mri csf (cfd) - ihiwgihiwg.org/wp-content/uploads/2012_bryn_martin.pdf · work in progress: a....
TRANSCRIPT
4D MRI quantification of CSF velocities with comparison to computational fluid dynamics (CFD) simulations
April 26, 2012
American Society of Neuroradiology, 50th Annual MeetingCSF and hydrocephalus study group
Bryn A. Martin1
Theresia I. Yiallourou1, Jan‐Robert Kroeger2, Nikolaos Stergiopulos1, A.C. Bunck2
1Laboratory of Hemodynamics and Cardiovascular Technology, EPFL, Switzerland2Department of Clinical Radiology, University Hospital of Muenster, Germany
3Department of Radiology, University of Basel, Switzerland
1
Clinical tools & measurements
CSF pulse wave velocity in craniospinaldisorders
CPAP: influence on cardiovascular and
CSF system(MRI and US)
Intraoperative spinal cord
hemodynamics
Modeling
Spinal cord blood flow
Hydrocephalus(FE brain)
CSF: in vitro models
CSF: in vitro models
Spinal cord fluid structure interaction
T. Yiallourou, EPFL LausanneAlexander Bunk, M.D. Muenster GKroeger, M.D. Muenster GMaintz, M.D. Muenster G
Dr. F. Santini, Basel, CHDr. Jelena Bock, Freiburg, G
L. Hirt, M.D., CHUV, CHC. Odier, M.D., CAR. Heinzer, M.D., CHUV, CHJ. Rubio, M.D., CHUV, CH
C. Marcucci, M.D., CHUV, CHB. Tozzi, M.D., CHUV, CH
A. Gottshalk, M.D., Ulm, GDr. E. Fonari, CHUV, CHDr. J. Oshinski, Emory U., USADr. F. Santini, Basel, CHDr. F. Loth, Akron, USA
Dr. F. Loth, Akron, USAMilan Radojicic, M.D. UCSF
Dr. J. M. Drezet, EPFL, CHS. Momjain, M.D., Geneva, CH
Dr. N. Stergiopulos, EPFL, CHDr. O. Baledent, Amien, FRDr. P. Reymond, EPFL, CH
J. Novy, M.D., CHUV, CH / UKDr. Pablo Bianco, Brazil
Dr. F. Loth, Akron, USADr. T. Royston, UIC, USAYifei Liu, UIC, USA
4D MRI in aqueduct of Sylvius and 3rd
ventricle
4D MRI in aqueduct of Sylvius and 3rd
ventricle
Coupled cardiovascular and CSF system
Coupled cardiovascular and CSF system
4D MRI in craniospinal disorders
Ongoing research in CSF hydrodynamics
Motivation• Abnormal CSF hydrodynamics have been implicated in many
craniospinal disorders
3ninds.nih.gov/disorders/
Hydrocephalus
Chiari malformation
Syringomyelia
Tethered cord / Spina Bifida
1 in 500
1 in 10001 in 1500
1 in 8000
Need for CSF biomechanics research• The underlying biomechanical forces that contribute to
craniospinal disorders are not well understood.• Engineers + medical doctors are helping to do in CSF
hydrodynamics what has been done in hemodynamics.
4
• Anterior / posterior flow differences• Nerve roots appear to impact flow
Chiari patient
Krueger et al. Martin et al.
Quigley et al.
Background: CSF flow in cervical spine
5
Linge SO et al.. CSF flow dynamics at the craniovertebra. AJNR American journal of neuroradiology. 2010;31(1):185‐92.Krueger KD, et al.. Peak CSF velocities. AJNR American journal of neuroradiology. 2010;31(10):1837‐41.Shaffer, et al.. Neurological research 33, 247‐260, doi:10.1179/016164111X12962202723805 (2011).Martin, B. A., et al. Syringomyelia hydrodynamics. J Biomech Eng 127, 1110‐1120 (2005).Lichtor, et al., Noncommunicating. Spine 30, 1335‐1340, doi:00007632‐200506010‐00019 [pii] (2005).Quigley, M. F., et al. Cerebrospinal fluid flow. Radiology 232, 229‐236 (2004).Brugieres, P. et al. CSF flow measurement in syringomyelia. AJNR Am J Neuroradiol 21, 1785‐1792 (2000).Bhadelia, R. A., et al. Cerebrospinal fluid pulsation. Neuroradiology 39, 258‐264 (1997).Enzmann, D. R. & Pelc, N. J. Normal flow... Radiology 178, 467‐474 (1991).
