This material may be protected by Copyright law (Title 17 U.S. Code)
`
`Hopewell EX1019
`
`1
`
`

`

`383
`
`
`
`treatment and the occurrence of signs or symptoms, leading to an
`increase in EDSS(clinicalactivity).
`MRIwasperformed on a 0.6 Tesla system, using a standard head
`coil. A standard imaging protocol was employed. To ensure re-
`producibility (irrespective of the patient’s position in the imager),
`we used two consecutive scout images (coronal, sagittal), the ob-
`lique scanning facility and internal landmarks of the brain. Axial
`(double oblique) images were obtained through a plane determined
`by the inferior borderof the pituitary gland and the fastigium of the
`4th ventricle for angulation and the inferior border of the splenium
`of the corpus callosum as Z-centre. This gives an acceptably small
`angulation error on serial studies [10]. T2-weighted spin echo ima-
`ges
`(2755/60, 120(reduced bandwidth)/2)
`[TR/TE/excitations]
`were obtained. Gd-DTPA wasthen given IV (0.2 mmol/kg), fol-
`lowed by 10 cc saline. T1-weighted spin-echo sequences were ob-
`tained (400/28/4) starting 510 min after Gd injection,with in-plane
`resolution 1.0 x 1.3 mm, andslice thickness 5 mm (gap 1.25 mm).
`Thetotal investigation time was 45 min. Halfway through the study,
`the imaging protocol was adjusted, and Gd wasinjectedat the out-
`set, and Tl-weighted images were then-obtained before the T2-
`
`Fig. 1a-j. Sequential MR images of patient 7 (see also Fig.2a)
`showing adequate repositioning. a-f MR imagesat level of basal
`ganglia. a T2-weighted image at the time ofa clinical relapse just
`before 4th course of methylprednisolone (see Fig.2.a) shows multi-
`ple high signal periventricular
`lesions.
`b Corresponding T1-
`weighted image showsthat only 2 of these are active, enhancing with
`gadolinium (Gd). In all, 8 active lesions were detected on this ex-
`amination. ¢,d Images 4 weeks later, during remission after me-
`thylprednisolone treatment, showed complete suppression of con-
`trast enhancement, but no change on T2-weighted image. e,f 3 Gd-
`enhancing lesions are visible 9 weeks later, two of which (right
`frontal and left internal capsule) are teactivatedlesions (visible on
`prior T2-weighted images), while a new lesion has appearedin left
`parietal lobe, and is visible in e. Ten active lesions were detected.
`The increase in MRIactivity (“MRI relapse”) occurred while the
`patientstill was clinically stable, and about 10 weeks before a new
`clinical relapse.
`g-—j Images at level of centrum semiovale. T2-
`weighted image (g) obtained at the same time as a shows multiple,
`confluent high-signal lesions, none of which enhances (in b). Four
`weeks later (compare with e,d), the T2-weighted image(i) is un-
`changed, while a Ti-weighted image (j) reveals two new areas of
`enhancement;this activity would have been missed without Gd
`
`weightedseries. The latter protocol reduces the investigation time
`to 35 min, precludes changesin repositioning during injection of Gd
`and does not reduce lesion conspicuity on T2-weighted images[11].
`The images were analysed by two of the authors in conference.
`The number, not the size, of Gd-enhancing lesions was noted. All
`new lesions were assigned to one of seven anatomical regions. We
`assessed whether areas of enhancement were represented by ab-
`normalities on T2-weighted images. In addition we looked for new
`unenhancing lesions. On each follow-up” image, we analysed
`whetherlesions persisted or had ceased to enhance, new areas of
`enhancementhad occurred, new unenhancinglesions had appeared
`and whetherlesions changedin size on serial T2-weighted images.
`Although meticulous efforts were made to ensure reproducibility,
`
`2
`
`

`

`384
`
` Location Numberoflesions (%)
`
`Table 1. Distribution of (549 enhancing and 18 nonenhancing) new
`lesions
`
`Supratentorial
`Periventricular
`Cerebral cortex
`Junction cortex/white matter
`Deep white matter
`Frontal lobe
`Parietal lobe
`Occipital lobe
`Temporal lobe
`Internal capsula
`Basal ganglia
`Infratentorial
`11 (2%)
`Brainstem
`13 (2%)
`Cerebellum
`
`Total 567 (100 %)
`
`92 (16%)
`29 (5%)
`44 (8 %)
`
`213 (38%)
`73 (13 %)
`24 (4%)
`51 (9%)
`4(1%)
`13 (2%)
`
`slight changes in positioning inevitably occurred, due to hardware
`and patient limitations. As a consequence,
`lesions sometimes
`changedin apparentsize. Only marked changes in size were there-
`fore recorded; disappearance of lesions (beyond the resolution of
`the imager) was assumed whena lesion wasnotvisible on two con-
`secutive images.
