- Email: [email protected]
Differentiating progressive supranuclear palsy from Parkinson's disease by MRI-based dynamic cerebrospinal fluid flow
Journal of the Neurological Sciences 357 (2015) 178–182
Contents lists available at ScienceDirect
Journal of the Neurological Sciences journal homepage: www.elsevier.com/locate/jns
Differentiating progressive supranuclear palsy from Parkinson's disease by MRI-based dynamic cerebrospinal fluid flow Yusuke Fukui, Nozomi Hishikawa, Kota Sato, Taijun Yunoki, Syoichiro Kono, Kosuke Matsuzono, Yumiko Nakano, Yasuyuki Ohta, Toru Yamashita, Kentaro Deguchi, Koji Abe ⁎ Department of Neurology, Graduate School of Medicine, Dentistry and Pharmaceutical Sciences, Okayama University, 2-5-1 Shikatacho, Kitaku, Okayama 700-8558, Japan
a r t i c l e
i n f o
Article history: Received 7 January 2015 Received in revised form 25 June 2015 Accepted 16 July 2015 Available online 21 July 2015 Keywords: Cerebrospinal fluid Midbrain aqueduct Neurodegenerative disorder Phase contrast-magnetic resonance imaging Parkinson's disease Progressive supranuclear palsy
a b s t r a c t Objective: The purpose of this study was to clarify the difference between PSP and PD from the viewpoint of dynamic cerebrospinal fluid (CSF) flow focusing on the midbrain aqueduct. Methods: Thirty-three PD patients (mean age 69.2 ± 7.9) and 35 PSP patients (mean age 70.5 ± 6.6) were included in this study. CSF flow was calculated by 15 images in an equidistant magnetic resonance imaging (MRI) sequence that was taken throughout a cardiac cycle. Results: Absolute values of the velocity (time points of 2–6 and 12–15, *p b 0.05), and the width of the CSF velocity (Vheight) (PSP, 5.1 ± 2.3 cm/s; PD, 6.0 ± 1.6 cm/s, p b 0.05) effectively discriminated PSP from PD patients. On the other hand, conventional MRI measurements discriminated well the midbrain aqueduct area (Area) (PSP, 7.7 ± 2.6 mm2; PD, 5.4 ± 1.8 mm2, p b 0.01). Two cutoff value lines (Vheight: 4.75, Area: 5.77) of the ROC curve analysis established two areas for discriminating PSP from PD. Conclusion: In the present dynamic CSF flow study, it was newly found that mean velocity of each time point and Vheight showed a more significant decline in PSP than in PD patients, providing a sensitive biomarker for differentiating them. The combination of Vheight and Area could further discriminate PSP from PD patients. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Parkinson's disease (PD) is a progressive neurodegenerative disorder that is mainly characterized by movement disorders such as resting tremor, rigidity, bradykinesia, and gait disturbance. These motor symptoms result from neuronal loss associated with the appearance of Lewy bodies (LBs) in the dopaminergic neurons of substantia nigra. Progressive supranuclear palsy (PSP) is also a progressive neurodegenerative disorder that is characterized by supranuclear gaze palsy, akinetic rigidity, gait disturbance, and dementia [1,2]. The neuropathology of PSP includes degeneration of the substantia nigra, globus pallidus, subthalamic nucleus and the dentate nucleus, accompanied by the accumulation of tau protein in the neuron and glial cells [3]. Since PSP patients show similar clinical manifestations as PD patients, especially at an early stage [4], it is sometimes difficult to discriminate PSP from PD. A diagnosis of PSP depends on the findings of clinical features, such as vertical eye movement limitation, ineffectiveness of levodopa, atrophy of the midbrain, dementia, and rapid clinical
progression [5,6]. In addition, recent studies have shown that PSP patients were associated with progressive atrophy of the midbrain tegmentum [7,8]. Accordingly, it is presumed that this leads to dilation of the aqueduct, and has a causal association with the dynamic cerebrospinal fluid (CSF) flow in the midbrain aqueduct. However, such clinical and conventional image analyses do not sufficiently or clearly discriminate these similar, but distinct, diseases. Recently, it has been possible to measure dynamic CSF motion due to technical improvement of magnetic resonance imaging (MRI). Dynamic CSF flow analysis is important in revealing diseases of the central nervous system and spinal cord disorders such as hydrocephalus, Chiari malformation, syringomyelia, multiple sclerosis, and amyotrophic lateral sclerosis [9–13]. However, no such study has yet applied dynamic CSF flow to differentiate PSP from PD patients. Thus, the purpose of this study was to clarify the difference between PSP and PD from the viewpoint of dynamic CSF flow focusing on the midbrain aqueduct. 2. Patients and methods 2.1. Participants
⁎ Corresponding author at: Department of Neurology, Okayama University Graduate School of Medicine, Dentistry and Pharmacy, 2-5-1 Shikata-cho, Okayama 700-8558, Japan. E-mail address: [email protected] (K. Abe).
http://dx.doi.org/10.1016/j.jns.2015.07.026 0022-510X/© 2015 Elsevier B.V. All rights reserved.
Patients, all of whom were Japanese and who attended the PD and PD-related diseases outpatient clinic of our Okayama University Hospital from November 2009 to March 2014, were the subjects of this study.
Y. Fukui et al. / Journal of the Neurological Sciences 357 (2015) 178–182
179
2.2. CSF flow imaging
Table 1 Patients profile. Subject
n
Gender (M/F)
Examined age (years)
Disease onset (years)
Disease duration (months)
PD PSP
33 35
21/12 13/22
69.2 ± 7.9 70.5 ± 6.6 (n.s.)
62.8 ± 9.1 66.3 ± 7.4 (n.s.)
77.9 ± 44.4 49.6 ± 35.8 ⁎⁎
n.s. indicates no significant difference vs PD patients. ⁎⁎ p b 0.01 vs PD patients.
