J7ournal ofNeurology, Neurosurgery, and Psychiatry 1994;57:51-57

Magnetic resonance imaging in degenerativeataxic disorders

I E C Ormerod, A E Harding, D H Miller, G Johnson, D MacManus,E P G H du Boulay, B E Kendall, I F Moseley, W I McDonald

Institute ofNeurology,London, UKNMR Research GroupI E C OrmerodD H MillerG JohnsonD MacManusW I McDonaldUniversityDepartment ofClinical NeurologyI E C OrmerodA E HardingD H MillerG JohnsonD MacManusW I McDonaldDepartment ofNeuroradiology,National Hospital forNeurology andNeurosurgery,London, UKE P G H du BoulayB E KendallI F MoseleyCorrespondence to:Professor A E Harding,University Department ofClinical Neurology, Instituteof Neurology, QueenSquare, London WC1N3BG, UK.Received 4 December 1992and in revised form16 March 1993.Accepted 15 April 1993

AbstractMRI of the brain was performed in 53patients with a variety of degenerativeataxias and related disorders and 96 con-trol subjects. Atrophy of intracranialstructures was not seen in patients withthe pure type of hereditary spastic para-plegia, or in early cases of Friedreich'sataxia. In advanced Friedreich's ataxiathere was atrophy of the vermis andmedulla. The MRI features of early onsetcerebellar ataxia with retained reflexeswere variable, and suggest heterogeneity.In autosomal dominant cerebellar atax-ias, most patients had cerebellar andbrainstem atrophy, probably reflectingthe pathological process of olivoponto-cerebellar atrophy; there was no clearlydefined group with both clinical andimaging features of isolated cerebellarinvolvement. The MRI abnormalities inidiopathic late onset cerebellar ataxiawere predoinantly those of cerebellarand brainstem atrophy or pure cerebel-lar atrophy. The clinical and imagigfeatures ofbrainstem abnormalities werediscordant in several patients. Pure cere-bellar atrophy was associated with slowerprogression of disability. Cerebral atro-phy was common in the late onset atax-ias. Cerebral white matter lesions,although usually few in number, wereobserved in significantly more patientsthan controls, particularly those agedover 50 years.

(7 Neurol Neurosurg Psychiatry 1994;57:5 1-57)

The degenerative ataxias are a heterogeneousgroup of disorders, many of which are geneti-cally determined. Their clinical features arediverse, with variable degrees of cerebralhemisphere, brainstem, spinal cord, andperipheral nerve dysfunction in addition tocerebellar ataxia. Classification of these dis-

orders is difficult; early attempts were basedexclusively on pathological findings, but it hasbeen suggested that disease categories can bemore usefully defined using clinical andgenetic criteria.' Autopsy data are relativelyscarce and a reliable method for definition ofinvolved structures during life could con-tribute to classification and more precisediagnosis. Computerised tomography has alimited role to play in this respect, as imagesof posterior fossa structures are usually of rel-atively poor quality compared with those ofthe cerebral hemispheres. Magnetic reso-nance imaging provides superior images ofthe brainstem and cerebellum,26 and hasbeen used to study small series of patientswith degenerative ataxic disorders with usefuldefinition of the distribution of atrophy in thebrain.79 In larger series it has been suggestedthat MRI offers a useful adjunct to clinicalfeatures for diagnostic and prognostic pur-poses.'01' We report on 53 patients with avariety of degenerative ataxias studied byMRI, comparing the findings in different dis-ease groups, and also with those seen in 96control subjects.

SubjectsCONTROLSMRI was performed in 96 control subjectsaged 18-73 years (table 1). Sixty-four wereaged less than 50 years and 32 were 50 yearsor over. Seventy-six of these were normal vol-unteers from the Salvation Army or from thestaff of the National Hospital who were notexamined clinically but who had no previousneurological history. The remainder (20)were neurological control subjects who wereinvestigated at the National Hospital for dis-orders of peripheral nerves or the spinal cordthat are not associated with brain pathology.These patients had no signs attributable toneurological dysfunction above the foramenmagnum.

