IMAGING IN MULTIPLE SCLEROSIS

Transcription

1 See end of artile for authors affiliations Correspondene to: Professor David H Miller, NMR Researh Unit, Department of Neuroinflammation, Institute of Neurology, University College London, Queen Square, London WC1N 3BG, UK; d.miller@ion.ul.a.uk IMAGING IN MULTIPLE SCLEROSIS MRI M S A Trip, D H Miller J Neurol Neurosurg Psyhiatry 2005; 76(Suppl III):iii11 iii18. doi: /jnnp ultiple slerosis (MS) is a ommon entral nervous system (CNS) disease haraterised pathologially by the development of multifoal inflammatory demyelinating white matter lesions. Magneti resonane imaging (MRI) is the gold standard imaging tehnique for the identifiation of demyelinating lesions whih an be used to support a linial diagnosis of MS, and MS an now be diagnosed in some patients after a linially isolated syndrome (CIS) using new MRI diagnosti riteria. Clinial trials of disease modifying treatments inlude MRI outome measures of disease ativity. The use of MR tehniques to study the evolution of MS pathology in vivo is an important tool for obtaining better insights into pathophysiologial mehanisms underlying the linial ourse. This artile will onentrate on the pratial appliation of MRI in the management of MS. CONTRAST HELPS TO IDENTIFY MS PATHOLOGY MR images are said to have ontrast if there are differing areas of signal intensity. The ontrast an be ontrolled to a ertain extent by the parameters used and an produe images whih ontain differing levels of ontrast depending on tissue ontent. For example, fat and water produe different image ontrast. Images an be weighted towards partiular ontrast mehanisms. In T1 weighted images, fat is bright and erebrospinal fluid (CSF) is dark; with T2 weighting, fat is dark and CSF is bright. Proton density (PD) weighting produes image ontrast on the basis of the PD of the tissue. Suh weighting of images has been exploited to identify inflammatory demyelinating lesions at different stages in their evolution. 1 T2 and PD weighted imaging T2 weighted imaging identifies MS lesions as high signal foi against the low signal bakground of white matter. However, periventriular lesions are often indistinguishable from the adjaent CSF whih is also of high signal with T2 weighting. Contrast from the lesion an be improved here by using PD weighting beause of the lower CSF signal with this sequene. Fortunately, T2 and PD weighted images an be aquired together in a single spin-eho sequene providing omplimentary information (fig 1). Fluid attenuated inversion reovery (FLAIR) sequene The problem of identifying lesions in the periventriular region, whih is a ommon site for MS lesions, an also be addressed by suppressing the signal from CSF yet maintaining heavy T2 weighting using a fluid attenuated inversion reovery (FLAIR) sequene (fig 1). This sequene is also superior at deteting ortial/juxtaortial lesions. FLAIR is therefore a ommonly used MR sequene on linial sanners when MS has been raised as a possible linial diagnosis. The only drawbak is inferior quality lesion detetion in the posterior fossa and spinal ord where PD and T2 weighting are preferred. T1 weighted imaging without ontrast Some high signal lesions on T2 weighted imaging will also be visible on T1 weighted images as areas of low signal ompared to the normal white matter and are ommonly known as blak holes (fig 2). Suh lesions when newly formed will either disappear with time, when it is thought they are aused by reversible oedema or demyelination, or persist as hroni blak holes, when it is thought they are aused by permanent axonal loss. T1 weighted imaging with gadolinium enhanement A gadolinium helate administered intravenously five minutes before T1 weighted imaging detets blood brain barrier breakdown in assoiation with ative inflammation. New lesions appear enhaned (fig 1) and usually persist for a month on average, making them a useful marker for monitoring disease ativity. Suh lesions play an important role in indiating dissemination in time within the new diagnosti riteria. Triple dose gadolinium or ombination with iii11 J Neurol Neurosurg Psyhiatry: first published as /jnnp on 16 August Downloaded from on 4 September 2018 by guest. Proteted by opyright.

