Brain Single-Photon Emission Computed Tomography and Magnetic Resonance Imaging in Machado-Joseph Disease

ORIGINAL CONTRIBUTION Brain Single-Photon Emission Computed Tomography and Magnetic Resonance Imaging in Machado-Joseph Disease

Elba C. S. C. Etchebehere, MD; Fernando Cendes, MD, PhD; Iscia Lopes-Cendes, MD, PhD; Juliana A. Pereira; Mariana C. L. Lima, MD; Carla R. Sansana; Cleide A. M. Silva; Maria F. A. G. Camargo; Allan O. Santos, MD; Celso D. Ramos, MD; Edwaldo E. Camargo, MD, PhD

Background: Machado-Joseph disease (MJD) is one of parietal lobes and vermis. Brain SPECT imaging identified the most frequently encountered spinocerebellar ataxias. (by visual analysis) more perfusion abnormalities in the in- However, few reports on brain single-photon emission com- ferior portion of the frontal lobes, mesial and lateral por- puted tomographic (SPECT) imaging (BSI) with hexyl- tions of the temporal lobes, basal ganglia, and cerebellar methylpropylene amineoxine labled with technetium hemispheres. Magnetic resonance imaging identified more Tc 99m and magnetic resonance imaging (MRI) have been abnormalities in the pons and superior portions of the fron- performed for the evaluation of patients with MJD. tal lobes. Olivary atrophy was identified by MRI. Semi- quantitative analysis showed a statistically significant dif- Objectives: To investigate possible abnormalities with ference of perfusion in the inferior and superior portions BSI and MRI in patients with MJD and to correlate these of the frontal lobes, lateral portion of the temporal lobes, findings with the duration of symptoms; cerebellar, ex- parietal lobes, left basal ganglia, cerebellar hemispheres, and trapyramidal, and pyramidal syndromes; and the mo- vermis when compared with the control group. A signifi- lecular characteristics of the MJD mutation. cant difference was noted between the vermis and cerebellar volumes on MRI when compared with the control group. Patients and Methods: Twelve patients (8 males and A significant relationship was observed between the per- 4 females [mean age, 39 years]) with genetically proven fusion of the left parietal lobe (P=.05) and extrapyramidal MJD were studied. The patients underwent BSI and MRI syndrome. There was a tendency toward an inverse rela- on the same day. Brain SPECT imaging was performed tionship between the duration of symptoms and the per- after an intravenous injection of 99mTc–hexylmethylpro- fusion of the cerebellar hemispheres (␳=−0.37; P=.24) and pylene amineoxine. The transaxial, coronal, and sagittal volume of the vermis (␳=−0.30; P=.34); between the length BSIs obtained were submitted to visual and semiquanti- of the expanded (CAG)n repeat and the perfusion of the left tative analyses. Magnetic resonance imaging was parietal lobe (␳=−0.32; P=.36), vermis (␳=−0.28; P=.43), obtained in a 2-T system with coronal, sagittal, trans- and pons (␳=−0.28; P=.42). A direct association was ob- axial, and 3-dimensional (volumetric) acquisitions. The served between the length of the expanded (CAG)n repeat volumes of the cerebellar hemispheres and vermis were and the perfusion of the lateral portion of the right tem- calculated. Control groups for BSI (22 female and 20 male poral lobe (␳=0.67; P=.03). subjects [mean age, 33 years]) and MRI (13 female and 4 male subjects [mean age, 32.2 years]) were included Conclusions: Brain SPECT imaging and MRI were ca- for comparison. pable of identifying subclinical abnormalities in individuals with MJD. These findings may be helpful for a better Results: Correlation was observed between the perfu- understanding of the pathophysiology of this disease. sion abnormalities identified by visual analysis in the BSI with the structural abnormalities observed on MRI in the Arch Neurol. 2001;58:1257-1263

ACHADO-JOSEPH dis- Patients with MJD frequently disease (MJD) is one of play progressive cerebellar ataxia, exter- the most frequently nal ophthalmoplegia, pyramidal and encountered spino- extrapyramidal syndromes, distal muscu- cerebellar ataxias. lar atrophy, eyelid retraction, and twitch- Machado-Joseph disease was first de- ing of the face and tongue.8 Early mani- M 1 scribed in 1972 by Nakano et al in a Por- festations of the disease occur in affected tuguese-American family, which de- individuals between the ages of 25 and scended from Guilherme Machado, and 55 years (mean age, 40 years). In 1993, 9 The affiliations of the authors emigrated from the Azores to Massachu- Takiyama et al localized the MJD gene appear in the acknowledgment setts. Since then, the disease has been de- in chromosome 14q. In 1994, Kawagushi section at the end of the article. scribed in other countries.2-7 et al10 identified the genetic defect as an

