J7ournal ofNeurology, Neurosurgery, and Psychiatry 1994;57:51-57 51 Magnetic resonance imaging in degenerative J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.57.1.51 on 1 January 1994. Downloaded from ataxic 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

Abstract orders is difficult; early attempts were based MRI of the brain was performed in 53 exclusively on pathological findings, but it has patients with a variety of degenerative been suggested that disease categories can be and related disorders and 96 con- more usefully defined using clinical and trol subjects. Atrophy of intracranial genetic criteria.' Autopsy data are relatively structures was not seen in patients with scarce and a reliable method for definition of the pure type of hereditary spastic para- involved structures during life could con- plegia, or in early cases of Friedreich's tribute to classification and more precise . In advanced Friedreich's ataxia diagnosis. Computerised tomography has a there was atrophy of the vermis and limited role to play in this respect, as images medulla. The MRI features of early onset of posterior fossa structures are usually of rel- cerebellar ataxia with retained reflexes atively poor quality compared with those of were variable, and suggest heterogeneity. the cerebral hemispheres. Magnetic reso- In autosomal dominant cerebellar atax- nance imaging provides superior images of ias, most patients had cerebellar and the and ,26 and has brainstem atrophy, probably reflecting been used to study small series of patients the pathological process of olivoponto- with degenerative ataxic disorders with useful of the distribution of in the Institute ofNeurology, cerebellar atrophy; there was no clearly definition atrophy London, UK defined group with both clinical and brain.79 In larger series it has been suggested NMR Research Group imaging features of isolated cerebellar that MRI offers a useful adjunct to clinical I E C Ormerod features for diagnostic and prognostic pur- D H Miller involvement. The MRI abnormalities in G Johnson idiopathic late onset cerebellar ataxia poses.'01' We report on 53 patients with a D MacManus were predoinantly those of cerebellar variety of degenerative ataxias studied by W I McDonald and brainstem atrophy or pure cerebel- MRI, comparing the findings in different dis- University lar atrophy. The clinical and imagig ease groups, and also with those seen in 96 Department of Clinical features of brainstem abnormalities were control subjects. I E C Ormerod discordant in several patients. Pure cere- A E Harding bellar atrophy was associated with slower D H Miller G Johnson progression of disability. Cerebral atro- Subjects in the late onset atax- CONTROLS D MacManus phy was common http://jnnp.bmj.com/ W I McDonald ias. Cerebral lesions, MRI was performed in 96 control subjects Department of although usually few in number, were aged 18-73 years (table 1). Sixty-four were Neuroradiology, more aged less than 50 years and 32 were 50 years National Hospital for observed in significantly patients Neurology and than controls, particularly those aged or over. Seventy-six of these were normal vol- Neurosurgery, over 50 years. unteers from the Salvation Army or from the London, UK staff of the National Hospital who were not E P G H du Boulay B E Kendall (7 Neurol Neurosurg Psychiatry 1994;57:5 1-57) examined clinically but who had no previous I F Moseley neurological history. The remainder (20) on September 29, 2021 by guest. Protected copyright. Correspondence to: were neurological control subjects who were Professor A E Harding, investigated at the National Hospital for dis- University Department of The degenerative ataxias are a heterogeneous Clinical Neurology, Institute group of disorders, many of which are geneti- orders of peripheral or the ofNeurology, Queen that are not associated with brain . Square, London WC1N cally determined. Their clinical features are 3BG, UK. diverse, with variable degrees of cerebral These patients had no signs attributable to Received 4 December 1992 hemisphere, brainstem, spinal cord, and neurological dysfunction above the foramen and in revised form magnum. 16 March 1993. peripheral dysfunction in addition to Accepted 15 April 1993 cerebellar ataxia. Classification of these dis- PATIENTS Of the 53 patients with ataxia (table 1), 30 were aged under 50 years and 23 were aged Table 1 Control subjects andpatients studied 50 years or more. They were subdivided into Age range Disease duration range six categories on clinical grounds (table 1). Disease category Number (years) (mean) (years) (mean) There were six patients with Friedreich's Controls 96 18-73 (41) N/A ataxia"2 and six with early onset cerebellar Friedreich's ataxia 6 13-38 (23) 9-25 (15) Early onset ataxia/retained reflexes 6 17-42 (28) 1-27 (14) ataxia with retained reflexes. 13 Fourteen Autosomal dominant cerebellar ataxia 14 22-66 (49) 2-25 (16) patients had autosomal dominant cerebellar Idiopathic late onset cerebellar ataia with other features 10 39-73 (61) 3-14 (7) ataxia (ADCA; table 2); seven had additional pure 10 38-64 (51) 2-30 (10) features such as supranuclear ophthalmople- Hereditary spastic paraplegia 7 13-39 (27) 2-33 (14) gia, pseudobulbar palsy, and mild dementia 52 Ormerod, Harding, Miller, J'ohnson, MacManus, du Boulay, Kendall, Moseley, McDonald