Thirty-three PD patients (21 male and 12 female; mean age 69.2 ± 7.9 years old) and 35 PSP patients (13 male and 22 female; mean age 70.5 ± 6.6 years old) were included in this study. There was no significant difference in the examined ages or in the disease onset ages between both groups, but the disease duration was shorter in PSP (49.6 ± 35.8 months, **p b 0.01) than in PD (77.9 ± 44.4 months). Clinical diagnoses were based on the consensus criteria for probable PD [14] and the National Institute of Neurological Disorders and Stroke and the Society for PSP, Inc. (NINDS–SPSP) criteria [5]. All PSP patients who enrolled were classified as Richardson-type PSP. The PSP patients were almost the same stage of Hoehn–Yahr scale scores of PD patients (2.97 ± 0.63). The pulsion tests of PSP patients were disturbed, but they were able to walk unaided. Clinical information of PD and PSP groups is summarized in Table 1. Ethical permission for this study was provided by the Ethics Committee on Epidemiological Studies of the Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Sciences (approval number 304) and informed consent was obtained from all participants prior to enrollment in this study.
CSF flow was measured in patients in the supine position essentially according to our previous report [13]. All flow images were evaluated by phase contrast-magnetic resonance imaging (PC-MRI) using a 1.5 T MRI unit (Archieva, Philips, The Netherlands). The imaging parameters were as follows: repetition time = 21 ms, echo time = 6.273 ms, flip angle = 10°, 15 image/cycle, acquisition matrix = 256 × 256, voxel size = 0.98 × 0.98 × 10 mm, cardiac synchronization. This study was performed using an oblique plane transversal to the midbrain aqueduct (Fig. 1A, B). Regions of interest (ROI) were drawn manually to encompass the midbrain aqueduct (Fig. 1C, D).
2.3. Flow analysis Flow analysis was also performed essentially according to our previous report [13]. On the assumption that CSF is considered to be an incompressible and viscous Newtonian fluid, we adopted a calculation program using Navier-Stokes and Darcy equations. Caudocranial flow was represented by positive values, while craniocaudal flow was represented by negative values. CSF flow, which summed up flow at each pixel, was calculated by the ROI of 15 images in the equidistant MRI sequence that was taken through a cardiac cycle. From that, approximate flow curves were then extracted. We obtained several quantitative parameters through these calculations: mean velocity of each time point (Velocity), width of the CSF velocity (Vheight), and the area of the midbrain aqueduct (Area). Furthermore, Vheight represented the difference between maximum and minimum velocity.
PD
PSP
A
B
C
D
Fig. 1. MR images for PD and PSP patients. Sagittal images (A, B) were taken to obtain the axial sections to the midbrain aqueduct, where the ROI were set for dynamic CSF flow measurement in coronal sections (C, D).
180
Y. Fukui et al. / Journal of the Neurological Sciences 357 (2015) 178–182
3
Statistical analyses were performed using statistical software (SPSS 22.0.0.0; SPSS Inc., Chicago, IL, USA). After having checked for normality, we performed Mann–Whitney tests to compare Velocity, Vheight, and Area between PD patients and PSP patients. P-values less than 0.05 were considered to be significantly different. Receiver-operator characteristic (ROC) curves were computed for both Vheight and Area. Cutoff levels were selected from the ROC curves as those values with optimum sensitivity and specificity for this study.
2
velocity (cm/s)
2.4. Statistical analysis
PD PSP
1 0 1
2
3
4
5
6
7
8
9
10 11 12 13 14 15
-1 -2
3. Results 3.1. CSF flow velocity
-3
Table 2 shows the mean values of Velocity (cm/s) and Fig. 2 shows the approximate curves. Compared with PD patients, absolute values of the velocity of PSP patients were reduced in time points 2–6 and 12–15 (*p b 0.05 vs PD patients, Table 2). No such significant differences were observed for the craniocaudal area (time points 7–11). Vheight (cm/s) was significantly reduced in PSP patients (5.1 ± 2.3 cm/s, *p b 0.05 vs PD) compared to PD patients (6.0 ± 1.6 cm/s). The Vheight of PSP patients showed a large dispersion, and the Vheight of PD patients overlapped with the dispersion of PSP patients. However, the minimum and median of Vheight of PSP patients were clearly smaller than the same parameters of PD patients (Fig. 3A). In contrast, the Area (mm2) of PSP patients (7.7 ± 2.6 mm2, **p b 0.01 vs PD patients) was significantly larger than that of PD patients (5.4 ± 1.8 mm2). The Area of PD and PSP patients showed partially overlapping dispersion but the minimum, maximum, median, and quartile of PSP patients were clearly larger than the same parameters of PD patients (Fig. 3B).
-4
To evaluate the diagnostic sensitivity and specificity of Vheight and Area, assessed by significant differences using Mann–Whitney tests, ROC curves were computed for PD and PSP patients (Fig. 4). When the cutoff value of Vheight was set to 4.75, the sensitivity and specificity to discriminate PD patients from PSP patients were 80.0% and 58.8%, respectively (predictive accuracy: 69.6%). In the same manner (cutoff: 5.77), the sensitivity and specificity of the Area were 68.6% and 67.6%, respectively (predictive accuracy: 68.1%) (Fig. 4). In a graph that represents Vheight (cm/s) on the vertical axis and Area (mm2) on the horizontal axis, the data of PD and PSP patients were plotted to create a scatter diagram (Fig. 5). When both cutoff value lines were drawn on the same diagram, 14 out of 15 patients (93.3%) were PSP patients in the lower right area (Vheight: 4.75 or
Fig. 2. Average curves of CSF velocity in PD (solid line) and PSP (dotted line) at 15 subpoints of a cardiac cycle. Compared with PD patients, absolute values of the velocity of PSP patients were reduced in time points 2–6 and 12–15 (*p b 0.05 vs PD patients).