2 iii12 magnetisation transfer imaging (see next setion) an both inrease ative lesion detetion further but are not required in linial pratie. NON-CONVENTIONAL MRI IN MS The onventional MR tehniques already disussed are restrited to the PD, T1, and T2 signal ontrast behaviour of MS lesions. While these have transformed the diagnosis of MS and provided information about the signifiane of lesions, not all the pathologial features of lesions an be studied using PD, T2, and T1 weighting, and CNS tissue that appears normal with suh images an show pathologial hanges using quantitative MR tehniques that are now disussed. 2 3 Magnetisation transfer imaging (MTI) Tissues ontain protons in the liquid phase (mobile pool) and protons whih are bound to maromoleules inluding proteins and lipids (bound pool). The latter have a very broad magneti resonane frequeny that normally deays too quikly for the sanner to detet. Bound pool protons are onstantly in a state of exhange with the mobile pool. If a strong radiofrequeny pulse is applied far enough away from the resonane of the mobile pool but is still able to exite the Figure 1 Axial magneti resonane imaging (MRI) of a 30 year old man with relapsing remitting multiple slerosis (MS) showing multiple periventriular lesions: (A) T2 weighted image; (B) proton density (PD) weighted image; (C) fluid attenuated inversion reovery (FLAIR) image; (D) T1 weighted image following administration of gadolinium (Gd) demonstrating enhaning lesions. bound pool, some of the magnetisation is transferred from the bound to the mobile pool. This produes a magnetisation transfer (MT) weighted image and the magnetisation transfer ratio (MTR) an be alulated from the MT image and an image without MT weighting. Magnetisation transfer imaging (MTI) has been shown to be a sensitive marker of pathologial hange in many neurologial disorders and, as a general rule, MTR dereases with inreasing pathologial hange. Correlative MTI pathologial studies have suggested that the myelin ontent and the axonal ount are the most relevant substrates of MTR hanges in patients with MS, espeially the former. MTI in MS has foused on three main areas: (1) using MTI with gadolinium to improve lesion detetion (disussed earlier); (2) distinguishing lesions of differing severity; (3) studying MTR in brain tissues that appear normal on onventional MRI. In vivo studies have shown that T1- hypointense lesions ( blak holes ) have a lower MTR than T1 isointense lesions, supporting the idea that these lesions our as a result of destrutive pathology. MTR values fall onsiderably when gadolinium enhanement ours in lesions, with a reovery of MTR over the following months, although not usually bak to normal. In some studies, brain MTR delines for several years. The severity of tissue damage J Neurol Neurosurg Psyhiatry: first published as /jnnp on 16 August Downloaded from on 4 September 2018 by guest. Proteted by opyright.

3 has some orrelation with the ourse of MS, as seondary progressive patients display lower lesion MTR than patients with benign MS. Whole brain MTR and segmented normal appearing white matter MTR are lower in patients than in healthy subjets, and independently predit subsequent disability. MTI an be used to study the time ourse of lesion development and has shown that some lesions develop in previously normal appearing tissue in whih there is delining MTR up to two years before gadolinium enhanement appears. Suh lesions are less likely to reover MTR values ompared to lesions developing in tissue with a previously normal MTR. Given the sensitivity of MTI to detet pathologial hange, guidelines for its use in future linial trials have been developed. Diffusion weighted imaging (DWI) Diffusion is the random motion of moleules in any fluid system inluding biologial tissue. Diffusion weighted imaging (DWI) refers to the proess of making MRI sensitive to the moleular motion of water moleules and is potentially a useful tehnique for studying white matter struture and pathology. Water diffusion an our in any diretion but ours preferentially along the orientation of axons beause their ell membranes at as barriers to diffusion. Suh diffusion is said to be anisotropi and is dependent on the strutural integrity of white matter trats. Any disruption to white matter trats or axonal membrane permeability should lead to an inrease in the apparent diffusion oeffiient (ADC) and mean diffusivity (MD), a measure of average moleular motion, and also to a derease in frational anisotropy (FA), a measure of the diretional preponderane