atrophy is not a late event. This would be consistent with current concepts about atrophy in MS.3 The link between disability and brain stem atrophy was explored by examining the relationship between EDSS, disease duration, and brain stem CSA. While there was no significant correlation with disease duration at any of the sampled sites, there was a significant correlation between EDSS and normalized CSA at the medulla oblongata. This, however, was a secondary finding requiring further work. We attempted to see if there was a correlation between T2 lesion load in the brain stem and the normalized brain stem CSA. Previous studies suggest a relationship between T2 lesion load and brain atrophy32 and it is possible this may also apply to the brain stem. A significant correlation was found in the midbrain but not at the pons or medulla oblongata. The observed reductions in mean CSA show very similar levels of atrophy in the pons and midbrain, but reduced levels of atrophy in the medulla oblongata. This might be due to measurement error; the medulla oblongata was more difficult to delineate than other parts of the brain stem as there was less contrast between the CSF and brain tissue. It could, alternatively, be due to partial volume averaging, an issue affecting accuracy when measuring smaller structures such as the medulla oblongata. A correction for partial volume averaging, such as the one used by Mann et al,23 might have increased the accuracy.
Spinal Cord Atrophy The patients from this study were previously investigated to determine if atrophy occurs in the upper cervical cord in RRMS.23 In that study no reduction in cord CSA was detected, and it was hypothesized that the atrophy might be masked by inflammation. However, a reduction in brain stem CSA has been detected in this study. This result may indicate a degree of independence between cord and brain stem pathology. Previous studies33–35 have also found this disconnection between brain and cord atrophy.
Normalization Normalization to the TECV was performed in order to remove some of the normal population variance and gender differences, hence improving the power of the study. Correlation between the TECV and CSA shows that between 22% and 33% (site
dependent) of the variance in CSA can be explained by TECV. This agrees with existing opinions about normalization.36 In the cervical cord study, the same method of normalization23 was used to show that 38% of the variation in CSA was explained by TICV. A similar level was achieved in this study using TECV. TECV might be a convenient normalization dimension in future atrophy studies.
Gender Differences We observed a significant difference in both the size of the TECV and brain stem of male and female controls. In all three sites sampled, CSA was significantly larger in male controls than female controls, and correlated with TECV, which was also significantly larger in males than females. This is expected17,23 and suggests that the method used in the experiment is sensitive. In the patients, there was also a gender difference in TECV. This, however, was not detectible in the brain stem CSA of the patients. The absence of this normal physiological difference may be the result of an external factor playing a role in determining the size of brain stem CSA; this might be inflammation, atrophy, or both in MS. Normalization removed all gender differences in both patients and controls with the exception of one site, in the pons of the patients. This apparently significantly larger, after normalization, pons CSA in females is interesting, but may represent a special case in our patient population.
Study Limitations Factors such as upgrades or repairs to the scanner could affect CSA readings over time.7 The images collected in this study were taken over a long period so it is possible that changes in the scanner could have affected the results. Normalization, however, should remove changes in volume due to scanner recalibration. The coefficient of variance was used to assess error on repeating the scan. It was found to be acceptably low in the midbrain and pons. There was, however, a slightly high level of error and weak level of significance in the medulla oblongata. It is possible that if the study was repeated with the increased sensitivity of 3T MRI it would be easier to delineate the medulla, and the level of error at this site would be more acceptable.37
Conclusions Normalization using the TECV was successfully used to reduce normal subject variance. A significantly reduced, normalized brain stem CSA was detected in all areas of the brain stem of the RRMS patients, when compared to age- and gender-matched controls. Lack of detectable upper cervical cord atrophy in the same patients suggests some independence of the MS pathology in these regions.
References 1. Miller DH, Grossman RI, Reingold SC, et al. The role of magnetic resonance techniques in understanding and managing multiple sclerosis. Brain 1998;121:3-24. 2. Bakshi R, Dandamudi VSR, Neema M, et al. Measurement of brain and spinal cord atrophy by magnetic resonance imaging as a tool to monitor multiple sclerosis. J Neuroimaging 2005;15:30S-45S. 3. Zivadinov R, Sepcic J, Nasuelli D, et al. A longitudinal study of brain atrophy and cognitive disturbances in the early phase of relapsing-remitting multiple sclerosis. J Neurol Neurosurg Psychiatry 2001;70:773-80.
