MRI for the differential diagnosis of neurodegenerative parkinsonism in clinical practice

Transcription

1 Parkinsonism & Related Disorders Parkinsonism and Related Disorders 13 (2007) S400 S405 MRI for the differential diagnosis of neurodegenerative parkinsonism in clinical practice Klaus Seppi* Department of Neurology, Medical University Innsbruck, Austria Abstract Parkinson s disease (PD) is the most common neurodegenerative cause of parkinsonism, followed by progressive supranuclear palsy (PSP) and multiple system atrophy (MSA). Despite the publication of consensus operational criteria for the diagnosis of PD and the various atypical parkinsonian disorders (APD) such as PSP, MSA and corticobasal degeneration, an accurate diagnosis of neurodegenerative parkinsonian syndromes remains a challenge for each neurologist. Particularly in the early disease stages the clinical separation of APDs from PD carries a high rate of misdiagnosis. However, an early differentiation between APD and PD, each characterized by completely different natural histories, is crucial for determining the prognosis and choosing a treatment strategy. MRI plays an important role in the exclusion of symptomatic parkinsonism due to other pathologies. Over the past two decades, conventional MRI and advanced MRI techniques, including proton magnetic resonance spectroscopy ( 1 H-MRS), diffusion-weighted imaging (DWI), magnetization transfer imaging (MTI), and magnetic resonance volumetry (MRV) have shown abnormalities in the substantia nigra and basal ganglia, especially in APD. Furthermore, in accordance with neuropathological studies suggesting that the olfactory system is an early target of the disease, recent studies using advanced MRI techniques have shown abnormalities in the olfactory system in the early disease stages of patients with PD. Given that olfactory deficits may be a premotor marker of the disease, such methods may eventually evolve into an early screening tool for PD Elsevier B.V. All rights reserved. Keywords: Parkinson s disease (PD); Progressive supranuclear palsy (PSP); Multiple system atrophy (MSA); Corticobasal degeneration (CBD); Atypical parkinsonism; Magnetic resonance imaging (MRI); Differential diagnosis 1. Introduction Despite the publication of consensus operational criteria for the diagnosis of Parkinson s disease (PD) and the various atypical parkinsonian disorders (APD) such as progressive supranuclear palsy (PSP) and multiple system atrophy (MSA) [1], an accurate diagnosis of neurodegenerative parkinsonian syndromes remains a challenge for each neurologist. Particularly in the early disease stages, both the clinical separation of APD from PD and the determination of the appropriate APD sub-type carry a high rate of misdiagnosis. However, an early differentiation between APD and PD, each characterized by completely different natural histories, is crucial in determining the prognosis and choosing a treatment strategy. 2. Conventional MRI (cmri) An important role of cmri in the differential diagnosis of parkinsonism is the differentiation between neurodegenera- * Correspondence: Anichstrasse 35, A-6020 Innsbruck, Austria. Tel.: ; fax: address: klaus.seppi@uki.at (K. Seppi). tive and symptomatic parkinsonism due to other pathologies like multiple sclerosis, brain tumors, normal pressure hydrocephalus, vascular etiology, or other causes [2,3]. Furthermore, cmri is believed to be usually normal in patients with PD, while it frequently shows characteristic abnormalities in patients with APD, thus offering the potential for objective criteria in the differential diagnosis among the different forms of neurodegenerative parkinsonism. Following the description of signal alteration and atrophy in the putamen in 1986 [4,5], several subsequent MRI studies have reported on further features found in patients with neurodegenerative parkinsonism. Not only PD patients, but also APD patients may show distinct abnormalities of the substantia nigra, including signal increase on T2-weighted MR images, smudging of the hypointensity in the substantia nigra towards the red nucleus or signal loss when using inversion recovery MRI [3,6]. Overall, specificity of atrophy and signal changes in the putamen as well as in infratentorial structures on cmri in patients with MSA [2,3,7] is considered quite high, while sensitivity seems to be suboptimal especially in the early disease stages [3]. Sensitivity may be moderately improved by modifying some technical aspects like slice thickness / $ see front matter 2007 Elsevier B.V. All rights reserved.