of diffusion whih an be obtained with diffusion tensor imaging (DTI). DWI has been applied qualitatively in linial pratie where it an be used to detet ishaemi hanges within minutes of aute stroke. DWI and DTI have supplemented onventional MRI tehniques for the quantitative study of MS pathology in vivo. DTI studies in MS have shown inreased ADC and MD with dereased FA in hroni T 1 hypointense lesions ompared to T 1 isointense lesions whih is ompatible with evidene that T 1 hypointense lesions Abbreviations ADC: apparent diffusion oeffiient ADEM: aute disseminated enephalomyelitis CADASIL: erebral autosomal dominant arteriopathy with subortial infarts and leuoenephalopathy CIS: linially isolated syndrome CNS: entral nervous system CSF: erebrospinal fluid DTI: diffusion tensor imaging DWI: diffusion weighted imaging EDSS: expanded disability status sale FA: frational anisotropy FLAIR: fluid attenuated inversion reovery MD: mean diffusivity MRI: magneti resonane imaging MRS: magneti resonane spetrosopy MS: multiple slerosis MTI: magnetisation transfer imaging MTR: magnetisation transfer ratio SLE: systemi lupus erythematosus PD: proton density PML: progressive multifoal leuoenephalopathy PPMS: primary progressive multiple slerosis VEP: visual evoked potential represent more extensive tissue loss. FA is lower in aute, gadolinium enhaning lesions ompared to non-enhaning lesions beause extraellular oedema alters the anisotropi pattern of diffusion. ADC and MD values are raised, but the extent may depend upon the lesion age. Changes in normal appearing brain tissue in MS patients have been deteted by diffusion measures. As with MTI hanges, these an our early in the linial ourse of MS and an preede new lesion formation. Several studies have reported a relation between DWI hanges in normal appearing brain tissue and disability in different MS subtypes. DWI has not yet been inluded in linial trials but may be useful in future studies of treatments whih may prevent axonal loss. The prodution of maps by DTI showing the prinipal diretion of diffusion on a voxel by voxel basis allows the path of white matter trats to be traed through the brain, Figure 2 Axial MRI of a 46 year old man with seondary progressive MS showing a large left sided periventriular lesion whih is hyperintense with (A) T2 weighted imaging and hypointense with (B) T1 weighted imaging ( blak hole ). iii13 J Neurol Neurosurg Psyhiatry: first published as /jnnp on 16 August Downloaded from on 4 September 2018 by guest. Proteted by opyright.

4 iii14 and this is known as tratography. This novel tehnique has emerging promise for investigation of white matter pathology in MS and other disorders. Magneti resonane spetrosopy (MRS) Proton magneti resonane spetrosopy (MRS) enables the in vivo investigation of metaboli alterations assoiated with brain pathology, and provides a quantitative method for investigating the abnormalities in the normal appearing white matter. At longer MRI eho times, N-aetyl-aspartate (a neuroaxonal marker), reatine and phosphoreatine, and holine ontaining ompounds may be quantified, while at shorter eho times additional metabolite peaks may be seen from myo-inositol (a potential marker of glial ells), glutamate and glutamine, and mobile lipids. MRS studies in patients with established MS have demonstrated signifiant abnormalities in the onentration of metabolites in the normal appearing white matter, with notably redued N-aetyl-aspartate, raised myo-inositol, and raised glutamate. A derease in N-aetyl-aspartate is thought to indiate axonal dysfuntion or loss, while an inrease in myo-inositol has been proposed to reflet an inrease in glial ell ativity or numbers, and raised glutamate may have a role autely in axonal injury. The stage at whih these abnormalities first appear is less lear but two studies of patients with a CIS have shown separately that N-aetylaspartate is redued and myo-inositol is raised. It is relevant to study this early phase of disease in order to searh for prognosti markers for the future linial ourse and to gain insights into early pathologial proesses. DISTRIBUTION OF LESIONS IN MS Brain Lesions an be seen anywhere in the brain parenhyma, but there are ertain areas in whih they are typial for MS. 1 Charateristi sites are in the periventriular region and juxtaortial white matter. The prediletion for these sites may be explained in part by the distribution of venules where blood brain barrier breakdown takes plae. This onept is illustrated by the desription of lesions arranged like fingers pointing away