Putamen Volume Reduction on Magnetic Resonance Imaging Exceeds Caudate Changes in Mild Huntington’s Disease Gordon J. Harris, PhD, Godfrey D. Pearlson, MB, BS, Carol E. Peyser, MD, Elizabeth H. Aylward, PhD, Joy Roberts, BA, Patrick E. Barta, MD, PhD, Gary A. Chase, PhD, and Susan E. Folstein, M D
The characteristic pathological features of Huntington’s disease (HD) are neostriatal atrophy and neuronal loss. Although neuroradiological studies often show caudate atrophy in patients with moderate HD, frequently no caudate atrophy is found early in the illness. There have been no quantitative reports to date on in vivo putamen volume measures in mild HD, although the structure is known to be neuropathologically involved in the illness. We measured volumes of caudate nucleus and putamen and bicaudate ratios (BCR) from magnetic resonance images, blind to diagnosis, in 15 patients with mild HD and 19 age- and sex-matched control subjects using a computerized image analysis system. The region showing greatest atrophy was the putamen, which was reduced 50.1% in mean volume in HD patients compared with control subjects @ < 0.000001). In contrast, caudate volume was reduced 27.7% ( p = 0.004). BCR was increased 28.5% in H D patients ( p = 0.0002). Discriminant function analysis was 94% effective in identifying the diagnostic group based on putamen volume alone, whereas caudate measures had considerable overlap. Correction of putamen volume for head size led to 100% separation by group. Putamen measures and BCR correlated with neurological examination scores but caudate volume did not. Volumetric measurement of putamen is a more sensitive indicator of brain abnormalities in mild HD than measures of caudate atrophy. Harris GJ, Pearlson GD, Peyser CE, Aylward EH, Roberts J, Barta PE, Chase GA, Folstein SE. Putamen volume reduction on magnetic resonance imaging exceeds caudate changes in mild Huntington’s disease. Ann Neurol 1992;31:69-75
Huntington’s disease (HD) is an autosomal dominant, neurodegenerative disorder characterized by motor, cognitive, and emotional abnormalities [ 1, 2). The characteristic neuropathological features are neostriatal atrophy and neuronal loss. O n the basis of analysis of a large series, Vonsattel and associates { 3 ) reported that neuronal loss in neostriatum proceeds from medial to lateral and from dorsal to ventral, a pattern confirmed by others [4). In less severely affected cases, Vonsattel and colleagues { 3 } reported that neuronal loss could generally be appreciated in caudate before putamen. One case report, however, of a man at risk for HD who suffered an accidental death describes neostriatal atrophy primarily in the putamen [5 } . Although postmortem neuropathological studies largely involve assessment of advanced cases, in vivo neuroimaging has the potential to provide answers about the pattern of neostriatal degeneration in mild HD. Both magnetic resonance imaging (MRI) and X-ray computed tomography (CT) scans are commonly
read as qualitatively normal in early HD cases, and are thus considered to be of marginal use in aiding clinical diagnosis. Caudate atrophy can be detected by quanrified means, such as with the bicaudate ratio (BCR; see Methods) using CT {b- 11). Differences between patients with early HD and control subjects as measured by BCR are small, however, with a substantial overlap between groups. In more advanced cases, abnormalities are usually detectable on clinical readings of MRIs or CTs, which often reveal widespread cortical and subcortical atrophy, such as described in one MRI study that reported qualitative subcortical atrophy in 4 patients with HD {IZ}. Volumetric measurement may be more sensitive than qualitative assessment or linear measures such as BCR for identifying brain abnormalities in mild HD. Volumes of caudate and putamen are difficult to quantify using CT, however, due to its relatively poor resolution of gray and white matter. MRI is superior for quantifying gray matter, but as yet there has been only
From the Division of Psychiatric Neuro-Imaging, The Department of Psychiatry and Behavioral Sciences, The Johns Hopkins Medical Institutions, Baltimore, MD.
Address correspondence to D r Harris, Division of Psychiatric Neuro-Imaging, Meyer 3-166, Department of Psychiatry, The Johns Hopkins Hospital, 600 North Wolfe St, Baltimore, MD 21205.
Received Apr 23, 1991. Accepred for publication Jun 18, 1991.
Copyright 0 1992 by the American Neurological Association 69
o n e volumetric MRI study in HD, which involved severely demented patients 1131. In these severely affected patients with HD, Jernigan and associates [13] found that the structure m o s t reduced in volume was the caudate nucleus. T h e y did not measure t h e putamen, per se, but instead measured t h e lentiform nucleus (i.e., putamen and globus pallidus). We report t h e results of volumetric measurement of caudate and putamen, as well as HCR, in 15 patients with HD with very mild m o t o r aiid cognitive symptoms.