(REPRINTED) ARCH NEUROL / VOL 58, AUG 2001 WWW.ARCHNEUROL.COM 1257

©2001 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/26/2021 PATIENTS AND METHODS normalized, a Metz filter and attenuation correction were applied. The images were reconstructed in the transaxial, coronal, and sagittal planes. Twelve patients (8 males, 4 females) from 5 families in the state of Sa˜o Paulo, Brazil, with genetically proven MJD were MRI STUDIES studied. Ages ranged from 22 to 67 years (mean age, 39 years). All patients were required to sign an informed con- Magnetic resonance images were obtained in a 2-T system. sent for the genetic analysis, BSIs, and MRIs. Both BSIs and Acquisition was performed in the coronal, sagittal, and trans- MRIs were performed on the same day. The length of the axial planes and in 3-dimensional mode (volumetric). Sag- CAG repeat was obtained in 10 of the 12 patients studied. ittal T1-weighted spin-echo images were used to guide the Table 1lists the characteristics of the 12 patients, accord- acquisition plane of the other images (slice thickness, 6 mm; ing to sex, age, duration of the disease, signs and symp- tip angle, 180°; repetition time, 430 milliseconds; echo time, toms, magnitude of cerebellar ataxia, and length of the ex- 12 milliseconds; acquistion matrix, 200ϫ350 pixels; and field ϫ panded (CAG)n repeat. The duration of the disease was of view, 25 25 cm). T2-weighed fast spin-echo images were determined from the moment the patient first noted the clini- obtained (slice thickness, 4 mm; tip angle, 120°; repetition cal manifestation of ataxic gait. Ataxic abnormalities were time, 5800 milliseconds; echo time, 129 milliseconds; ac- classified as follows: 1 indicates mild; 2, mild to moderate; quistion matrix, 252ϫ320 pixels; and field of view, 18ϫ18 3, moderate to severe; and 4, severe. cm). Three-dimensional acquisition was obtained on the sag- The control group for BSI consisted of 42 normal vol- ittal plane, gradient-echo T1-weighted (slice thickness, 1.5 unteers (22 females, 20 males) whose ages ranged from 22 mm; tip angle, 35°; repetition time, 22 milliseconds; echo to 66 years (mean age, 33 years). The control group for the time, 9 milliseconds; acquistion matrix, 256ϫ220 pixels; field MRI studies consisted of 17 normal volunteers (13 fe- of view, 230ϫ250 cm; and pixel size, 1ϫ1). Volumetric males, 4 males) whose ages ranged from 21 to 62 years analyses of the cerebellar hemispheres and vermis were per- (mean age, 32.2 years). formed using a semiautomatic manual drawing.

GENETIC ANALYSIS OF PATIENTS WITH MJD ANALYSIS OF BSI AND MRI STUDIES

Genomic DNA was isolated from peripheral lymphocytes Two nuclear medicine physicians (E.C.S.C.E. and E.E.C.) through conventional methods. The fragment containing performed visual and semiquantitative analyses of BSIs. One the (CAG)n repeat of the MJD gene was amplified by poly- neuroradiologist (F.C.) performed visual and quantitative merase chain reaction using the primers MJD52 and MJD25. analyses of MRIs. Polymerase chain reaction was performed in a final vol- Visual analysis of BSI was performed using a 4-grade ume of 12.5 µL, containing 100 ng of genomic DNA, 10mM hypoperfusion scale from normal to severely abnormal, us- of Tris-hydrochloride (pH, 8.8), 15mM of magnesium chlo- ing brain perfusion of the control group as reference. The ride, 50mM of potassium chloride, 2% formamide, 250µM perfusion abnormalities were graded as follows: 1, normal of dCTP, dGTP and dTTP, each, 25µM of adenosine tri- or mildly abnormal; 2, mild to moderately abnormal; 3, mod- phosphate, 1.5 µCi of radiolabled [35S]dATP, 100 ng of each erate to severely abnormal; and 4, severely abnormal. Semi- primer and1UofTaq polimerase. The DNA was denatur- quantitative analysis of BSI was performed using the thalami ated at 94°C for 5 minutes, after 32 cycles at 94°C for 1 as reference, by placing regions of interest (ROIs) on the minute, 60°C for 1 minute, and 72°C for 5 minutes, fol- cerebral and cerebellar cortices. Counts per pixel of each lowed by a final extension at 72°C for 5 minutes. ROI were obtained, divided by counts per pixel of the To determine the size of the allele, the products of the thalami, and compared with the control group. polymerase chain reaction were analyzed in gels of poly- Visual analysis of MRIs was performed using the same acrylamide at 6% in parallel with a marker of molecular 4-grade scale from normal to severely abnormal. Quanti- weight and visualized through autoradiography. The size tative analysis of MRI was performed by placing ROIs on of the alleles was determined by comparison with a se- the cerebral and cerebellar cortices. The volume of each ROI quence of the marker and the numbers converted into units was obtained and compared with the control group. of CAG repeat (N), using the equation N=[(T−121)/3], In both the BSI and MRI studies, the regions ana- where T is the size of the fragments in base pairs, assum- lyzed were the frontal lobes (inferior and superior por- ing that the variation of the size of the polymerase chain tions), the temporal lobes (lateral and mesial portions), the reaction product occurred in the repetitive CAG repeat. parietal lobes, the cerebellar hemispheres, vermis, and pons. On BSIs the basal ganglia and primary visual cortex were BSI STUDIES also analyzed. The inferior olives were evaluated on MRIs.