`Not only enhancing lesions (new areas of enhancement, re-
`current enhancement in previously enhancing areas and persis-
`tently enhancing areas), but also new nonenhancinglesions, and
`enlarging lesions without enhancement were considered as lesions
`showingdisease activity, and called “active” lesions. The duration of
`effect of methylprednisolone on the developmentof active lesions
`on MRIwas defined as the period between the end of the therapy
`and an increase in the numberofactive lesions (MRIactivity).
`
`Results
`
`Clinical
`
`Four patients suffered no new relapse during the study
`and therefore were treated once only, at the beginning.
`All the others suffered clinical relapses, for which they
`received additional courses of methylprednisolone:
`4 patients one, 2 patients two, 2 patients three, and
`1 patient four. We thus gave a total of 31 courses of me-
`thylprednisolone. One patient was treated twice at the
`time of development of numerous new lesions on MRI,
`although she had minimal accompanying signs or symp-
`toms. When 29 courses of methylprednisolone treatment
`were given on clinical grounds,clinical improvement was
`observedafter 26 (90 %), with a decrease in EDSS of 0.5
`after 5 courses (18 %) and of 1.0 or more points with 21
`(72%).
`
`Magnetic resonance imaging
`
`Even with optimal repositioning,it is sometimesdifficult
`to detect all changes on the T2-weighted images, and Gd-
`enhancement greatly facilitates this. In some cases new
`activity would have been missed completely on sequen-
`tial T2-weighted images, especially when new areas of
`enhancement occurred within confluent abnormalities
`
`on T2-weighted images or when enhancement recurred
`(Fig. 1).
`We performed 195 MRI examinations (7-24 per pa-
`tient) and 549 new Gd-enhancinglesions were detected
`(enhancing lesions seen on the first examination were
`assumed to be new), 55 of which were not seen on the
`corresponding T2-weighted images (mean 41 per pa-
`tient; range, 1-105); in addition, 18 new nonenhancing
`lesions (3 % of all new lesions) were seen in 7/13 patients,
`and 4 enlarging lesions without enhancement (enlarge-
`ment occurred far more often, but was usually accom-
`panied by Gd-enhancement) were observed. Enhance-
`ment recurred in 38 lesions in 10/13 patients: in 12 it re-
`curred on the second or third MRI examination (1 or
`2 months after treatment), and in 26 it appeared in-
`dependentof treatment (Fig. 1). Thus, 609 active lesions
`were detected during the study, 93 (15 %) of which would
`have been overlooked without Gd-enhancement. Most
`of the new lesions (543, 96 %) were supratentorial. The
`majority of those (365, 64 %) lay in the deep white mat-
`ter; 15 % were related to grey matter (cortex,junction of
`cortex and white matter, or basal ganglia) (Table 1).
`Before the courses of treatment, 293 active lesions
`were found. A decrease in the numberof active lesions
`(Fig. 2) was seen after every course but the secondin pa-
`tient 12. After treatment 229 (78 %) were no longerac-
`tive, while 17 new nonenhancing lesions were detected
`and 1 lesion enlarged without enhancement; thus, 82 ac-
`tive lesions were seen after treatment. Withoutthe use of
`Gd, the persistent enhancementin 60 lesions would not
`have been appreciated; 71 % of persistent activity would
`not have been evident. The decrease in the numberof ac-
`tive lesions from before treatment ranged from 0-100 %
`for the individual courses. The percentage of decrease,
`per patient, ranged from 35 to 100%, mean 78 %. The
`mean interval between pre- and post-treatment MRI was
`22 days (SD 12 days). The last two courses in patient 11,
`given because of an increase in the numberof active le-
`sions, led to complete suppression of enhancement di-
`rectly after treatment, but one month later a return to the
`pretreatment numberof active lesions was observed.