Methods Subjects Fifteen patients (6 women, 9 men) with mild HD attending the Johns Hopkins Hospital Huntington’s Disease Clinic were included in the study. Mean age (t standard deviation {SD]) was 43.2 t 12.9 years (range, 29-72 yr). Table 1 presents the demographics and clinical scores of each group. Diagnostic criteria for definite HD 114) were (1) chorea or the characreristic impairment of voluntary movement, which was not present at birth, was insidious at onset, and had gradually become worse; and (2) a family history of at least one other member with these typical symptoms of HD. We selected patients who had the lowest scores on their Quantified Neurcilogic Exam (QNE), compatible with a confident diagnosis. ‘The QNE is a coded, reliable, and valid version of the clinical neurological examination of the motor system ClS}. Included in the Q N E is a motor impairment subscale (MIS1, which measures voluntary mot-or impairment, and a chorea subscale. Patients were excluded who had Q N E scores greater than 45; most patients scored below 30 (25.3 10.2; range, 15-43). Patients had mean Mini-Mental State Exam (MICISE) El61 scores of 27.7 ? 1.9 out of a possible 10 (range, 25-30), indicating, at most, mild cognitive deficit. The control group consisted of 19 age- and sex-matched control subjects (6 women, 13 men), age 45.2 I+_ 15.9 years
*
Table I . Demographic Infomiation for Study G’roups“ Huntington’s Disease
Variable n Age Sex (FIM) Q N E (max
MIS (max
=
-=
105) 27)
Chorea (max = 2 5 ) MMSE (max = 30)
15 43.2 ? 12.yt’ (29- 7 2) 619 25.3 2 10.2 (15-43) 5.0 2 3.0 (1-10) 7.3 I 3.4 (3-14) 27.7 t 1.9 (25-30)
Control Subjects 19
45.2
15.9
2
(23-73)
6113 NIA NIA
NIA NIA ~~
.‘Neurological rests and MMSE were given to the liuntington’s disease group o n l y . ‘Values are mean f standard deviation (min-max).
Q N E = Quitntitied Neurologic Exam; MIS = motor impairment suhscale; MMSE = Mini-Mental State Exam; NIA = not applicable.
70
Annals of Neurology
Vol 31 N o 1 January 1992
(range, 23-73 yr), who had no history of neurological illness, including head trauma causing unconsciousness for more than 1 hour. Exclusion criteria for the control subjects also included history of any psychiatric illness (assessed using the SCID 117)). This project was approved by the Johns Hopkins Joint Committee on Clinical Investigations, and informed consent was obtained from all subjects.
Magnetic Resonance Imaging All patients and control subjects were scanned with tine same 1.5 tesla GE Signa MRI scanner. A set of contiguous 5-mm thick transaxial images was obtained parallel to the anteriorcommissurelposterior-commissure (AC-PC) line for each patient and control subject, which included the entire cerebral cortex and cerebellum. Number of excitations was one. Both T2-weighted and proton-weighted (repetition time = 2,500; echo time = 20180) images were acquired simultaneously. Images were stored on %track magnetic tape and transferred to a DECsystem 5400 RISC server computer, where data were archived on SONY readwrite magneto-optical discs.
Quantitative Measurements MRI slices containing structures of interest were identified blind to diagnosis from images displayed on a DECstation 3100 RISC workstation with a color graphics monitor. Graphics software was developed in-house in “C”/”XWindows” in the ULTRIX ( U N I X ) environment. Structures of interest were outlined interactively, blind to diagnosis, using proton-weighted images on all slices on which they appeared. Areas over all slices were summed and multiplied by slice thickness to compute volumes. All MRI basal ganglia volume ratings were made by the same rater based on a standardized computer-based atlas of MRI basal ganglia neuroanaromy (obtainable from G. D. Pearlson by request). Total caudate (head of caudate only) and total putamen volumes were determined. Examples of caudate and putamen measurements for a patient with HD and a control subject at one slice level are shown in Figure 1. Validity and interrater reliability of this method haw been described previously for several brain structures [ 18). W e determined intrarater and interrater reliability for caudate and putamen volume calculations as further validation. Intrarater reliability for subcortical volume measurements made on two separate occasions gave a correlation coefficient greater than 0.99 for both caudate and pucamen. Interrater reliability yielded a correlation coefficient o f 0.97 for c,ludate and 0.94 for putamen volume. All correlations were :;ignificant (p < 0.001). There were no statistically signifiant differences between means on either interrater or intrarattr reliability trials. T o compare the subjects in our group with those of other studies that used BCR as a measure of caudate atrophy, we also measured BCRs. This technique involves measuring the minimal distance between the caudate indentations of the frontal horns divided by the distance between [he oui-er tables of the skull along the same line, then multiplied by 1 0 0 C11!. Whole brain volume (includes brain and all cerebrospinal fluid [CSF}), total CSF volume (includes ventricular and subdural), and total CSF percentage were calculated using a
F i g I . Transaxial magnetic resonance images at the level of the basal ganglia of a patient with early Huntington’s disease (HD) (left: 30-year-old man; Quantified Neurologic Exam = 18; motor impairment subsrale = 3; chorea = 6; MiniMental State Exam = 27) and control subject (right: 45-yearold man). Caudztes and putamen are manually outlined, as was done in this study. Even though this is a wevy early case of HD with minimaf neurological abnormality, the putamen is drastically reduced in size in the patient with HD. Caudates, however, appear similar. semiautomatic method described previously by our group [ 191. This method uses T2-weighted minus proton-weighted MRI images to produce CSF highlighting, then segments the CSF based on thresholding, and uses semiautomated outlining to calculate whole brain volume. CSF percentage is calculated by dividing CSF volume by whole brain volume. Whole brain volume could not be calculated for 5 patients with HD and 1 control subject because these image sets did not include slices near the apex of the head. The area of the brain slice at the level of the BCR measurement, however, (“brain area”) was determined for all cases using rhe semiautomated outlining method described. This area measurement was then used to normalize all volumetric measurements to account for head size differences between individuals. Significance testing was carried out using two-tailed Student’s t-tests to determine differences between groups. Discriminant function analysis (DFA) was used to ascertain which brain measures best identified diagnostic groups. All DFA scores reported refer to analysis based on each individ-
ual variable assessed separately. Regression analysis was used to determine which structural measures best correlated with clinical state exam scores and to control for the effects of age.