Table 1 Control subjects and patients studied

Age range Disease duration rangeDisease category Number (years) (mean) (years) (mean)Controls 96 18-73 (41) N/AFriedreich's ataxia 6 13-38 (23) 9-25 (15)Early onset ataxia/retained reflexes 6 17-42 (28) 1-27 (14)Autosomal dominant cerebellar ataxia 14 22-66 (49) 2-25 (16)Idiopathic late onset cerebellar ataia

with other features 10 39-73 (61) 3-14 (7)pure 10 38-64 (51) 2-30 (10)

Hereditary spastic paraplegia 7 13-39 (27) 2-33 (14)

PATIENTSOf the 53 patients with ataxia (table 1), 30were aged under 50 years and 23 were aged50 years or more. They were subdivided intosix categories on clinical grounds (table 1).There were six patients with Friedreich'sataxia"2 and six with early onset cerebellarataxia with retained reflexes. 13 Fourteenpatients had autosomal dominant cerebellarataxia (ADCA; table 2); seven had additionalfeatures such as supranuclear ophthalmople-gia, pseudobulbar palsy, and mild dementia

51

on March 26, 2020 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.57.1.51 on 1 January 1994. D

ownloaded from

Ormerod, Harding, Miller, J'ohnson, MacManus, du Boulay, Kendall, Moseley, McDonald

Table 2 Clinical and imagingfeatures ofpatients with autosomal dominant cerebellar ataxia

AtrophyAge Duration

No. (years) (years) Type Features CH Vermis Medulla Pons Midbrain Cerebrum WML

1 55 11 I B,N - + + - - + + +2 48 6 I SNO + + + + ++ -3 61 20 I SNO,D - + + - - - - +4* 61 15 I SNO +++ + ++ ++ + + + + + + + +5* 60 25 I SNO + + + + ++ ++ ++- + + +6 34 6 I SNO,D,B + + + + + + + ++ - - -7 51 7 I D ++ +++ +++++ ++ ++ -8 29 3 II R - + + + + + + + + + ++ -9 56 20 II R + + + + + + + + + + + + ++ _10 66 3 III - + + + +++ - + + + +11 61 5 III - + + + + + + + - - +12 34 8 D,My,N - - - - - + + +13 53 25 withET - - ++ - - - ++ -14 20 2 withdeafness - +++ - + -

*Siblings.B = bulbar dysfunction; SNO = supranuclear ophthalmoplegia; D = dementia; N = neuropathy; R = retinopathy; M = myoclonus; ET = essential type of tremor;CH = cerebellar hemispheres; WML = white matter lesions.+, ++, +++ = mild, moderate, and severe atrophy, respectively.

(ADCA type I), two had maculopathy(ADCA type II), three a later onset pure cere-bellar syndrome (ADCA type III),14 and threemore unusual dominant ataxias, including asyndrome of ataxia, dementia, andmyoclonus exhibiting paternal transmission.Ten patients had idiopathic late onset cere-bellar ataxia (table 3)15 with other clinical fea-tures such as supranuclear ophthalmoplegia,peripheral neuropathy, mild dementia, opticatrophy, and parkinsonism (ILOCA/O). Allpresented with ataxia and this remained thepredominant feature. Only one had clinical orinvestigative evidence of autonomic failureand thus fulfilled the criteria for a diagnosis ofmultiple system atrophy'6. Ten had pureILOCA (ILOCA/P)-that is, no neurologicaldysfunction other than a cerebellar syn-drome'5. The last group consisted of sevenpatients with hereditary spastic paraplegia,either with the pure form (five cases) or withadditional clinical features.'7

Each patient was examined by one of theauthors (IECO or AEH). The case notes ofall patients were individually reviewed byAEH for the purpose of diagnostic classifica-tion, which was made according to previously

published criteria.' For patients with ILOCA,disability was assessed on a four point scale:(1) mild ataxia, able to work; (2) unable towork, able to walk and perform activities ofdaily living; (3) as (2), but unable to walkunassisted or chairbound; (4) chairbound,dependent on others for activities of dailyliving. A severity score related to diseaseduration was obtained by dividing disabilityscore by disease duration in years and multi-plying by 100.

MethodsMAGNETIC RESONANCE IMAGINGAll subjects were examined on a Picker 0 5 TMR imaging system. Multi-slice, contiguous5 mm thick axial T2-weighted spin echo (SE2000/60) images were taken throughout thebrain, to optimise the detection of white mat-ter lesions. Axial inversion recovery (IR2000/40/500) images were performed in mostsubjects to facilitate assessment of cerebralatrophy. Tl-weighted (IR 2000/500/40 or SE500/40) sagittal images were obtained in mostcontrol subjects and all patients, to assess thedegree of atrophy of the posterior fossa struc-tures.

Age Duration DisabilityNo. (years) (years) score

Atrophy

Features CH Vermis Medulla Pons Midbrain Cerebrum WML

P,ADD,SNODOApSNONNN

++

++

++

++

++

_ +++

_ +++

_ +++

- ++

++

+++

+++

++

++ - + ++

+ +

+++++ +++ + ++

_-

_-

+ +- ++

++ ++++ ++

++ ++

- ++

+

++Abbreviations as in table 2, plus P = parkinsonism; A = autonomic failure; OA = optic atrophy. ILOCA(O) = idiopathic late onset cerebellar ataxia with other

neurological dysfunctions; ILOCA(P) = ataxia alone.