2 K. Seppi / Parkinsonism and Related Disorders 13 (2007) S400 S405 S401 and the use of conventional spin-echo or T2 -weighted gradient echo sequences [3]. A number of findings suggestive of PSP have been described, such as midbrain atrophy with enlargement of the third ventricle, reduced anteroposterior midbrain diameter and tegmental atrophy, signal increase in midbrain and inferior olives, as well as frontal and temporal lobe atrophy [3,7]. Only few studies have investigated the role of cmri in patients with corticobasal degeneration (CBD), showing cortical atrophy especially in frontoparietal areas, which usually seems to be asymmetric, putaminal hypointensity, and hyperintense signal changes in the motor cortex or subcortical white matter on T2-weighted images [3,7]. By reviewing 40 autopsy cases presenting with the clinical diagnosis of a corticobasal syndrome (CBS) during life including CBD and other neurodegenerative causes such as PSP and frontotemporal degenerations and having at least one MRI examination, similar patterns of regional atrophy as well as subcortical and periventricular white matter signal changes were found in patients with either post-mortem confirmed CBD or other neurodegenerative conditions presenting as CBS [8]. 3. Quantitative assessment of regional cerebral atrophy using MRI planimetry and volumetry Using MRV with semi-automatic segmentation techniques, volume loss of different supratentorial and infratentorial brain structures in patients with MSA and PSP has been reported [9 12]. Patients with MSA showed significant reductions in mean striatal, brainstem and cerebellar volumes [9,10], whereas patients with PSP had significant reductions in whole brain, striatal, brainstem especially midbrain and frontal cortical volumes [9,11,12]. In the only study using the volumetric approach in CBD, the pattern of atrophic changes was compared between 18 CBD patients, 33 patients with PSP, and 22 controls. Patients with CBD showed atrophy of the parietal cortex and corpus callosum [12]. The strength of the latter study is the construction of a mathematical model derived from a discriminant analysis using only post-mortem confirmed cases of PSP (n = 8) and CBD (n = 7) as well as controls. The volumes of midbrain, parietal white matter, temporal grey matter, brainstem, frontal white matter, and pons were identified to separate best between groups, predicting the diagnosis correctly in 95% of controls as well as in 76% of all PSP and 83% of all CBD patients. More recently, voxel-based morphometry (VBM) has been used to study neurodegenerative parkinsonian disorders including PD, MSA, and PSP. While approaches with semi-automatic segmentation techniques are operator dependent and only include pre-selected brain areas, VBM permits an unbiased, operator-independent and semiautomated detection of significant differences in different tissue types of the whole brain avoiding a-priori ROI selection [13]. Whereas ROI-based MRV studies did not detect any differences between PD patients and controls [9 11], VBM revealed grey matter loss of frontal cortical areas including the motor areas in PD patients across the different VBM studies published [14 16]. In confirmation of previous ROI-based MRV studies [9, 10], a VBM study showed basal ganglia and infratentorial volume loss in patients with the parkinsonian variant of MSA (MSA-P) compared to PD patients and controls [17]. This study also revealed prominent cortical volume loss in MSA-P, mainly comprising the cortical targets of basal ganglia output pathways such as the primary sensorimotor, lateral premotor cortices and the prefrontal cortex [17]. Recently, voxel-based relaxometry (VBR), a technique based on voxel-by-voxel calculation of whole brain T2 relaxation rate (R2) maps derived from multi-echo T2-weighted images to assess signal abnormalities, was used to study brain morphology in 14 patients with the cerebellar variant of MSA (MSA-C) versus 13 healthy volunteers [18]. The results of the VBR analysis were then compared with those obtained by VBM in the same cohort published earlier by the same group [19]. The reduction of R2 in the cerebellum and brainstem, reflecting infratentorial brain atrophy, largely corresponded