Results Regional Measurements
Although all three subcortical measures were significantly different between patients with HD and control subjects (Table 2), putamen volume was by far the best discriminator. The putamen had lost its characteristic convex lens shape in patients with H D , appearing rather as thin strips, as shown in Figure 1. Mean total putamen volume was reduced 50.1% in patients with HD compared with control subjects. DFA of total putamen volume alone had 93.3% sensitivity (14 of 15) and 94.7% specificity (18 of 19). There was overlap between one patient with HD and one control subject, resulting in 94.1% correct classification. All but 1 patient with HD had a putamen volume below 6.0 ml, whereas all but 1 control subject had a putamen volume greater than 7.8 ml. The 2 misclassified subjects were a 73-year-old male control subject, and a 45year-old male patient with very early HD (QNE = 16; MMSE = 30; duration of illness = 2 yr). After correcting for head size by dividing putamen volume ( x 100) by brain area, there was a 50.4% difference Harris et al: Reduced MRI Putamen in Mild HD
71
Table 2. Regional Measurements by Diagnojtic Groupa Huntington’s Disease
DFA%
- Control
Brain area
4.43 f 1.47b (1.36-7.47) 2.67 f 0.87 (0.75-4.22) 2.76 ? 0.63 (1.69-3.77) 1.66 +- 0.38 (0.97-2.48) 12.8 +- 2.3 (8.0-16.1) 166.4 i_ 12.4
8.88 k 1.19 (6.46-11.29) 5.39 2 0.65 (4.31-6.54) 3.82 ? 1.18 (1.5 5-6.03) 2.31 k 0.69 ( 1.04-3.54) 9.0 f 1.7 (6.5-1 3.9) 165.9 i 11.2
Brain volume
1438 f 133
1397
163
...
CSF volume
158.8
124.6 ? 57.3
...
CSF percent
11.1
8.8 +- 3.1
...
Putamen (ml) Putmenibrain Caudate (ml) Caudate/brain BCR
f
73.3
* 5.3
f
94%
P
t
9.76
*
ll
*
It B
, 10
20
30
40
CAUDATE r = -0.11 p = 0.71
...................
v -
50
QNE
Fig 5. Regression plots of putamen and cuuabe volumes with Quatzti$ed Neurologic Exam (QNE)scores for patients with Huntington’s diseuxe. Putamen volume is significantly correlated with QNE (r = -0.60; n = 15; p = 0.018), whereus cuudate wlume i.c not (r = - 0.11: p = 0 . ; 7 1 )
Consistent with prior CT imaging studies, we found significant differences in BCRs in HD with a degree of overlap between patients with HD and control subjects. The degree of overlap and the group means for BCR in our study were comparable to the results reported in prior CT studies of BCR in HD [6-111, indicating that MRI produces similar results as CT for the calculation of BCR. Only 9 of 15 (6096) patients with early HD had BCR values that were more than 2 SDs above the control mean. Caudate volume, as measured using MRI in this study, proved to be relatively insensitive at discriminating patients with HD from control subjects. None of the subjects in either group had caudate volumes more than 2 SDs below the mean control value. BCR was a better measure of the changes in the caudate nucleus than was overall caudate volume for classifying patients with HD. One possible explanation for this finding is that caudate atrophy in HD occurs in a specific pattern (e.g., medial to lateral), and only the medial portion is affected early in the disease [3].Therefore, the relative change in overall caudate size may be small, whereas the BCR (a measure of the distance between the medial caudates) is more sensitive to medial caudate atrophy. There has been one previous study using quantitative MRI in HD [13]. Jernigan and colleagues [13} studied severely demented patients with advanced H D , a stage of the illness when severe subcortical and cortical atrophy is usually clearly visible on clinical inspection of CT or MRI scans. These investigators measured volumes of caudate and lenticular nucleus (putamen and globus pallidus) rather than putamen alone, as we did. They found that in advanced HD, caudate volume reduction was greater than that of lenticular nucleus. Yonsattel and associates [3:t, in a large autopsy study, found that globus pallidus atrophy is modest compared to the caudate or putamen. Therefore, it is 74
Annals of Neurology
Vol 31 No 1 January 1992
not surprising that in Jernigan and colleagues’ [ 131 study, volumes of putamen and globus pallidus measured together were less reduced than caudate volumes. O n the basis of the neuropathology literaturc, especially the study of Vonsattel and colleagues {13}, o u r findings were not anticipated. Although there is general agreement that in advanced HD atrophy arid neuronal loss in caudate and putamen is approx.imately equal [20, 211, in the less-advanced cases studied by Vonsattel and colleagues [I37 (Grades 0 and l), no gross atrophy was appreciated in neostriatum, and microscopic neuronal depletion was noted in caudate but not putamen (the cellular population of the head of the caudate, but not the putamen, was assessed quantitatively and found to be reduced) [ 3 ] . Morphometric study of putamen was not done on these patients, nor were putamen masses reported. It is possible that visual impressions of gross atrophy are not as accurate and sensitive as volumetric or morphometric assessments, even in the hands of highly experienced observers. It is also possible that our patients with early HD were biased toward a population who tended to have relatively more damage to putamen than caudat’e. Currently, a clinical diagnosis of HD cannot be made until patients show clear motor signs, even when they seem to be experiencing disabling cognitive and emotional symptoms that are typical of HD. It is possible that such patients might have relatively more caudate than putamen damage in early illness. We follow a number of such “at-risk” patients in our clinic, but they were not included in this study. Longitudinal volumetricMRI studies of populations of patients with HD and presymptomatic subjects with the HD gene might provide answers to these questions. It is not clear at what point in HD subtle putamen atrophy begins; it is even possible that putamen might be congenitally smaIIer in those inheriting the HD gene. MRI volumetric studies of gene-positive, presymptomatic subjects can address these issues. Neurological exam scores correlated significantly with putamen but not caudate volume. This finding is in accordance with studies indicating that a variety of motor circuits involve putamen but not caudate, whereas cognitive circuits are thought to involve caudate but not putamen [22, 231. Our patients witlh early HD were only mildly cognitively impaired, with a small range of MMSE scores, so the lack of correlation seen in our study between MMSE and linear or volume measures is not surprising. In summary, MRI volumetric measurements of putamen in our sample of 15 patients with mild H D , normalized for brain size as measured by brain area at the level of the caudate head, completely discriminated patients with HD from control subjects. Normalized volumetric measurements of caudate discriminated
only 68% of patients with HD from control subjects, whereas BCR discriminated 82%. Our study suggests either that put-men is congenitally small in those inheriting the HD gene, or that putamen atrophy is evident by the time a clinical diagnosis can be made. This research was supported in part by MH-40391 and MH-43775 from the National Institute of Mental Health to G. D. Pearlson, by grant RR-0722 from the Outpatient Clinical Research Center, National Institutes of Health Division of Research Resources to G. D. Pearlson and S. E. Folstein, by National Institute of Communicative Disorders and Stroke grant NS-16375, and by the Foundation for the Care and Cure of Huntington’s disease, to S. E. Folstein. We would like to thank R. Nick Bryan, MD, PhD, and John C. Hedreen, MD, for their critiques of the manuscript.