A 12 * p < 0.05 10
Vheight (cm/s)
3.2. ROC curve analysis
* p<0.05 vs PD patients
8 6 4 2 0
PD
PSP
B 12
** p < 0.01
Table 2 Mean velocity of each time point (cm/s). Time point 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 ⁎ p b 0.05 vs PD patients.
PD
PSP
0.0 ± 0.0 0.8 ± 0.3 1.3 ± 0.5 1.7 ± 0.6 2.1 ± 0.7 2.3 ± 0.8 2.1 ± 0.7 1.6 ± 1.0 0.4 ± 1.5 −1.1 ± 1.6 −2.5 ± 1.3 −3.2 ± 1.2 −3.0 ± 1.1 −2.2 ± 0.9 −1.0 ± 0.4
0.0 ± 0.0 0.6 ± 0.4 ⁎ 1.1 ± 0.8 ⁎ 1.4 ± 1.0 ⁎ 1.6 ± 1.0 ⁎ 1.8 ± 1.1 ⁎ 1.8 ± 1.1 1.3 ± 1.1 0.4 ± 1.1 −1.0 ± 1.3 −2.2 ± 1.5 −2.6 ± 1.4 ⁎ −2.4 ± 1.1 ⁎ −1.7 ± 0.8 ⁎ −0.8 ± 0.5 ⁎
Area (mm2)
10 8 6 4 2 0
PD
PSP
Fig. 3. Comparison of Vheight (cm/s) (A) and Area (mm2) (B) between PD and PSP patients. Note a significant reduction of Vheight in PSP patients (5.1 ± 2.3 cm/s, *p b 0.05 vs PD) relative to PD patients (6.0 ± 1.6 cm/s), and a significant enlargement of the Area in PSP patients (7.7 ± 2.6 mm2, **p b 0.01 vs PD patients) relative to PD patients (5.4 ± 1.8 mm2).
Y. Fukui et al. / Journal of the Neurological Sciences 357 (2015) 178–182
1.0 Cutoff 4.75 Cutoff 5.77
Sensitivity
0.8
0.6
0.4 Vheight
0.2
0
Area
0.2
0.4 0.6 1-Specificity
0.8
1.0
Fig. 4. ROC curve analysis shows 80.0% sensitivity and 58.8% specificity to discriminate PSP from PD patients with the cutoff value for Vheight being 4.75 (predictive accuracy: 69.6%), and 68.6% sensitivity and 67.6% specificity with the cutoff value for Area being 5.77 (predictive accuracy: 68.1%).
less, Area: 5.77 or more) (Fig. 5). In the upper left area (Vheight: more than 4.75, Area: less than 5.77) (Fig. 5), 18 out of 23 patients (78.3%) were PD patients. 4. Discussion The present study confirmed that PSP patients are associated with progressive atrophy of the midbrain tegmentum (Fig. 1D) [7,8], and with dilation of the midbrain aqueduct (Fig. 3B). The present study newly found that Velocity and Vheight showed more significant declines in PSP than in PD patients (Table 2, Fig. 3A). In addition, the
12 PD PSP
Vheight (cm/s)
10
combination of Vheight cutoff values and Areas could distinguish PSP patients from PD patients (Fig. 5), although each cutoff value was not very sensitive (Fig. 4). It is sometimes difficult to clinically distinguish PSP from PD patients especially at an early stage as they both share similar symptoms [4]. PSP may generally be characterized by limitations in vertical eye movement, ineffectiveness of levodopa, atrophy of the midbrain, dementia, and rapid clinical progression [5,6]. Consistent with our present study, recent MRI studies [7,15–17] showed progressive atrophy of the midbrain tegmentum leading to dilation of the aqueduct, as was shown in the Area (**p b 0.01, Fig. 3B). However, such structural dilation may follow after functional differences between PSP and PD related to CSF flow are detected. In a previous study, we reported that CSF flow dynamics of amyotrophic lateral sclerosis (ALS) facilitated the differentiation of ALS from cervical spondylosis [13]. Bernoulli's principle originally showed an important relationship between the diameter and speed of non-viscous fluid. However, CSF is considered to be an incompressible and viscous Newtonian fluid primarily composed of glucose and various proteins, which may be different in PSP and PD [18,19]. Since a larger aqueduct was accompanied by a slower velocity in PSP than in PD (Table 2, Fig. 3), the present calculation with Navier-Stokes and Darcy equations [20] may provide a useful diagnostic biomarker of CSF flow for PSP (Figs. 4, 5). However, the results might not be directly connected with early stage diagnosis because we conducted a retrospective study (i.e., we found differences in patients with a probable degree of diagnostic criteria). A prospective study including a prediagnostic stage is necessary to validate the usefulness of the results at an early stage. In summary, the present study demonstrated that Velocity, Vheight, and Area effectively discriminated PSP with a mean disease duration of 49.6 M from PD patients with a mean disease duration of 77.9 M. When the two cutoff value lines of the ROC curve analysis were drawn on the same diagram, two areas could discriminate PSP from PD (Fig. 5). Thus, this CSF flow analysis may be a sensitive biomarker for differentiating PSP from PD patients. Acknowledgments This work was partly supported by a Grant-in-Aid for Scientific Research (B) 21390267 from the Ministry of Education, Science, Culture, and Sports of Japan and by Grants-in Aid from the Research Committee of CNS Degenerative Diseases (H23-NANJI-IPPAN-015) (I. Nakano) and grants (H. Mizusawa, M. Nishizawa, H. Sasaki, G. Sobue) from the Ministry of Health, Labour and Welfare of Japan (H23-NANJI-IPPAN-039, H25-NANJITOU-SHITEI-002). References
8 6 4 2 0
181
2.5
5.0 7.5 Area (mm2)
10.0
12.5
Fig. 5. A scatter diagram of Area and Vheight for PD and PSP patients shows 93.3% attribution to PSP in the lower right area (Area: ≥5.77, Vheight: ≤4.75), and 78.3% attribution to PD patients in the upper left area (Vheight: N4.75, Area: b5.77).