Brain SPECT imaging was performed using 99mTc-HMPAO. STATISTICAL ANALYSIS Patients were intravenously injected with 99mTc-HMPAO 15 minutes after preparation of the radioactive material. Pa- To compare the variables measured between groups the tients and control subjects were required to remain resting Mann-Whitney test was applied. To verify associations or in a dark quiet room for 10 minutes. While at rest, they re- compare proportions the ␹2 or the Fisher exact test were ceived an intravenous injection of 30 mCi of 99mTc-HMPAO performed. Statistical significance was set at PϽ.05. To iden- and were required to rest for 10 additional minutes. tify the magnitude of relationship between 2 measures, Brain SPECT images were acquired in a scintillation cam- Spearman linear correlation coefficient was applied. To verify era equipped with a fan beam collimator. Sixty images were the concordance between BSI and MRI, the ␬ coefficient acquired, at 6°-intervals for a total of 360°. Images were was applied.

(REPRINTED) ARCH NEUROL / VOL 58, AUG 2001 WWW.ARCHNEUROL.COM 1258

Signs and Symptoms Patient No./ Magnitude of Length of Expanded

Sex/Age, y Duration of MJD, y Cerebellar Pyramidal Extrapyramidal Cerebellar Ataxia (CAG)n Repeat 1a/M/36 3 + + − 1 73 2b/M/42 9 + + − 3 73 3b/F/43 7 + + − 2 70 4c/F/22 7 + + − 1 75 5c/M/28 16 + + + 4 77 6a/M/31 4 + + + 1 ND 7/F/25 7 + − + 4 78 8/M/65 19 + − + 4 68 9d/M/67 15 + − + 4 70 10d/M/35 5 + + + 3 68 11a/F/30 3 + + − 1 ND 12b/M/44 9 + − − 2 71 Mean/. . ./39 8.7 ...... 2.6 72.3 SD/. . ./14.4 5.3 ...... 1.2 3.5

*The superscript letters a, b, c, and d in the stub indicate individuals belonging to the same family; +, presence; −, absence; ataxia abnormality grades: 1, mild; 2, mild to moderate; 3, moderate to severe; and 4, severe; ND, not done; and ellipses, does not apply.

expansion of a repetitive sequence of a CAG trinucleo- tide (or [CAG] repeat). With modern molecular genetic Table 2. Correlation Between the Visual Analyses of BSI n and MRI in Patients With Machado-Joseph Disease technology, it is possible to establish the diagnosis of * MJD with high sensitivity and specificity.6,11 Region of Interest r † There are only a few reports on brain single-photon emission computed tomographic (SPECT) imaging (BSI) Frontal lobes with hexylmethylpropylene amineoxine radiolabeled with Inferior 0.05 99m Superior 0.25 technium Tc 99m ( Tc-HMPAO) and magnetic reso- Temporal lobes nance imaging (MRI) for the evaluation of patients with Lateral 0.16 MJD. Most of these studies were performed before ge- Mesial 0.16 netic testing was available.12 Brain SPECT imaging with Parietal lobes 0.40 99mTc-HMPAO is a sensitive method for evaluating cere- Cerebellar hemispheres 0.00 bral perfusion with the advantage of a lower cost when com- Vermis 0.74 Pons 0.01 pared with positron emission tomographic scans. The objectives of this study were to evaluate neu- *BSI indicates brain single-photon emission computed tomographic ronal perfusion and function in patients with MJD who imaging; MRI, magnetic resonance imaging. underwent visual and semiquantitative BSI analyses and †␬ Coefficient. Values greater than or equal to 0.75 indicate excellent correlation; values between 0.75 and 0.40, good correlation; and values visual and volumetric MRI analyses, and to determine the below 0.40, poor correlation. possible relationship between BSI and MRI abnormalities with the duration of the disease and the length of expanded CAG repeat. There was no correlation between the BSI and MRI abnormalities in other areas, and in general, the abnormali- RESULTS ties were more severe on BSI than on MRI. Brain SPECT imaging identified more functional abnormalities in Analysis of the length of the expanded (CAG)n repeat was the inferior portion of the frontal lobes, temporal lobes not performed in 2 patients (patients 6 and 11) because (mesial and lateral portions), and cerebellar hemi- they belong to the same family as patient 1, who was geno- spheres. Magnetic resonance imaging identified more struc- typed and positive for the MJD mutation (Table 1). In tural abnormalities in the pons and superior portion of the patients with MJD, the expanded (CAG)n repeat ranged frontal lobes. Hypoperfusion of the basal ganglia was ob- from 68 to 78. No statistical significance was noted for served on BSI. Atrophy of the olives was identified in most age distribution differences between the controls and the patients on MRI and was not, because of the limitations patients with MJD (P=.09, Mann-Whitney test for BSI; of the method, for those studied using BSI. Perfusion of P=.09, Mann-Whitney test for MRI). the primary visual cortex was normal in all patients ex- cept for patient 5 who presented hypoperfusion of this re- VISUAL ANALYSIS OF BSI AND MRI gion on BSI but in whom this region was normal on MRI. No significant relationship was observed between the Table 2 gives the ␬ coefficients obtained for the visual magnitude of the ataxia and perfusion abnormalities of the analyses of BSI and MRI in the 12 patients with MJD. Cor- cerebellar hemispheres or vermis in the patients with MJD. relation was excellent between BSI and MRI abnormali- No correlation was observed between the magnitude of ties in the vermis and good correlation in the parietal lobes. the ataxia and the anatomical abnormalities of the cer-