Table 2 Clinical and imagingfeatures ofpatients with autosomal dominant cerebellar ataxia J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.57.1.51 on 1 January 1994. Downloaded from Atrophy Age Duration No. (years) (years) Type Features CH Vermis Medulla 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 with ET - - ++ - - - ++ - 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 published criteria.' For patients with ILOCA, (ADCA type II), three a later onset pure cere- disability was assessed on a four point scale: bellar syndrome (ADCA type III),14 and three (1) mild ataxia, able to work; (2) unable to more unusual dominant ataxias, including a work, able to walk and perform activities of syndrome of ataxia, dementia, and daily living; (3) as (2), but unable to walk myoclonus exhibiting paternal transmission. unassisted or chairbound; (4) chairbound, Ten patients had idiopathic late onset cere- dependent on others for activities of daily bellar ataxia (table 3)15 with other clinical fea- living. A severity score related to disease tures such as supranuclear ophthalmoplegia, duration was obtained by dividing disability peripheral neuropathy, mild dementia, optic score by disease duration in years and multi- atrophy, and parkinsonism (ILOCA/O). All plying by 100. presented with ataxia and this remained the predominant feature. Only one had clinical or Methods investigative evidence of autonomic failure MAGNETIC RESONANCE IMAGING and thus fulfilled the criteria for a diagnosis of All subjects were examined on a Picker 0 5 T multiple system atrophy'6. Ten had pure MR imaging system. Multi-slice, contiguous ILOCA (ILOCA/P)-that is, no neurological 5 mm thick axial T2-weighted spin echo (SE dysfunction other than a cerebellar syn- 2000/60) images were taken throughout the drome'5. The last group consisted of seven brain, to optimise the detection of white mat- patients with hereditary spastic paraplegia, ter lesions. Axial inversion recovery (IR either with the pure form (five cases) or with 2000/40/500) images were performed in most additional clinical features.'7 subjects to facilitate assessment of cerebral Each patient was examined by one of the atrophy. Tl-weighted (IR 2000/500/40 or SE http://jnnp.bmj.com/ authors (IECO or AEH). The case notes of 500/40) sagittal images were obtained in most all patients were individually reviewed by control subjects and all patients, to assess the AEH for the purpose of diagnostic classifica- degree of atrophy of the posterior fossa struc- tion, which was made according to previously tures.

Table 3 Clinical and imagingfeatures ofidiopathic late onset cerebellar ataxia on September 29, 2021 by guest. Protected copyright. Atrophy Age Duration Disability No. (years) (years) score Features CH Vermis Medulla Pons Midbrain Cerebrum WML ILOCA (0) 15 62 6 33 P,A + + 16 66 14 21 D ++ - + ++ 17 57 4 50 D,SNO + + 18 68 4 50 D + + + + 19 73 8 25 OA + ++ 20 64 3 67 p + + 21 64 6 33 SNO + + + ++ ++++ 22 62 7 28 N + +++ + ++ + + 23 39 9 22 N _- 24 54 9 22 N + + _- ILOCA (P) 25 41 3 33 + + 26 38 2 100 _ +++ - ++ + 27 64 4 11 28 62 30 67 _ +++ 29 56 2 150 ++ ++ ++ ++ 30 39 8 12 ++ ++ 31 41 5 22 - ++ 32 60 11 27 _ +++ 33 56 6 50 - ++ 34 56 27 11 ++

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. MRI in degenerative ataxic disorders 53