`Of the 567 enhancing and nonenhancing new lesions
`103 (18 %) eventually disappeared on the T2-weighted
`images. Of the 357 lesions first seen directly before me-
`thylprednisolone treatment was initiated, 87 (24%)
`eventually disappeared. Because lesions which devel-
`oped before treatment had a longer follow-up period
`than lesions developing after treatment, lesionsfirst seen
`on the pretreatment MRI study presumably had the
`same disappearance rate as those seen to develop after
`treatment. Lesions which persisted on follow-up T2-
`weighted images frequently became confluent with pre-
`existing lesions, or could not be identified individually in
`the first place.
`
`Clinical and MRIcorrelation
`
`Figure 2 showsthe relation between the numberof active
`lesions and the EDSSasa function of timefor patients 7,
`8 and 9.
`
`3
`
`

`

`385
`
`
`
`
`
`a
`=
`MR imaging
`aisease
`activity
`
`Clinical
`Aisease
`activity
`
`
`
`
`So
`40
`30
`20
`10
`G
`weeks after end of therapy
`
`Fig.3. Kaplan-Meier curve shows percentage of patients free of
`disease activity on clinical and MRI groundsagainst time. The per-
`centageof patients free of new MRIdiseaseactivity decreases more
`rapidly than thatof patients free of newclinical diseaseactivity
`
`100
`
`80
`
`6a
`
`40
`
`20
`
`6
`
`U
`
`aq0 :
`
`% 40
`
`0
`y
`u
`
`0 ¥
`
`G
`0
`Pm
`
`u Q
`
`B
`
`oO
`3
`Las
`Ww
`oO
`2
`Ww
`
`10
`
`19
`
`78
`A
`6
`
`5
`
`4
`
`3 2 4
`
`0
`
`10
`
`9 8
`
`20
`
`15
`
`g
`P=
`6
`Et
`=
`°
`3
`i
`&
`Zz
`
`a
`
`Patient 7
`
`Months
`
`Patient 8
`
`a
`&
`8
`=
`3
`5
`i
`5
`
`
`
`
`Cc
`
`7
`
`6
`
`©
`8
`a
`5
`a
`wn
`sg
`
`32 1
`
`10
`
`9 8
`
`7
`6
`65
`
`d
`17)
`
`4
`
`3 2 1
`
`0
`
`6 and 13) neither clinical nor MRIactivity was detected
`before the end of the follow-up period. With most cour-
`ses (89%), however, the effect on MRIlasted for a
`shorter period than the clinical effect. After 11 courses
`
`Months
`new MRI activity was detected without any newclinical
`activity during follow-up, whereasafter the 12 courses in
`which any newclinical activity was detected, new MRI
`activity was detected first (Fig.2). When new MRIactiv-
`ity was detected, usually more than one new lesion (an
`“MRI relapse”) appeared; new incidental lesions were
`sometimesdetected slightly earlier (1 week on average)
`than such “MRI relapses”.
`Suppression of MRI activity by methylprednisolone
`treatment had a mean duration of 9.7 weeks, while the
`clinical effects lasted 17.1 weeks. Thus, MRI showed
`new activity on average 7.4 weeksearlier than newclin-
`ical activity was recorded; this is probably an under-
`estimate, as in all but two courses new MRIactivity was
`detected before the end of follow-up, while after most
`courses no newclinical activity was detected. In other
`words, with regard to new clinical activity, the observa-
`tion time had been censored. The actual duration of the
`clinical effect can be estimatedstatistically by the max-
`imum likelihood estimation for censored observations
`[12]. Assuming that the distribution of the duration of
`the clinical effect is approximately Gaussian,the actual
`average duration of the clinical effect can be estimated
`at 20.3 weeks, indicating that the MRI effect was on
`average 10.6 weeks shorter, A Kaplan-Meier curve
`(Fig. 3) shows that the percentage of patients free of dis-
`ease activity is clearly lower if assessed by MRI then by
`clinical examination.
`
`Death iit iiliieee 0
`
`Patient 9
`
`
`wa
`s
`:
`
`i
`=
`=|
`Zz
`
`Months
`
`Fig.2a-c. Comparison of clinical and MRI disease activity in
`3 patients. Relation between the number of active lesions (con-
`tinuouslines andleft Y-axis) and EDSSscore (dotted lines and right
`Y-axis) as a function of time (X-axis, stacks at monthly intervals).