[1] J.C. Steele, J.C. Richardson, J. Olszewski, Progressive Supranuclear Palsy. A heterogeneous degeneration involving the brain stem, basal ganglia and cerebellum with vertical gaze and pseudobulbar palsy, nuchal dystonia and dementia, Arch. Neurol. 10 (1964) 333–359. [2] M.L. Albert, R.G. Feldman, A.L. Willis, The ‘subcortical dementia’ of progressive supranuclear palsy, J. Neurol. Neurosurg. Psychiatry 37 (2) (1974) 121–130. [3] K. Liao, J. Wagner, A. Joshi, I. Estrovich, M.F. Walker, M. Strupp, et al., Why do patients with PSP fall? Evidence for abnormal otolith responses, Neurology 70 (10) (2008) 802–809. [4] E. Tolosa, G. Wenning, W. Poewe, The diagnosis of Parkinson's disease, Lancet Neurol. 5 (1) (2006) 75–86. [5] I. Litvan, Y. Agid, D. Calne, G. Campbell, B. Dubois, R.C. Duvoisin, et al., Clinical research criteria for the diagnosis of progressive supranuclear palsy (Steele–Richardson– Olszewski syndrome): report of the NINDS–SPSP international workshop, Neurology 47 (1) (1996) 1–9. [6] I. Litvan, G. Campbell, C.A. Mangone, M. Verny, A. McKee, K.R. Chaudhuri, et al., Which clinical features differentiate progressive supranuclear palsy (Steele–Richardson– Olszewski syndrome) from related disorders? A clinicopathological study, Brain 120 (Pt 1) (1997) 65–74. [7] M. Warmuth-Metz, M. Naumann, I. Csoti, L. Solymosi, Measurement of the midbrain diameter on routine magnetic resonance imaging: a simple and accurate method of differentiating between Parkinson disease and progressive supranuclear palsy, Arch. Neurol. 58 (7) (2001) 1076–1079.
182
Y. Fukui et al. / Journal of the Neurological Sciences 357 (2015) 178–182
[8] C. Brenneis, K. Seppi, M. Schocke, T. Benke, G.K. Wenning, W. Poewe, Voxel based morphometry reveals a distinct pattern of frontal atrophy in progressive supranuclear palsy, J. Neurol. Neurosurg. Psychiatry 75 (2) (2004) 246–249. [9] N. Watabe, T. Tominaga, H. Shimizu, K. Koshu, T. Yoshimoto, Quantitative analysis of cerebrospinal fluid flow in patients with cervical spondylosis using cine phasecontrast magnetic resonance imaging, Neurosurgery 44 (4) (1999) 779–784. [10] E. Hofmann, M. Warmuth-Metz, M. Bendszus, L. Solymosi, Phase-contrast MR imaging of the cervical CSF and spinal cord: volumetric motion analysis in patients with Chiari I malformation, AJNR Am. J. Neuroradiol. 21 (1) (2000) 151–158. [11] T. Tominaga, N. Watabe, T. Takahashi, H. Shimizu, T. Yoshimoto, Quantitative assessment of surgical decompression of the cervical spine with cine phase contrast magnetic resonance imaging, Neurosurgery 50 (4) (2002) 791–795 (discussion 6). [12] P. Sundstrom, A. Wahlin, K. Ambarki, R. Birgander, A. Eklund, J. Malm, Venous and cerebrospinal fluid flow in multiple sclerosis: a case-control study, Ann. Neurol. 68 (2) (2010) 255–259. [13] K. Sato, N. Morimoto, T. Matsuura, Y. Ohta, M. Tsunoda, Y. Ikeda, et al., CSF flow dynamics in motor neuron disease, Neurol. Res. 34 (5) (2012) 512–517. [14] A.J. Hughes, S.E. Daniel, L. Kilford, A.J. Lees, Accuracy of clinical diagnosis of idiopathic Parkinson's disease: a clinico-pathological study of 100 cases, J. Neurol. Neurosurg. Psychiatry 55 (3) (1992) 181–184.
[15] A. Righini, A. Antonini, R. De Notaris, E. Bianchini, N. Meucci, G. Sacilotto, et al., MR imaging of the superior profile of the midbrain: differential diagnosis between progressive supranuclear palsy and Parkinson disease, AJNR Am. J. Neuroradiol. 25 (6) (2004) 927–932. [16] K. Seppi, W. Poewe, Brain magnetic resonance imaging techniques in the diagnosis of parkinsonian syndromes, Neuroimaging Clin. N. Am. 20 (1) (2010) 29–55. [17] K. Tsukamoto, E. Matsusue, Y. Kanasaki, S. Kakite, S. Fujii, T. Kaminou, et al., Significance of apparent diffusion coefficient measurement for the differential diagnosis of multiple system atrophy, progressive supranuclear palsy, and Parkinson's disease: evaluation by 3.0-T MR imaging, Neuroradiology 54 (9) (2012) 947–955. [18] M. Shoji, E. Matsubara, T. Murakami, Y. Manabe, K. Abe, M. Kanai, et al., Cerebrospinal fluid tau in dementia disorders: a large scale multicenter study by a Japanese study group, Neurobiol. Aging 23 (3) (2002) 363–370. [19] R. Constantinescu, H. Zetterberg, B. Holmberg, L. Rosengren, Levels of brain related proteins in cerebrospinal fluid: an aid in the differential diagnosis of parkinsonian disorders, Parkinsonism Relat. Disord. 15 (3) (2009) 205–212. [20] A.A. Linninger, B. Sweetman, R. Penn, Normal and hydrocephalic brain dynamics: the role of reduced cerebrospinal fluid reabsorption in ventricular enlargement, Ann. Biomed. Eng. 37 (7) (2009) 1434–1447.