Table 4 Frequency ofcerebral white matter lesions mild atrophy were over the age of 50 years, J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.57.1.51 on 1 January 1994. Downloaded from Age (years) No. No. with lesions (5I) and 10 of these also had white matter lesions in the cerebral hemispheres. Five of seven Control subjects 18-49 63 3(5) with moderate atrophy were over the age of 50-59 19 4(21) 50 years and four of these had cerebral white 60-69 10 4(40) 70-79 2 2(100) matter lesions. Lesions within the cerebral white matter Category (no.) No. with lesions No. aged < 50 with lesions were seen in three subjects (5%) aged less Patients than 50 years and in 10 (30%) subjects aged Friedreich's ataxia (6) 0 0 Early onset cerebellar ataxia with retained 50 years or over (table 4). This difference was reflexes (6) 1 1 highly significant p < 0-001). In the three Autosomal dominant cerebellar ataxia (14) 7 1/5 X2t Idiopathic late onset cerebellar ataxia with subjects in the younger group there were only other neurological dysfunction (10) 7 0 minor lesions within the cerebral white matter Idiopathic late onset cerebellar atasia alone (10) 2 0 in two and smooth periventricular signal Hereditary spastic paraplegia (7) 3 3 change in one. Of the 10 older subjects with cerebral lesions, four were aged 50-59 years; three of these showed minor changes but one had extensive periventricular and white mat- ASSESSMENT OF MIf ter lesions. Six of the subjects aged 60 years The MR images of both controls and patients or over had white matter abnormalities. were intermingled and examined for the pres- These were minor in four (with some smooth ence of atrophy. Atrophy was qualitatively periventricular change and a few small scat- graded on a four point scale (0 = no atrophy, tered lesions in the cerebral white matter) and 1 = mild, 2 = moderate and 3 = severe) by two had more extensive abnormalities. two experienced neuroradiologists (BEK and IFM) who reported by consensus without MMI IN PATENTS prior Inowledge of the diagnosis. Atrophy of Only two of the patients with hereditary spas- cerebellar hemispheres, cerebellar vermis, tic paraplegia, both with additional cerebellar medulla, pons, midbrain, and cerebral hemi- signs and aged less than 50 years, had mild or spheres was rated in this way for each patient. moderate atrophy of intracranial structures, Cerebral white matter lesions were defined as the brainstem, cerebellum, and cerebrum in areas of hyperintensity on T2-weighted one, and the vermis and cerebrum in the images of at least three pixels, excluding small other. Both of these patients had white matter caps at the frontal or occipital poles. lesions, as did another with the pure disease. MRI was normal in the four other patients with pure disease. Results MRI was normal in two patients with MRI IN NORMAL CONTROL SUBJECTS Friedreich's ataxia (fig la). Four of the six Mild atrophy of the brainstem was found in patients, all with disease duration greater six subjects and of these three were aged over than 15 years, had moderate medullary atro- 50 years. In two subjects more than one part phy which was associated with lesser changes

of the brainstem was considered atrophic. in the pons and midbrain. Only one of these http://jnnp.bmj.com/ Mild atrophy of the cerebellar hemispheres patients had marked was seen in four subjects and three were aged atrophy (fig ib) and two had moderate or under 50 years. The vermis was atrophic in severe atrophy of the vermis. None of the 10 subjects, all of whom were over 50 years; patients with Friedreich's ataxia had cerebral in nine of these there was also atrophy of the atrophy or white matter lesions. cerebral hemispheres. The cerebral hemi- The MRI features of early onset cerebellar spheres were atrophic in 42 control subjects; ataxia were variable. MRI was normal in one the degree of atrophy was mild in 35 and patient with recent onset (within two years). on September 29, 2021 by guest. Protected copyright. moderate in seven cases. Nineteen of 35 with Only one had severe cerebellar hemisphere

Figure 1 Ti weighted (SE 500/40) sagital MRI of two patients with Friedreich's ataxia. (A) a 14 year old boy with symptoms ofseven years' duration. MRI is normal; (B) a 38year old woman who had had symptomsfor 25years. MRI shows atrophy ofthe cerebellum, brainstem, and upper cervical cord. 54 Ormerod, Harding, Miller, Johnson, MacManus, du Boulay, Kendall, Moseley, McDonald

Figure 2 Tl-weighted J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.57.1.51 on 1 January 1994. Downloaded from (SE 500140) sagittal MRI of a 41 year old woman with pure idiopathic late onset cerebellar ataxia (case 25) showing marked atrophy ofthe cerebellar hemispheres (A) and vermis (B). The brainstem is normal.