`While the right Y-axis has a fixed scale, the scales ofthe left Y-axis
`and the X-axis vary. After most methylprednisolone courses MRI
`activity is suppressed, but than an increase in MRIactivity (“MR
`relapse”) occursin all cases, and far precedesclinical relapses,if any
`(see also Fig.3). Arrows, courses of methylprednisolone
`
`After two courses of methylprednisolone (the last
`courses for patients 7 and 8) no adequate follow-up was
`available. Only following one course (patient 10) was the
`clinical effect shorter than the MRI effect (although the
`MRimages were never completely negative after me-
`thylprednisolone), while after three courses (patients 2,
`
`4
`
`

`

`386
`
`Discussion
`
`Corticosteroids and most other drugs used to treat MS
`are employed because of their immunosuppressive ef-
`fect. As blood-brain barrier disruption in MSindicates
`active inflammation, Gd~enhancementis well suited to
`monitor immunosuppressive treatment, such as methy!-
`prednisolone.In our study “active”lesions also included
`new or enlarging nonenhancing lesions. Our rationale
`was that such lesions represent progression of the dis-
`ease, even in the absence of (detectable) blood-brain
`barrier disruption. By combining enhancing (96 %) and
`nonenhancing (4%) active lesions, every type of pro-
`gression of the disease on MRIis included.
`The beneficial short term clinical effect of methyl-
`prednisolone is well established [2-4], and the drug is
`widely used to shorten the duration and reduce the in-
`tensity of relapses. The effect is associated with an im-
`proved integrity of the blood-brain barrier [5-7, 13].
`There is a correlation between clinical improvement,
`suppression of Gd-enhancement and decrease in myelin
`breakdown products in the CSF [14]. It therefore may be
`assumed that restoration of the blood-brain barrier plays
`an important role in clinical improvement, by reducing
`cellular infiltration, oedema,the effects of cytokines or
`complement
`[7] or by reducing demyelination [14].
`Whether methylprednisolone has any long term bene-
`ficial effect on the clinical course is debatable [15-17]. We
`showed with MRIthat the effect of treatment is tempor-
`ary and probably will havelittle long term consequences.
`Although Gd-enhancement
`is
`transient,
`its dis-
`appearanceis accelerated by methylprednisolone [5—7].
`The suppression by methylprednisolone of Gd-enhance-
`ment seen prior to treatment seems to be effectively
`permanent, as enhancement recurred in only 12 of
`293 lesions (4%) active before treatment with the dose
`used. If a lower dose of is used (e.g. 1 g for 3 days) en-
`hancement may recur more often (39%) [7], although
`the timing of the treatment and MRI examinations in
`other studies differed from ours. Unfortunately, the
`rather effective suppression of Gd-enhancement does
`not lead to increased disappearance of related abormal-
`ities on T2-weighted images, when compared tothe dis-
`appearancerate of lesions developing after treatment or
`to the natural disappearancerate in studies of untreated
`patients [10, 18]. Thus, although inflammatiom is sup-
`pressed, the vast majority of lesions cause persistent T2-
`abnormalities. Apparently, the temporary openingof the
`blood-brain barrier has already been accompanied by
`considerable tissue changes. Alternatively, the suppres-
`sion of Gd-enhancement may bethe result of reduced
`cytokine production by inflammatorycells, while the in-
`flammatory process inside the blood-brain barrier con-
`tinues.
`The most important indication that there is no long
`term effect of methylprednisoloneis that an increase in
`new lesions (“MRI relapse”) occurs at some stage in
`most patients after cessation of therapy. Given sufficient
`follow-up, new MRIactivity preceded new,if any, clinical
`activity after all courses but one. MRI showed an average
`reduction in duration of effect of 52 % comparedtoclin-
`
`ical criteria. The reduction in the percentage of patients
`free from disease activity on MRI as comparedto clinical
`assessment is shown in Fig.3. One could argue that the
`effect of methylprednisolone could be prolonged by
`adding a tapering oral dose, or by prophylactic monthly
`intravenouspulse doses[19, 20].