(REPRINTED) ARCH NEUROL / VOL 58, AUG 2001 WWW.ARCHNEUROL.COM 1259

©2001 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/26/2021 Table 3. Values Obtained From Semiquantitative Analysis of the Cerebral and Cerebellar Cortices Perfusion Data in Patients With Machado-Joseph Disease (MJD) and Control Subjects

Mean (SD) P Value, Region of Interest Side Control Subjects Patients With MJD Mann-Whitney Test* Frontal lobe Inferior R 0.95 (0.08) 0.83 (0.09) .0004 L 0.93 (0.08) 0.82 (0.09) .0012 Superior R 0.98 (0.08) 0.88 (0.05) .0002 L 0.97 (0.07) 0.86 (0.07) .0002 Temporal lobe Lateral R 0.96 (0.09) 0.90 (0.07) .03 L 0.92 (0.09) 0.83 (0.09) .0051 Mesial R 0.84 (0.07) 0.85 (0.06) .93 L 0.84 (0.07) 0.82 (0.07) .33 Parietal lobe R 0.92 (0.09) 0.74 (0.10) .0001 L 0.91 (0.08) 0.75 (0.08) .0001 Cerebellar hemispheres R 1.09 (0.10) 0.95 (0.06) .0001 L 1.07 (0.09) 0.92 (0.11) .0001 Basal ganglia R 1.03 (0.07) 1.00 (0.10) .40 L 1.03 (0.06) 0.97 (0.08) .04 Cingulate gyrus (anterior portion) NA† 1.03 (0.08) 1.00 (0.06) .12 Primary visual cortex NA 1.01 (0.09) 0.98 (0.12) .72 Vermis NA 1.08 (0.09) 0.78 (0.08) .0001 Pons NA 0.98 (0.10) 0.92 (0.13) .24

*Statistical significance was set at PϽ.05. †NA indicates does not apply. SEMIQUANTITATIVE ANALYSIS OF BSI Table 4. Correlation Between Length of the Expanded (CAG)n Repeat and Perfusion Abnormalities on BSI* Table 3 gives the mean (SD) of semiquantitative analysis of perfusion data for the controls and patients Right Side Left Side with MJD. A statistically significant difference in perfu- ␳ † P Value ␳ † P Value sion was observed in the inferior and superior portions Frontal lobe of the frontal lobes, lateral portion of the temporal Inferior 0.43 .22 0.12 .74 lobes, parietal lobes, cerebellar hemispheres, left basal Superior 0.32 .37 0.35 .31 ganglia, and vermis of the patients with MJD compared Temporal lobe with the controls. No statistically significant difference Lateral 0.67 .03‡ 0.56 .09 in perfusion between the groups was observed in the Mesial 0.54 .11 0.22 .54 mesial portion of the temporal lobes, right basal gan- Parietal lobe −0.12 .75 −0.32§ .36 Cerebellar hemispheres 0.02 .96 0.11 .77 glia, anterior portion of the cingulate gyrus, pons, and Basal ganglia 0.13 .72 0.27 .45 primary visual cortex. Cingulate gyrus 0.42 .22 ...... No significant difference was observed between the Primary visual cortex 0.17 .65 duration of the disease and perfusion of the vermis (Spear- Vermis −0.28§ .43 ...... man ␳=0.10, P=.77). There was a tendency toward an Pons −0.29§ .42 inverse relationship between cerebellar hemisphere per- ␳ *BSI indicates brain single-photon emission tomographic imaging; fusion and the duration of symptoms (Spearman =−0.37, ellipses, does not apply. P=.24). †Determined using the Spearman linear correlation coefficient. There was a direct relationship between cerebellar ‡PϽ.05. syndrome and the perfusion of the vermis and cerebel- §Tendency toward an inverse relationship between the length of expanded