atrophy, associated with mild atrophy of the Cerebral hemisphere atrophy was present vermis, pons and cerebral hemispheres. Two in 50 per cent or more of patients in these patients had mild or moderate vermis atro- three groups of late onset ataxic disorders, phy, one with mild cerebral atrophy. One had being least common in ILOCA/P. Dementia atrophy of the whole of the brainstem, and was observed in three patients with ADCA; the sixth had vermis, brainstem, and cerebral one had gross cerebral atrophy without white atrophy with scanty periventricular white matter lesions, one had periventricular and matter lesions. white matter lesions but no significant cere- Of the 14 patients with ADCA, eight had bral atrophy, and in the third the cerebral moderate or severe cerebellar hemisphere hemispheres appeared normal. Three patients atrophy, and all but one atrophy of the vermis with ILOCA/O were demented. All had cere- (table 2). Five of seven of those with ADCA bral atrophy and two had white matter type I had brainstem atrophy, predominantly lesions. involving the pons, although in one it was There was no significant effect (Kruskall- confined to the medulla; six of these patients Wallis analysis of variance) of disease dura- had clinical evidence of brainstem dysfunc- tion in any of the seven diagnostic groups tion. Brainstem involvement was seen in both which might explain the distribution or sever- patients with ADCA type II, and both with ity of atrophy in these patients. Moderate or type III, a pure cerebellar syndrome, although severe cerebral, vermis, and cerebellar atro- disease duration was relatively short in the phy was more frequent (p < 0 003, p < 0x001 latter. and p < 0x002, respectively) in patients aged Marked cerebellar hemisphere and vermis over 50 years than in patients under 50 years, atrophy was frequent in patients with but all the patients with early onset ataxia and http://jnnp.bmj.com/ ILOCA/O, a disorder clinically similar to Friedreich's ataxia were in the younger group. ADCA type I, but almost universal in the A significant difference was found between pure type of ILOCA (fig 2). Seven of the 10 the frequency of lesions within the cerebral patients with ILOCA/O had brainstem atro- hemispheres (fig 4), either periventricular or phy, again maximally involving the pons; in discrete within the cerebral white matter, in two this was confined to the midbrain or

medulla. Two of the patients with supranu- on September 29, 2021 by guest. Protected copyright. clear ophthalmoplegia had normal brainstem images. Two patients with ILOCA/O had extrapyramidal features and both of these had moderate or severe pontine atrophy. Brainstem atrophy without clinical correlate was seen in five of the 10 patients with ILOCA/P, including all those with disease durations of less than five years (fig 3). Disability score was significantly higher in the 11 patients with brainstem atrophy on MRI (excluding case 18 who only had mild medullary atrophy, p < 0-01, Wilcoxon's rank sum test). A major contribution to this result came from four patients, two with a pure cerebellar syndrome, clinically and radiologi- cally, who had a very benign course (cases 28 and 34), and two with a rapidly progressive pure cerebellar syndrome of short duration Figure 3 Tl-weighted (SE 500/40) sagittal MRI ofa 59 year old woman with pure idiopathic late onset cerebellar who had striking brainstem atrophy on MRI ataxia oftwo years' duration (case 29). There is marked (cases 26 (fig 3) and 29). atrophy of the cerebellum and brainstem. MRI in degenerative ataxic disorders 55