to those regions in which VBM showed reductions of grey and white matter. On the other hand, the inverted VBR analysis revealed increased R2 in the putamen that was not detected by VBM and was clearly different from the increase of white matter along the pyramidal tract observed using VBM [18,19]. In PSP patients, grey matter loss was reported in frontotemporal cortical areas including the prefrontal cortex, the insular region comprising the frontal opercula, the supplementary motor areas and the left mediotemporal area, in comparison to healthy volunteers [20]. White matter loss was found in frontotemporal areas and in the midbrain, including the cerebral peduncles and the central midbrain. When testing the clinical utility of the VBM results as a guide for the differentiation of PSP from PD and controls [21], based upon neuroradiological review of the T1-weighted MR images, the study participants were classified as either PSP or non-psp guided by the presence or absence of midbrain tissue loss in the PSP group as highlighted using VBM. With these regional differences on VBM as a guide, the neuroradiological diagnosis achieved a sensitivity of 83% and a specificity of 79% [21]. Even though VBM has been used to investigate the progression of cortical and subcortical atrophy patterns in MSA-P compared to PD [22] and led to the suggestion that early degeneration of the basal ganglia drives late onset cortical atrophy, up to now it cannot be applied for routine diagnostic work-up of individual patients. More recently, different groups have applied simple quantitative measures of diameters and areas of different structures on MRI for the differential diagnosis of MSA from other neurodegenerative parkinsonian disorders. Oba and co-workers [23] measured the surface area of the midbrain and pons at 1.5 T in patients with PSP (n = 21),

3 S402 K. Seppi / Parkinsonism and Related Disorders 13 (2007) S400 S405 PD (n = 23), MSA-P (n = 25) and 31 age-matchedcontrols. The surface area of the pons was significantly smaller in the MSA-P group than in the PSP group, the PD group, and the normal control group. However, individual values of the pontine surface area overlapped in the different groups studied. The mean ratio of the surface area of the midbrainto the surface area of pons in the patients with PSP was significantly smaller than that in PD, MSA-P, and in normal control subjects, allowing complete discrimination of PSP from MSA-P, while there was an overlap between patients with MSA-P and PD and controls. Nicoletti and co-workers [24] studied middle cerebellar peduncle (MCP) atrophy in 16 patients with MSA (13 of them having MSA-P), 26 patients with PD, and 14 healthy controls by measuring the MCP width on T1-weighted volumetric spoiled gradient-echo images at 1.5 T. MCP widths were then correlated with abnormalities on MRI including putaminal atrophy, putaminal hypointensity, slit-like hyperintensity in the posterolateral margin of the putamen, brainstem atrophy, hyperintensity of the MCP, and cruciform hyperintensity of the pons. All patients with MSA had at least one of the abnormal features commonly observed in MSA on MRI, whereas control subjects and all but one patient with PD (having a putaminal rim sign) had normal MRI. The average MCP width was significantly smaller in patients with MSA than in those with PD or control subjects, without any overlap between the MSA group and the PD or control group when using a cut-off value of 8.0 mm. 4. Quantitative structural MR-based techniques: Diffusion-weighted imaging and magnetization transfer imaging Diffusion-weighted (DWI) and magnetization transfer (MTI) imaging may represent useful diagnostic tools that can provide additional support in the differential diagnosis of patients with neurodegenerative parkinsonism [3]. MTI is based on the interactions between highly bound protons within structures such as myelin or cell membranes and the very mobile protons of free water [25]. By application of irradiation that selectively saturates the energy level of bound protons, exchange of magnetization between bound and free protons is induced and the signal intensity of bound protons is reduced. The difference between the signal intensities with