References 1. Bruyn GW. Huntington’s chorea: historical, clinical and labora-
2. 3.
4.
5.
6. 7.
8.
9.
tory synopsis. In: Vinken PJ, Bruyn GW, eds. Handbook of clinical neurology, vol 6. Amsterdam: North-Holland Publishing, 1968:298-378 Folstein SE. Huntington’s disease: a disorder of families. Baltimore: The Johns Hopkins University Press, 1989 Vonsattel JP, Meyers RH, Stevens TJ, et al. Neuropathologic classification of Huntington’s disease. J Neuropathol Exp Neurol 1985;44:559-577 Roos RAC, Pruyt JFM, deVries J, Bots GTA. Neuronal distribution in the putamen in Huntington’s disease. J Neurol Neurosurg Psychiatry 1985;48:422-425 Carrosco LH, Mukherji CS. Atrophy of corpus striatum in normal male at risk of Huntington’s chorea. Lancet 1986;l: 1388-1389 Starkstein SE, Folstein SE, Brandt J, et al. Brain atrophy in Huntington’s disease: a CT scan study. Neuroradiology 1989; 3 1:156-1 59 Starkstein SE, Brandt J, Folstein SE, et al. Neuropsychological and neuroradiological correlates in Huntington’s disease. J Neurol Neurosurg Psychiatry 1988;51:1259-1263 Oepen G, Ostertag CH. Diagnostic value of CT in patients with Huntington’s chorea and their offspring. J Neurol 1981;225: 189-196 Sax DS, ODonnell B, Butters N, et al. Computed tomographic,
neurologic and neuropsychological correlates of Huntington’s disease. Int J Neurosci 1983;18:21-36 10. Bamford KA, Caine ED, Kid0 DK, et al. Clinical-pathologic correlation in Huntington’s disease: a neuropsychological and computed tomography study. Neurology 1989;39:796-801 11. Barr AN, Heinze WJ, Dobben GD, et al. Bicaudate index in computerized tomography of Huntington’s disease and cerebral atrophy. Neurology 1978;28:1196-1200 12. Simmons JT, Pastakia B, Chase TN, Shults CW. Magnetic resonance imaging in Huntington’s disease. AJNR 1986;7:25-28 13. Jernigan TL, Salmon DP, Butters N, Hesselink JR. Cerebral structure on MRI, part 11: specific changes in Alzheimer’s and Huntington’s diseases. Biol Psychiatry 1991;29:68-8 1 14. Folstein SE, Leigh JR, Parhad IM, Folstein MF. The diagnosis of Huntington’s disease. Neurology 1986;36:1279-1283 15. Folstein SE, Jensen B, Leigh JR, Folstein MF. The measurement of abnormal movement: methods developed for Huntington’s disease. Neurobehav Toxic01 Teratol 1983;5:605-609 16. Folstein MF, Folstein SE, McHugh PR. “Mini Mental State.” A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975;12:189-198 17. Spitzer RL, Williams JBW, Gibbon N, et al. Structured clinical interview for DSM-IIIR. New York: New York State Psychiatric Institute, Biometrics Research, 1989 18. Barta PE, Pearlson GD, Powers RE, et al. Auditory hallucinations and smaller superior temporal gyral volume in schizophrenia. Am J Psychiatry 1990;147:1457-1462 19. Harris GJ, Rhew EH, Noga JT, Pearlson GD. A user-friendly method for rapid brain and CSF volume calculation using transaxial MRI images. Psychiatr Res: Neuroimag 1991;40: 61-68 20. Lange H , Thorner G , Hopf A, Schroder KF. Morphometric studies of the neuropathological changes in choreatic diseases. J Neurol Sci 1976;28:401-425 2 1. De la Monte SM, Vonsattel JP, Richardson EP. Morphometric demonstration of atrophic changes in the cerebral cortex, white matter, and neostriatum in Huntington’s disease. J Neuropathol Exp Neurol 1988;47:516-525 22. DeLong MR, Alexander GE, Miller WC, Crutcher MD. Anatomical and functional aspects of basal ganglia-thalamocortical circuits. In: Franks AJ, Ironside JW, Mindham RHS, et al, eds. Function and dysfunction in the basal ganglia. Manchester and New York: Manchester University Press, 1990:3-34 23. Alexander GE, DeLong MR, Strick PL. Parallel organization of functionally segregated circuits linking basal ganglia and cortex. Ann Rev Neurosci 1986;9:357-381
Harris et al: Reduced MRI Putamen in Mild HD
75
The characteristic pathological features of Huntington’s disease (HD) are neostriatal atrophy and neuronal loss. Although neuroradiological studies often show caudate atrophy in patients with moderate HD, frequently no caudate atrophy is found early in the illness. There have been no quantitative reports to date on in vivo putamen volume measures in mild HD, although the structure is known to be neuropathologically involved in the illness. We measured volumes of caudate nucleus and putamen and bicaudate ratios (BCR) from magnetic resonance images, blind to diagnosis, in 15 patients with mild HD and 19 age- and sex-matched control subjects using a computerized image analysis system. The region showing greatest atrophy was the putamen, which was reduced 50.1% in mean volume in HD patients compared with control subjects @ < 0.000001). In contrast, caudate volume was reduced 27.7% ( p = 0.004). BCR was increased 28.5% in H D patients ( p = 0.0002). Discriminant function analysis was 94% effective in identifying the diagnostic group based on putamen volume alone, whereas caudate measures had considerable overlap. Correction of putamen volume for head size led to 100% separation by group. Putamen measures and BCR correlated with neurological examination scores but caudate volume did not. Volumetric measurement of putamen is a more sensitive indicator of brain abnormalities in mild HD than measures of caudate atrophy. Harris GJ, Pearlson GD, Peyser CE, Aylward EH, Roberts J, Barta PE, Chase GA, Folstein SE. Putamen volume reduction on magnetic resonance imaging exceeds caudate changes in mild Huntington’s disease. Ann Neurol 1992;31:69-75