the seven groups (p < 0 05, analysis of vari- Friedreich's ataxia; they also described J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.57.1.51 on 1 January 1994. Downloaded from ance; table 4). As stated, these were uncom- prominent atrophy of the cervical cord which mon in patients with Friedreich's ataxia, early we did not assess in detail. onset ataxia, and spastic paraplegia. In the These authors compared the imaging find- diagnostic categories which included more ings in Friedreich's ataxia with those in early older patients (ADCA, ILOCA/O and onset cerebellar ataxia with retained reflexes, ILOCA/P), the lesions occurred mainly in and found a high incidence of cerebellar patients over the age of 50 years. The effect hemisphere and vermis atrophy in the latter, of age on the frequency of the lesions was with less prominent brainstem atrophy. The therefore analysed and comparison was made MRI features in our cases of this syndrome between the patients and the normal control were rather variable. Neuropathological subjects. Significantly more patients over the studies in such patients are extremely scarce, age of 50 years had lesions than younger showing olivopontocerebellar atrophy (OPCA) patients (2, p < 0 005). More patients had in two cases.2'22 The MRI features reported lesions than controls, and this difference was here and by others" do not suggest OPCA in more marked for patients over the age of 50 most cases, and indicate that this syndrome is years (e p < 003) than for those aged less pathologically and genetically heterogeneous. than 50 years (X2, p < 0 05). When the fre- Pathological heterogeneity is well recog- quency of lesions was analysed with respect nized in ADCA. Most, but not all, patients to disease duration, no significant association with the clinical features of ADCA types I was found (Kruskall-Wallis one way analysis and II have OPCA at necropsy. There is of variance). Only marked atrophy of the some variation in pathological findings within cerebral hemispheres was associated with the large families.' Six of our nine patients in presence of lesions (X;2, p < 0-02). When these two categories had MRI findings sug- the groups were subdivided further to com- gesting OPCA. The same applied to two of pare the effect of age (<50 years and >50 three with ADCA type III, a later onset pure years) and atrophy of the brain on the cerebellar syndrome in which one patho- presence of lesions, no significant differences logical report has shown cerebello-olivary were found. atrophy.23 Thus brainstem involvement is by The had been examined no means always manifest clinically, regard- in 26 patients but none, of the control sub- less of disease duration. Most patients with jects. One patient had oligoclonal clear clinical evidence of brainstem dysfunc- immunoglobulin. This was a 60-year-old tion showed brainstem atrophy on MRI. Five woman with an 11 year history of a progres- of six patients with ADCA type I and sive ataxic syndrome which was clinically supranuclear ophthalmoplegia had significant classified as a pure type of ILOCA. The MRI atrophy of the pons or midbrain, or both. In demonstrated some atrophy of the cerebellar these patients with ADCA, there was no hemispheres and a more marked atrophy of clearly defined group with both clinical and the cerebellar vermis but there were no imaging features of isolated cerebellar lesions in the cerebral hemispheres. Two involvement. patients, with ADCA and a pure type of spas- ILOCA has been divided on pathological

tic paraplegia respectively, showed elevated grounds into OPCA and cerebello-olivary http://jnnp.bmj.com/ cerebrospinal immunoglobulin:albumin ratios atrophy, the latter with sparing of the brain- at 12% and 18%. stem apart from the inferior olives. Atrophy of the cerebellar hemispheres is seen in both groups, which may represent ILOCA/O and Discussion ILOCA/P respectively. These disorders are In this series of patients with degenerative difficult to differentiate clinically early in the ataxias, the presence of atrophy of brainstem course of the disease as both usually present and cerebellar structures on MRI scans was with an isolated cerebellar syndrome. For on September 29, 2021 by guest. Protected copyright. generally in accord with previous pathological these reasons it has been suggested that the observations. The normal findings in pure two disease groups cannot be distinguished hereditary spastic paraplegia reflect the fact for at least the first four years after onset of that degenerative changes in this disease are symptoms.'0 In our patients severe atrophy of usually confined to the spinal cord.'8 In the cerebellum (hemispheres and vermis) was Friedreich's ataxia, necropsy studies have seen in nearly all patients. Seven patients shown marked atrophy of the spinal cord but (four with ILOCA/P) had moderate or the brainstem, cerebellum and cerebrum are marked atrophy of the pons, vermis, and usually normal, apart from some atropy of the cerebellar hemispheres compatible with a medulla and degenerative changes in the morphological diagnosis of OPCA. In two cerebellar peduncles.'9 CT studies have others with ILOCA/P, there was cerebellar reported mild atrophy of the cerebrum and atrophy combined with atrophy of the more marked vermis atrophy in some, usually medulla or midbrain, sparing the pons. advanced, cases.20 Some of our patients, who Several patients with non-cerebellar features had had the disease for a long time, had atro- had normal brainstem images. In both ADCA phy of the vermis, medulla, pons, or midbrain and ILOCA, it is well recognized that patho- with changes in the medulla being most fre- logical findings do not always have clinical quent and striking. Klockgether and col- correlates and vice versa,' and the same leagues" reported essentially similar MRI appears to apply to imaging abnormalities. findings to those described here in Chida and colleagues24 assessed pontine 56 Ormerod, Harding, Miller, Johnson, MacManus, du Boulay, Kendall, Moseley, McDonald