Chivers et al: MRI-Based Measurement of Brain Stem CSA in RRMS
1005
4. Davie CA, Barker GJ, Webb S, et al. Persistent functional deficit in multiple sclerosis and autosomal dominant cerebellar ataxia is associated with axon loss. Brain 1995;118:1583-92. 5. Miller DH, Barkhof F, Frank JA, et al. Measurement of atrophy in multiple sclerosis: pathological basis, methodological aspects and clinical relevance. Brain 2002;125:1676-95. 6. Zivadinov R, Bakshi R. Role of MRI in multiple sclerosis II: brain and spinal cord atrophy. Front Biosci. 2004;9:647-64. 7. Minagar A, et al. Central nervous system atrophy in multiple sclerosis: mechanisms. In: Zivadinov R, Bakshi R, eds. Brain and Spinal Cord Atrophy in Multiple Sclerosis. New York: Nova, 2004;9:5-13. 8. Brex PA, Jenkins R, Fox NC, et al. Detection of ventricular enlargement in patients at the earliest clinical stage of MS. Neurology 2000;54:1689-91. 9. Simon JH. Brain and spinal cord atrophy in multiple sclerosis. Neuroimaging Clin N Am. 2000;10:753-70. 10. Losseff NA, Webb SL, O’Riordan JI, et al. Spinal cord atrophy and disability in multiple sclerosis. A new reproducible and sensitive MRI method with potential to monitor disease progression. Brain 1996;119:701-8. 11. Lin X, Tench CR, Turner B, et al. Spinal cord atrophy and disability in multiple sclerosis over four years: application of a reproducible automated technique in monitoring disease progression in a cohort of the interferon beta-1a (Rebif) treatment trial. J Neurol Neurosurg Psychiatr 2003;74:1090-4. 12. Valsasina P, Rocca MA, Horsfield MA, et al. Regional cervical cord atrophy and disability in multiple sclerosis: a voxel-based analysis. Radiology 2013;266:853-61. 13. Ormerod IE, Miller DH, McDonald WI, et al. The role of NMR imaging in the assessment of multiple sclerosis and isolated neurological lesions. A quantitative study. Brain 1987;110:1579616. 14. Magnano I, Pes GM, Pilurzi G, et al. Exploring brainstem function in multiple sclerosis by combining brainstem reflexes, evoked potentials, clinical and MRI investigations. Clin Neurophysiol 2014;125:2286-96. 15. Preziosa P, Rocca MA, Mesaros S, et al. Relationship between damage to the cerebellar peduncles and clinical disability in multiple sclerosis. Radiology 2014;271:822-30. 16. Loizou LA, Rolfe EB, Hewazy H. Cranial computed tomography in the diagnosis of multiple sclerosis. J Neurol Neurosurg Psychiatr 1982;45:905-12. 17. Liptak Z, Berger AM, Sampat MP, et al. Medulla oblongata volume: a biomarker of spinal cord damage and disability in multiple sclerosis. AJNR Am J Neuroradiol 2008;29:1465-70. 18. Liu C, Edwards S, Gong Q, et al. Three dimensional MRI estimates of brain and spinal cord atrophy in multiple sclerosis. J Neurol Neurosurg Psychiatry 1999;66:323-30. 19. Lin X, Blumhardt LD, Constantinescu CS. The relationship of brain and cervical cord volume to disability in clinical subtypes of multiple sclerosis: a three-dimensional MRI study. Acta Neurol Scand 2003;108:401-6. 20. Lansley J, Mataix-Cols D, Grau M, et al. Localized grey matter atrophy in multiple sclerosis: a meta-analysis of voxel-based morphometry studies and associations with functional disability. Neurosci Biobehav Rev 2013;37:819-30.