Table 3 Clinical and imagingfeatures of idiopathic late onset cerebellar ataxia

ILOCA (0)15 6216 6617 5718 6819 7320 6421 6422 6223 3924 54ILOCA (P)25 4126 3827 6428 6229 5630 3931 4132 6033 5634 56

6 3314 214 504 508 253 676 337 289 229 22

3 332 1004 11

30 672 1508 125 22

11 276 50

27 11

52 on M

arch 26, 2020 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.57.1.51 on 1 January 1994. Dow

nloaded from

MRI in degenerative ataxic disorders

Table 4 Frequency of cerebral white matter lesions

Age (years) No. No. with lesions (5I)

Control subjects18-49 63 3(5)50-59 19 4(21)60-69 10 4(40)70-79 2 2(100)

Category (no.) No. with lesions No. aged < 50 with lesions

PatientsFriedreich's ataxia (6) 0 0Early onset cerebellar ataxia with retained

reflexes (6) 1 1Autosomal dominant cerebellar ataxia (14) 7 1/5Idiopathic late onset cerebellar ataxia with

other neurological dysfunction (10) 7 0Idiopathic late onset cerebellar atasia

alone (10) 2 0Hereditary spastic paraplegia (7) 3 3

ASSESSMENT OF MIfThe MR images of both controls and patientswere intermingled and examined for the pres-ence of atrophy. Atrophy was qualitativelygraded on a four point scale (0 = no atrophy,1 = mild, 2 = moderate and 3 = severe) bytwo experienced neuroradiologists (BEK andIFM) who reported by consensus withoutprior Inowledge of the diagnosis. Atrophy ofcerebellar hemispheres, cerebellar vermis,medulla, pons, midbrain, and cerebral hemi-spheres was rated in this way for each patient.Cerebral white matter lesions were defined asareas of hyperintensity on T2-weightedimages of at least three pixels, excluding smallcaps at the frontal or occipital poles.

ResultsMRI IN NORMAL CONTROL SUBJECTSMild atrophy of the brainstem was found insix subjects and of these three were aged over50 years. In two subjects more than one partof the brainstem was considered atrophic.Mild atrophy of the cerebellar hemisphereswas seen in four subjects and three were agedunder 50 years. The vermis was atrophic in10 subjects, all of whom were over 50 years;in nine of these there was also atrophy of thecerebral hemispheres. The cerebral hemi-spheres were atrophic in 42 control subjects;the degree of atrophy was mild in 35 andmoderate in seven cases. Nineteen of 35 with

mild atrophy were over the age of 50 years,and 10 of these also had white matter lesionsin the cerebral hemispheres. Five of sevenwith moderate atrophy were over the age of50 years and four of these had cerebral whitematter lesions.

Lesions within the cerebral white matterwere seen in three subjects (5%) aged lessthan 50 years and in 10 (30%) subjects aged50 years or over (table 4). This difference washighly significant X2t p < 0-001). In the threesubjects in the younger group there were onlyminor lesions within the cerebral white matterin two and smooth periventricular signalchange in one. Of the 10 older subjects withcerebral lesions, four were aged 50-59 years;three of these showed minor changes but onehad extensive periventricular and white mat-ter lesions. Six of the subjects aged 60 yearsor over had white matter abnormalities.These were minor in four (with some smoothperiventricular change and a few small scat-tered lesions in the cerebral white matter) andtwo had more extensive abnormalities.

MMI IN PATENTSOnly two of the patients with hereditary spas-tic paraplegia, both with additional cerebellarsigns and aged less than 50 years, had mild ormoderate atrophy of intracranial structures,the brainstem, cerebellum, and cerebrum inone, and the vermis and cerebrum in theother. Both of these patients had white matterlesions, as did another with the pure disease.MRI was normal in the four other patientswith pure disease.MRI was normal in two patients with

Friedreich's ataxia (fig la). Four of the sixpatients, all with disease duration greaterthan 15 years, had moderate medullary atro-phy which was associated with lesser changesin the pons and midbrain. Only one of thesepatients had marked cerebellar hemisphereatrophy (fig ib) and two had moderate orsevere atrophy of the vermis. None of thepatients with Friedreich's ataxia had cerebralatrophy or white matter lesions.The MRI features of early onset cerebellar

ataxia were variable. MRI was normal in onepatient with recent onset (within two years).Only one had severe cerebellar hemisphere

Figure 1 Ti weighted(SE 500/40) sagital MRIof two patients withFriedreich's ataxia.(A) a 14 year old boy withsymptoms ofseven years'duration. MRI is normal;(B) a 38year old womanwho had had symptoms for25 years. MRI showsatrophy of the cerebellum,brainstem, and uppercervical cord.

