Diffusion Tensor MRI in MultipleSclerosis

Marco Rovaris, MD

Massimo Filippi, MD

A B S T R A C T

During the last 10 years, thanks to the development of sophis-ticated acquisition schemes and the application of novel im-age analysis and postprocessing, diffusion tensor (DT) mag-netic resonance imaging (MRI) has increasingly been applied tothe study of multiple sclerosis (MS). DT MRI proved to be able todetect and quantify tissue damage within and outside T2-visibleMS lesions. In addition, DT MRI has been shown to be sen-sitive to the evolution of MS damage over short-term periodsof time, and therefore holds promise to provide us with in vivocorrelates of MS clinical severity, as well as predictors of long-term disease evolution. Recent developments of the technique,such as DT tractography, are likely to improve dramatically ourunderstanding of the mechanisms associated to the accumula-tion of MS disability. Unresolved issues to be addressed includethe definition of the actual features underlying diffusion changesin MS and the potential of DT MRI in the differential diagnosisbetween MS and other demyelinating conditions. The best ac-quisition and postprocessing strategies for DT MRI studies ofMS also remain a matter of debate. Moreover, the precision andaccuracy of DT MRI scans in detecting longitudinal, MS-relatedchanges need to be further investigated. This is a pivotal is-sue for a future application of DT MRI to the monitoring of MSevolution in large-scale clinical trials and, possibly, in individualpatients.

Key words: magnetic resonance imaging, multiple sclerosis,diffusion tensor MRI.

Rovaris M, Filippi M.Diffusion tensor MRI in multiple sclerosis.

J Neuroimaging 2007;17:27S-30S.DOI: 10.1111/j.1552-6569.2007.00133.x

From the Neuroimaging Research Unit, Department ofNeurology, San Raffaele Scientific Institute, Milan, Italy.

Address correspondence to Massimo Filippi, MD, Neu-roimaging Research Unit, Department of Neurology,Scientific Institute and University Ospedale San Raf-faele, via Olgettina 60, 20132 Milan, Italy. E-mail:filippi.massimo@hsr.it.

Introduction

Soon after its first in vivo applications, diffusion magneticresonance imaging (MRI) has been used to study multiplesclerosis (MS), because of its ability to detect and quantifydisease-related pathology of the central nervous system(CNS). The characteristics of diffusion, which can be de-fined as the random translational motion of moleculesin a fluid system, are influenced by several CNS tissuecomponents, including cell membranes and organelles.The MRI-measured diffusion coefficient of healthy CNStissues is, therefore, lower than that in free water and it isnamed apparent diffusion coefficient (ADC).1 Patholog-ical processes, which result in a decrease of “restricting”barriers, can determine an increase of the ADC values.Since the magnitude of diffusion is also dependent onthe direction in which it is measured, its full characteriza-tion can only be obtained in terms of a diffusion tensor(DT),2 a matrix that accounts for the correlation exist-ing between molecular displacement along orthogonaldirections. From the DT, it is possible to derive the meandiffusivity (MD), which is independent of the spatial ori-entation of tissue structures, and some other dimension-less indexes of anisotropic diffusion, including fractionalanisotropy (FA), which reflect the prevalence of diffusiv-ity along one spatial direction (eg, along axonal fibersrather than perpendicular to them).

Earlier studies of MS3 consistently reported increasedADC values both within T2-visible lesions and in thenormal-appearing white matter (NAWM), without a clearrelationship with patients’ clinical features. These studieswere, however, affected by several technical constraintsof the MR sequences in use, such as limited brain cov-erage and susceptibility to motion artifacts. During thelast 10 years, thanks to the development of more sophis-ticated acquisition schemes and the application of novelimage analysis and postprocessing techniques, DT MRIhas increasingly been applied to the study of MS withpromising results.

DT MRI and MS: The Lessons We Learned During

the Last Decade

The major achievements and the most relevant contribu-tions of DT MRI in the study of MS can be summarizedas follows:

Copyright ◦C 2007 by the American Society of Neuroimaging 27S

Fig 1. 3D projections of DT tracking of pyramidal tract (A) and corpus callosum (B) fibers in a right-handed, 34-year-oldhealthy man.