from the lateral ventriular walls whih is also seen in pathologial speimens where they are alled Dawson s fingers. These elliptial or ovoid lesions are believed to develop from demyelination along straight medullary venules arranged perpendiular to the ventriular walls; they are best seen on sagittal images where the ommon finding of lesions affeting the orpus allosum is also best seen. Juxtaortial lesions are often seen and are more suggestive of MS as they are unommon with aging. Although urrent sanners do not ommonly pik up grey matter lesions in the ortex, suh lesions are seen in a signifiant number of postmortem brains from MS patients. Below the tentorium, the erebellar pedunles and white matter, superfiial pons, and floor of the fourth ventrile are ommon sites of lesions. Opti nerves The ausative lesion of opti neuritis an usually be identified as an area of high signal on T2 or PD weighted imaging when appropriate fat and CSF suppression is applied. In theory, a linially silent lesion in an opti nerve ould provide evidene of dissemination in spae, but this is not sought routinely beause additional dediated sequenes are required to detet suh nerve lesions reliably. Spinal ord Spinal ord lesions are ommonly seen in MS. 4 In fat, the presene of ord lesions is more speifi for demyelination than in the brain beause age related or non-speifi ishaemi lesions are rare. T2 weighted imaging typially demonstrates small and irumsribed high signal lesions (fig 3). They tend to be aligned with the long axis of the ord and are usually less than two vertebral segments in length and involve less than half the axial ord area. On axial sans they may display a wedge appearane and reah the outer surfae of the ord (fig 3). The lesions are not onfined to white matter trats and may also involve entral grey matter. The ervial ord is most frequently affeted and most lesions involve the dorsal or lateral ord. Blak holes on T1 weighted imaging are not usually seen in the spinal ord but gadolinium enhanement of aute ord lesions does our. Foal ord swelling an our autely but this invariably resolves. More diffuse abnormalities an be seen in disabled patients, partiularly those with primary progressive disease whih has a higher disease burden in the spinal ord. Lesion dynamis T2 hyperintense lesions an alter in size over the ourse of weeks and a proportion of their volume disappears beause of resolution of oedema, although omplete resolution is rare. Lesions an also ause loal atrophy, a finding best appreiated in the opti nerve or spinal ord. About 50% of autely T1 hypointense lesions resolve and this may be partly aused by remyelination. Gadolinium enhanement of aute lesions may appear as pathy, homogenous (involving the whole lesions, whih is usually small), or ring shaped (often the ring is inomplete). VARIANTS OF MS There are more aggressive forms of MS suh as so alled tumefative MS whih presents as a large high signal mass lesion sometimes requiring biopsy. Other terms that have in the past been inluded within the spetrum of inflammatory demyelination in the CNS inlude: Marburg s disease, Balo s onentri slerosis, Shilder s disease, aute disseminated enephalomyelitis (ADEM), and Devi s neuromyelitis optia. Current evidene suggest that Marburg s disease is MS in an unusually aggressive form and that the onentri zones of demyelination and myelination seen in Balo s onentri slerosis are an oasional finding in otherwise typial MS. Cases previously onsidered to have Shilder s disease almost ertainly had posterior hemispheri white matter demyelination aused by MS or adrenoleuodystrophy. Monophasi ADEM is not unommon in hildren, and oasionally ours in adults although muh less frequently than MS. We doubt that reurrent ADEM exists in adults as an entity separate from MS and are onerned that suh a diagnosti label may prelude patients reeiving MS disease modifying treatments. In its lassial form, patients with Devi s neuromyelitis optia have severe episodes of opti neuritis and transverse myelitis with variable reovery, longitudinally extensive spinal ord MRI lesions, normal brain MRI, and a moderate to pronouned CSF pleoytosis without oligolonal bands. The suggestion that Devi s syndrome is immunopathogenially distint from MS omes from the reent observation by Lennon et al 5 that 70% of patients with the former ondition have a serum autoantibody against CNS vasular and meningeal tissue omponents. J Neurol Neurosurg Psyhiatry: first published as /jnnp on 16 August Downloaded from on 4 September 2018 by guest. Proteted by opyright.