1006
21. Ceccarelli A, Jackson JS, Arora A, et al. Abstract #7: Voxel Based Morphometry with DARTEL in Multiple Sclerosis. Neurotherapeutics 2010;7:330. 22. Whitwell JL. Voxel-based morphometry: an automated technique for assessing structural changes in the brain. J Neurosci 2009;29:9661-4. 23. Mann RS, Constantinescu CS, Tench CR. Upper cervical spinal cord cross-sectional area in relapsing-remitting multiple sclerosis: application of a new technique for measuring cross-sectional area on magnetic resonance images. J Magn Reson Imaging 2007;26: 61-5. 24. Poser CM, Paty DW, Scheinberg L, et al. New diagnostic criteria for multiple sclerosis: guidelines for research protocols. Ann Neurol 1983;13:227-31. 25. Kurtzke JF. Rating neurologic impairment in multiple sclerosis: an expanded disability status scale (EDSS). Neurology 1983;33:144452. 26. Tench C. NeuRoi [Internet]. Nottingham University Department Clinical Neurology. www.nottingham.ac.uk/scs/divisions/ clinicalneurology/software/neuroi/aspx. Accessed March 4, 2013. 27. Nakashima I, Fujihara K, Miyazawa I, et al. Clinical and MRI features of Japanese patients with multiple sclerosis positive for NMO-IgG. J Neurol Neurosurg Psychiatr 2006;77:1073-5. 28. Edwards SG, Gong QY, Liu C, et al. Infratentorial atrophy on magnetic resonance imaging and disability in multiple sclerosis. Brain 1999;122:291-301. 29. Jack CR Jr, Twomey CK, Zinsmeister AR, et al. Anterior temporal lobes and hippocampal formations: normative volumetric measurements from MR images in young adults. Radiology 1989;172:54954. 30. Inglese M, Rovaris M, Giacomotti L, et al. Quantitative brain volumetric analysis from patients with multiple sclerosis: a follow-up study. J Neurol Sci 1999;171:8-10. 31. Ge Y, Grossman RI, Udupa JK, et al. Brain atrophy in relapsingremitting multiple sclerosis and secondary progressive multiple sclerosis: longitudinal quantitative analysis. Radiolog. 2000;214: 665-70. 32. Chard DT, Brex PA, Ciccarelli O, et al. The longitudinal relation between brain lesion load and atrophy in multiple sclerosis: a 14 year follow up study. J Neurol Neurosurg Psychiatry 2003;74:1551-4. 33. Kearney H, Rocca MA, Valsasina P, et al. Magnetic resonance imaging correlates of physical disability in relapse onset multiple sclerosis of long disease duration. Mult Scler 2014;20:72-80. 34. Bonati U, Fisniku LK, Altmann DR, et al. Cervical cord and brain grey matter atrophy independently associate with long-term MS disability. J Neurol Neurosurg Psychiatry 2011;82:471-2. 35. Cohen AB, Neema M, Arora A, et al. The relationships among MRI-defined spinal cord involvement, brain involvement, and disability in multiple sclerosis. J Neuroimaging 2012;22:122-8. 36. Sormani MP, Rovaris M, Valsasina P, et al. Measurement error of two different techniques for brain atrophy assessment in multiple sclerosis. Neurology 2004;62:1432-4. 37. Stankiewicz JM, Glanz BI, Healy BC, et al. Brain MRI lesion load at 1.5T and 3T versus clinical status in multiple sclerosis. J Neuroimaging 2011;21:e50-6.
Journal of Neuroimaging Vol 25 No 6 November/December 2015
ABSTRACT PURPOSE: To determine if patients with relapsing-remitting multiple sclerosis (RRMS) have a reduced brain stem crosssectional area (CSA) compared to age- and sex-matched controls. The brain stem is a common site of involvement in MS. However, relatively few imaging studies have investigated brain stem atrophy. METHODS: Brain magnetic resonance imaging (MRI) was performed on patients and controls using a 1.5T MRI scanner with a quadrature head coil. Three-dimensional magnetization-prepared rapid acquisition gradient-echo (MPRAGE) images with 128 contiguous slices, covering the whole brain and brain stem and a T2-weighted image with 3 mm transverse contiguous images were acquired. We measured the brain stem CSA at three sites, the midbrain, the pons, and the medulla oblongata in 35 RRMS patients and 35 controls using a semiautomated algorithm. CSA readings were normalized using the total external cranial volume to reduce normal population variance and increase statistical power. RESULTS: A significant CSA reduction was found in the midbrain (P ࣘ .001), pons (P ࣘ .001), and the medulla oblongata (P = .047) postnormalization. A CSA reduction of 9.3% was found in the midbrain, 8.7% in the pons, and 6.5% in the medulla oblongata. CONCLUSIONS: A significantly reduced, normalized brain stem CSA was detected in all areas of the brain stem of the RRMS patients, when compared to age- and gender-matched controls. Lack of detectable upper cervical cord atrophy in the same patients suggests some independence of the MS pathology in these regions.