Contents lists available at ScienceDirect
Journal of the Neurological Sciences journal homepage: www.elsevier.com/locate/jns
Differentiating progressive supranuclear palsy from Parkinson's disease by MRI-based dynamic cerebrospinal fluid flow Yusuke Fukui, Nozomi Hishikawa, Kota Sato, Taijun Yunoki, Syoichiro Kono, Kosuke Matsuzono, Yumiko Nakano, Yasuyuki Ohta, Toru Yamashita, Kentaro Deguchi, Koji Abe ⁎ Department of Neurology, Graduate School of Medicine, Dentistry and Pharmaceutical Sciences, Okayama University, 2-5-1 Shikatacho, Kitaku, Okayama 700-8558, Japan
a r t i c l e
i n f o
Article history: Received 7 January 2015 Received in revised form 25 June 2015 Accepted 16 July 2015 Available online 21 July 2015 Keywords: Cerebrospinal fluid Midbrain aqueduct Neurodegenerative disorder Phase contrast-magnetic resonance imaging Parkinson's disease Progressive supranuclear palsy
a b s t r a c t Objective: The purpose of this study was to clarify the difference between PSP and PD from the viewpoint of dynamic cerebrospinal fluid (CSF) flow focusing on the midbrain aqueduct. Methods: Thirty-three PD patients (mean age 69.2 ± 7.9) and 35 PSP patients (mean age 70.5 ± 6.6) were included in this study. CSF flow was calculated by 15 images in an equidistant magnetic resonance imaging (MRI) sequence that was taken throughout a cardiac cycle. Results: Absolute values of the velocity (time points of 2–6 and 12–15, *p b 0.05), and the width of the CSF velocity (Vheight) (PSP, 5.1 ± 2.3 cm/s; PD, 6.0 ± 1.6 cm/s, p b 0.05) effectively discriminated PSP from PD patients. On the other hand, conventional MRI measurements discriminated well the midbrain aqueduct area (Area) (PSP, 7.7 ± 2.6 mm2; PD, 5.4 ± 1.8 mm2, p b 0.01). Two cutoff value lines (Vheight: 4.75, Area: 5.77) of the ROC curve analysis established two areas for discriminating PSP from PD. Conclusion: In the present dynamic CSF flow study, it was newly found that mean velocity of each time point and Vheight showed a more significant decline in PSP than in PD patients, providing a sensitive biomarker for differentiating them. The combination of Vheight and Area could further discriminate PSP from PD patients. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Parkinson's disease (PD) is a progressive neurodegenerative disorder that is mainly characterized by movement disorders such as resting tremor, rigidity, bradykinesia, and gait disturbance. These motor symptoms result from neuronal loss associated with the appearance of Lewy bodies (LBs) in the dopaminergic neurons of substantia nigra. Progressive supranuclear palsy (PSP) is also a progressive neurodegenerative disorder that is characterized by supranuclear gaze palsy, akinetic rigidity, gait disturbance, and dementia [1,2]. The neuropathology of PSP includes degeneration of the substantia nigra, globus pallidus, subthalamic nucleus and the dentate nucleus, accompanied by the accumulation of tau protein in the neuron and glial cells [3]. Since PSP patients show similar clinical manifestations as PD patients, especially at an early stage [4], it is sometimes difficult to discriminate PSP from PD. A diagnosis of PSP depends on the findings of clinical features, such as vertical eye movement limitation, ineffectiveness of levodopa, atrophy of the midbrain, dementia, and rapid clinical
progression [5,6]. In addition, recent studies have shown that PSP patients were associated with progressive atrophy of the midbrain tegmentum [7,8]. Accordingly, it is presumed that this leads to dilation of the aqueduct, and has a causal association with the dynamic cerebrospinal fluid (CSF) flow in the midbrain aqueduct. However, such clinical and conventional image analyses do not sufficiently or clearly discriminate these similar, but distinct, diseases. Recently, it has been possible to measure dynamic CSF motion due to technical improvement of magnetic resonance imaging (MRI). Dynamic CSF flow analysis is important in revealing diseases of the central nervous system and spinal cord disorders such as hydrocephalus, Chiari malformation, syringomyelia, multiple sclerosis, and amyotrophic lateral sclerosis [9–13]. However, no such study has yet applied dynamic CSF flow to differentiate PSP from PD patients. Thus, the purpose of this study was to clarify the difference between PSP and PD from the viewpoint of dynamic CSF flow focusing on the midbrain aqueduct. 2. Patients and methods 2.1. Participants
⁎ Corresponding author at: Department of Neurology, Okayama University Graduate School of Medicine, Dentistry and Pharmacy, 2-5-1 Shikata-cho, Okayama 700-8558, Japan. E-mail address: [email protected] (K. Abe).
http://dx.doi.org/10.1016/j.jns.2015.07.026 0022-510X/© 2015 Elsevier B.V. All rights reserved.
Patients, all of whom were Japanese and who attended the PD and PD-related diseases outpatient clinic of our Okayama University Hospital from November 2009 to March 2014, were the subjects of this study.
Y. Fukui et al. / Journal of the Neurological Sciences 357 (2015) 178–182
179
2.2. CSF flow imaging
Table 1 Patients profile. Subject
n
Gender (M/F)
Examined age (years)
Disease onset (years)
Disease duration (months)
PD PSP
33 35
21/12 13/22
69.2 ± 7.9 70.5 ± 6.6 (n.s.)
62.8 ± 9.1 66.3 ± 7.4 (n.s.)
77.9 ± 44.4 49.6 ± 35.8 ⁎⁎
n.s. indicates no significant difference vs PD patients. ⁎⁎ p b 0.01 vs PD patients.