5 Figure 3 Cervial spine MRI of a 49 year old woman with a spinal ord linially isolated syndrome (CIS) demonstrating a lesion at C3/4 on (A) sagittal imaging and (B) axial imaging whih shows that the lesion is on the right side of the ord. DIFFERENTIAL RADIOLOGICAL DIAGNOSIS OF WHITE MATTER LESIONS White matter T2 hyperintensities in the brain are not speifi to MS and are seen in a number of other disorders. 1 They an even be seen in otherwise normal individuals, partiularly with inreasing age. The hallenge is to reognise the typial lesion morphology and distribution of MS lesions (brain and spinal ord) in onjuntion with the harateristi linial presentations. Mirovasular ishaemi lesions These lesions are a ommon form of T2 hyperintense white matter lesions in the general population and are thought to be aused by ishaemi demyelination of the white matter. In addition, small vessel pathology, enlarged perivasular spaes, and sometimes infarts an all give rise to T2 hypertintense white matter lesions. The frequeny of mirovasular ishaemi lesions inreases with age. Of ourse, it is not simply the number but also the distribution and linial ontext whih determines their signifiane. The inrease with age makes it diffiult to assess MR imaging in the older patient who presents with a linial piture suggestive of MS, partiularly if lesions are present in sites typial of MS. In the young and the old, it is important to look for a history of risk fators for erebral mirovasular disease suh as smoking, hypertension, diabetes, migraine, or potential for ardiovasular emboli disease. The lesions an be puntate or pathy and are more ommon in the supratentorial white matter where they an be present at any site, though most often are subortial rather than periventriular. Infratentorial lesions are less frequent but an our for example, in the entral pons as an area of high signal. Confluent high signal hange is more ommon in the periventriular areas and the subortial U fibres are typially spared. Mirovasular ishaemi lesions are not seen in the spinal ord even with aging or vasular risk fators. Although not related to ishaemi risk fators, erebral autosomal dominant arteriopathy with subortial infarts and leuoenephalopathy (CADASIL) an produe onfluent subortial white matter high signal whih is usually symmetrial and typially involves temporal poles and external apsules. These appearanes are therefore not typial of MS and the linial piture of strokes, migraine, ognitive impairment, and a dominant family history will also distinguish this ondition. Other inflammatory and vasuliti disorders There are numerous disorders in this group that have MR appearanes that ould be mistaken for or even indistinguishable from MS. Saroidosis an affet the CNS produing multiple lesions affeting the ortex, subortial regions, ranial nerves, and rarely the spinal ord. The key distinguishing feature of this ondition from MS is the presene of meningeal enhanement whih is due to the granulomatous pathology. This finding highlights the importane of performing any repeat MR imaging with ontrast before lumbar punture as the latter an result in false positive meningeal enhanement on MRI. Systemi lupus erythematosus (SLE) an present similarly to MS and may also have multiple small T2 signal intensities in the white matter. The MRI appearane is more suggestive of SLE when the lesions are not onfined to periventriular white matter but favour the grey white interfae or even involve grey matter itself. The additional systemi linial features and blood tests will help to distinguish SLE from MS. Behçet s disease typially affets the dienephalon and brainstem but an also affet supratentorial white matter. Infetious disorders Two disorders in partiular whih are aused by an infetious agent an mimi MRI appearanes of MS. Progressive multifoal leuoenephalopathy (PML) is aused by reativation of latent JC polyomavirus in the immunosuppressed and is a multifoal demyelinating white matter disease. It auses onfluent white matter lesions on MRI whih do not ause mass effet or enhane. It is relentlessly progressive and usually fatal. iii15 J Neurol Neurosurg Psyhiatry: first published as /jnnp on 16 August Downloaded from on 4 September 2018 by guest. Proteted by opyright.

6 iii16 Neurologial manifestations of Lyme disease an our in assoiation with white matter lesions similar to MS, although the usual linial presentation is with a meningoradiulitis and/or ranial nerve involvement and an inflammatory CSF. The history of tik bite in an endemi area and serologial investigation will help to distinguish Lyme disease from MS. Neoplasti disorders MS an usually be easily distinguished from the appearanes of a tumour, but oasionally large MS lesions an mimi a neoplasti disorder suh as primary CNS lymphoma and even metastases or glioma. The linial ourse and follow up imaging usually larify the diagnosis, but biopsy may be required. Metaboli disorders The inherited leuodystrophies are a group of rare white matter disorders whih result in differing patterns of dysmyelination or demyelination. The MR appearanes differ from MS in that the white matter hanges are typially onfluent and fairly symmetrial without the disrete, disseminated, foal lesions seen in MS. Adult onset disease ours and an show regional hanges on MRI. Adrenomyeloneuropathy often shows hanges in