Huntington’s disease (HD) is an autosomal dominant, neurodegenerative disorder characterized by motor, cognitive, and emotional abnormalities [ 1, 2). The characteristic neuropathological features are neostriatal atrophy and neuronal loss. O n the basis of analysis of a large series, Vonsattel and associates { 3 ) reported that neuronal loss in neostriatum proceeds from medial to lateral and from dorsal to ventral, a pattern confirmed by others [4). In less severely affected cases, Vonsattel and colleagues { 3 } reported that neuronal loss could generally be appreciated in caudate before putamen. One case report, however, of a man at risk for HD who suffered an accidental death describes neostriatal atrophy primarily in the putamen [5 } . Although postmortem neuropathological studies largely involve assessment of advanced cases, in vivo neuroimaging has the potential to provide answers about the pattern of neostriatal degeneration in mild HD. Both magnetic resonance imaging (MRI) and X-ray computed tomography (CT) scans are commonly
read as qualitatively normal in early HD cases, and are thus considered to be of marginal use in aiding clinical diagnosis. Caudate atrophy can be detected by quanrified means, such as with the bicaudate ratio (BCR; see Methods) using CT {b- 11). Differences between patients with early HD and control subjects as measured by BCR are small, however, with a substantial overlap between groups. In more advanced cases, abnormalities are usually detectable on clinical readings of MRIs or CTs, which often reveal widespread cortical and subcortical atrophy, such as described in one MRI study that reported qualitative subcortical atrophy in 4 patients with HD {IZ}. Volumetric measurement may be more sensitive than qualitative assessment or linear measures such as BCR for identifying brain abnormalities in mild HD. Volumes of caudate and putamen are difficult to quantify using CT, however, due to its relatively poor resolution of gray and white matter. MRI is superior for quantifying gray matter, but as yet there has been only
From the Division of Psychiatric Neuro-Imaging, The Department of Psychiatry and Behavioral Sciences, The Johns Hopkins Medical Institutions, Baltimore, MD.
Address correspondence to D r Harris, Division of Psychiatric Neuro-Imaging, Meyer 3-166, Department of Psychiatry, The Johns Hopkins Hospital, 600 North Wolfe St, Baltimore, MD 21205.
Received Apr 23, 1991. Accepred for publication Jun 18, 1991.
Copyright 0 1992 by the American Neurological Association 69
o n e volumetric MRI study in HD, which involved severely demented patients 1131. In these severely affected patients with HD, Jernigan and associates [13] found that the structure m o s t reduced in volume was the caudate nucleus. T h e y did not measure t h e putamen, per se, but instead measured t h e lentiform nucleus (i.e., putamen and globus pallidus). We report t h e results of volumetric measurement of caudate and putamen, as well as HCR, in 15 patients with HD with very mild m o t o r aiid cognitive symptoms.
Methods Subjects Fifteen patients (6 women, 9 men) with mild HD attending the Johns Hopkins Hospital Huntington’s Disease Clinic were included in the study. Mean age (t standard deviation {SD]) was 43.2 t 12.9 years (range, 29-72 yr). Table 1 presents the demographics and clinical scores of each group. Diagnostic criteria for definite HD 114) were (1) chorea or the characreristic impairment of voluntary movement, which was not present at birth, was insidious at onset, and had gradually become worse; and (2) a family history of at least one other member with these typical symptoms of HD. We selected patients who had the lowest scores on their Quantified Neurcilogic Exam (QNE), compatible with a confident diagnosis. ‘The QNE is a coded, reliable, and valid version of the clinical neurological examination of the motor system ClS}. Included in the Q N E is a motor impairment subscale (MIS1, which measures voluntary mot-or impairment, and a chorea subscale. Patients were excluded who had Q N E scores greater than 45; most patients scored below 30 (25.3 10.2; range, 15-43). Patients had mean Mini-Mental State Exam (MICISE) El61 scores of 27.7 ? 1.9 out of a possible 10 (range, 25-30), indicating, at most, mild cognitive deficit. The control group consisted of 19 age- and sex-matched control subjects (6 women, 13 men), age 45.2 I+_ 15.9 years
*
Table I . Demographic Infomiation for Study G’roups“ Huntington’s Disease
Variable n Age Sex (FIM) Q N E (max
MIS (max
=
-=
105) 27)
Chorea (max = 2 5 ) MMSE (max = 30)
15 43.2 ? 12.yt’ (29- 7 2) 619 25.3 2 10.2 (15-43) 5.0 2 3.0 (1-10) 7.3 I 3.4 (3-14) 27.7 t 1.9 (25-30)
Control Subjects 19
45.2
15.9
2
(23-73)
6113 NIA NIA
NIA NIA ~~
.‘Neurological rests and MMSE were given to the liuntington’s disease group o n l y . ‘Values are mean f standard deviation (min-max).