volume on CT scans and correlated this with sel disease.2'27 Their increasing frequency J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.57.1.51 on 1 January 1994. Downloaded from the clinical picture in a mixed group of adult with age could be a manifestation of minor onset ataxias. They found that pontine atro- vascular insults accrued by the brain over the phy was-significantly correlated with the pres- passage of time. Other studies have found ence of autonomic and extrapyramidal such abnormalities in a similar proportion of features, but this differentiation was not older subjects.'>'0 absolute. Klockgether and coworkers'0 used Our patients had significantly more lesions MRI to study 28 patients with ILOCA, of than the age matched control subjects and whom 13 had additional non-cerebellar fea- this difference was more apparent in those tures. They were able to identify a group of aged over 50 years. The reason for this excess patients with clinical and imaging features of of cerebral white matter lesions in patients a pure cerebellar syndrome and proposed that with degenerative ataxic disorders is unclear. this diagnosis should be reserved for patients There was no evidence that any of these with a disease duration greater than four patients had ; the only one years, in whom there is a pure cerebellar syn- with oligoclonal immunoglobulin in her cere- drome and where imaging shows atrophy only brospinal fluid did not have white matter affecting the cerebellum. Of the patients with lesions. The presence of white matter lesions additional features, 11 had brainstem atro- was correlated with the radiological appear- phy. The authors acknowledged that some ance of severe cerebral atrophy. Neuronal fall patients with pure cerebellar atrophy had out and in the place of the may non-cerebellar features but that the clinical be the explanation for these, as the fibre and imaging abnormalities were not those of tracts connecting the cortical , partic- OPCA, as is true of five patients in this series; ularly from the frontal lobes, course the they suggested that these patients had some* periventricular areas. Alternatively, perhaps other undefined multisystem degeneration. compliance changes in the presence of atro- Of our patients with ILOCA/P, four had a phy, and perivascular foci of degeneration are disease duration of four years or less at the more likely to occur, this being the major time of assessment and three of these had pathological substrate of MRI white matter atrophy of brainstem structures as well as lesions in asymptomatic individuals." Such cerebellum. The remaining six patients with mechanisms remain speculative in the ILOCA/P had been symptomatic for up to 30 absence of definitive pathology, but it is years; two had atrophy of the brainstem and of interest that a similar excess of white four had atrophy of only the cerebellum. matter lesions is seen in another degenerative Thus in our patients the combined clinical (and atrophic) disorder, Alzheimer's dis- and imaging parameters did not allow an ease.32 absolute distinction between OPCA and dis- Other MRI studies6 81011 in patients with ease restricted largely to the cerebellum, cerebellar ataxias have not commented upon regardless of disease duration. white matter changes in the cerebral hemi- The importance of detecting clinical or spheres. Savoiardo and colleagues8 studied 23 radiological evidence of brainstem dysfunc- patients with OPCA and noted signal change tion was stressed by Klockgether and col- within the cerebellum and brainstem and

leagues,'0 who suggested that this indicated a some changes in the putamen in patients with http://jnnp.bmj.com/ relatively poor prognosis. It is notable that all a clinical picture of multiple system atrophy. their 13 patients with combined cerebellar These abnormalities were not observed in our and brainstem disease had additional parkin- study, but in the previous one some of the sonism, several had bulbar dysfunction, and patients were examined on a high field system five were incontinent of urine. It seems highly (1.5 T) and coronal images of the brainstem likely that a significant proportion of these and cerebellum were made routinely, which patients had multiple system atrophy, even the authors found helpful in demonstrating though none had overt orthostatic hypoten- these changes.8 on September 29, 2021 by guest. Protected copyright. sion. Only one of our patients (case 15) had The frequency of the white matter lesions this condition, diagnosable on the basis of the in the present series of patients is very much presence of autonomic failure, and autonomic less than that seen in patients with multiple function tests were normal in the rest. The sclerosis. In patients with clinically definite other patients in this series with ILOCA and disease," periventricular lesions were demon- the MRI appearances of brainstem atrophy strated in 112/114 patients on the same imag- did exhibit a more rapid disease course than ing system.9 The number of white matter those with pure cerebellar atrophy, although lesions in our ataxic patients was also usually this was not as striking as in the study of small. An isolated progressive cerebellar syn- Klockgether and colleagues.'0 This difference drome is uncommon in multiple sclerosis, but can be explained by the lower proportion of the diagnosis is often considered in this con- cases in our series with probable multiple sys- text, particularly early in the course of the tem atrophy. This disease has a poor progno- disease and in younger patients. MRI is of sis compared with ILOCA." 16 diagnostic assistance in such cases, as the Asymptomatic periventricular and white presence of extensive periventricular and dis- matter lesions were visible on MRI in a num- crete white matter lesions is much more in ber of our normal control subjects, with favour of multiple sclerosis, particularly in increasing frequency in the older age groups. patients under the age of 50 years. Also, the Such changes have been correlated with degree of cerebellar and brainstem atrophy pathological changes secondary to small ves- seen in many patients described here is MRI in degenerative ataxic disorders 57