53 on M

arch 26, 2020 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.57.1.51 on 1 January 1994. Dow

nloaded from

Ormerod, Harding, Miller, Johnson, MacManus, du Boulay, Kendall, Moseley, McDonald

Figure 2 Tl-weighted(SE 500140) sagittalMRIof a 41 year old womanwith pure idiopathic lateonset cerebellar ataxia(case 25) showing markedatrophy of the cerebellarhemispheres (A) andvermis (B). The brainstemis normal.

atrophy, associated with mild atrophy of thevermis, pons and cerebral hemispheres. Twopatients had mild or moderate vermis atro-phy, one with mild cerebral atrophy. One hadatrophy of the whole of the brainstem, andthe sixth had vermis, brainstem, and cerebralatrophy with scanty periventricular whitematter lesions.Of the 14 patients with ADCA, eight had

moderate or severe cerebellar hemisphereatrophy, and all but one atrophy of the vermis(table 2). Five of seven of those with ADCAtype I had brainstem atrophy, predominantlyinvolving the pons, although in one it wasconfined to the medulla; six of these patientshad clinical evidence of brainstem dysfunc-tion. Brainstem involvement was seen in bothpatients with ADCA type II, and both withtype III, a pure cerebellar syndrome, althoughdisease duration was relatively short in thelatter.Marked cerebellar hemisphere and vermis

atrophy was frequent in patients withILOCA/O, a disorder clinically similar toADCA type I, but almost universal in thepure type of ILOCA (fig 2). Seven of the 10patients with ILOCA/O had brainstem atro-phy, again maximally involving the pons; intwo this was confined to the midbrain ormedulla. Two of the patients with supranu-clear ophthalmoplegia had normal brainstemimages. Two patients with ILOCA/O hadextrapyramidal features and both of these hadmoderate or severe pontine atrophy.Brainstem atrophy without clinical correlatewas seen in five of the 10 patients withILOCA/P, including all those with diseasedurations of less than five years (fig 3).Disability score was significantly higher in the11 patients with brainstem atrophy on MRI(excluding case 18 who only had mildmedullary atrophy, p < 0-01, Wilcoxon's ranksum test). A major contribution to this resultcame from four patients, two with a purecerebellar syndrome, clinically and radiologi-cally, who had a very benign course (cases 28and 34), and two with a rapidly progressivepure cerebellar syndrome of short durationwho had striking brainstem atrophy on MRI(cases 26 (fig 3) and 29).

Cerebral hemisphere atrophy was presentin 50 per cent or more of patients in thesethree groups of late onset ataxic disorders,being least common in ILOCA/P. Dementiawas observed in three patients with ADCA;one had gross cerebral atrophy without whitematter lesions, one had periventricular andwhite matter lesions but no significant cere-bral atrophy, and in the third the cerebralhemispheres appeared normal. Three patientswith ILOCA/O were demented. All had cere-bral atrophy and two had white matterlesions.

There was no significant effect (Kruskall-Wallis analysis of variance) of disease dura-tion in any of the seven diagnostic groupswhich might explain the distribution or sever-ity of atrophy in these patients. Moderate orsevere cerebral, vermis, and cerebellar atro-phy was more frequent (p < 0 003, p < 0x001and p < 0x002, respectively) in patients agedover 50 years than in patients under 50 years,but all the patients with early onset ataxia andFriedreich's ataxia were in the younger group.A significant difference was found betweenthe frequency of lesions within the cerebralhemispheres (fig 4), either periventricular ordiscrete within the cerebral white matter, in

Figure 3 Tl-weighted (SE 500/40) sagittalMRI ofa 59year old woman with pure idiopathic late onset cerebellarataxia of two years' duration (case 29). There is markedatrophy of the cerebellum and brainstem.