Keywords: Multiple sclerosis, brain stem, atrophy, MRI. Acceptance: Received February 20, 2015. Accepted for publication February 27, 2015. Correspondence: Address correspondence to Tomos R. Chivers, Division of Clinical Neurology, C Floor, South Block, University Hospital NHS Trust, Queen’s Medical Centre, Nottingham, NG7 2HU, UK. E-mail: [email protected]. Funding: This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. J Neuroimaging 2015;25:1002-1006. DOI: 10.1111/jon.12244
Introduction Multiple Sclerosis (MS) is a disease that affects the central nervous system (CNS) causing chronic demyelination, inflammation, and axonal loss. Its distinguishing feature is the development of lesions throughout the CNS. Progressive atrophy of the CNS is a recognized feature of MS, and is thought to be a better correlate of permanent disability than T2 lesion load, which varies more unpredictably compared with disability.1–3 Magnetic resonance imaging (MRI) studies of atrophy in MS have shown that there is a correlation between axonal loss and CNS atrophy.4 Further studies have also shown that demyelination also plays a role in atrophy, but demyelination is potentially reversible whereas axonal loss is not.5,6 Many MRI studies have compared the brain and spinal cord of relapsing-remitting MS (RRMS) patients with those of unaffected controls. The correlation between brain atrophy and progression of RRMS is widely accepted,7 and detectable atrophy appears early in the disease.8,9 Spinal cord atrophy appears to exist in both progressive forms of MS and relapsing-remitting forms of MS.10–12 While there have been many studies focused on atrophy of the brain and spinal cord, relatively little research has focused on the brain stem. The brain stem is a common site for pathology in MS with one study of 114 clinically definite MS patients
1002
Copyright
finding brain stem lesions in 68%.13 Pathology of the brain stem due to MS can cause eye movement disorders, vertigo, tremor, urinary problems, hemiplegia, facial pain, paralysis, and hearing problems. Many of these manifestations may develop early in MS. In a recent study of MS patients, brain stem reflexes were abnormal in 87% and evoked potentials in 83%.14 Other studies have shown a strong correlation between brain stem lesion load, diffusion abnormalities, and disability.15 Early studies looking at brain stem atrophy utilized computer tomography (CT) to examine the brain stem. One study of 200 MS patients found brain stem/cerebellar atrophy in 53 of the patients and found there was a strong correlation between brain stem atrophy and brain stem dysfunction.16 CT images, however, lack the contrast and resolution necessary to accurately take measurements, so recent studies have utilized MRI.5 Of the MRI studies that have looked at brain stem atrophy in MS, there has been some inconsistency. Those that have detected atrophy in MS patients, including RRMS, when compared with controls, have found levels of volume reduction that vary from 11%17 to 20.6%.18 However, other studies, such as one looking at 49 RRMS patients, have found no noticeable atrophy compared with age-matched controls.19 Modern techniques such as voxel-based morphometry (VBM) have been used to show atrophy in brain structures.20
◦ 2015 by the American Society of Neuroimaging C
VBM has been used to show white matter atrophy in the brain stem and that this correlates with lesion volume.21 VBM, however, is better suited to studies with large groups of subjects where it is not practical to take manual measurements.22 In this study we measure the cross-sectional area (CSA) of three brain stem structures: the midbrain, pons, and medulla oblongata. We use a semiautomated algorithm with good reproducibility in 35 RRMS patients, and 35 age- and gendermatched control subjects; these subjects were previously used to study atrophy in the upper cervical cord.23 Intersubject variance is reduced by normalizing results to the total external cranial volume (TECV), a dimension measure we assume to be unaffected by multiple sclerosis.
Methods Patients and Controls Thirty-five clinically definite RRMS patients24 (21 females and 14 males) and 35 control subjects (21 females and 14 males) were recruited to the study. The median age (interquartile range) for patients was 39 (32-44) years and the median age of control subjects was 34 (29-42) years. Patients were excluded if they had received immunosuppressive or immunomodulatory drugs in the last year or corticosteroids in the last 3 months so as to avoid any effect treatment may have upon the processes of inflammation or atrophy. The median disease duration was 9 (5-13) years and the median expanded disability status scale score (EDSS)25 was 3 (2-4). Approval was obtained from the local research ethics committees. All patients gave informed written consent.
MRI Acquisition
sites using a semiautomated flood fill algorithm. A seed point was selected within each of the structures and a semiautomated flood fill algorithm used to grow the seed to capture the region of interest. This was performed on all 35 patients and 35 controls. Previous studies have employed the total intracranial volume (TICV) to normalize results.5,7,18,23,28 Despite using a similar method to previous studies to normalize TICV,18,23,28 preliminary work found that using the external volume of the cranium gives a better correlation to brain stem CSA than TICV, possibly because TICV was more difficult to measure reproducibly. TECVs were recorded to allow normalization of the CSA to reduce the interindividual variations in subject size. To measure the TECV, the sagittal view of the cranium was first rotated until the frontal pole of the cerebrum was horizontal with the posterior surface of the cerebellum. The external surface of the calvaria was outlined on a number of axial slices starting with those where the cerebral tissue was first visible and ending when cerebellar tissue was no longer visible. Once these regions of interest (ROI) were established, the slice volumes were added together to calculate the TECV. TECVs were compared for males and females to ensure the method was sensitive to known differences. T2 lesions in the brain stem and cerebellum were measured on axial T2 weighted images for all of the 35 patients. All T2 hyperintense lesions were identified as ROIs and the individual volumes were added together to give T2 lesion load.