(CAG)n repeat and perfusion on BSI. lar hemispheres. Extrapyramidal and pyramidal syndromes were correlated with the perfusion of the frontal, temporal, and parietal lobes and the basal ganglia. A ebellar hemispheres (P=.93, Fisher exact test for BSI; P=.54, statistically significant relationship was observed be- Fisher exact test for MRI) or vermis (P=.54, Fisher exact tween the perfusion of the left parietal lobe (P=.05) and test for BSI; P=.68, Fisher exact test for MRI) in the pa- the presence of extrapyramidal syndrome. tients with MJD. In general, the BSI and MRI abnormali- Table 4 gives the Spearman association coeffi- ties in the vermis and cerebellar hemispheres were more cients and P values between the length of the expanded severe than the magnitude of the ataxia. The exceptions (CAG)n repeat and the perfusion abnormalities on BSI. were BSI and MRI abnormalities in the vermis of patients There was a tendency toward an inverse association be- 7, 8, and 11. The magnitude of ataxia was more severe than tween the hypoperfusion of the left parietal lobe, ver- the perfusion and anatomical abnormalities in the cer- mis, and pons and the length of the expanded (CAG)n ebellar hemispheres in 5 patients (patients 5, 7-9, and 11). repeat. A direct association was observed between the

(REPRINTED) ARCH NEUROL / VOL 58, AUG 2001 WWW.ARCHNEUROL.COM 1260

©2001 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/26/2021 A B A BSI 20 MRI 18 000 3 160 000 3 ∗ 17 000 150 000 10 16 000 140 000 15 000 130 000 0 14 000 120 000 13 000 –10 12 000 110 000 100 000 11 000 –20

Volume of the Vermis, mm of the Vermis, Volume 10 000 90 000

Patients Control of the Cerebellum, mm Volume Control Patients

With MJD Subjects Subjects With MJD % Loss in the Vermis, Volume –30

Figure 1. Vermis (A) and cerebellar hemisphere (B) volumes on magnetic –40 resonance imaging of a patient with Machado-Joseph disease (MJD) and a 12345678910 11 12 normal volunteer. B

0 length of the expanded (CAG)n repeat and the perfusion of the lateral portion of the right temporal lobe. –10 MRI VOLUMETRIC ANALYSIS –20 Patients with MJD presented marked reduction of the vermis and cerebellar volumes, with a statistically signifi- –30 cant difference when compared with the controls (P=.004 and .001, respectively, Mann-Whitney test). Figure 1 –40 compares the volumes of the vermis and cerebellar hemi- –50 spheres, respectively, between both study groups. There Loss in the Cerebellar Hemisphere, % Volume 12345678910 11 12 was a tendency toward an inverse relationship between Patients the duration of the disease and the volume of the vermis ␳ Figure 2. Percentage of volume loss of vermis (A) and cerebellar (Spearman =−0.30, P=.34). No statistically significant hemispheres (B) on magnetic resonance imaging (MRI) and hypoperfusion difference was observed between the volume of the cer- on brain single-photon emission computed tomography (BSI). ebellar hemispheres and the duration of the disease (Spear- man ␳=−0.25, P=.42) or the length of the expanded demonstrated atrophy of the pons, olives, cerebellar hemi- ␳ 16 (CAG)n repeat (Spearman =−0.07, P=.85), as well as spheres, vermis, temporal and frontal lobes, and glo- 17 between the length of the expanded (CAG)n repeat and bus pallidus. To our knowledge, this report is the first the volume of the vermis (Spearman ␳=−0.15, P=.69). to describe perfusion abnormalities in patients with MJD who undergo BSI using 99mTc-HMPAO. SEMIQUANTITATIVE BSI vs VOLUMETRIC MRI Positron emission tomographic studies with differ- ent radiolabeled tracers such as 18F-dopamine (FD),11 C- No correlation was observed between the perfusion of raclopride (RAC), and 18F-fluorodeoxyglucose (FDG) have the vermis and the volume of the vermis (P=.81, Spear- been performed in patients with MJD. Shinotoh et al,18 man linear correlation test) as well as the perfusion of in 1997, demonstrated variable uptake of FD and nor- the cerebellar hemispheres and the volume of the cer- mal uptake of RAC in the striatum of these patients, which ebellar hemispheres (P=.33, Spearman linear correla- did not correlate with the phenotype, length of the ex- tion test). Figure 2 shows the percentage of volume loss panded (CAG)n repeat, duration, or age of onset of symp- of vermis and cerebellar hemispheres on MRI and per- toms. Reduced FDG metabolism in the cerebellum,19 fusion loss on BSI, respectively. brainstem, striatum, and cerebral cortex in patients with The percentage loss for BSI and MRI was calculated MJD has also been described.20 Positron emission tomo- by the following equation: graphic studies of patients with MJD21 have contributed to identifying cortical lesions not identified by anatomi- Percentage of Loss={[Patient−Control cal imaging modalities. Unfortunately, positron emis- (Mean Value of the Control Group)]/Control sion tomographic scanners are unavailable in most nuclear (Mean Value of the Control Group)}ϫ100%. medicine laboratories. Brain SPECT imaging with 99mTc- There was a tendency to identify more perfusion ab- HMPAO has the advantage of lower cost and high sen- normalities than volumetric abnormalities, since the per- sitivity in identifying brain perfusion abnormalities. centage of loss was greater on BSI than on MRI. Overall, In this study, the thalami were used as a reference BSI detected more functional abnormalities than MRI de- because in many neuropathological,22,23 anatomical,16,17 tected structural abnormalities, as seen in the vermis of and functional12,14,15,18,20,21,24 studies, these structures have patients 1 through 6, 10, and 12 (Figure 3 and Figure 4). been reported as normal. Despite the description of the uptake reduction of 123I-iomazenil in the thalami of 4 pa- 13 COMMENT tients with MJD by Ishibashi et al, this is the only finding in the literature. The same authors, while evaluating There are a few reports on BSI12-15 and MRI findings16,17 regional cerebral blood flow with this tracer, did not ob- in patients with MJD. Magnetic resonance imaging has serve perfusion abnormalities in the thalami.