Thirty-three PD patients (21 male and 12 female; mean age 69.2 ± 7.9 years old) and 35 PSP patients (13 male and 22 female; mean age 70.5 ± 6.6 years old) were included in this study. There was no significant difference in the examined ages or in the disease onset ages between both groups, but the disease duration was shorter in PSP (49.6 ± 35.8 months, **p b 0.01) than in PD (77.9 ± 44.4 months). Clinical diagnoses were based on the consensus criteria for probable PD [14] and the National Institute of Neurological Disorders and Stroke and the Society for PSP, Inc. (NINDS–SPSP) criteria [5]. All PSP patients who enrolled were classified as Richardson-type PSP. The PSP patients were almost the same stage of Hoehn–Yahr scale scores of PD patients (2.97 ± 0.63). The pulsion tests of PSP patients were disturbed, but they were able to walk unaided. Clinical information of PD and PSP groups is summarized in Table 1. Ethical permission for this study was provided by the Ethics Committee on Epidemiological Studies of the Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Sciences (approval number 304) and informed consent was obtained from all participants prior to enrollment in this study.
CSF flow was measured in patients in the supine position essentially according to our previous report [13]. All flow images were evaluated by phase contrast-magnetic resonance imaging (PC-MRI) using a 1.5 T MRI unit (Archieva, Philips, The Netherlands). The imaging parameters were as follows: repetition time = 21 ms, echo time = 6.273 ms, flip angle = 10°, 15 image/cycle, acquisition matrix = 256 × 256, voxel size = 0.98 × 0.98 × 10 mm, cardiac synchronization. This study was performed using an oblique plane transversal to the midbrain aqueduct (Fig. 1A, B). Regions of interest (ROI) were drawn manually to encompass the midbrain aqueduct (Fig. 1C, D).
2.3. Flow analysis Flow analysis was also performed essentially according to our previous report [13]. On the assumption that CSF is considered to be an incompressible and viscous Newtonian fluid, we adopted a calculation program using Navier-Stokes and Darcy equations. Caudocranial flow was represented by positive values, while craniocaudal flow was represented by negative values. CSF flow, which summed up flow at each pixel, was calculated by the ROI of 15 images in the equidistant MRI sequence that was taken through a cardiac cycle. From that, approximate flow curves were then extracted. We obtained several quantitative parameters through these calculations: mean velocity of each time point (Velocity), width of the CSF velocity (Vheight), and the area of the midbrain aqueduct (Area). Furthermore, Vheight represented the difference between maximum and minimum velocity.
PD
PSP
A
B
C
D
Fig. 1. MR images for PD and PSP patients. Sagittal images (A, B) were taken to obtain the axial sections to the midbrain aqueduct, where the ROI were set for dynamic CSF flow measurement in coronal sections (C, D).
180
Y. Fukui et al. / Journal of the Neurological Sciences 357 (2015) 178–182
3
Statistical analyses were performed using statistical software (SPSS 22.0.0.0; SPSS Inc., Chicago, IL, USA). After having checked for normality, we performed Mann–Whitney tests to compare Velocity, Vheight, and Area between PD patients and PSP patients. P-values less than 0.05 were considered to be significantly different. Receiver-operator characteristic (ROC) curves were computed for both Vheight and Area. Cutoff levels were selected from the ROC curves as those values with optimum sensitivity and specificity for this study.
2
velocity (cm/s)
2.4. Statistical analysis
PD PSP
1 0 1
2
3
4
5
6
7
8
9
10 11 12 13 14 15
-1 -2
3. Results 3.1. CSF flow velocity
-3
Table 2 shows the mean values of Velocity (cm/s) and Fig. 2 shows the approximate curves. Compared with PD patients, absolute values of the velocity of PSP patients were reduced in time points 2–6 and 12–15 (*p b 0.05 vs PD patients, Table 2). No such significant differences were observed for the craniocaudal area (time points 7–11). Vheight (cm/s) was significantly reduced in PSP patients (5.1 ± 2.3 cm/s, *p b 0.05 vs PD) compared to PD patients (6.0 ± 1.6 cm/s). The Vheight of PSP patients showed a large dispersion, and the Vheight of PD patients overlapped with the dispersion of PSP patients. However, the minimum and median of Vheight of PSP patients were clearly smaller than the same parameters of PD patients (Fig. 3A). In contrast, the Area (mm2) of PSP patients (7.7 ± 2.6 mm2, **p b 0.01 vs PD patients) was significantly larger than that of PD patients (5.4 ± 1.8 mm2). The Area of PD and PSP patients showed partially overlapping dispersion but the minimum, maximum, median, and quartile of PSP patients were clearly larger than the same parameters of PD patients (Fig. 3B).
-4
To evaluate the diagnostic sensitivity and specificity of Vheight and Area, assessed by significant differences using Mann–Whitney tests, ROC curves were computed for PD and PSP patients (Fig. 4). When the cutoff value of Vheight was set to 4.75, the sensitivity and specificity to discriminate PD patients from PSP patients were 80.0% and 58.8%, respectively (predictive accuracy: 69.6%). In the same manner (cutoff: 5.77), the sensitivity and specificity of the Area were 68.6% and 67.6%, respectively (predictive accuracy: 68.1%) (Fig. 4). In a graph that represents Vheight (cm/s) on the vertical axis and Area (mm2) on the horizontal axis, the data of PD and PSP patients were plotted to create a scatter diagram (Fig. 5). When both cutoff value lines were drawn on the same diagram, 14 out of 15 patients (93.3%) were PSP patients in the lower right area (Vheight: 4.75 or
Fig. 2. Average curves of CSF velocity in PD (solid line) and PSP (dotted line) at 15 subpoints of a cardiac cycle. Compared with PD patients, absolute values of the velocity of PSP patients were reduced in time points 2–6 and 12–15 (*p b 0.05 vs PD patients).