Data Analysis The brain stem CSA were normalized to the TECV using a method previously summarized by Jack et al.29 The equation used for normalization was: CSAn = CSA + α(TECVmean − TECV),
Brain MRI was performed on patients and controls using a 1.5T Siemens Vision (Siemens, Erlangen, Germany) MRI scanner with a quadrature head coil. This was used to acquire 3-dimensional magnetization-prepared rapid acquisition gradient-echo (MPRAGE) images with 128 contiguous slices, covering the whole brain and brain stem (voxel size = .98 × .98 × 1.25 mm, matrix size = 256 × 256 × 128, repetition time (TR) = 9.7 ms, echo time (TE) = 4 ms, inversion time (TI) = 300 ms, flip angle = 10°), and a T2-weighted image with 3 mm transverse contiguous images was acquired (TR 4,100 ms, TE 90 ms, matrix 256 × 256, field of view 25 cm).
where CSAn is the normalized CSA and TECVmean is the average TECV for the control subjects. The coefficient α is the gradient of the least-squares straight line fit of CSA to TECV in the controls (Fig 2). Statistical tests were performed on pre- and postnormalized data using Pearson’s or Spearman’s correlation coefficient, twosample t-tests for comparison of mean values, and a MannWhitney U test for comparison of median values. Ten subjects had repeat scans to assess scan-rescan reproducibility.
MRI Analysis
Results
Image analysis was performed using NeuRoi.26 Three slices of the brain stem, one in each of the midbrain, pons and medulla oblongata were chosen as sites to measure CSA on the T1 images (see Fig 1). The midbrain was identified as the axial slice where the superior colliculi were at their largest. The pons was identified as the most inferior of the axial slices where the superior cerebellar peduncles have yet to attach to the cerebellum. The medulla oblongata was identified as the axial slice at the superior origin of the pyramids, just inferior to the floor of the fourth ventricle. These were chosen due to the ease of consistently identifying them and they were largely surrounded by sufficient cerebrospinal fluid (CSF) to give high contrast between brain stem and surroundings. All three sites are known areas of MS pathology.27 The brain stem was rotated in the sagittal view until vertical. It was then delineated on the axial view at the three chosen
There was no significant difference in the median age of the controls and the patients (P = .2). The coefficient of variance was calculated for nonnormalized CSA using scan-rescan images and was found to be 1.7% in the midbrain, 1.9% in the pons, and 5.5% in the medulla oblongata.
Brain stem Atrophy After normalization to the TECV, the mean CSA of the midbrain, pons, and medulla oblongata in the patients was significantly reduced relative to the control group (P ࣘ .001, P = .001 and P = .047, respectively); see Table 1. The differences between patient and control mean normalized CSA of the midbrain, pons, and medulla oblongata were 50.2 mm2 (CI 48.0–52.4 mm2 ; 9.5%), 45.1 mm2 (CI 44.3–45.9 mm2 ; 8.9%), and 12.8 mm2 (CI 12.0–13.6 mm2 ; 6.7%), respectively.
Chivers et al: MRI-Based Measurement of Brain Stem CSA in RRMS
1003
A
B
C
Fig 1. Examples of an outlined midbrain (A), pons (B), and medulla oblongata (C).
Fig 2. Normalizing for TECV. Gradient of the least-squares straight line fit of CSA to TECV in the control = coefficient α. Table 1. CSAs in Patients and Controls Postnormalization
Midbrain Pons Medulla
Control Mean CSA (mm2 )
Patient Mean CSA (mm2 )
Mean CSA Difference (%)
530.7 ± 8.9 509.6 ± 9.8 190.8 ± 4.8
480.5 ± 7.8 464.5 ± 9.4 178.0 ± 3.8
9.5%*** 8.9%*** 6.7%*
Significantly reduced CSA in patients compared to control is indicated by asterisk. * P ࣘ .05, ** P ࣘ .01, *** P ࣘ .001.
There was no significant correlation between normalized CSA and disease duration. There was no significant correlation between EDSS and CSA in the midbrain (r = .18; P = .28) and pons (r = .07; P = .67). There was, however, a significant correlation between EDSS and CSA in the medulla oblongata (r = .36; P = .03). Mean T2 lesion load in the cerebellum and brain stem was 1.3 mL (range 0-7.9 mL). There was a significant correlation between T2 lesion load and normalized CSA in the midbrain (Pearson’s r = –.39, P = .02, CI –.67 to –.01) but there was no significant correlation in the pons (P = .22, Pearson’s r = –.21, CI –.55 to .18) and medulla oblongata (P = .52, Pearson’s r = –.11, CI –.47 to .28).