the parietooipital lobe white matter, and Krabbe s disease (globoid ell leuodystrophy) an seletively affet the ortiospinal trats. Central pontine myelinolyis results in T2 high signal in the entral pontine fibres and ours following an aute metaboli disturbane, typially rapid orretion of hyponatraemia. Extra-pontine myelinolyis an also our. ROLE OF MRI IN THE DIAGNOSTIC CRITERIA FOR MS The linial diagnosis of MS is based on a history of at least two separate episodes of foal neurologial dysfuntion supported by evidene of at least two lesions within CNS white matter whih have been disseminated in spae and time. The first set of riteria formalising the diagnosis of MS by Poser et al 6 in 1983 were reated before the advent of MRI and the degree of diagnosti ertainty was identified by ategories ranging from a linially definite diagnosis to laboratory supported definite MS, linially probable MS, and laboratory supported probable MS. During the 1980s and 1990s, MRI beame inreasingly available and was integrated into the investigation and diagnosis of patients thought to have MS. MRI riteria for the diagnosis of MS were developed by Barkhof et al 7 in 1997 and were later modified by Tintoré et al 8 in The Barkhof riteria 7 were developed in order to improve the predition of a CIS onverting to linially definite MS on the basis of one MRI san. A umulative hane model with four riteria was used: (1) presene of at least one juxtaortial lesion; (2) at least one gadolinium enhaning lesion; (3) at least one infratentorial lesion; and (4) at least three periventriular lesions. These riteria were found to have a sensitivity of 82%, a speifiity of 78%, and an auray of 80%. Tintoré et al 8 modified these riteria by allowing a substitution of one enhaning lesion by nine T2 lesions whih was also preditive of onversion of a CIS to MS. In 2001, the International Panel on the Diagnosis of MS published new guidelines for the diagnosis of MS whih for the first time utilised MRI evidene of lesion dissemination in time and spae whih ould potentially allow the diagnosis of MS at an earlier stage. 9 These are ommonly known as the MDonald riteria and inorporate the two previous sets of MRI riteria by Barkhof 7 and Tintoré. 8 The MDonald riteria, 9 like the Poser riteria, 6 rely on objetive evidene of dissemination in spae and time but aept that speifi MRI abnormalities an provide evidene in patients with a CIS. The modified Barkhof-Tintoré riteria are used to provide evidene of dissemination in spae. If these imaging riteria are not fulfilled, two T2 lesions plus the presene of oligolonal bands an also provide evidene of dissemination in spae, but this may lead to a redution in auray of the diagnosis. The MRI riteria for the dissemination in time require the presene of a gadolinium enhaned lesion on an MRI san done at least three months after the CIS, or a new T2 lesion onfirmed on two MRI sans done at least three months after the linial episode. These riteria have been shown in natural history studies to have a high speifiity in prediting at three months and one year whether a patient with a CIS will develop linially definite MS within three years. 10 If a new T2 lesion is allowed as evidene for dissemination in time at three months, speifiity is maintained and sensitivity is inreased. The MDonald riteria were also applied to the plaebo arm of a CIS treatment trial (ETOMS trial) where it was found that the riteria for dissemination in spae were assoiated with an inreased likelihood of developing linially definite MS. Figure 4 provides an approah for diagnosing MS after presentation with a CIS. In brainstem CIS, the dissemination in spae riteria are less speifi than in other types of CIS in prediting onversion to MS beause the presene of the symptomati infratentorial lesion is not used to show dissemination in spae. The potential for spinal ord MRI to aid diagnosis of MS is of interest beause, as disussed earlier, the presene of asymptomati ord lesions is rare in other disorders, and abnormalities in the spinal ord are ommon in patients with a new diagnosis of MS and thus are useful in providing evidene of dissemination in spae. The MDonald riteria allow for the substitution of one brain lesion by one ord lesion. Diagnosti riteria for primary progressive multiple slerosis (PPMS) The linial presentation of primary progressive multiple slerosis (PPMS) makes the diagnosis more diffiult than relapsing-remitting MS beause the defining linial relapses are absent, making it harder to define dissemination in time and spae, and there are often fewer brain lesions with a greater disease burden in the spinal ord. Criteria for the diagnosis for PPMS are inluded in the MDonald riteria 9 and differ in that the presene of isolated oligolonal bands in the CSF is a requirement. PPMS an be diagnosed when there is additional evidene of dissemination in spae, using MRI or visual evoked potentials (VEPs), and in time, using MRI or progression of disability over one year. Revisions to diagnosti riteria As this artile goes to press, we are aware that the MDonald riteria are being revised; one published, the newly revised riteria will need to be onsidered in future diagnosti workup algorithms. J Neurol Neurosurg Psyhiatry: first published as /jnnp on 16 August Downloaded from on 4 September 2018 by guest. Proteted by opyright.