and without magnetization transfer (MT) is measured by calculating the MT ratios (MTRs), which correlate with the degree of myelinisation and with axonal density [26]. DWI visualises the random movement of the water molecules in the tissue by applying diffusion-sensitised gradients between two radiofrequency (rf) pulses [27]. The signal of the water molecules decreases with the extent of diffusion between the two rf pulses. An absolute quantification of the diffusivity is achieved by applying diffusion-sensitized gradients of different degrees in three orthogonal directions and calculating the apparent diffusion coefficient (ADC) for each direction, which forms the trace of the diffusion tensor, Trace(D) [28]. The CNS is highly organised in numerous tracts of myelinated fibre bundles, whereby the movement of the water molecules is restricted perpendicular to these fibre bundles. The resulting anisotropic diffusion is quantified by the fractional anisotropy (FA), which is determined by diffusion-sensitised gradients in at least six directions. Both the diffusivity and the FA form the diffusion tensor [27]. By using MTI in patients with neurodegenerative parkinsonian disorders, abnormalities of the basal ganglia and substantia nigra (SN) have been reported in patients with PD, MSA and PSP [29 31]. The most recent of these studies investigated the potential of MTI in the differential diagnosis of neurodegenerative parkinsonism, including 37 patients with different parkinsonian syndromes and 20 agematched controls [31]. The main finding in this study was a change in the MTR in the globus pallidus, putamen, caudate nucleus, substantia nigra, and white matter in PD, MSA, and PSP patients, matching the pathological features of the underlying disorder. By application of stepwise discriminant analysis, there was a good discrimination of PD patients and controls from the MSA and PSP patients, with only one MSA patient wrongly classified into the control group. On the other hand separation between PD patients and controls as well as between MSA and PSP patients was suboptimal [31]. Pathological DWI findings in PD patients are very rare. One recent study detected a decrease in FA in the SN and in 11 ROIs along a line between SN and striatum segmented manually on transverse slices of seven patients with early PD [32]. The authors interpreted this finding as a sign of the well-known degeneration of the nigrostriatal projection in early PD. Additionally, reduced FA values were found in the white matter of the premotor area of the advanced PD cases in this study [32]. A more recent study using DWI and statistical parametric mapping (SPM), [33] the latter to objectively localize focal changes of structural neuronal integrity without having to make an a priori hypothesis as to its location, localized significant increases of diffusivity in the region of both olfactory tracts in patients with PD compared to healthy controls. This observation is in line with the well-established clinical finding [34] of hyposmia in Parkinson s disease. Several recent studies have shown that DWI permits to discriminate between MSA-P in early disease stages and PD as well as healthy humans on the basis of increased putaminal ADC and Trace(D) values, which also correlate with disease severity [3,28,35 38]. A further study investigated DWI in MSA-C patients versus controls, describing an increase of the ADC in the pons, the middle cerebellar peduncle, the cerebellar white matter and the putamen in the MSA-C patients [39]. A more recent study in 15 patients with MSA-P and 17 patients with PD compared DWI with dopamine D2 receptor (D2R) binding, using