1. The main post mortem correlates of diffusivity changes inMS are demyelination and axonal loss. This is more ev-ident for anisotropy indexes than for ADC.4 The notionthat DT MRI has the potential to quantify the severity ofdegenerative changes in the MS brain is also supported bythe results of studies of normal aging5,6 and Alzheimer’sdisease.7

2. DT MRI reflects the variety of pathological featureswithin T2-visible MS lesions. In the different types oflesions, diffusion abnormalities are always more pro-nounced than those found in the normal-appearing braintissues, but their values are highly heterogeneous (see ref.8 for a review). The highest degrees of abnormalities havebeen found in T1 “black holes,”9−11 which represent areasof irreversible tissue disruption and axonal loss. On theother hand, conflicting results have been obtained whencomparing the findings in enhancing vs. nonenhancinglesions,9-11 thus raising the unsolved issue of how muchDT MRI-detectable tissue disorganization within the for-mer ones is permanent (ie, related to neurodegeneration)and how much is transient (ie, related to edema, demyeli-nation, and remyelination).

3. DT MRI discloses the presence of diffusion abnormali-ties in the NAWM and gray matter (GM) of patients withMS (see ref. 8 for a review), even before the develop-ment of T2-visible lesions.12 Diffusion abnormalities be-come more pronounced with increasing disease durationand neurological impairment,13−15 thus supporting thenotion that DT MRI is particularly sensitive to the moredisabling features of MS pathology. The observed limitedcorrelations between DT MRI findings, the correspond-ing T2-visible lesion burden and other MRI measures oftissue damage in the NAWM/GM suggest that: (a) diffu-sivity changes do not merely depend upon neuroaxonalretrograde degeneration, but may reflect the presence ofpathological features (eg, microscopic lesions), which goundetected when using conventional MRI; (b) DT MRImay complement other techniques when assessing theactual burden of the disease in MS patients.

4. DT MRI is sensitive to the evolution of MS damage overshort-term periods of time.15,16 Changes of DT MRI-derived measures seem to be widely independent of theconcomitant T2-visible lesion and brain tissue volumechanges.15,16 Moreover, DT MRI might be more sensi-tive to the accrual of GM damage than to that of NAWMdamage.15,16

5. DT MRI and clinical findings in MS are correlated,the strongest relationships being those for the diffusioncharacteristics of T2-visible lesions10 and GM.13,17,18 DTMRI-derived quantities can contribute to the generationof composite MR-based scores able to explain a largepart of the variance of MS-related disability.19,20 Somestudies17,21 also indicated that the pathologic damage de-tected by DT MRI in clinically eloquent brain regionsmay be a significant factor contributing to clinical fea-tures in MS patients. Moreover, the results of a recentprospective study of patients with primary progressiveMS22 suggest that the severity of diffusion abnormalitiesin the GM predict disability worsening five years later.

6. DT MRI of the optic nerve and spinal cord, albeit pre-senting more technical difficulties than that of the brain,is now feasible23−26 and may allow us to study in vivo thecharacteristics of MS pathology in these clinically elo-quent CNS regions. We will not detail DT MRI findingsin the optic nerve and spinal cord, since they are exten-sively described elsewhere in this supplement.

7. The examination of DT eigenvectors,27 which representthe diffusion coefficients along the axes of the diffusionellipsoid, may provide a signature of Wallerian degenera-tion in the NAWM of MS patients. DT tractography meth-ods28 can also be used to segment the different functionalwhite matter structures in MS patients, such as the corti-cospinal tracts and the corpus callosum (Fig 1). Thanks toDT tractography, probing functional pathways with DTMRI measures may enable us to achieve a better reflec-tion of the disease status.29−33

DT MRI and MS: Unresolved Issues

and Future Perspectives

Little is still known about the actual features underly-ing diffusion changes in MS. On the other hand, how-ever, several pieces of evidence suggest that the various,and often concomitant, pathological changes occurring inthe MS brain might affect the diffusivity and anisotropycharacteristics of tissues in opposite ways, thereby re-ducing the sensitivity and specificity of DT MRI find-ings.34 These aspects have to be carefully considered

28S Journal of Neuroimaging Vol 17 No 2 (Supplement) April 2007

before interrogating the potential of DT MRI in the dif-ferential diagnosis between MS and other demyelinatingdisorders.35,36

The best acquisition and postprocessing strategies forMS studies remain a matter of debate (see ref. 37 for a re-view). The interscanner variability of DT MRI measuresand the issues related to high field scanner DT MRI ac-quisition (such as the increased chemical shift and suscep-tibility artifacts) warrant further investigation. Moreover,some limitations of DT tractography, which may lead toerroneous fiber tracing, still need to be overcome by fur-ther technical developments. The contribution of newerand more sophisticated techniques to DT MRI investiga-tions of MS also needs to be further evaluated. Amongthese techniques, high b-value q-space imaging38 holdspromise to increase the sensitivity of “conventional,” lowb-value DT MRI in the detection of NAWM abnormali-ties, even though it has a complex acquisition scheme.