54 on M

arch 26, 2020 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.57.1.51 on 1 January 1994. Dow

nloaded from

MRI in degenerative ataxic disorders

the seven groups (p < 0 05, analysis of vari-ance; table 4). As stated, these were uncom-mon in patients with Friedreich's ataxia, earlyonset ataxia, and spastic paraplegia. In thediagnostic categories which included moreolder patients (ADCA, ILOCA/O andILOCA/P), the lesions occurred mainly inpatients over the age of 50 years. The effectof age on the frequency of the lesions wastherefore analysed and comparison was madebetween the patients and the normal controlsubjects. Significantly more patients over theage of 50 years had lesions than youngerpatients (2, p < 0 005). More patients hadlesions than controls, and this difference wasmore marked for patients over the age of 50years (e p < 003) than for those aged lessthan 50 years (X2, p < 0 05). When the fre-quency of lesions was analysed with respectto disease duration, no significant associationwas found (Kruskall-Wallis one way analysisof variance). Only marked atrophy of thecerebral hemispheres was associated with thepresence of lesions (X;2, p < 0-02). Whenthe groups were subdivided further to com-pare the effect of age (<50 years and >50years) and atrophy of the brain on thepresence of lesions, no significant differenceswere found.The cerebrospinal fluid had been examined

in 26 patients but none, of the control sub-jects. One patient had oligoclonalimmunoglobulin. This was a 60-year-oldwoman with an 11 year history of a progres-sive ataxic syndrome which was clinicallyclassified as a pure type of ILOCA. The MRIdemonstrated some atrophy of the cerebellarhemispheres and a more marked atrophy ofthe cerebellar vermis but there were nolesions in the cerebral hemispheres. Twopatients, with ADCA and a pure type of spas-tic paraplegia respectively, showed elevatedcerebrospinal immunoglobulin:albumin ratiosat 12% and 18%.

DiscussionIn this series of patients with degenerativeataxias, the presence of atrophy of brainstemand cerebellar structures on MRI scans wasgenerally in accord with previous pathologicalobservations. The normal findings in purehereditary spastic paraplegia reflect the factthat degenerative changes in this disease areusually confined to the spinal cord.'8 InFriedreich's ataxia, necropsy studies haveshown marked atrophy of the spinal cord butthe brainstem, cerebellum and cerebrum areusually normal, apart from some atropy of themedulla and degenerative changes in thecerebellar peduncles.'9 CT studies havereported mild atrophy of the cerebrum andmore marked vermis atrophy in some, usuallyadvanced, cases.20 Some of our patients, whohad had the disease for a long time, had atro-phy of the vermis, medulla, pons, or midbrainwith changes in the medulla being most fre-quent and striking. Klockgether and col-leagues" reported essentially similar MRIfindings to those described here in

Friedreich's ataxia; they also describedprominent atrophy of the cervical cord whichwe did not assess in detail.

These authors compared the imaging find-ings in Friedreich's ataxia with those in earlyonset cerebellar ataxia with retained reflexes,and found a high incidence of cerebellarhemisphere and vermis atrophy in the latter,with less prominent brainstem atrophy. TheMRI features in our cases of this syndromewere rather variable. Neuropathologicalstudies in such patients are extremely scarce,showing olivopontocerebellar atrophy (OPCA)in two cases.2'22 The MRI features reportedhere and by others" do not suggest OPCA inmost cases, and indicate that this syndrome ispathologically and genetically heterogeneous.

Pathological heterogeneity is well recog-nized in ADCA. Most, but not all, patientswith the clinical features of ADCA types Iand II have OPCA at necropsy. There issome variation in pathological findings withinlarge families.' Six of our nine patients inthese two categories had MRI findings sug-gesting OPCA. The same applied to two ofthree with ADCA type III, a later onset purecerebellar syndrome in which one patho-logical report has shown cerebello-olivaryatrophy.23 Thus brainstem involvement is byno means always manifest clinically, regard-less of disease duration. Most patients withclear clinical evidence of brainstem dysfunc-tion showed brainstem atrophy on MRI. Fiveof six patients with ADCA type I andsupranuclear ophthalmoplegia had significantatrophy of the pons or midbrain, or both. Inthese patients with ADCA, there was noclearly defined group with both clinical andimaging features of isolated cerebellarinvolvement.ILOCA has been divided on pathological

grounds into OPCA and cerebello-olivaryatrophy, the latter with sparing of the brain-stem apart from the inferior olives. Atrophyof the cerebellar hemispheres is seen in bothgroups, which may represent ILOCA/O andILOCA/P respectively. These disorders aredifficult to differentiate clinically early in thecourse of the disease as both usually presentwith an isolated cerebellar syndrome. Forthese reasons it has been suggested that thetwo disease groups cannot be distinguishedfor at least the first four years after onset ofsymptoms.'0 In our patients severe atrophy ofthe cerebellum (hemispheres and vermis) wasseen in nearly all patients. Seven patients(four with ILOCA/P) had moderate ormarked atrophy of the pons, vermis, andcerebellar hemispheres compatible with amorphological diagnosis of OPCA. In twoothers with ILOCA/P, there was cerebellaratrophy combined with atrophy of themedulla or midbrain, sparing the pons.Several patients with non-cerebellar featureshad normal brainstem images. In both ADCAand ILOCA, it is well recognized that patho-logical findings do not always have clinicalcorrelates and vice versa,' and the sameappears to apply to imaging abnormalities.