4 K. Seppi / Parkinsonism and Related Disorders 13 (2007) S400 S405 S403 single-photon emission computed tomography (SPECT) with [ 123 I]-iodobenzamide (IBZM), which in the past has been shown to contribute to the differential diagnosis of MSA from PD by showing reduced striatal D2R binding in patients with MSA-P compared to PD patients. Striatal ADCs were found to have a significantly higher predictive accuracy than D2R binding with IBZM (97% versus 75%) suggesting that DWI may be more accurate than IBZM SPECT in the differential diagnosis of MSA-P versus PD [40]. In agreement with the above findings, Nicoletti and co-workers [24] reported a complete separation of MSA-P patients from both controls and PD patients on the basis of putaminal Trace(D) values at 1.5 T. These workers examined 16 MSA-P patients, 16 patients with PSP, and 16 patients with PD, as well as 15 healthy volunteers. Whereas putaminal Trace(D) values showed overlap between MSA-P and PSP patients, MCP Trace(D) values provided complete separation between MSA-P patients and patients with PD as well as healthy controls. Similar to a previous study [35], there was overlap of putaminal ADC values between patients with PSP and PD. A more recent study [38] compared ADC and fractional anisotropy (FA) values in the pons, cerebellum and putamen of 61 subjects (MSA, n = 20; MSA-C and MSA-P each n = 10; PD, n = 21 and healthy controls, n = 20) using a 3.0 T MR system. ADC values in the pons, cerebellum and putamen were significantly higher, and FA values lower, in MSA compared to PD and controls. Sensitivity and specificity of FA values for the differentiation of MSA-P versus PD were 70.0% and 100.0% in the pons, 70.0% and 63.6% in the cerebellum, 70.0% and 87.5% in the putamen; sensitivity and specificity for ADC values were 70.0% and 70.0% in the pons, 60.0% and 87.5% in the cerebellum, 70.0% and 63.6% in the putamen. In addition, all patients who had significant low FA and high ADC values in each of these three areas were MSA-P cases, whereas those who had both normal FA and ADC values in the pons were all PD cases. Normal FA and ADC values were found in the cerebellum and putamen of 2 and 1 MSA-P cases, respectively. Based on these results, the authors devised an algorithm for differentiating probable MSA-P from PD. Using this algorithm in the 31 cases (PD: 21, MSA-P: 10), all patients who had both normal FA and ADC values in all three areas were PD cases (12 PD cases, 57%). In addition, all patients who had both low FA and high ADC values in any of the three areas were MSA-P cases (9 MSA-P cases, 90%), while 10 cases (9 PD cases and 1 MSA-P case) had low FA but normal ADC values or normal FA but high ADC values in any of the three areas. When studying 11 patients with MSA-P, 20 patients with PSP, 12 patients with PD and 7 healthy controls with DWI at 1.5 T, Paviour and co-workers [41] found that MCP ADC values discriminated MSA-P patients from all other conditions (PSP, PD, healthy controls) with a sensitivity of 91% and a specificity of 82%. Significant group differences of ADCs were found also in the rostral pons. On the other hand, ADCs in the putamen or in the SCP did not differ between groups. This clearly contrasts to previous reports showing significantly increased putaminal diffusivity in patients with MSA-P and significantly decreased ADCs in the SCP in PSP patients. Methodological and demographic reasons might explain this discrepancy. Disease duration especially of the PD group in the paper by Paviour et al. [41] was longer than in the previous papers [28,35 37]. The DWI MR protocol with a slice thickness of 7 mm compared to 3 5 mm used in other studies [28,35 37] might be inadequate. Furthermore, segmentation errors might have occurred, as indicated by the published figure displaying the ROIs: the delineated ROI of the putamen clearly contains parts of the globus pallidus; also the actual thalamus is larger than segmented in the figure. In another study, changes in FA and mean diffusivity in the MCP, decussation of the SCP and pons were measured in 17 patients with MSA (MSA-P, n = 10; MSA-C, n = 7), 17 with PSP, 12 with PD, and 12 healthy volunteers, without giving measures of sensitivity and specificity [42]. In MSA, FA was significantly reduced in the MCP and mean diffusivity was increased both in the MCP and in the pons compared with the other groups. In PSP, mean diffusivity was strikingly increased in the decussation of the SCPs compared with the other groups. 