A 12 * p < 0.05 10
Vheight (cm/s)
3.2. ROC curve analysis
* p<0.05 vs PD patients
8 6 4 2 0
PD
PSP
B 12
** p < 0.01
Table 2 Mean velocity of each time point (cm/s). Time point 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 ⁎ p b 0.05 vs PD patients.
PD
PSP
0.0 ± 0.0 0.8 ± 0.3 1.3 ± 0.5 1.7 ± 0.6 2.1 ± 0.7 2.3 ± 0.8 2.1 ± 0.7 1.6 ± 1.0 0.4 ± 1.5 −1.1 ± 1.6 −2.5 ± 1.3 −3.2 ± 1.2 −3.0 ± 1.1 −2.2 ± 0.9 −1.0 ± 0.4
0.0 ± 0.0 0.6 ± 0.4 ⁎ 1.1 ± 0.8 ⁎ 1.4 ± 1.0 ⁎ 1.6 ± 1.0 ⁎ 1.8 ± 1.1 ⁎ 1.8 ± 1.1 1.3 ± 1.1 0.4 ± 1.1 −1.0 ± 1.3 −2.2 ± 1.5 −2.6 ± 1.4 ⁎ −2.4 ± 1.1 ⁎ −1.7 ± 0.8 ⁎ −0.8 ± 0.5 ⁎
Area (mm2)
10 8 6 4 2 0
PD
PSP
Fig. 3. Comparison of Vheight (cm/s) (A) and Area (mm2) (B) between PD and PSP patients. Note a significant reduction of Vheight in PSP patients (5.1 ± 2.3 cm/s, *p b 0.05 vs PD) relative to PD patients (6.0 ± 1.6 cm/s), and a significant enlargement of the Area in PSP patients (7.7 ± 2.6 mm2, **p b 0.01 vs PD patients) relative to PD patients (5.4 ± 1.8 mm2).
Y. Fukui et al. / Journal of the Neurological Sciences 357 (2015) 178–182
1.0 Cutoff 4.75 Cutoff 5.77
Sensitivity
0.8
0.6
0.4 Vheight
0.2
0
Area
0.2
0.4 0.6 1-Specificity
0.8
1.0
Fig. 4. ROC curve analysis shows 80.0% sensitivity and 58.8% specificity to discriminate PSP from PD patients with the cutoff value for Vheight being 4.75 (predictive accuracy: 69.6%), and 68.6% sensitivity and 67.6% specificity with the cutoff value for Area being 5.77 (predictive accuracy: 68.1%).
less, Area: 5.77 or more) (Fig. 5). In the upper left area (Vheight: more than 4.75, Area: less than 5.77) (Fig. 5), 18 out of 23 patients (78.3%) were PD patients. 4. Discussion The present study confirmed that PSP patients are associated with progressive atrophy of the midbrain tegmentum (Fig. 1D) [7,8], and with dilation of the midbrain aqueduct (Fig. 3B). The present study newly found that Velocity and Vheight showed more significant declines in PSP than in PD patients (Table 2, Fig. 3A). In addition, the
12 PD PSP
Vheight (cm/s)
10
combination of Vheight cutoff values and Areas could distinguish PSP patients from PD patients (Fig. 5), although each cutoff value was not very sensitive (Fig. 4). It is sometimes difficult to clinically distinguish PSP from PD patients especially at an early stage as they both share similar symptoms [4]. PSP may generally be characterized by limitations in vertical eye movement, ineffectiveness of levodopa, atrophy of the midbrain, dementia, and rapid clinical progression [5,6]. Consistent with our present study, recent MRI studies [7,15–17] showed progressive atrophy of the midbrain tegmentum leading to dilation of the aqueduct, as was shown in the Area (**p b 0.01, Fig. 3B). However, such structural dilation may follow after functional differences between PSP and PD related to CSF flow are detected. In a previous study, we reported that CSF flow dynamics of amyotrophic lateral sclerosis (ALS) facilitated the differentiation of ALS from cervical spondylosis [13]. Bernoulli's principle originally showed an important relationship between the diameter and speed of non-viscous fluid. However, CSF is considered to be an incompressible and viscous Newtonian fluid primarily composed of glucose and various proteins, which may be different in PSP and PD [18,19]. Since a larger aqueduct was accompanied by a slower velocity in PSP than in PD (Table 2, Fig. 3), the present calculation with Navier-Stokes and Darcy equations [20] may provide a useful diagnostic biomarker of CSF flow for PSP (Figs. 4, 5). However, the results might not be directly connected with early stage diagnosis because we conducted a retrospective study (i.e., we found differences in patients with a probable degree of diagnostic criteria). A prospective study including a prediagnostic stage is necessary to validate the usefulness of the results at an early stage. In summary, the present study demonstrated that Velocity, Vheight, and Area effectively discriminated PSP with a mean disease duration of 49.6 M from PD patients with a mean disease duration of 77.9 M. When the two cutoff value lines of the ROC curve analysis were drawn on the same diagram, two areas could discriminate PSP from PD (Fig. 5). Thus, this CSF flow analysis may be a sensitive biomarker for differentiating PSP from PD patients. Acknowledgments This work was partly supported by a Grant-in-Aid for Scientific Research (B) 21390267 from the Ministry of Education, Science, Culture, and Sports of Japan and by Grants-in Aid from the Research Committee of CNS Degenerative Diseases (H23-NANJI-IPPAN-015) (I. Nakano) and grants (H. Mizusawa, M. Nishizawa, H. Sasaki, G. Sobue) from the Ministry of Health, Labour and Welfare of Japan (H23-NANJI-IPPAN-039, H25-NANJITOU-SHITEI-002). References
8 6 4 2 0
181
2.5
5.0 7.5 Area (mm2)
10.0
12.5
Fig. 5. A scatter diagram of Area and Vheight for PD and PSP patients shows 93.3% attribution to PSP in the lower right area (Area: ≥5.77, Vheight: ≤4.75), and 78.3% attribution to PD patients in the upper left area (Vheight: N4.75, Area: b5.77).