Chida and colleagues24 assessed pontine

55

on March 26, 2020 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.57.1.51 on 1 January 1994. D

ownloaded from

Ormerod, Harding, Miller, Johnson, MacManus, du Boulay, Kendall, Moseley, McDonald

volume on CT scans and correlated this withthe clinical picture in a mixed group of adultonset ataxias. They found that pontine atro-phy was -significantly correlated with the pres-ence of autonomic and extrapyramidalfeatures, but this differentiation was notabsolute. Klockgether and coworkers'0 usedMRI to study 28 patients with ILOCA, ofwhom 13 had additional non-cerebellar fea-tures. They were able to identify a group ofpatients with clinical and imaging features ofa pure cerebellar syndrome and proposed thatthis diagnosis should be reserved for patientswith a disease duration greater than fouryears, in whom there is a pure cerebellar syn-drome and where imaging shows atrophy onlyaffecting the cerebellum. Of the patients withadditional features, 11 had brainstem atro-phy. The authors acknowledged that somepatients with pure cerebellar atrophy hadnon-cerebellar features but that the clinicaland imaging abnormalities were not those ofOPCA, as is true of five patients in this series;they suggested that these patients had some*other undefined multisystem degeneration.Of our patients with ILOCA/P, four had adisease duration of four years or less at thetime of assessment and three of these hadatrophy of brainstem structures as well ascerebellum. The remaining six patients withILOCA/P had been symptomatic for up to 30years; two had atrophy of the brainstem andfour had atrophy of only the cerebellum.Thus in our patients the combined clinicaland imaging parameters did not allow anabsolute distinction between OPCA and dis-ease restricted largely to the cerebellum,regardless of disease duration.The importance of detecting clinical or

radiological evidence of brainstem dysfunc-tion was stressed by Klockgether and col-leagues,'0 who suggested that this indicated arelatively poor prognosis. It is notable that alltheir 13 patients with combined cerebellarand brainstem disease had additional parkin-sonism, several had bulbar dysfunction, andfive were incontinent of urine. It seems highlylikely that a significant proportion of thesepatients had multiple system atrophy, eventhough none had overt orthostatic hypoten-sion. Only one of our patients (case 15) hadthis condition, diagnosable on the basis of thepresence of autonomic failure, and autonomicfunction tests were normal in the rest. Theother patients in this series with ILOCA andthe MRI appearances of brainstem atrophydid exhibit a more rapid disease course thanthose with pure cerebellar atrophy, althoughthis was not as striking as in the study ofKlockgether and colleagues.'0 This differencecan be explained by the lower proportion ofcases in our series with probable multiple sys-tem atrophy. This disease has a poor progno-sis compared with ILOCA." 16

Asymptomatic periventricular and whitematter lesions were visible on MRI in a num-ber of our normal control subjects, withincreasing frequency in the older age groups.Such changes have been correlated withpathological changes secondary to small ves-

sel disease.2'27 Their increasing frequencywith age could be a manifestation of minorvascular insults accrued by the brain over thepassage of time. Other studies have foundsuch abnormalities in a similar proportion ofolder subjects.'>'0Our patients had significantly more lesions

than the age matched control subjects andthis difference was more apparent in thoseaged over 50 years. The reason for this excessof cerebral white matter lesions in patientswith degenerative ataxic disorders is unclear.There was no evidence that any of thesepatients had multiple sclerosis; the only onewith oligoclonal immunoglobulin in her cere-brospinal fluid did not have white matterlesions. The presence of white matter lesionswas correlated with the radiological appear-ance of severe cerebral atrophy. Neuronal fallout and gliosis in the place of the axons maybe the explanation for these, as the fibretracts connecting the cortical neurons, partic-ularly from the frontal lobes, course theperiventricular areas. Alternatively, perhapscompliance changes in the presence of atro-phy, and perivascular foci of degeneration aremore likely to occur, this being the majorpathological substrate of MRI white matterlesions in asymptomatic individuals." Suchmechanisms remain speculative in theabsence of definitive pathology, but it isof interest that a similar excess of whitematter lesions is seen in another degenerative(and atrophic) disorder, Alzheimer's dis-ease.32