5. Proton magnetic resonance spectroscopy ( 1 H-MRS) In vivo 1 H-MRS visualises signals from carbon-bound, non-exchangeable protons, showing the highest information density in the spectral region from one to five parts per million [43]. Principal metabolite signs detected by 1 H-MRS include N-acetylaspartate (NAA) as an indirect expression of the integrity of neurons, choline (Cho) containing compounds as metabolites involved in phospholipid membrane synthesis and marker for glial activity, creatine (Cr) including phosphocreatine as a marker for energy metabolism, and lactate as an indicator for anaerobic glycolysis detected under pathological conditions [44,45]. Although the first 1 H-MRS study published in patients with neurodegenerative parkinsonism suggested that 1 H-MRS of striatal structures might differentiate parkinsonian syndromes by virtue of reduced NAA/Cr ratios in MSA but not PD [46], further MRS studies have shown reduced NAA/Cr and NAA/Cho ratios in the lentiform nucleus or striatum not only in APD, but also in PD [44,47]. Technical factors such as MRS technique including different echoand relaxation times, voxel sizes and pulse sequences used in the different studies, may be responsible for some of the variation of results seen in the published literature on 1 H-MRS for the differential diagnosis of neurodegenerative parkinsonism [44,47]. The development of 1 H-MRS at higher magnetic field strengths may lead 1 H-MRS to a more important role as an imaging tool in the differential diagnosis of parkinsonian

5 S404 K. Seppi / Parkinsonism and Related Disorders 13 (2007) S400 S405 disorders. This is illustrated by a recent study applying multiple regional single voxel 1 H-MRS including putamen, pontine base and cerebral white matter at 3.0 T in 24 patients with MSA compared to 11 PD patients and 18 controls [48]. Significant NAA/Cr reductions were found in the pontine base of both patients with MSA-C and MSA-P, while putaminal NAA/Cr was only reduced in the patients with MSA-P. Eight of the 11 MSA-P patients compared to none of the PD and control group were classified correctly by combining individual NAA/Cr reductions in the pontine base and in the putamen. These results suggest that combined assessment of NAA/Cr in the pontine base and putamen may be effective in differentiating MSA-P from PD in terms of the high specificity of reduced NAA/Cr in the pontine base or in the putamen in patients with MSA-P [48]. 6. Conclusions This article has reviewed applications of MRI in the differential diagnosis of neurodegenerative parkinsonism. Conventional MRI is a precious tool for the exclusion of symptomatic causes of parkinsonism. Nonetheless, it has been established that cmri is not sufficiently sensitive for the differential diagnosis among neurodegenerative parkinsonian disorders. The role of 1 H-MRS for the differential diagnosis of neurodegenerative parkinsonian disorders is inconclusive and unclear, possibly having increased accuracy when using higher field strengths. MRV, MTI, and DWI may help to differentiate between APD and PD and provide close insight into the pathophysiology of APD. Whether these techniques will emerge as standard investigations in the work-up of patients presenting with parkinsonism requires further prospective MR studies during early disease stages. Conflict of Interest statement None declared. References [1] Litvan I, Bhatia KP, Burn DJ, Goetz CG, Lang AE, McKeith I, et al. Movement Disorders Society Scientific Issues Committee report: SIC Task Force appraisal of clinical diagnostic criteria for Parkinsonian disorders. Mov Disord 2003;18: [2] Heiss WD, Hilker R. The sensitivity of 18-fluorodopa positron emission tomography and magnetic resonance imaging in Parkinson s disease. Eur J Neurol 2004;11:5 12. [3] Seppi K, Schocke MF. An update on conventional and advanced magnetic resonance imaging techniques in the differential diagnosis of neurodegenerative parkinsonism. Curr Opin Neurol 2005;18: [4] Drayer BP, Olanow W, Burger P, Johnson GA, Herfkens R, Riederer S. Parkinson plus syndrome: diagnosis using high field MR imaging of brain iron. Radiology 1986;159: [5] Pastakia B, Polinsky R, Di Chiro G, Simmons JT, Brown R, Wener L. Multiple system atrophy (Shy Drager syndrome): MR imaging. Radiology 1986;159: [6] Brooks DJ. Morphological and functional imaging studies on the diagnosis and progression of Parkinson s disease. J Neurol 2000; 247(Suppl 2):II11 8. [7] Savoiardo M. Differential diagnosis of Parkinson s disease and atypical parkinsonian disorders by magnetic