uncommon other than in advanced cases of 9 Ormerod EEC, Miller DH, McDonald WI, etal.The role J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.57.1.51 on 1 January 1994. Downloaded from of NMR imaging in the assessment of multiple sclerosis multiple sclerosis. and isolated neurological lesions. A quantitative study. In is of value in dis- Brain 1987;11O:1579-1616. conclusion, MRI great 10 Klockgether J, Schroth G, Diener H-C, Dichgans J. tinguishing degenerative disorders from those Idiopathic cerebellar ataxia of late onset: natural history lesions in and MRI morphology. J Neurol Neurosurg Psychiatry due to structural or inflammatory 1990;53:297-305. the patient presenting with a cerebellar syn- 11 Klockgether T, Peterson D, Grodd W, Dichgans J. Early drome who has no affected relatives or spe- onset cerebellar ataxia with retained tendon reflexes. Clinical, electrophysiological and MRI observations in cific diagnostic features. It is clearly the comparison with Friedreich's ataxia. Brain 1991; examination in this con- 114:1559-73. preferred imaging 12 Harding AE. Friedreich's ataxia: a clinical and genetic text. MRI also has a role to play in the precise study of 90 families with an analysis of early diagnostic The present criteria and intrafamilial clustering of clinical features. diagnosis of these disorders. Brain 1981;104:589-620. study indicates that moderate to severe brain- 13 Harding AE. Early onset cerebellar ataxia with retained atrophy at the tendon reflexes: a clinical and genetic study of a disor- stem with or without cerebellar der distinct from Friedreich's ataxia. J Neurol Neurosurg time of presentation should point against a Psychiatry 1981;44:503-8. hereditary 14 Harding AE. Theclinical features and classification of the diagnosis of Friedreich's ataxia or late onset autosomal dominant cerebellar ataxias: a spastic paraplegia. The presence of brainstem study of eleven families, including descendants of the in a with a late onset pure 'Drew family of Walworth'. Brain 1982;105:1-28. atrophy patient 15 Harding AE. 'Idiopathic' late onset cerebellar ataxia. A cerebellar syndrome predicts more rapid pro- clinical and genetic study of 36 cases. J Neurol Sci than if the brainstem is spared,10 and 1981;51:259-71. gression 16 Quinn N. Multiple system atrophy-the nature of the should also suggest investigation for other beast. J Neurol Neurosurg Psychiatry 1989;52(suppl.): features of atrophy. MRI has 78-89. multiple system 17 Harding AE. Hereditary 'pure' spastic paraplegia: a clini- not yet, however, solved all the long estab- cal and genetic study of 22 families. J7 Neurol Neurosurg lished in the degenera- Psychiatry 1981;44:871-83. difficulties classifying 18 Behan WMH, Maia M. Strumpell's familial spastic para- tive ataxias. MRI technology has made major plegia: genetics and neuropathology. J Neurol Neurosurg advances while this series patients was Psychiatry 1974;37:8-20. of 19 Oppenheimer DR. Brain lesions in Friedreich's ataxia. being studied. With high field systems, it is Can JNeurolSci 1979;6:173-6. now to obtain much better resolu- 20 Claus D, Aschoff JC. Cranial computerised tomography possible in spinocerebellar atrophies. Ann NY Acad Sci tion, and there are methods for allowing 1981;374:831-8. accurate of area/volume of 21 Fickler A. Klinische und pathologisch-anatomische quantitation Beitrage zu den Erkrankungen des Kleinhirns. Dtsche Z regions of interest.34 Newer strategies such as Nervenheilkd 191 1;41:306-75 diffusion transfer 22Jellinger K, Tarnowska-Dziduszko E. Die ZNS- imaging, magnetisation Veranderungen bei den olivo-ponto-cerebellaren imaging, and spectroscopy offer the possibil- Atrophien. ZNeurol 1971;199:192-214. of A fur- 23 Hoffmann PM, Stuart WH, Earle KM, Brody JA. ity greater pathological specificity. Hereditary late onset cerebellar degeneration. Neurology ther study of degenerative ataxias using such 1971;21:771-7. developments would be worthwhile. 