Other MRI studies6 81011 in patients withcerebellar ataxias have not commented uponwhite matter changes in the cerebral hemi-spheres. Savoiardo and colleagues8 studied 23patients with OPCA and noted signal changewithin the cerebellum and brainstem andsome changes in the putamen in patients witha clinical picture of multiple system atrophy.These abnormalities were not observed in ourstudy, but in the previous one some of thepatients were examined on a high field system(1.5 T) and coronal images of the brainstemand cerebellum were made routinely, whichthe authors found helpful in demonstratingthese changes.8The frequency of the white matter lesions

in the present series of patients is very muchless than that seen in patients with multiplesclerosis. In patients with clinically definitedisease," periventricular lesions were demon-strated in 112/114 patients on the same imag-ing system.9 The number of white matterlesions in our ataxic patients was also usuallysmall. An isolated progressive cerebellar syn-drome is uncommon in multiple sclerosis, butthe diagnosis is often considered in this con-text, particularly early in the course of thedisease and in younger patients. MRI is ofdiagnostic assistance in such cases, as thepresence of extensive periventricular and dis-crete white matter lesions is much more infavour of multiple sclerosis, particularly inpatients under the age of 50 years. Also, thedegree of cerebellar and brainstem atrophyseen in many patients described here is

56 on M

arch 26, 2020 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.57.1.51 on 1 January 1994. Dow

nloaded from

MRI in degenerative ataxic disorders

uncommon other than in advanced cases ofmultiple sclerosis.

In conclusion, MRI is of great value in dis-tinguishing degenerative disorders from thosedue to structural or inflammatory lesions inthe patient presenting with a cerebellar syn-drome who has no affected relatives or spe-cific diagnostic features. It is clearly thepreferred imaging examination in this con-text. MRI also has a role to play in the precisediagnosis of these disorders. The presentstudy indicates that moderate to severe brain-stem with or without cerebellar atrophy at thetime of presentation should point against adiagnosis of Friedreich's ataxia or hereditaryspastic paraplegia. The presence of brainstematrophy in a patient with a late onset purecerebellar syndrome predicts more rapid pro-gression than if the brainstem is spared,10 andshould also suggest investigation for otherfeatures of multiple system atrophy. MRI hasnot yet, however, solved all the long estab-lished difficulties in classifying the degenera-tive ataxias. MRI technology has made majoradvances while this series of patients wasbeing studied. With high field systems, it isnow possible to obtain much better resolu-tion, and there are methods for allowingaccurate quantitation of area/volume ofregions of interest.34 Newer strategies such asdiffusion imaging, magnetisation transferimaging, and spectroscopy offer the possibil-ity of greater pathological specificity. A fur-ther study of degenerative ataxias using suchdevelopments would be worthwhile.

We wish to thank the Multiple Sclerosis Society of GreatBritain and Northern Ireland, the Medical Research Council,and the Brain Research Trust for financial support, and all theclinicians who allowed us to study their patients.

1 Harding AE. The hereditary ataxias and related disorders.Edinburgh: ChurchillLivingstone, 1984.

2 Han JS, Bonstelle CT, Kaufman B, et al. Magnetic reso-nance imaging in the evaluation of the brainstem.Radiology 1984;105:705-12.

3 Bodley TM, Cohen DA, Schatz NJ, et al. Comparison ofmetizamide computed tomography and magnetic reso-nance imaging in the evaluation of lesions at the cervi-comedullary junction. Neurology 1985;35:485-92.

4 Flannagan BD, Bradley WG, Mazziota JC, et al. Magneticresonance imaging of the brainstem: normal structureand basic functional anatomy. Radiology 1985;154:375-83.

5 Ormerod IEC, Bronstein A, Rudge P, et al. Magnetic res-onance in clinically isolated lesions of the brain stem.J Neurol Neurosurg Psychiatry 1986;49:737-43.

6 Bydder GM, Steiner RE, Thomas DJ, et al. Nuclear reso-nance of the posterior fossa: 50 cases. Clin Radiol1983;34: 173-88.

7 Nabatame H, Fukuyama H, Akiguchi I, et al. Spino-cerebellar degeneration: qualitative and quantitative MRanalysis of atrophy. Jf Comput Assist Tomogr 1988;12:298-303.

8 Savoiardo M, Strada L, Girotti F, et al. Olivoponto-cerebellar atrophy: MR diagnosis and relationship tomultisystem atrophy. Radiology 1990;174:693-6.

9 Ormerod EEC, Miller DH, McDonald WI, etal.The roleof NMR imaging in the assessment of multiple sclerosisand isolated neurological lesions. A quantitative study.Brain 1987;11O:1579-1616.

10 Klockgether J, Schroth G, Diener H-C, Dichgans J.Idiopathic cerebellar ataxia of late onset: natural historyand MRI morphology. J Neurol Neurosurg Psychiatry1990;53:297-305.