7 After diagnosis of a CIS, explain that MS is the ause for approximately 50 60% of people. A brief disussion of the nature of MS is appropriate, inluding its potential to be mild and non-disabling OPTS FOR MRI Explain that brain MRI an provide a more aurate predition of the likelihood of having MS: an abnormal san predits a 60 80% hane, whereas a normal san suggests a 20% hane. GIVE PATIENT THE OPTION OF HAVING A SCAN OR NOT T2 and gadolinium enhaned brain MRI should be performed MRI SCAN > 3 MONTHS FROM ATTACK? YES FULFILS MDONALD CRITERIA FOR DISSEMINATION IN SPACE AND TIME? YES MS DIAGNOSED NO NO YES MRI ASSESSMENT OF PROGNOSIS Patients presenting with a CIS will often ask about the likelihood of developing MS in the future and also about the development of disability. The latter question is also relevant in patients who have reently been diagnosed with relapsing remitting MS who have no signifiant disability. Several fators have a modest influene on prognosis inluding type of initial presentation and appearanes on MRI san. The likelihood of subsequent development of MS has been addressed in natural history studies of CIS. 10 Based on data DECLINES MRI Wait for a seond linial episode before diagnosing MS If MRI abnormal but the riteria for MS are not fulfilled, a repeat T2 brain san between 3 and 6 months after the original study has a signifiant prospet of providing MRI evidene for dissemination in spae and time providing that a new T2 lesion is allowed as evidene of dissemination in time NEW T2 OR ENHANCING LESION? NO CONSIDER EXAMINING CSF AND PERFORMING EVOKED POTENTIALS No benefit from further repeat MRI sanning as the disease beomes more likely to manifest linially Figure 4 Approah to the diagnosis of multiple slerosis (MS) in a patient with a typial presentation of a linially isolated syndrome (CIS). CSF, erebrospinal fluid; MRI, magneti resonane imaging. from these studies, approximately 50 60% of all CIS patients will go on to develop MS in the long term. If an MRI san at presentation is abnormal the likelihood of developing MS is inreased to 60 80%, and if the san is normal the risk is redued substantially to approximately 20%. Patients with opti neuritis seem to have a lower risk of onverting to MS in some studies where they have shown a lower frequeny of MRI lesion than other types of CIS. The predition of subsequent disability has been assessed by omparing MRI findings at different time points with the iii17 J Neurol Neurosurg Psyhiatry: first published as /jnnp on 16 August Downloaded from on 4 September 2018 by guest. Proteted by opyright.

8 iii18 expanded disability status sale (EDSS). A normal MRI appearane at presentation with CIS is assoiated with a lower level of disability many years later. The number of abnormal lesions at baseline has been modestly related to the severity of subsequent disability, and there are also data to suggest that the distribution of lesions may be important, with infratentorial lesions being better preditors of disability. In one CIS ohort, the final level of disability after 14 years was orrelated with the total lesion volume after five years and the inrease in lesion volume ourring over the first five years. Subsequent hanges in lesion volume were less informative. It has been onluded from these data that early lesion aumulation on MRI is related to future disability from MS, although the prognosti information is relatively weak. MONITORING DISEASE ACTIVITY IN CLINICAL TRIALS USING MRI As well as being a diagnosti tool, serial MRI sanning an be useful as an objetive measure to monitor lesion ativity, lesion load, and response to treatment with disease modifying treatments in patients with MS. T2 lesions and blak holes on T1 weighted imaging are used to measure lesion load, and gadolinium enhaning lesions measure new lesion ativity. The number and volume of lesions are the most ommonly used measures, but brain and spinal ord atrophy an also be used to monitor disease progression and may be more relevant when onsidering disability outome in the longer term. Exploratory studies of novel therapeuti agents normally have monthly sanning with gadolinium enhanement as the primary end point of treatment effiay beause linial relapses will be relatively infrequent and disability progression will be minimal during short follow up periods. In larger phase III trials, T2 lesion number, T2 lesion load and atrophy, as well as enhaning lesions, are used as seondary outome measures. T1 blak hole lesion load an also be used and may be a better marker for axonal loss seondary to lesions if only persistent blak holes are measured. Trials of patients with a CIS