24 Chida K, Tamura M, Kamikura I, Takasu T, Goto N. A quantitative evaluation of pontine volume by computed tomography in patients with cerebellar degeneration. We wish to thank the Multiple Sclerosis Society of Great Neurology 1990;40:1241-5. Britain and Northern Ireland, the Medical Research Council, 25 Chimowitz MI, Awad IA, Furlan AJ. Periventricular and the Brain Research Trust for financial support, and all the lesions on MRI. Facts and theories. 1989;24: clinicians who allowed us to study their patients. 7-12. 26 Fazekas F, Kleinert R, Offenbacher H, et al.The morpho- logical correlate of incidental punctate white matter hyperintensities on MR images. Am J Neuroradiol 1991; 12:915-22. 1 Harding AE. The hereditary ataxias and related disorders. 27 Kirkpatrick JB, HaymanLA. White-matter lesions inMR http://jnnp.bmj.com/ Edinburgh: ChurchillLivingstone, 1984. imaging of clinically healthy brains of elderly subjects: 2 Han JS, Bonstelle CT, Kaufman B, et al. Magnetic reso- possible pathological basis. Radiology 1987;162:509-1 1. nance imaging in the evaluation of the brainstem. 28 Bradley WG, Waluch V, Wycoff RR. Differential diagno- Radiology 1984;105:705-12. sis of periventricular abnormalities in MRI of the brain. 3 Bodley TM, Cohen DA, Schatz NJ, et al. Comparison of Noninv Med Imag 1984;1:35-41. metizamide computed tomography and magnetic reso- 29 Brant-Zawardski M, Fein G, van Dyke C, et al. MR imag- nanceimaging in the evaluation of lesions at the cervi- ing of the ageing brain: patchy white matter lesions and comedullary junction. Neurology 1985;35:485-92. dementia. AmYNeuroradiol 1985;6:675-82. 4 Flannagan BD, Bradley WG, Mazziota JC, et al. Magnetic 30 Gerard G, Weisberg IA. MRI periventricular lesions in resonance imaging of the brainstem: normal structure adults. Neurology 1986;36:998-1001. and basic functional anatomy. Radiology 1985;154: 31 Awad IA, Johnson PC, Spetzler RF, Hodak JA. on September 29, 2021 by guest. Protected copyright. 375-83. Subcortical lesions identified on magnetic resonance 5 Ormerod IEC, Bronstein A, RudgeP, et al. Magnetic res- imaging in the elderly. 2. Post mortem pathological onance in clinically isolated lesions of the brain stem. correlations. Stroke 1986;17:1090-7. J Neurol Neurosurg Psychiatry 1986;49:737-43. 32 Scheltens P, Barkhof F, Valk J, et al. White matter lesions 6 Bydder GM, Steiner RE, Thomas DJ, et al. Nuclear reso- on magnetic resonance imaging in clinically diagnosed nance of the posterior fossa: 50 cases. Clin Radiol Alzheimer's disease. Evidence for heterogeneity. Brain 1983;34: 173-88. 1992;115:735-48. 7 Nabatame H, Fukuyama H, Akiguchi I, et al. Spino- 33 Poser CM, Paty DW, Scheinberg L, et al. New diagnostic cerebellar degeneration: qualitative and quantitative MR criteria for multiple sclerosis: guidelines for research analysis of atrophy. Jf Comput Assist Tomogr 1988;12: protocols. Ann Neurol 1983;13:227-31. 298-303. 34 Wicks DAG, Tofts PS, Miller DH, et al. Volume mea- 8 Savoiardo M, Strada L, Girotti F, et al. Olivoponto- surements of multiple sclerosis lesions with magnetic cerebellar atrophy: MR diagnosis and relationship to resonance images: a preliminary study. Neuroradiology multisystem atrophy. Radiology 1990;174:693-6. 1992;34:475-9.