Q N E = Quitntitied Neurologic Exam; MIS = motor impairment suhscale; MMSE = Mini-Mental State Exam; NIA = not applicable.
70
Annals of Neurology
Vol 31 N o 1 January 1992
(range, 23-73 yr), who had no history of neurological illness, including head trauma causing unconsciousness for more than 1 hour. Exclusion criteria for the control subjects also included history of any psychiatric illness (assessed using the SCID 117)). This project was approved by the Johns Hopkins Joint Committee on Clinical Investigations, and informed consent was obtained from all subjects.
Magnetic Resonance Imaging All patients and control subjects were scanned with tine same 1.5 tesla GE Signa MRI scanner. A set of contiguous 5-mm thick transaxial images was obtained parallel to the anteriorcommissurelposterior-commissure (AC-PC) line for each patient and control subject, which included the entire cerebral cortex and cerebellum. Number of excitations was one. Both T2-weighted and proton-weighted (repetition time = 2,500; echo time = 20180) images were acquired simultaneously. Images were stored on %track magnetic tape and transferred to a DECsystem 5400 RISC server computer, where data were archived on SONY readwrite magneto-optical discs.
Quantitative Measurements MRI slices containing structures of interest were identified blind to diagnosis from images displayed on a DECstation 3100 RISC workstation with a color graphics monitor. Graphics software was developed in-house in “C”/”XWindows” in the ULTRIX ( U N I X ) environment. Structures of interest were outlined interactively, blind to diagnosis, using proton-weighted images on all slices on which they appeared. Areas over all slices were summed and multiplied by slice thickness to compute volumes. All MRI basal ganglia volume ratings were made by the same rater based on a standardized computer-based atlas of MRI basal ganglia neuroanaromy (obtainable from G. D. Pearlson by request). Total caudate (head of caudate only) and total putamen volumes were determined. Examples of caudate and putamen measurements for a patient with HD and a control subject at one slice level are shown in Figure 1. Validity and interrater reliability of this method haw been described previously for several brain structures [ 18). W e determined intrarater and interrater reliability for caudate and putamen volume calculations as further validation. Intrarater reliability for subcortical volume measurements made on two separate occasions gave a correlation coefficient greater than 0.99 for both caudate and pucamen. Interrater reliability yielded a correlation coefficient o f 0.97 for c,ludate and 0.94 for putamen volume. All correlations were :;ignificant (p < 0.001). There were no statistically signifiant differences between means on either interrater or intrarattr reliability trials. T o compare the subjects in our group with those of other studies that used BCR as a measure of caudate atrophy, we also measured BCRs. This technique involves measuring the minimal distance between the caudate indentations of the frontal horns divided by the distance between [he oui-er tables of the skull along the same line, then multiplied by 1 0 0 C11!. Whole brain volume (includes brain and all cerebrospinal fluid [CSF}), total CSF volume (includes ventricular and subdural), and total CSF percentage were calculated using a
F i g I . Transaxial magnetic resonance images at the level of the basal ganglia of a patient with early Huntington’s disease (HD) (left: 30-year-old man; Quantified Neurologic Exam = 18; motor impairment subsrale = 3; chorea = 6; MiniMental State Exam = 27) and control subject (right: 45-yearold man). Caudztes and putamen are manually outlined, as was done in this study. Even though this is a wevy early case of HD with minimaf neurological abnormality, the putamen is drastically reduced in size in the patient with HD. Caudates, however, appear similar. semiautomatic method described previously by our group [ 191. This method uses T2-weighted minus proton-weighted MRI images to produce CSF highlighting, then segments the CSF based on thresholding, and uses semiautomated outlining to calculate whole brain volume. CSF percentage is calculated by dividing CSF volume by whole brain volume. Whole brain volume could not be calculated for 5 patients with HD and 1 control subject because these image sets did not include slices near the apex of the head. The area of the brain slice at the level of the BCR measurement, however, (“brain area”) was determined for all cases using rhe semiautomated outlining method described. This area measurement was then used to normalize all volumetric measurements to account for head size differences between individuals. Significance testing was carried out using two-tailed Student’s t-tests to determine differences between groups. Discriminant function analysis (DFA) was used to ascertain which brain measures best identified diagnostic groups. All DFA scores reported refer to analysis based on each individ-
ual variable assessed separately. Regression analysis was used to determine which structural measures best correlated with clinical state exam scores and to control for the effects of age.