resonance imaging. Neurol Sci 2003;24(Suppl 1):S35 7. [8] Josephs KA, Tang-Wai DF, Edland SD, Knopman DS, Dickson DW, Parisi JE, et al. Correlation between antemortem magnetic resonance imaging findings and pathologically confirmed corticobasal degeneration. Arch Neurol 2004;61: [9] Schulz JB, Skalej M, Wedekind D, Luft AR, Abele M, Voigt K, et al. Magnetic resonance imaging-based volumetry differentiates idiopathic Parkinson s syndrome from multiple system atrophy and progressive supranuclear palsy. Ann Neurol 1999;45: [10] Ghaemi M, Hilker R, Rudolf J, Sobesky J, Heiss WD. Differentiating multiple system atrophy from Parkinson s disease: contribution of striatal and midbrain MRI volumetry and multi-tracer PET imaging. J Neurol Neurosurg Psychiatry 2002;73: [11] Cordato NJ, Pantelis C, Halliday GM, Velakoulis D, Wood SJ, Stuart GW, et al. Frontal atrophy correlates with behavioural changes in progressive supranuclear palsy. Brain 2002;125: [12] Groschel K, Hauser TK, Luft A, Patronas N, Dichgans J, Litvan I, et al. Magnetic resonance imaging-based volumetry differentiates progressive supranuclear palsy from corticobasal degeneration. Neuroimage 2004;21: [13] Ashburner J, Friston KJ. Voxel-based morphometry the methods. Neuroimage 2000;11: [14] Burton EJ, McKeith IG, Burn DJ, Williams ED, O Brien JT. Cerebral atrophy in Parkinson s disease with and without dementia: a comparison with Alzheimer s disease, dementia with Lewy bodies and controls. Brain 2004;127: [15] Nagano-Saito A, Washimi Y, Arahata Y, Kachi T, Lerch JP, Evans AC, et al. Cerebral atrophy and its relation to cognitive impairment in Parkinson disease. Neurology 2005;64: [16] Summerfield C, Junque C, Tolosa E, Salgado-Pineda P, Gomez-Anson B, Marti MJ, et al. Structural brain changes in Parkinson disease with dementia: a voxel-based morphometry study. Arch Neurol 2005;62: [17] Brenneis C, Seppi K, Schocke MF, Muller J, Luginger E, Bosch S, et al. Voxel-based morphometry detects cortical atrophy in the Parkinson variant of multiple system atrophy. Mov Disord 2003;18: [18] Specht K, Minnerop M, Muller-Hubenthal J, Klockgether T. Voxelbased analysis of multiple-system atrophy of cerebellar type: complementary results by combining voxel-based morphometry and voxel-based relaxometry. Neuroimage 2005;25: [19] Specht K, Minnerop M, Abele M, Reul J, Wullner U, Klockgether T. In vivo voxel-based morphometry in multiple system atrophy of the cerebellar type. Arch Neurol 2003;60: [20] Brenneis C, Seppi K, Schocke M, Benke T, Wenning GK, Poewe W. Voxel based morphometry reveals a distinct pattern of frontal atrophy in progressive supranuclear palsy. J Neurol Neurosurg Psychiatry 2004;75: [21] Price S, Paviour D, Scahill R, Stevens J, Rossor M, Lees A, et al. Voxel-based morphometry detects patterns of atrophy that help differentiate progressive supranuclear palsy and Parkinson s disease. Neuroimage 2004;23: [22] Brenneis C, Egger K, Scherfler C, Seppi K, Schocke M, Poewe W, et al. Progression of brain atrophy in multiple system atrophy. A longitudinal VBM study. J Neurol 2007;254: [23] Oba H, Yagishita A, Terada H, Barkovich AJ, Kutomi K, Yamauchi T, et al. New and reliable MRI diagnosis for progressive supranuclear palsy. Neurology 2005;64(12): [24] Nicoletti G, Fera F, Condino F, Auteri W, Gallo O, Pugliese P, et al. MR imaging of middle cerebellar peduncle width: differentiation of

6 K. Seppi / Parkinsonism and Related Disorders 13 (2007) S400 S405 S405 multiple system atrophy from Parkinson disease. Radiology 2006; 239: [25] Wolff SD, Balaban RS. Magnetization transfer contrast (MTC) and tissue water proton relaxation in vivo. Magn Reson Med 1989;10: [26] Van Buchem MA, McGowan JC, Grossman RI. Magnetization transfer histogram methodology: its clinical and neuropsychological correlates. Neurology 1999;53(Suppl 3):S23 8. [27] Le Bihan D. Looking into the functional architecture of the brain with diffusion MRI. Nat Rev Neurosci 2003;4: [28] Schocke MF, Seppi K, Esterhammer R, Kremser C, Mair KJ, Czermak BV, et al. Trace of diffusion tensor differentiates the Parkinson variant of multiple system atrophy and Parkinson s disease. Neuroimage 2004;21: [29] Hanyu H, Asano T, Sakurai H, Takasaki M, Shindo H, Abe K. Magnetisation transfer