TECV The TECV in the controls was significantly smaller (P ࣘ .001) in the females (2,296 ± 51 mL) than the males (2,648 ± 42 mL). There was significant correlation between the control TECV 1004
and the CSA of the midbrain (Pearson’s r = .58; P ࣘ .001), pons (Pearson’s r = .54; P ࣘ .001), and medulla oblongata (Pearson’s r = .46; P = .005) of the controls. The gradient of the least-squares straight line fit (α) for the midbrain, pons, and medulla oblongata was .1378, .1378, and .0554, respectively. The TECV in the patient group was significantly smaller (P ࣘ .001) in the females (2,329 ± 41 mL) than the males (2,722 ± 66 mL). In the patients, there was a significant correlation between the TECV and the CSA of the midbrain (Pearson’s r = .44; P = .008), but not the pons (Pearson’s r = .23; P = .18) or medulla oblongata (Pearson’s r = .24; P = .17) of the patients. There was no significant difference (P = .46) in the mean TECV between the patients and the normal controls.
Gender In the controls, the mean CSA of midbrain (P = .02), pons (P = .001), and medulla oblongata (P = .003) was significantly smaller in females than in males. There was no difference in the mean of the normalized CSA of control males and females. In the patients, there was no significant difference between the CSA of the males and females in the midbrain (P = .06), pons (P = .72), and medulla oblongata (P = .13) (see Table 1). There was no significant difference between male and female mean normalized CSA of the midbrain and medulla oblongata; however, the female mean normalized CSA was significantly larger in the pons (P = .01).
Journal of Neuroimaging Vol 25 No 6 November/December 2015
Discussion Brain stem Atrophy This study adds to relatively sparse literature showing reduced brain stem volume17,18,28 in MS by demonstrating that three structures in the brain stem are significantly atrophied when normalized to account for natural variance and compared with matched controls. After normalization, a CSA reduction of 9.3% was found in the midbrain, 8.7% in the pons, and 6.5% in the medulla oblongata. Several longitudinal studies have suggested that brain volume decreases by .6% to 1% a year in patients with MS.5,30,31 The median disease duration of the patients in this study was 9 years, which may indicate that the hypothesized rate of atrophy in the brain could also apply to the brain stem. Further work would be necessary to confirm this. Given the disease duration and the fact the patients were free of disease modifying treatment, it might also suggest that atrophy is not a late event. This would be consistent with current concepts about atrophy in MS.3 The link between disability and brain stem atrophy was explored by examining the relationship between EDSS, disease duration, and brain stem CSA. While there was no significant correlation with disease duration at any of the sampled sites, there was a significant correlation between EDSS and normalized CSA at the medulla oblongata. This, however, was a secondary finding requiring further work. We attempted to see if there was a correlation between T2 lesion load in the brain stem and the normalized brain stem CSA. Previous studies suggest a relationship between T2 lesion load and brain atrophy32 and it is possible this may also apply to the brain stem. A significant correlation was found in the midbrain but not at the pons or medulla oblongata. The observed reductions in mean CSA show very similar levels of atrophy in the pons and midbrain, but reduced levels of atrophy in the medulla oblongata. This might be due to measurement error; the medulla oblongata was more difficult to delineate than other parts of the brain stem as there was less contrast between the CSF and brain tissue. It could, alternatively, be due to partial volume averaging, an issue affecting accuracy when measuring smaller structures such as the medulla oblongata. A correction for partial volume averaging, such as the one used by Mann et al,23 might have increased the accuracy.
Spinal Cord Atrophy The patients from this study were previously investigated to determine if atrophy occurs in the upper cervical cord in RRMS.23 In that study no reduction in cord CSA was detected, and it was hypothesized that the atrophy might be masked by inflammation. However, a reduction in brain stem CSA has been detected in this study. This result may indicate a degree of independence between cord and brain stem pathology. Previous studies33–35 have also found this disconnection between brain and cord atrophy.
Normalization Normalization to the TECV was performed in order to remove some of the normal population variance and gender differences, hence improving the power of the study. Correlation between the TECV and CSA shows that between 22% and 33% (site
dependent) of the variance in CSA can be explained by TECV. This agrees with existing opinions about normalization.36 In the cervical cord study, the same method of normalization23 was used to show that 38% of the variation in CSA was explained by TICV. A similar level was achieved in this study using TECV. TECV might be a convenient normalization dimension in future atrophy studies.