(REPRINTED) ARCH NEUROL / VOL 58, AUG 2001 WWW.ARCHNEUROL.COM 1261

B B

Figure 4. Brain single-photon emission computed tomographic imaging (BSI) (A) and magnetic resonance imaging (B) of a patient with Figure 3. Brain single-photon emission computed tomographic imaging Machado-Joseph disease. C, The BSI of a normal volunteer. The BSI of the (BSI) of a normal volunteer (A) and of a patient with Machado-Joseph patient shows marked hypoperfusion of the vermis (arrows), while on disease (B). Note the hypoperfusion of the vermis (arrow), the parietal lobe magnetic resonance imaging the vermis is normal (dotted arrows). The BSI (arrowhead), and the inferior portion of the frontal lobe (dotted arrow) in the of the normal volunteer shows normal perfusion of the vermis (arrowhead). patient with Machado-Joseph disease.

Brain SPECT imaging may have underestimated pons tant atrophy, which agrees with the findings in this abnormalities because of the size of this structure, which study: patients 5, 8 through 10, and 12 had more hy- is difficult to visualize because of the resolution of the poperfusion than atrophy. These authors also noted scintillation camera. Underestimation of abnormalities severe cerebellar atrophy with the normal regional ce- in the superior portion of the frontal lobes may have been rebral blood flow in 1 patient, which was also seen in the due to the intense perfusion of the anterior portion of present study: patients 3, 4, 7, and 11 showed more signs the cingulate gyrus, producing a partial volume effect. of atrophy than hypoperfusion. Ataxia was present in all patients and, therefore, the Hypoperfusion of the frontal lobes, observed in these perfusion and structural findings in the cerebellar hemi- patients, may be due to deaferentation of 2 anatomical path- spheres and vermis are important. Patients with grade 4 ways: the dento-rubro-thalamo-cortical-ponto-cerebellar ataxia had grade 3 or 4 hypoperfusion of the vermis, while (dopaminergic pathway); or the pathway originating from the cerebellar hemispheres were less affected, indicat- the cerebellar hemispheres that sends connections to the ing that the ataxia was mainly of the trunk and not ap- thalami and from there to the frontal cortex. Hypoperfu- pendicular. In patients 2 through 4 and 6 the severity of sion and atrophy of the parietal lobes may be caused by hypoperfusion of the vermis was greater than the mag- deafferentation of the synapses originating from the fron- nitude of the ataxia observed. Also, patient 4 had hy- tal lobes or from the association pathway of the parietal poperfusion grade 2 of the vermis while atrophy was cortex with the cerebellum through the nuclei in the base graded as 1 on MRI. This may represent preclinical, func- of the pons. Hypoperfusion and atrophy of the temporal tional abnormalities preceding structural damage. These lobes may also be due to deafferentation of the commu- findings suggest that BSI may be more useful in the early nication pathways with the frontal lobes and parietal lobes. stages of disease for identification of preclinical func- It may also be due to deafferentation of the mesocortical tional changes before structural damages occur. Hypoper- limbic pathway secondary to lesions in the ventral teg- fusion and atrophy of the cerebellar hemispheres and ver- mentum of the mesencephalon, causing reduction of the mis encountered in these patients are in agreement with function of the mesial portion of the temporal lobes. Botez the neuropathologic findings in patients with MJD.22 This et al12 and Taniwaki et al20 described hypoperfusion with may be related not only to the deaferentation of the dento- 99mTc-HMPAO and hypometabolism with FDG, respec- rubro-thalamo-cortical-ponto-cerebellar pathway, but to tively, in the frontal, temporal, and parietal lobes. Ishiba- atrophy of the superior and medial cerebellar pe- shi et al,13 using 123I-iomazenil, noted hypoperfusion of the duncles, and the anterior and posterior spinocerebellar left temporal lobe. Murata et al17 identified atrophy mainly tracts, structures known to be affected in patients with in the frontal and temporal lobes, pons, and cerebellum MJD. Vermis hypoperfusion and atrophy may be caused of patients with MJD. These findings may be valuable for by neuronal loss in the anterior and posterior spinocer- a better understanding of the pathways of the corticocerebellar tracts and the Clark column. ebellar connections and may explain behavioral changes The percentage volume loss of the cerebellar hemi- that occur in these patients. spheres on MRI and of hypoperfusion on BSI did not cor- Left basal ganglia hypoperfusion may be caused relate. This is probably because of the high sensitivity of by lesion of the substantia nigra with consequent deaf- BSI for detecting functional abnormalities before ana- ferentation of the afferent and efferent fibers of the stria- tomical changes occur. Taniwaki et al20 observed hypo- tum. Hypometabolism of these structures has been metabolism of FDG in the cerebral and cerebellar corti- described.13,15,18,20 ces of patients with MJD while little or no structural Olivary atrophy was identified on MRI in most pa- abnormalities were found on MRI. Takahashi et al14 noted tients in this study. Lesion of this structure has not been a reduction of the regional cerebral blood flow in the cer- described as a characteristic finding of patients with MJD22 ebellum of 2 patients with MJD who had no concomi- although this has been previously described.16,25