Results Regional Measurements
Although all three subcortical measures were significantly different between patients with HD and control subjects (Table 2), putamen volume was by far the best discriminator. The putamen had lost its characteristic convex lens shape in patients with H D , appearing rather as thin strips, as shown in Figure 1. Mean total putamen volume was reduced 50.1% in patients with HD compared with control subjects. DFA of total putamen volume alone had 93.3% sensitivity (14 of 15) and 94.7% specificity (18 of 19). There was overlap between one patient with HD and one control subject, resulting in 94.1% correct classification. All but 1 patient with HD had a putamen volume below 6.0 ml, whereas all but 1 control subject had a putamen volume greater than 7.8 ml. The 2 misclassified subjects were a 73-year-old male control subject, and a 45year-old male patient with very early HD (QNE = 16; MMSE = 30; duration of illness = 2 yr). After correcting for head size by dividing putamen volume ( x 100) by brain area, there was a 50.4% difference Harris et al: Reduced MRI Putamen in Mild HD
71
Table 2. Regional Measurements by Diagnojtic Groupa Huntington’s Disease
DFA%
- Control
Brain area
4.43 f 1.47b (1.36-7.47) 2.67 f 0.87 (0.75-4.22) 2.76 ? 0.63 (1.69-3.77) 1.66 +- 0.38 (0.97-2.48) 12.8 +- 2.3 (8.0-16.1) 166.4 i_ 12.4
8.88 k 1.19 (6.46-11.29) 5.39 2 0.65 (4.31-6.54) 3.82 ? 1.18 (1.5 5-6.03) 2.31 k 0.69 ( 1.04-3.54) 9.0 f 1.7 (6.5-1 3.9) 165.9 i 11.2
Brain volume
1438 f 133
1397
163
...
CSF volume
158.8
124.6 ? 57.3
...
CSF percent
11.1
8.8 +- 3.1
...
Putamen (ml) Putmenibrain Caudate (ml) Caudate/brain BCR
f
73.3
* 5.3
f
94%
P
t
9.76
*
ll
*
It B
, 10
20
30
40
CAUDATE r = -0.11 p = 0.71
...................
v -
50
QNE
Fig 5. Regression plots of putamen and cuuabe volumes with Quatzti$ed Neurologic Exam (QNE)scores for patients with Huntington’s diseuxe. Putamen volume is significantly correlated with QNE (r = -0.60; n = 15; p = 0.018), whereus cuudate wlume i.c not (r = - 0.11: p = 0 . ; 7 1 )
Consistent with prior CT imaging studies, we found significant differences in BCRs in HD with a degree of overlap between patients with HD and control subjects. The degree of overlap and the group means for BCR in our study were comparable to the results reported in prior CT studies of BCR in HD [6-111, indicating that MRI produces similar results as CT for the calculation of BCR. Only 9 of 15 (6096) patients with early HD had BCR values that were more than 2 SDs above the control mean. Caudate volume, as measured using MRI in this study, proved to be relatively insensitive at discriminating patients with HD from control subjects. None of the subjects in either group had caudate volumes more than 2 SDs below the mean control value. BCR was a better measure of the changes in the caudate nucleus than was overall caudate volume for classifying patients with HD. One possible explanation for this finding is that caudate atrophy in HD occurs in a specific pattern (e.g., medial to lateral), and only the medial portion is affected early in the disease [3].Therefore, the relative change in overall caudate size may be small, whereas the BCR (a measure of the distance between the medial caudates) is more sensitive to medial caudate atrophy. There has been one previous study using quantitative MRI in HD [13]. Jernigan and colleagues [13} studied severely demented patients with advanced H D , a stage of the illness when severe subcortical and cortical atrophy is usually clearly visible on clinical inspection of CT or MRI scans. These investigators measured volumes of caudate and lenticular nucleus (putamen and globus pallidus) rather than putamen alone, as we did. They found that in advanced HD, caudate volume reduction was greater than that of lenticular nucleus. Yonsattel and associates [3:t, in a large autopsy study, found that globus pallidus atrophy is modest compared to the caudate or putamen. Therefore, it is 74
Annals of Neurology
Vol 31 No 1 January 1992
not surprising that in Jernigan and colleagues’ [ 131 study, volumes of putamen and globus pallidus measured together were less reduced than caudate volumes. O n the basis of the neuropathology literaturc, especially the study of Vonsattel and colleagues {13}, o u r findings were not anticipated. Although there is general agreement that in advanced HD atrophy arid neuronal loss in caudate and putamen is approx.imately equal [20, 211, in the less-advanced cases studied by Vonsattel and colleagues [I37 (Grades 0 and l), no gross atrophy was appreciated in neostriatum, and microscopic neuronal depletion was noted in caudate but not putamen (the cellular population of the head of the caudate, but not the putamen, was assessed quantitatively and found to be reduced) [ 3 ] . Morphometric study of putamen was not done on these patients, nor were putamen masses reported. It is possible that visual impressions of gross atrophy are not as accurate and sensitive as volumetric or morphometric assessments, even in the hands of highly experienced observers. It is also possible that our patients with early HD were biased toward a population who tended to have relatively more damage to putamen than caudat’e. Currently, a clinical diagnosis of HD cannot be made until patients show clear motor signs, even when they seem to be experiencing disabling cognitive and emotional symptoms that are typical of HD. It is possible that such patients might have relatively more caudate than putamen damage in early illness. We follow a number of such “at-risk” patients in our clinic, but they were not included in this study. Longitudinal volumetricMRI studies of populations of patients with HD and presymptomatic subjects with the HD gene might provide answers to these questions. It is not clear at what point in HD subtle putamen atrophy begins; it is even possible that putamen might be congenitally smaIIer in those inheriting the HD gene. MRI volumetric studies of gene-positive, presymptomatic subjects can address these issues. Neurological exam scores correlated significantly with putamen but not caudate volume. This finding is in accordance with studies indicating that a variety of motor circuits involve putamen but not caudate, whereas cognitive circuits are thought to involve caudate but not putamen [22, 231. Our patients witlh early HD were only mildly cognitively impaired, with a small range of MMSE scores, so the lack of correlation seen in our study between MMSE and linear or volume measures is not surprising. In summary, MRI volumetric measurements of putamen in our sample of 15 patients with mild H D , normalized for brain size as measured by brain area at the level of the caudate head, completely discriminated patients with HD from control subjects. Normalized volumetric measurements of caudate discriminated
only 68% of patients with HD from control subjects, whereas BCR discriminated 82%. Our study suggests either that put-men is congenitally small in those inheriting the HD gene, or that putamen atrophy is evident by the time a clinical diagnosis can be made. This research was supported in part by MH-40391 and MH-43775 from the National Institute of Mental Health to G. D. Pearlson, by grant RR-0722 from the Outpatient Clinical Research Center, National Institutes of Health Division of Research Resources to G. D. Pearlson and S. E. Folstein, by National Institute of Communicative Disorders and Stroke grant NS-16375, and by the Foundation for the Care and Cure of Huntington’s disease, to S. E. Folstein. We would like to thank R. Nick Bryan, MD, PhD, and John C. Hedreen, MD, for their critiques of the manuscript.