Gender Differences We observed a significant difference in both the size of the TECV and brain stem of male and female controls. In all three sites sampled, CSA was significantly larger in male controls than female controls, and correlated with TECV, which was also significantly larger in males than females. This is expected17,23 and suggests that the method used in the experiment is sensitive. In the patients, there was also a gender difference in TECV. This, however, was not detectible in the brain stem CSA of the patients. The absence of this normal physiological difference may be the result of an external factor playing a role in determining the size of brain stem CSA; this might be inflammation, atrophy, or both in MS. Normalization removed all gender differences in both patients and controls with the exception of one site, in the pons of the patients. This apparently significantly larger, after normalization, pons CSA in females is interesting, but may represent a special case in our patient population.
Study Limitations Factors such as upgrades or repairs to the scanner could affect CSA readings over time.7 The images collected in this study were taken over a long period so it is possible that changes in the scanner could have affected the results. Normalization, however, should remove changes in volume due to scanner recalibration. The coefficient of variance was used to assess error on repeating the scan. It was found to be acceptably low in the midbrain and pons. There was, however, a slightly high level of error and weak level of significance in the medulla oblongata. It is possible that if the study was repeated with the increased sensitivity of 3T MRI it would be easier to delineate the medulla, and the level of error at this site would be more acceptable.37
Conclusions Normalization using the TECV was successfully used to reduce normal subject variance. A significantly reduced, normalized brain stem CSA was detected in all areas of the brain stem of the RRMS patients, when compared to age- and gender-matched controls. Lack of detectable upper cervical cord atrophy in the same patients suggests some independence of the MS pathology in these regions.
References 1. Miller DH, Grossman RI, Reingold SC, et al. The role of magnetic resonance techniques in understanding and managing multiple sclerosis. Brain 1998;121:3-24. 2. Bakshi R, Dandamudi VSR, Neema M, et al. Measurement of brain and spinal cord atrophy by magnetic resonance imaging as a tool to monitor multiple sclerosis. J Neuroimaging 2005;15:30S-45S. 3. Zivadinov R, Sepcic J, Nasuelli D, et al. A longitudinal study of brain atrophy and cognitive disturbances in the early phase of relapsing-remitting multiple sclerosis. J Neurol Neurosurg Psychiatry 2001;70:773-80.
Chivers et al: MRI-Based Measurement of Brain Stem CSA in RRMS
1005
4. Davie CA, Barker GJ, Webb S, et al. Persistent functional deficit in multiple sclerosis and autosomal dominant cerebellar ataxia is associated with axon loss. Brain 1995;118:1583-92. 5. Miller DH, Barkhof F, Frank JA, et al. Measurement of atrophy in multiple sclerosis: pathological basis, methodological aspects and clinical relevance. Brain 2002;125:1676-95. 6. Zivadinov R, Bakshi R. Role of MRI in multiple sclerosis II: brain and spinal cord atrophy. Front Biosci. 2004;9:647-64. 7. Minagar A, et al. Central nervous system atrophy in multiple sclerosis: mechanisms. In: Zivadinov R, Bakshi R, eds. Brain and Spinal Cord Atrophy in Multiple Sclerosis. New York: Nova, 2004;9:5-13. 8. Brex PA, Jenkins R, Fox NC, et al. Detection of ventricular enlargement in patients at the earliest clinical stage of MS. Neurology 2000;54:1689-91. 9. Simon JH. Brain and spinal cord atrophy in multiple sclerosis. Neuroimaging Clin N Am. 2000;10:753-70. 10. Losseff NA, Webb SL, O’Riordan JI, et al. Spinal cord atrophy and disability in multiple sclerosis. A new reproducible and sensitive MRI method with potential to monitor disease progression. Brain 1996;119:701-8. 11. Lin X, Tench CR, Turner B, et al. Spinal cord atrophy and disability in multiple sclerosis over four years: application of a reproducible automated technique in monitoring disease progression in a cohort of the interferon beta-1a (Rebif) treatment trial. J Neurol Neurosurg Psychiatr 2003;74:1090-4. 12. Valsasina P, Rocca MA, Horsfield MA, et al. Regional cervical cord atrophy and disability in multiple sclerosis: a voxel-based analysis. Radiology 2013;266:853-61. 13. Ormerod IE, Miller DH, McDonald WI, et al. The role of NMR imaging in the assessment of multiple sclerosis and isolated neurological lesions. A quantitative study. Brain 1987;110:1579616. 14. Magnano I, Pes GM, Pilurzi G, et al. Exploring brainstem function in multiple sclerosis by combining brainstem reflexes, evoked potentials, clinical and MRI investigations. Clin Neurophysiol 2014;125:2286-96. 15. Preziosa P, Rocca MA, Mesaros S, et al. Relationship between damage to the cerebellar peduncles and clinical disability in multiple sclerosis. Radiology 2014;271:822-30. 16. Loizou LA, Rolfe EB, Hewazy H. Cranial computed tomography in the diagnosis of multiple sclerosis. J Neurol Neurosurg Psychiatr 1982;45:905-12. 17. Liptak Z, Berger AM, Sampat MP, et al. Medulla oblongata volume: a biomarker of spinal cord damage and disability in multiple sclerosis. AJNR Am J Neuroradiol 2008;29:1465-70. 18. Liu C, Edwards S, Gong Q, et al. Three dimensional MRI estimates of brain and spinal cord atrophy in multiple sclerosis. J Neurol Neurosurg Psychiatry 1999;66:323-30. 19. Lin X, Blumhardt LD, Constantinescu CS. The relationship of brain and cervical cord volume to disability in clinical subtypes of multiple sclerosis: a three-dimensional MRI study. Acta Neurol Scand 2003;108:401-6. 20. Lansley J, Mataix-Cols D, Grau M, et al. Localized grey matter atrophy in multiple sclerosis: a meta-analysis of voxel-based morphometry studies and associations with functional disability. Neurosci Biobehav Rev 2013;37:819-30.