The precision and accuracy of DT MRI scans in de-tecting longitudinal, MS-related changes also need to befurther investigated. This is a pivotal issue for a future ap-plication of DT MRI to the monitoring of MS evolutionin large-scale clinical trials and, eventually, individual pa-tients.

References

1. Le Bihan D, Breton E, Lallemand D, et al. MR imaging ofintravoxel incoherent motions: application to diffusion andperfusion in neurologic disorders. Radiology 1986;161:401-407.

2. Pierpaoli C, Jezzard P, Basser PJ, et al. Diffusion tensor MRimaging of the human brain. Radiology 1996;201:637-648.

3. Horsfield MA, Larsson HB, Jones DK, et al. Diffusion mag-netic resonance imaging in multiple sclerosis. J Neurol Neu-rosurg Psychiatry 1998;64(suppl 1):S80-S84.

4. Mottershead JP, Schmierer K, Clemence M, et al. High fieldMRI correlates of myelin content and axonal density inmultiple sclerosis—a post-mortem study of the spinal cord.J Neurol 2003;250:1293-1301.

5. Rovaris M, Iannucci G, Cercignani M, et al. Age-relatedchanges in conventional, magnetization transfer, anddiffusion-tensor MR imaging findings: study with whole-brain tissue histogram analysis. Radiology 2003;227:731-738.

6. Benedetti B, Charil A, Rovaris M, et al. Influence of agingon brain gray and white matter changes assessed by con-ventional, MT, and DT MRI. Neurology 2006;66:535-539.

7. Bozzali M, Falini A, Franceschi M, et al. White matter dam-age in Alzheimer’s disease assessed in vivo using diffusiontensor magnetic resonance imaging. J Neurol Neurosurg Psy-chiatry 2002;72:742-746.

8. Rovaris M, Gass A, Bammer R, et al. Diffusion MRI inmultiple sclerosis. Neurology 2005;65:1526-1532.

9. Filippi M, Iannucci G, Cercignani M, et al. A quantitativestudy of water diffusion in multiple sclerosis lesions andnormal-appearing white matter using echo-planar imaging.

Arch Neurol 2000;57:1017-1021.10. Filippi M, Cercignani M, Inglese M, et al. Diffusion tensor

magnetic resonance imaging in multiple sclerosis. Neurology2001;56:304-311.

11. Bammer R, Augustin M, Strasser-Fuchs S, et al. Magneticresonance diffusion tensor imaging for characterizing dif-fuse and focal white matter abnormalities in multiple scle-rosis. Magn Reson Med 2000;44:583-589.

12. Rocca MA, Cercignani M, Iannucci G, et al. Weeklydiffusion-weighted imaging of normal-appearing whitematter in MS. Neurology 2000;55:882-884.

13. Bozzali M, Cercignani M, Sormani MP, et al. Quantifica-tion of brain gray matter damage in different MS pheno-types by use of diffusion tensor MR imaging. AJNR Am JNeuroradiol 2002;23:985-988.

14. Rovaris M, Bozzali M, Iannucci G, et al. Assessmentof normal-appearing white and grey matter in patientswith primary progressive multiple sclerosis. Arch Neurol2002;59:1406-1412.

15. Rovaris M, Gallo A, Valsasina P, et al. Short-term accrualof gray matter pathology in patients with progressive mul-tiple sclerosis: an in vivo study using diffusion tensor MRI.Neuroimage 2005;24:1139-1146.

16. Oreja-Guevara C, Rovaris M, Iannucci G, et al. Progressivegrey matter damage in patients with relapsing-remittingMS: a longitudinal diffusion tensor MRI study. Arch Neurol2005;2005:62-578.

17. Rovaris M, Iannucci G, Falautano M, et al. Cognitivedysfunction in patients with mildly disabling relapsing-remitting multiple sclerosis: an exploratory study withdiffusion tensor MR imaging. J Neurol Sci 2002;195:103-109.

18. Vrenken H, Pouwels PJ, Geurts JJ, et al. Altered diffusiontensor in multiple sclerosis normal-appearing brain tissue:cortical diffusion changes seem related to clinical deterio-ration. J Magn Reson Imaging 2006;23:628-636.

19. Mainero C, DeStefano N, Iannucci G, et al. Correlates ofMS disability assessed in vivo using aggregates of MR quan-tities. Neurology 2001;56:1331-1134.