References 1. Bruyn GW. Huntington’s chorea: historical, clinical and labora-
2. 3.
4.
5.
6. 7.
8.
9.
tory synopsis. In: Vinken PJ, Bruyn GW, eds. Handbook of clinical neurology, vol 6. Amsterdam: North-Holland Publishing, 1968:298-378 Folstein SE. Huntington’s disease: a disorder of families. Baltimore: The Johns Hopkins University Press, 1989 Vonsattel JP, Meyers RH, Stevens TJ, et al. Neuropathologic classification of Huntington’s disease. J Neuropathol Exp Neurol 1985;44:559-577 Roos RAC, Pruyt JFM, deVries J, Bots GTA. Neuronal distribution in the putamen in Huntington’s disease. J Neurol Neurosurg Psychiatry 1985;48:422-425 Carrosco LH, Mukherji CS. Atrophy of corpus striatum in normal male at risk of Huntington’s chorea. Lancet 1986;l: 1388-1389 Starkstein SE, Folstein SE, Brandt J, et al. Brain atrophy in Huntington’s disease: a CT scan study. Neuroradiology 1989; 3 1:156-1 59 Starkstein SE, Brandt J, Folstein SE, et al. Neuropsychological and neuroradiological correlates in Huntington’s disease. J Neurol Neurosurg Psychiatry 1988;51:1259-1263 Oepen G, Ostertag CH. Diagnostic value of CT in patients with Huntington’s chorea and their offspring. J Neurol 1981;225: 189-196 Sax DS, ODonnell B, Butters N, et al. Computed tomographic,
neurologic and neuropsychological correlates of Huntington’s disease. Int J Neurosci 1983;18:21-36 10. Bamford KA, Caine ED, Kid0 DK, et al. Clinical-pathologic correlation in Huntington’s disease: a neuropsychological and computed tomography study. Neurology 1989;39:796-801 11. Barr AN, Heinze WJ, Dobben GD, et al. Bicaudate index in computerized tomography of Huntington’s disease and cerebral atrophy. Neurology 1978;28:1196-1200 12. Simmons JT, Pastakia B, Chase TN, Shults CW. Magnetic resonance imaging in Huntington’s disease. AJNR 1986;7:25-28 13. Jernigan TL, Salmon DP, Butters N, Hesselink JR. Cerebral structure on MRI, part 11: specific changes in Alzheimer’s and Huntington’s diseases. Biol Psychiatry 1991;29:68-8 1 14. Folstein SE, Leigh JR, Parhad IM, Folstein MF. The diagnosis of Huntington’s disease. Neurology 1986;36:1279-1283 15. Folstein SE, Jensen B, Leigh JR, Folstein MF. The measurement of abnormal movement: methods developed for Huntington’s disease. Neurobehav Toxic01 Teratol 1983;5:605-609 16. Folstein MF, Folstein SE, McHugh PR. “Mini Mental State.” A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975;12:189-198 17. Spitzer RL, Williams JBW, Gibbon N, et al. Structured clinical interview for DSM-IIIR. New York: New York State Psychiatric Institute, Biometrics Research, 1989 18. Barta PE, Pearlson GD, Powers RE, et al. Auditory hallucinations and smaller superior temporal gyral volume in schizophrenia. Am J Psychiatry 1990;147:1457-1462 19. Harris GJ, Rhew EH, Noga JT, Pearlson GD. A user-friendly method for rapid brain and CSF volume calculation using transaxial MRI images. Psychiatr Res: Neuroimag 1991;40: 61-68 20. Lange H , Thorner G , Hopf A, Schroder KF. Morphometric studies of the neuropathological changes in choreatic diseases. J Neurol Sci 1976;28:401-425 2 1. De la Monte SM, Vonsattel JP, Richardson EP. Morphometric demonstration of atrophic changes in the cerebral cortex, white matter, and neostriatum in Huntington’s disease. J Neuropathol Exp Neurol 1988;47:516-525 22. DeLong MR, Alexander GE, Miller WC, Crutcher MD. Anatomical and functional aspects of basal ganglia-thalamocortical circuits. In: Franks AJ, Ironside JW, Mindham RHS, et al, eds. Function and dysfunction in the basal ganglia. Manchester and New York: Manchester University Press, 1990:3-34 23. Alexander GE, DeLong MR, Strick PL. Parallel organization of functionally segregated circuits linking basal ganglia and cortex. Ann Rev Neurosci 1986;9:357-381
Harris et al: Reduced MRI Putamen in Mild HD
75