`Whatis the clinical implication of new clinically silent
`brain lesions on MRI? First, disability shows a varying
`but usually positive correlation with overall extent of
`MRIabnormalities [21]. A closer correlation can be ob-
`served for individual symptomsscores with specific brain
`areas [22]. Secondly, follow-up revealed a significantly
`higher number of new MRIlesionsin patients with clin-
`ical relapses than in those without [23]. Apparently, not
`all new lesions increase clinical disability, as they are at
`nonstrategic sites (but do correlate with neuropsycholo-
`gical impairment[24]). The EDSS, for example, evalu-
`ates mainly motor function, which probably correlates
`best with the presence of spinal lesions. Thirdly, post-
`mortem MRIfindings correlate well with pathological
`findings, indicating that brain lesions on MRI do re-
`present the demyelinated plaques characteristic of MS
`[25, 26], although Lumdsden [27] notes a poor correla-
`tion of pathological findings with the clinical picture.
`Thus, new brain lesions on MRIindicate progression of
`the disease, although their exact relation to clinical dis-
`ability remains to be elucidated. Our finding of MRIre-
`lapses relatively soon after cessation of methylpredniso-
`lone treatment, indicates clinically silent, but never-
`theless significant progression of the disease.
`Serial MRI studies frequently detect clinically silent
`disease activity in patients with relapsing-remitting MS
`[1]. Therefore, the criteria in this study for the duration
`of effect on MRI have beenratherstrict: it was defined
`as the disease-free period before an increase in the
`number of active lesions
`(“MRI
`relapse”). After
`7 courses, however, the MR images were never without
`active lesions, although fewer lesions were seen than
`before treatment. If any new clinical activity was de-
`tected it was usually preceded by an increase in MRI
`activity. Thelatter apparently leads to clinical worsening
`only if enough active lesions are present, orif active le-
`sions develop in strategically important sites (internal
`capsule, brainstem, cerebellum, or spinal cord). These
`observations underline the value of serial MRI in asses-
`sing disease activity.
`Mostrelapses are self-limiting in relapsing-remitting
`MS;after an exacerbation, a remission phasewill set in by
`definition. Perhaps methylprednisolone is even less ef-
`fective then we believe; in two courses given solely be-
`cause of an increase in enhancing lesions its effect was
`extremely short. If we accept that MRIactivity and clin-
`ical activity are comparable measures, this observation
`showsthat disease activity is actually tempered for only
`a very short time, and that clinical improvement in most
`cases will be the result of (acceleration of) the natural
`course. It would be of particular interest to study in a
`placebo-controlled fashion whether steroids, given not
`on clinical grounds, but on MRIcriteria, could alter the
`evolution of the MRI pattern and subsequently the clin-
`ical course.
`
`5
`
`

`

`Acknowledgements. This work was supported by grants from the
`“Praeventiefonds” (28-1453), and the Nolet foundation. We are
`grateful to Ton Schweigmann for excellent technical support.
`
`References
`
`1. Miller DH, Barkhof F, Berry I, Kappos L, Scotti G, Thompson
`AJ (1991) Magnetic resonance imaging in monitoring the treat-
`ment of multiple sclerosis: Concerted Action guidelines. J Neu-
`rol Neurosurg Psychiatry 54: 683-688
`. Thompson AJ, Kennard C, Swash M etal (1989) Relative effi-
`cacy of intravenous methylprednisolone and ACTH in the
`treatmentof acute relapse in MS. Neurology 39: 969-971
`. Milligan NM, Newcombe R, Compston DAS(1987) A double-
`blind controlled trial of high dose methylprednisolone in pa-
`tients with multiple sclerosis: 1. Clinical effects. J Neurol Neu-
`rosurg Psychiatry 50: 511-516
`. Durelli L, Cocito D, Riccio A, Barile C, Bergamasco B, Baggio
`CFet al (1986) High-dose intravenous methylprednisolone in
`the treatment of multiple sclerosis: clinical-immunological cor-
`relations. Neurology 36: 238-243
`. Barkhof K HommesOR,ScheltensP, Valk J (1991) Quantitative
`MRIchanges in gadolinium-DTPA enhancementafter high-
`dose intravenous methylprednisolone treatment in multiple
`sclerosis. Neurology 41: 1219-1222
`. Burnham JA, Wright RR, Dreisbach J, Murray RS (1991) The
`effect of high-dose steroids on MRI gadolinium enhancement in
`acute demyelinating lesions. Neurology 41: 1349-1354
`. Miller DH, Thompson AJ, Morrissey SP et al (1992) High dose
`steroidsin acute relapses of multiple sclerosis: MRI evidence for
`a possible mechanism of therapeutic effect. J Neurol Neurosurg
`Psychiatry 55: 450-453
`. Poser CM, Paty DW,Scheinberg L et al (1983) New diagnostic
`criteria for multiple sclerosis: guidelines for research protocols.