also employ MRI to selet individuals at high risk of onversion to MS for enrolment. The first trial to show a treatment effet using MRI outome measures in patients with MS was the interferon beta-1b trial whih found signifiant redutions in both new lesion ativity and total T2 lesion load. MRI outome measures have subsequently been inluded in all the major linial trials evaluating the effiay of disease modifying agents inluding trials of glatiramer aetate, interferon beta- 1a, and mitoxantrone whih is used in patients with aggressive disease. The most reent large trial was the AFFIRM study of natalizumab whih demonstrated substantial redutions in relapse rate, disability, and MRI outome measures whih were of a higher magnitude than those seen in previous trials of existing disease modifying treatment. Although onventional MRI outome measures are used extensively in linial trials, they must be used in onjuntion with linial and other para-linial measures as they have been shown to be only modestly orrelated with disease progression. It is possible that the relation between MRI hanges and linial status may be onfounded by the adaptive apability of unaffeted areas of the CNS. Disease progression may only our when pathology affets funtionally important regions for example, the spinal ord and when the extent of pathology exeeds a ritial threshold. The onventional MRI outome lesion measures only assess pathology in white matter affeted by lesions. The nononventional MRI tehniques disussed earlier have promise in assessing the pathologial hanges in normal appearing white and grey matter and ould be inorporated into future linial trials. Measures of brain atrophy are also now routinely inluded in treatment trials monitoring. An exiting potential area for future study is to use MRI to monitor remyelination in experimental trials. MTR is higher in remyelinated than demyelinated lesions and may be a good tool for use in suh trials. Another tehnique that shows promise in deteting myelin is the short T2 omponent of multi-exponential T2 relaxation. ACKNOWLEDGEMENTS We thank Drs Li, Fernando, Swanton, and Lanyon for their assistane with the reprodution of the MR images. The NMR Researh Unit is supported by The Multiple Slerosis Soiety of Great Britain and Northern Ireland.... Authors affiliations S A Trip, D H Miller, NMR Researh Unit, Department of Neuroinflammation, Institute of Neurology, University College London, London, UK REFERENCES 1 Pretorius PM, Quaghebeur G. The role of MRI in the diagnosis of MS. Clin Radiol 2003;58: A useful pitorial review with 35 MRI images whih inludes the differential radiologial diagnosis of MS. 2 Symms M, Jager HR, Shmierer K, et al. A review of strutural magneti resonane neuroimaging. J Neurol Neurosurg Psyhiatry 2004;75: From the Neurosiene for Neurologists series. Reviews nononventional MRI tehniques as applied to neurologial diseases inluding MS. 3 Miller D, Barkhof F, Montalban X, et al. Clinially isolated syndromes suggestive of multiple slerosis, part 2: non-onventional MRI, reovery proesses, and management. Lanet Neurol 2005;4: Part 2 of a omprehensive review of CIS inluding findings from nononventional MRI. 4 Lyklama G, Thompson A, Filippi M, et al. Spinal-ord MRI in multiple slerosis. Lanet Neurol 2003;2: A omprehensive review of all aspets of spinal ord imaging in MS. 5 Lennon VA, Wingerhuk DM, Kryzer TJ, et al. A serum autoantibody marker of neuromyelitis optia: distintion from multiple slerosis. Lanet 2004;364: Poser CM, Paty DM, Sheinberg L, et al. New diagnosti riteria for multiple slerosis: guidelines for researh protools. Ann Neurol 1983;13: Barkhof F, Filippi M, Miller DH, et al. Comparison of MRI riteria at first presentation to predit onversion to linially definite multiple slerosis. Brain 1997;120: Tintore M, Rovira A, Martinez MJ, et al. Isolated demyelinating syndromes: omparison of different MR imaging riteria to predit onversion to linially definite multiple slerosis. AJNR Am J Neuroradiol 2000;21: MDonald WI, Compston A, Edan G, et al. Reommended diagnosti riteria for multiple slerosis: guidelines from the international panel on the diagnosis of multiple slerosis. Ann Neurol 2001;50: The urrent diagnosti riteria for MS. 10 Miller D, Barkhof F, Montalban X, et al. Clinially isolated syndromes suggestive of multiple slerosis, part I: natural history, pathogenesis, diagnosis, and prognosis. Lanet Neurol 2005;4: Part 1 of a omprehensive review of CIS inluding the role of onventional MRI in diagnosis and prediting prognosis. J Neurol Neurosurg Psyhiatry: first published as /jnnp on 16 August Downloaded from on 4 September 2018 by guest. Proteted by opyright.