measurements of the subcortical grey and white matter in Parkinson s disease with and without dementia and in progressive supranuclear palsy. Neuroradiology 2001;43: [30] Naka H, Imon Y, Ohshita T, Honjo K, Kitamura T, Miyachi T, et al. Magnetization transfer measurements of brain structures in patients with multiple system atrophy. Neuroimage 2002;17: [31] Eckert T, Sailer M, Kaufmann J, Schrader C, Peschel T, Bodammer N, et al. Differentiation of idiopathic Parkinson s disease, multiple system atrophy, progressive supranuclear palsy, and healthy controls using magnetization transfer imaging. Neuroimage 2004;21: [32] Yoshikawa K, Nakata Y, Yamada K, Nakagawa M. Early pathological changes in the parkinsonian brain demonstrated by diffusion tensor MRI. J Neurol Neurosurg Psychiatry 2004;75: [33] Scherfler C, Schocke MF, Seppi K, Esterhammer R, Brenneis C, Jaschke W, et al. Voxel-wise analysis of diffusion weighted imaging reveals disruption of the olfactory tract in Parkinson s disease. Brain 2006;129: [34] Katzenschlager R, Lees AJ. Olfaction and Parkinson s syndromes: its role in differential diagnosis. Curr Opin Neurol 2004;17: [35] Seppi K, Schocke MF, Esterhammer R, Kremser C, Brenneis C, Mueller J, et al. Diffusion-weighted imaging discriminates progressive supranuclear palsy from PD, but not from the parkinson variant of multiple system atrophy. Neurology 2003;60: [36] Schocke MF, Seppi K, Esterhammer R, Kremser C, Jaschke W, Poewe W, et al. Diffusion-weighted MRI differentiates the Parkinson variant of multiple system atrophy from PD. Neurology 2002;58: [37] Nicoletti G, Lodi R, Condino F, Tonon C, Fera F, Malucelli E, et al. Apparent diffusion coefficient measurements of the middle cerebellar peduncle differentiate the Parkinson variant of MSA from Parkinson s disease and progressive supranuclear palsy. Brain 2006; 129: [38] Ito M, Watanabe H, Kawai Y, Atsuta N, Tanaka F, Naganawa S, et al. Usefulness of combined fractional anisotropy and apparent diffusion coefficient values for detection of involvement in multiple system atrophy. J Neurol Neurosurg Psychiatry 2007;78: [39] Kanazawa M, Shimohata T, Terajima K, Onodera O, Tanaka K, Tsuji S, et al. Quantitative evaluation of brainstem involvement in multiple system atrophy by diffusion-weighted MR imaging. J Neurol 2004;251: [40] Seppi K, Schocke MF, Donnemiller E, Esterhammer R, Kremser C, Scherfler C, et al. Comparison of diffusion-weighted imaging and [(123)I]IBZM-SPECT for the differentiation of patients with the Parkinson variant of multiple system atrophy from those with Parkinson s disease. Mov Disord 2004;19: [41] Paviour DC, Thornton JS, Lees AJ, Jager HR. Diffusion-weighted magnetic resonance imaging differentiates Parkinsonian variant of multiple-system atrophy from progressive supranuclear palsy. Mov Disord 2007;22: [42] Blain CR, Barker GJ, Jarosz JM, Coyle NA, Landau S, Brown RG, et al. Measuring brain stem and cerebellar damage in parkinsonian syndromes using diffusion tensor MRI. Neurology 2006;67: [43] Trabesinger AH, Meier D, Boesiger P. In vivo 1 H NMR spectroscopy of individual human brain metabolites at moderate field strengths. Magn Reson Imaging 2003;21: [44] Firbank MJ, Harrison RM, O Brien JT. A comprehensive review of proton magnetic resonance spectroscopy studies in dementia and Parkinson s disease. Dement Geriatr Cogn Disord 2002;14: [45] Schocke MF, Berger T, Felber SR, Wolf C, Deisenhammer F, Kremser C, et al. Serial contrast-enhanced magnetic resonance imaging and spectroscopic imaging of acute multiple sclerosis lesions under highdose methylprednisolone therapy. Neuroimage 2003;20: [46] Davie CA, Wenning GK, Barker GJ, Tofts PS, Kendall BE, Quinn N, et al. Differentiation of multiple system atrophy from idiopathic Parkinson s disease using proton magnetic resonance spectroscopy. Ann Neurol 1995;37: [47] Clarke CE, Lowry M. Systematic review of proton magnetic resonance spectroscopy of the striatum in parkinsonian syndromes. Eur J Neurol 2001;8: [48] Watanabe H, Fukatsu H, Katsuno M, Sugiura M, Hamada K, Okada Y, et al. Multiple regional 1 H-MR spectroscopy in multiple system atrophy: NAA/Cr reduction in pontine base as a valuable diagnostic marker. J Neurol Neurosurg Psychiatry 2004;75:103 9.