`Ann Neurol 13: 227-231
`. Kurtzke JF (1983) Rating neurological impairment in multiple
`sclerosis: an expandeddisability status scale (EDSS). Neurology
`33: 1444-1452
`10. Barkhof F, Scheltens P, Frequin STFM,Nauta JJP, Tas MW,Valk
`J, Hommes OR (1992) Relapsing-remitting multiple scleroris:
`sequential enhanced MR imaging vs clinical findings in de-
`termining disease activity. AJR 159: 1041-1047
`BarkhofF, Valk J, Hommes OR,Scheltens P, Nauta JIP (1992)
`Gadolinium enhancement of multiple sclerosis lesions on long
`TR imagesat 0.6 T. AJNR 13: 1257-1259
`12. Cox DR, Oaks D (1984) Analysis of survival data. Chapman and
`Hall, London, pp 33-34
`
`11.
`
`387
`
`14.
`
`15
`
`16.
`
`Li
`
`18.
`
`19.
`
`20.
`
`21.
`
`22.
`
`13.
`Troiano R, Hafstein M, Ruderman M, Dowling P, Cook S (1984)
`' Effect of high-dose intravenous steroid administration on con-
`trast-enhancing computed tomographic scan lesions in multiple
`sclerosis. Ann Neurol 15: 257-263
`Barkhof F, Frequin STFM, HommesORetal (1992) A correla-
`tive triad of gadolinium-DTPA MRI, EDSS, and CSF-MBPin
`relapsing/remitting multiple sclerosis patients treated with high-
`dose intravenous methylprednisolone. Neurology 42: 63-67
`. Menken M (1989) Consensus and controversy in neurologic
`practice. The case of steroid treatment in multiple sclerosis.
`Arch Neurol46; 322
`Goodin DS (1991) The use of immunosuppressive agents in the
`treatmentof multiple sclerosis. Neurology 41: 980-985
`Compston A (1988) Methylprednisolone and multiple sclerosis.
`Arch Neurol 45: 669-670
`Thompson AJ, Miller DH, YoulB et al (1992) Serial gadolintum
`enhanced MRIin relapsing remitting multiple sclerosis of vary-
`ing disease duration. Neurology 42: 60-63
`Whitham RH, Bourdette DN (1989) Treatment of multiple
`sclerosis with high-dose methylprednisolone pulse therapy.
`Neurology 39 (S 1, abstract): 357
`Polman CH,van der Wiel HE, Teule GJJ, Koetsier JC (1991) A
`commentary on steroid treatment in multiple sclerosis. Arch
`Neurol 48: 1011-1012
`Wallace CJ, Seland TP, Fong TC (1992) Multiple sclerosis: the
`impact of MR imaging. AJR 158: 849-857
`Truyen L, GheuensJ, van de Vyver FL, Parizel PM, Peersman
`GV, Martin JJ (1990) Improved correlation of magnetic re-
`sonance imaging (MRI) with clinical status in multiple sclerosis
`by use of an extensive standardized imaging protocol. J Neurol
`Sci 96: 173-182
`Truyen L, GheuensJ, Parizel PM, van de Vyver FL, Martin JJ
`(1991) Long term follow-up of multiple sclerosis by standar-
`dized, non-contrast-enhanced magnetic resonance imaging. J
`Neurol Sci 106: 35-40
`Rao SM, Leo GJ, Haughton VM,St Aubin-Fleubert P, Bernar-
`din L (1989) Correlation of magnetic resonance imaging with
`neuropsychologicaltesting in multiple sclerosis. Neurology 32:
`161-166
`Nagara H,Inoue T, Koga T, KitaguchiT,Tateishi J, Goto I (1987)
`Formalin fixed brains are useful for magnetic resonance imaging
`(MRI) study. J Neurol Sci 81: 67-77
`Stewart WA, Hall LD, Berry K, Paty DW (1984) Correlation
`between NMRscan and brain slice data in multiple sclerosis.
`Lancetii: 412
`Lumsden CE (1970) The neuropathology of multiple sclerosis.
`In: Vinken PJ, Bruyn GW (eds). Handbookofclinical neurology,
`vol9. North-Holland Publishing Company, Amsterdam,
`pp 217-309
`
`23.
`
`24.
`
`25.
`
`26.
`
`aT:
`
`6
`
`