(REPRINTED) ARCH NEUROL / VOL 58, AUG 2001 WWW.ARCHNEUROL.COM 1262

©2001 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/26/2021 The tendency toward an inverse relationship be- REFERENCES tween the duration of the disease and cerebellar perfusion and volume of the vermis did not reach statistical 1. Nakano KK, Dawson DM, Spence A. Machado disease: a hereditary ataxia in Por- significance probably because of the size of the popula- tuguese emigrants to Massachusetts. Neurology. 1972;22:49-55. 2. Radvany J, Avila JO, Gabbai AA, Bacheschi LA. Doença de Machado-Joseph no tion studied. The presence of extrapyramidal syndrome Brazil: o relato das primeiras duas famı´lias [abstract]. Arq Neuropsiquiatr (Saõ correlated with hypoperfusion of the left parietal lobe. Paulo). 1988;46:46B. 3. Radvany J, Camargo CH, Costa ZM, Fonseca NC, Nascimento ED. Machado- There was also a tendency toward an inverse relation- Joseph disease of Azorean ancestry in Brazil: the Catarina kindred: neurological, ship between the length of (CAG)n repeat and perfusion neuroimaging, psychiatric and neuropsychological findings in the largest known of the left parietal lobe, vermis, and pons; and a direct family, the “Catarina” kindred. Arq Neuropsiquiatr. 1993;51:21-30. 4. Burt T, Currie B, Kilburn C, et al. Machado-Joseph disease in east Arnhem Land, relationship with the perfusion of the lateral portion of Australia: chromosome 14q32.1 expanded repeat confirmed in four families. Neu- the right temporal lobe. To our knowledge, this is the rology. 1996;46:1118-1122. 5. Lopes-Cendes I, Teive HG, Cardoso F, et al. Molecular characteristics of Machado- first study to correlate the length of the expanded (CAG)n 99m Joseph disease mutation in 25 newly described Brazilian families. Braz J Genet. repeat with BSI abnormalities using Tc-HMPAO or MRI. 1997;20:717-724. Despite the report by Takiyama et al22 that the ce- 6. Lopes-Cendes I, Teive HG, Calcagnotto ME, et al. Frequency of the distant mu- tations causing spinocerebellar ataxia (SCA1, SCA2, MJD/SCA3 and DRPLA)in rebral cortex of patients with MJD is almost never af- a large group of Brazilian patients. Arq Neuropsiquiatr. 1997;55:519-529. fected, it is clearly evident that cortical lesions do occur 7. Zhou YX, Takiyama Y, Igarashi S, et al. Machado-Joseph disease in four Chi- and can be identified by BSI and MRI. Cortical hypoper- nese pedigrees: molecular analysis of 15 patients including two juvenile cases and clinical correlations. Neurology. 1997;48:482-485. fusion may be caused by deafferentation of anatomical 8. Lima L, Coutinho P. Clinical criteria for diagnosis of Machado-Joseph disease: pathways and not necessarily mean that anatomical le- report of a non-Azorena Portuguese family. Neurology. 1980;30:319-322. sions in these cortical regions actually exist. There may 9. Takiyama Y, Nishizawa M, Tanaka H, et al. The gene for Machado-Joseph disease maps to human chromosome 14q. Nat Genet. 1993;4:300-304. be a reduction of cortical function because of a lesion in 10. Kawaguchi Y, Okamoto T, Taniwaki M, et al. CAG expansions in a novel gene the cerebellum and brainstem. from Machado-Joseph disease at chromosome 14q32.1. Nat Genet. 1994;8: 221-228. The evaluation of the regional cerebral metabolism 11. Lopes-Cendes I, Maciel P, Kish S, et al. Somatic mosaicism in the central ner- is a noninvasive method capable of identifying preclini- vous system in spinocerebellar ataxia type 1 and Machado-Joseph disease. Ann cal changes and an objective marker of the disease ac- Neurol. 1996;40:199-206. 