11 Klockgether T, Peterson D, Grodd W, Dichgans J. Earlyonset cerebellar ataxia with retained tendon reflexes.Clinical, electrophysiological and MRI observations incomparison with Friedreich's ataxia. Brain 1991;114:1559-73.

12 Harding AE. Friedreich's ataxia: a clinical and geneticstudy of 90 families with an analysis of early diagnosticcriteria and intrafamilial clustering of clinical features.Brain 1981;104:589-620.

13 Harding AE. Early onset cerebellar ataxia with retainedtendon reflexes: a clinical and genetic study of a disor-der distinct from Friedreich's ataxia. J Neurol NeurosurgPsychiatry 1981;44:503-8.

14 Harding AE. Theclinical features and classification of thelate onset autosomal dominant cerebellar ataxias: astudy of eleven families, including descendants of the'Drew family of Walworth'. Brain 1982;105:1-28.

15 Harding AE. 'Idiopathic' late onset cerebellar ataxia. Aclinical and genetic study of 36 cases. J Neurol Sci1981;51:259-71.

16 Quinn N. Multiple system atrophy-the nature of thebeast. J Neurol Neurosurg Psychiatry 1989;52(suppl.):78-89.

17 Harding AE. Hereditary 'pure' spastic paraplegia: a clini-cal and genetic study of 22 families. J7 Neurol NeurosurgPsychiatry 1981;44:871-83.

18 Behan WMH, Maia M. Strumpell's familial spastic para-plegia: genetics and neuropathology. J Neurol NeurosurgPsychiatry 1974;37:8-20.

19 Oppenheimer DR. Brain lesions in Friedreich's ataxia.Can JNeurol Sci 1979;6:173-6.

20 Claus D, Aschoff JC. Cranial computerised tomographyin spinocerebellar atrophies. Ann NY Acad Sci1981;374:831-8.

21 Fickler A. Klinische und pathologisch-anatomischeBeitrage zu den Erkrankungen des Kleinhirns. Dtsche ZNervenheilkd 191 1;41:306-75

22Jellinger K, Tarnowska-Dziduszko E. Die ZNS-Veranderungen bei den olivo-ponto-cerebellarenAtrophien. ZNeurol 1971;199:192-214.

23 Hoffmann PM, Stuart WH, Earle KM, Brody JA.Hereditary late onset cerebellar degeneration. Neurology1971;21:771-7.

24 Chida K, Tamura M, Kamikura I, Takasu T, Goto N. Aquantitative evaluation of pontine volume by computedtomography in patients with cerebellar degeneration.Neurology 1990;40:1241-5.

25 Chimowitz MI, Awad IA, Furlan AJ. Periventricularlesions on MRI. Facts and theories. Stroke 1989;24:7-12.

26 Fazekas F, Kleinert R, Offenbacher H, et al.The morpho-logical correlate of incidental punctate white matterhyperintensities on MR images. Am J Neuroradiol 1991;12:915-22.

27 Kirkpatrick JB, HaymanLA. White-matter lesions inMRimaging of clinically healthy brains of elderly subjects:possible pathological basis. Radiology 1987;162:509-1 1.

28 Bradley WG, Waluch V, Wycoff RR. Differential diagno-sis of periventricular abnormalities in MRI of the brain.Noninv Med Imag 1984;1:35-41.

29 Brant-Zawardski M, Fein G, van Dyke C, et al. MR imag-ing of the ageing brain: patchy white matter lesions anddementia. AmYNeuroradiol 1985;6:675-82.

30 Gerard G, Weisberg IA. MRI periventricular lesions inadults. Neurology 1986;36:998-1001.

31 Awad IA, Johnson PC, Spetzler RF, Hodak JA.Subcortical lesions identified on magnetic resonanceimaging in the elderly. 2. Post mortem pathologicalcorrelations. Stroke 1986;17:1090-7.

32 Scheltens P, Barkhof F, Valk J, et al. White matter lesionson magnetic resonance imaging in clinically diagnosedAlzheimer's disease. Evidence for heterogeneity. Brain1992;115:735-48.

33 Poser CM, Paty DW, Scheinberg L, et al. New diagnosticcriteria for multiple sclerosis: guidelines for researchprotocols. Ann Neurol 1983;13:227-31.

34 Wicks DAG, Tofts PS, Miller DH, et al. Volume mea-surements of multiple sclerosis lesions with magneticresonance images: a preliminary study. Neuroradiology1992;34:475-9.

57

on March 26, 2020 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.57.1.51 on 1 January 1994. D

ownloaded from