[1] J.C. Steele, J.C. Richardson, J. Olszewski, Progressive Supranuclear Palsy. A heterogeneous degeneration involving the brain stem, basal ganglia and cerebellum with vertical gaze and pseudobulbar palsy, nuchal dystonia and dementia, Arch. Neurol. 10 (1964) 333–359. [2] M.L. Albert, R.G. Feldman, A.L. Willis, The ‘subcortical dementia’ of progressive supranuclear palsy, J. Neurol. Neurosurg. Psychiatry 37 (2) (1974) 121–130. [3] K. Liao, J. Wagner, A. Joshi, I. Estrovich, M.F. Walker, M. Strupp, et al., Why do patients with PSP fall? Evidence for abnormal otolith responses, Neurology 70 (10) (2008) 802–809. [4] E. Tolosa, G. Wenning, W. Poewe, The diagnosis of Parkinson's disease, Lancet Neurol. 5 (1) (2006) 75–86. [5] I. Litvan, Y. Agid, D. Calne, G. Campbell, B. Dubois, R.C. Duvoisin, et al., Clinical research criteria for the diagnosis of progressive supranuclear palsy (Steele–Richardson– Olszewski syndrome): report of the NINDS–SPSP international workshop, Neurology 47 (1) (1996) 1–9. [6] I. Litvan, G. Campbell, C.A. Mangone, M. Verny, A. McKee, K.R. Chaudhuri, et al., Which clinical features differentiate progressive supranuclear palsy (Steele–Richardson– Olszewski syndrome) from related disorders? A clinicopathological study, Brain 120 (Pt 1) (1997) 65–74. [7] M. Warmuth-Metz, M. Naumann, I. Csoti, L. Solymosi, Measurement of the midbrain diameter on routine magnetic resonance imaging: a simple and accurate method of differentiating between Parkinson disease and progressive supranuclear palsy, Arch. Neurol. 58 (7) (2001) 1076–1079.
182
Y. Fukui et al. / Journal of the Neurological Sciences 357 (2015) 178–182
[8] C. Brenneis, K. Seppi, M. Schocke, T. Benke, G.K. Wenning, W. Poewe, Voxel based morphometry reveals a distinct pattern of frontal atrophy in progressive supranuclear palsy, J. Neurol. Neurosurg. Psychiatry 75 (2) (2004) 246–249. [9] N. Watabe, T. Tominaga, H. Shimizu, K. Koshu, T. Yoshimoto, Quantitative analysis of cerebrospinal fluid flow in patients with cervical spondylosis using cine phasecontrast magnetic resonance imaging, Neurosurgery 44 (4) (1999) 779–784. [10] E. Hofmann, M. Warmuth-Metz, M. Bendszus, L. Solymosi, Phase-contrast MR imaging of the cervical CSF and spinal cord: volumetric motion analysis in patients with Chiari I malformation, AJNR Am. J. Neuroradiol. 21 (1) (2000) 151–158. [11] T. Tominaga, N. Watabe, T. Takahashi, H. Shimizu, T. Yoshimoto, Quantitative assessment of surgical decompression of the cervical spine with cine phase contrast magnetic resonance imaging, Neurosurgery 50 (4) (2002) 791–795 (discussion 6). [12] P. Sundstrom, A. Wahlin, K. Ambarki, R. Birgander, A. Eklund, J. Malm, Venous and cerebrospinal fluid flow in multiple sclerosis: a case-control study, Ann. Neurol. 68 (2) (2010) 255–259. [13] K. Sato, N. Morimoto, T. Matsuura, Y. Ohta, M. Tsunoda, Y. Ikeda, et al., CSF flow dynamics in motor neuron disease, Neurol. Res. 34 (5) (2012) 512–517. [14] A.J. Hughes, S.E. Daniel, L. Kilford, A.J. Lees, Accuracy of clinical diagnosis of idiopathic Parkinson's disease: a clinico-pathological study of 100 cases, J. Neurol. Neurosurg. Psychiatry 55 (3) (1992) 181–184.
[15] A. Righini, A. Antonini, R. De Notaris, E. Bianchini, N. Meucci, G. Sacilotto, et al., MR imaging of the superior profile of the midbrain: differential diagnosis between progressive supranuclear palsy and Parkinson disease, AJNR Am. J. Neuroradiol. 25 (6) (2004) 927–932. [16] K. Seppi, W. Poewe, Brain magnetic resonance imaging techniques in the diagnosis of parkinsonian syndromes, Neuroimaging Clin. N. Am. 20 (1) (2010) 29–55. [17] K. Tsukamoto, E. Matsusue, Y. Kanasaki, S. Kakite, S. Fujii, T. Kaminou, et al., Significance of apparent diffusion coefficient measurement for the differential diagnosis of multiple system atrophy, progressive supranuclear palsy, and Parkinson's disease: evaluation by 3.0-T MR imaging, Neuroradiology 54 (9) (2012) 947–955. [18] M. Shoji, E. Matsubara, T. Murakami, Y. Manabe, K. Abe, M. Kanai, et al., Cerebrospinal fluid tau in dementia disorders: a large scale multicenter study by a Japanese study group, Neurobiol. Aging 23 (3) (2002) 363–370. [19] R. Constantinescu, H. Zetterberg, B. Holmberg, L. Rosengren, Levels of brain related proteins in cerebrospinal fluid: an aid in the differential diagnosis of parkinsonian disorders, Parkinsonism Relat. Disord. 15 (3) (2009) 205–212. [20] A.A. Linninger, B. Sweetman, R. Penn, Normal and hydrocephalic brain dynamics: the role of reduced cerebrospinal fluid reabsorption in ventricular enlargement, Ann. Biomed. Eng. 37 (7) (2009) 1434–1447.