12. Botez MI, Le´ville´ J, Lambert R, Botez T. Single photon emission computed to- tivity. The genetic abnormality may produce functional mography (SPECT) in cerebellar disease: cerebello-cerebellar diaschisis. Eur Neu- abnormalities seen before any anatomical changes have rol. 1991;31:405-412. occurred. Other studies using BSI with 99mTc-HMPAO in 13. Ishibashi M, Sakai T, Matsuishi T, et al. Decreased benzodiazepine receptor binding in Machado-Joseph disease. J Nucl Med. 1988;39:1518-1520. a larger population of patients with MJD would be im- 14. Takahashi N, Odano I, Nishihara M, Yuasa T, Sakai K. Regional cerebral blood portant to test the power of this method in identifying flow measured with N-isopropyl-p-[123]iodoamphetamine single-photon emission tomography in patients with Joseph disease. Eur J Nucl Med. 1994;21:615- abnormalities early in the course of the disease. This 620. method may be of value for a better understanding of the 15. Yen TC, Lu CS, Tzen KY, et al. Decreased dopamine transporter binding in Machado- pathways of the corticocerebellar connections in these Joseph disease. J Nucl Med. 2000;41:994-998. 16. Lopes-Cendes I, Silveira I, Maciel P, et al. Limits of clinical assessment in the ac- patients and may eventually be used to identify patients curate diagnosis of Machado-Joseph disease. Arch Neurol. 1996;53:1168-1174. who would be adequate candidates for therapeutic in- 17. Murata Y, Yamaguchi S, Kawakami H, et al. Characteristic magnetic resonance terventions and to evaluate response to treatment. Fol- imaging findings in Machado-Joseph disease. Arch Neurol. 1998;55:33-37. 18. Shinotoh H, Thiessen B, Snow BJ, et al. Fluorodopa and raclopride PET analysis low-up BSI and MRI of patients with MJD should be per- of patients with Machado-Joseph disease. Neurology. 1997;49:1133-1136. formed to evaluate disease progression and to eventually 19. Gilman S, St Laurent RT, Koeppe RA, Junck L, Kluin KJ, Lohman M. A comparison of cerebral blood flow and glucose metabolism in olivopontocerebellar at- give an insight on patient prognosis. rophy using PET. Neurology. 1995;45:1345-1352. 20. Taniwaki T, Sakai T, Kobayashi T, et al. Positron emission tomography (PET) in Accepted for publication February 26, 2001. Machado-Joseph disease. J Neurol Sci. 1997;145:63-67. From the Division of Nuclear Medicine, Departments 21. Soong BW, Cheng C, Liu RS, Shan D. Machado-Joseph disease: clinical, molecular and metabolic characterization in Chinese kindreds. Ann Neurol. 1997; of Radiology (Drs Etchebehere, Lima, Santos, Ramos, and 41:446-452. E. E. Camargo and Mmes Pereira and Sansana), Neurol- 22. Takiyama Y, Oyanagi S, Kawashima S, et al. A clinical and pathologic study of a ogy (Dr Lopez-Cendes), and Genetics (Dr Lopes-Cendes and large Japanese family with Machado-Joseph disease tightly linked to the DNA markers on chromosome 14q. Neurology. 1994;44:1302-1308. Mrs M. F. A. G. Camargo), and the Research Commission, 23. Robitaille Y, Lopes-Cendes I, Becher M, Rouleau G, Clark AW. The neuropathol- School of Medical Sciences (Mrs Silva), Campinas State Uni- ogy of CAG repeat diseases: review and update of genetic and molecular fea- tures [review]. Brain Pathol. 1997;7:901-926. versity, UNICAMP, Campinas, Brazil. 24. Soong BW, Liu RS. Positron emission tomography in asymptomatic gene car- Corresponding author and reprints: Elba C. S. C. Etche- riers of Machado-Joseph disease. J Neurol Neurosurg Psychiatry. 1998;64:499- behere, MD, Serviço de Medicina Nuclear, Hospital das Clıńi- 504. 25. Haberhausen G, Damian MS, Leweke F, Mu¨ller U. Spinocerebellar ataxia, type 3 cas da UNICAMP, Caixa Postal 6142, 13081-970 Campi- (SCA3) is genetically identical to Machado-Joseph disease (MJD). J Neurol Sci. nas, Brazil (e-mail: [email protected]). 1995;132:71-75.

(REPRINTED) ARCH NEUROL / VOL 58, AUG 2001 WWW.ARCHNEUROL.COM 1263