1006
21. Ceccarelli A, Jackson JS, Arora A, et al. Abstract #7: Voxel Based Morphometry with DARTEL in Multiple Sclerosis. Neurotherapeutics 2010;7:330. 22. Whitwell JL. Voxel-based morphometry: an automated technique for assessing structural changes in the brain. J Neurosci 2009;29:9661-4. 23. Mann RS, Constantinescu CS, Tench CR. Upper cervical spinal cord cross-sectional area in relapsing-remitting multiple sclerosis: application of a new technique for measuring cross-sectional area on magnetic resonance images. J Magn Reson Imaging 2007;26: 61-5. 24. Poser CM, Paty DW, Scheinberg L, et al. New diagnostic criteria for multiple sclerosis: guidelines for research protocols. Ann Neurol 1983;13:227-31. 25. Kurtzke JF. Rating neurologic impairment in multiple sclerosis: an expanded disability status scale (EDSS). Neurology 1983;33:144452. 26. Tench C. NeuRoi [Internet]. Nottingham University Department Clinical Neurology. www.nottingham.ac.uk/scs/divisions/ clinicalneurology/software/neuroi/aspx. Accessed March 4, 2013. 27. Nakashima I, Fujihara K, Miyazawa I, et al. Clinical and MRI features of Japanese patients with multiple sclerosis positive for NMO-IgG. J Neurol Neurosurg Psychiatr 2006;77:1073-5. 28. Edwards SG, Gong QY, Liu C, et al. Infratentorial atrophy on magnetic resonance imaging and disability in multiple sclerosis. Brain 1999;122:291-301. 29. Jack CR Jr, Twomey CK, Zinsmeister AR, et al. Anterior temporal lobes and hippocampal formations: normative volumetric measurements from MR images in young adults. Radiology 1989;172:54954. 30. Inglese M, Rovaris M, Giacomotti L, et al. Quantitative brain volumetric analysis from patients with multiple sclerosis: a follow-up study. J Neurol Sci 1999;171:8-10. 31. Ge Y, Grossman RI, Udupa JK, et al. Brain atrophy in relapsingremitting multiple sclerosis and secondary progressive multiple sclerosis: longitudinal quantitative analysis. Radiolog. 2000;214: 665-70. 32. Chard DT, Brex PA, Ciccarelli O, et al. The longitudinal relation between brain lesion load and atrophy in multiple sclerosis: a 14 year follow up study. J Neurol Neurosurg Psychiatry 2003;74:1551-4. 33. Kearney H, Rocca MA, Valsasina P, et al. Magnetic resonance imaging correlates of physical disability in relapse onset multiple sclerosis of long disease duration. Mult Scler 2014;20:72-80. 34. Bonati U, Fisniku LK, Altmann DR, et al. Cervical cord and brain grey matter atrophy independently associate with long-term MS disability. J Neurol Neurosurg Psychiatry 2011;82:471-2. 35. Cohen AB, Neema M, Arora A, et al. The relationships among MRI-defined spinal cord involvement, brain involvement, and disability in multiple sclerosis. J Neuroimaging 2012;22:122-8. 36. Sormani MP, Rovaris M, Valsasina P, et al. Measurement error of two different techniques for brain atrophy assessment in multiple sclerosis. Neurology 2004;62:1432-4. 37. Stankiewicz JM, Glanz BI, Healy BC, et al. Brain MRI lesion load at 1.5T and 3T versus clinical status in multiple sclerosis. J Neuroimaging 2011;21:e50-6.
Journal of Neuroimaging Vol 25 No 6 November/December 2015