20. Pulizzi A, Rovaris M, Judica E, et al. Determinants of dis-ability in multiple sclerosis at various disease stages: a mul-tiparametric magnetic resonance study. Arch Neurol 2007(in press).

21. Oh J, Henry RG, Genain C, et al. Mechanisms of normalappearing corpus callosum injury related to pericallosalT1 lesions in multiple sclerosis using directional diffusiontensor and 1H MRS imaging. J Neurol Neurosurg Psychiatry2004;75:1281-1286.

22. Rovaris M, Judica E, Gallo A, et al. Grey matter damagepredicts the evolution of primary progressive multiple scle-rosis at 5 years. Brain 2006;129:2628-2634.

23. Hickman SJ, Wheeler-Kingshott CAM, Jones SJ, etal. Optic nerve diffusion measurement from diffusionweighted imaging in optic neuritis. AJNR Am J Neuroradiol2005;26:951-956.

24. Valsasina P, Rocca MA, Agosta F, et al. Mean diffusivityand fractional anisotropy histogram analysis of the cervicalcord in MS patients. Neuroimage 2005;26:822-828.

25. Agosta F, Benedetti B, Rocca MA, et al. Quantification ofcervical cord pathology in primary progressive MS usingdiffusion tensor MRI. Neurology 2005;64:631-635.

Rovaris and Filippi: DT MRI in MS 29S

26. Hesseltine SM, Law M, Babb J, et al. Diffusion tensor imag-ing in multiple sclerosis: assessment of regional differencesin the axial plane within normal-appearing cervical spinalcord. AJNR Am J Neuroradiol 2006;27:1189-1193.

27. Henry RG, Oh J, Nelson SJ, et al. Directional diffusionin relapsing-remitting multiple sclerosis: a possible in vivosignature of Wallerian degeneration. J Magn Reson Imaging2003;18:420-426.

28. Mori S, Kaufmann WE, Davatzikos C, et al. Imaging cor-tical association tracts in the human brain using diffusion-tensor-based axonal tracking. Magn Reson Med 2002;47:215-223.

29. Wilson M, Tench CR, Morgan PS, et al. Pyramidal tractmapping by diffusion tensor magnetic resonance imagingin multiple sclerosis: improving correlations with disability.J Neurol Neurosurg Psychiatry 2003;74:203-207.

30. Vaithianathar L, Tench CR, Morgan PS, et al. T1 relaxationtime mapping of white matter tracts in multiple sclerosis de-fined by diffusion tensor imaging. J Neurol 2002;249:1272-1278.

31. Pagani E, Filippi M, Rocca MA, et al. A method for ob-taining tract-specific diffusion tensor MRI measurementsin the presence of disease: application to patients with clin-ically isolated syndromes suggestive of multiple sclerosis.Neuroimage 2005;26:258-265.

32. Lin X, Tench CR, Morgan PS, et al. ‘Importance sam-pling’ in MS: use of diffusion tensor tractography to quan-tify pathology related to specific impairment. J Neurol Sci2005;237:13-19.

33. Lowe MJ, Horenstein C, Hirsch JG, et al. Functionalpathway-defined MRI diffusion measures reveal increasedtransverse diffusivity of water in multiple sclerosis. Neuroim-age 2006;32:1127-1133.

34. Rosso C, Remy P, Creange A, et al. Diffusion-weighted MRimaging characteristics of an acute strokelike form of mul-tiple sclerosis. AJNR Am J Neuroradiol 2006;27:1006-1008.

35. Benedetti B, Valsasina P, Judica E, et al. Grading cervicalcord damage in neuromyelitis optica and MS by diffusiontensor MRI. Neurology 2006;67:161-163.

36. Lin F, Yu C, Jiang T, et al. Discriminative analysis of re-lapsing neuromyelitis optica and relapsing-remitting mul-tiple sclerosis based on two-dimensional histogram fromdiffusion tensor imaging. Neuroimage 2006;31:543-549.

37. Pagani E, Bammer R, Horsfield MA, et al. Diffusion MRI inmultiple sclerosis: technical aspects and challenges. AJNRAm J Neuroradiol (in press).

38. Assaf Y, Chapman J, Ben-Bashat D, et al. White matterchanges in multiple sclerosis: correlation of q-space diffu-sion MRI and 1H MRS. Magn Reson Imaging 2005;23:703-710.

30S Journal of Neuroimaging Vol 17 No 2 (Supplement) April 2007