Healthy
Shaffer et al.Linge et al.
Beledent et al.
Syrinxflow
Lichtor et al.
Brugieres et al.
Bhadelia et al.
Enzmann
• Recently employed technique to visualize CSF flow• Not yet quantitatively analyzed against CFD
4D MRI in cervical spine
6
Healthy Chiari patient
Bunck, A. C. et al.Magnetic resonance 4D flow characteristics of cerebrospinal fluid at the craniocervical junction and the cervical spinal canal. European radiology 21, 1788‐1796, doi:10.1007/s00330‐011‐2105‐7 (2011).Santini, F., et al. Time‐resolved three‐dimensional... Magnetic resonance in medicine : 62, 966‐974, doi:10.1002/mrm.22087 (2009).
Bunck, A. C.
Santini et al.
Background: CFD in spine• What level of CFD simulation complexity is needed?
7
a. b. c. d. e. f. g.
AuthorYear
GeometryTissue motion
Arachnoid trebeculae
Nerve roots / ligaments
a. Gupta et al. 20093D subject specific No Yes No
b. Stockman 2007 2D idealized No Yes Yesc. Linge et al.Roldan et al.Present study
2010?3D subject specific No No No
d. Linge et al. 2009 3D idealized No No Noe. Loth et al. 2001 2D concentric ellipse No No Nof. Bertram 2005‐12 2D idealized axi‐symmetric Yes No Nog. CirovicElliott/CarpenterLockey?
2005‐122D concentric tube with constant diameter
Yes No No
Objective1. Assess 4D MRI measurements of cervical CSF flow in healthy
and patients with CSF blockage at the foramen magnum.2. Compare 4D MRI measurements to subject specific 3D CFD.
Our study:• 3 healthy subjects• 4 patients with craniospinal instabilities at the foramen
magnum
8Work in progress: A. Bunck, J. R. Kröger, Bryn A. Martin, et al. “Magnetic resonance velocity mapping of 4D cerebrospinal fluid flow in Chiari I malformation with and without syringomyelia,” European Journal of Radiology
Methods
A. C. Bunck et al. "Magnetic resonance 4D flow characteristics of cerebrospinal fluid at the craniocervical junction and the cervical spinal canal," European Radiology, vol. 21, pp. 1788‐96, Aug 2011.
• 4D MRI CSF flow measurements by A. C. Bunck et al.1
• Processing & visualization of 4D MRI with GT Flow software
• Reconstruction of the 3D geometry (ITK snap)
• CFD simulation (rigid wall / newtonian fluid (water) /
Unsteady flow / ANSYS CFX)
9
4D MRI in cervical spine
10
Healthy Chiari patient
Bunck, A. C. et al.Magnetic resonance 4D flow characteristics of cerebrospinal fluid at the craniocervical junction and the cervical spinal canal. European radiology 21, 1788‐1796, doi:10.1007/s00330‐011‐2105‐7 (2011).
Comparison of 4D MRI and 3D CFDCompare in terms of:1. Velocity profiles2. Peak flow velocities
11
Method for velocity profile comparison
12
High res. MRI scan
4D M
RI3D
CFD
*Noise in 4DMR required masking
*No masking needed for CFD
C1
C2M
C2P
C3
C4
C5
C6
C7
3D reconstruction
Flow boundary condition• Due to noise in 4D MRI, we used the 4D MRI location with the
greatest peak flow rate (best signal) for flow BC.• Flow was input at the caudal end of the model (P=0 at outlet)• Blunt velocity profile w/ geometry extended at flow inlet.
13
4D MRI
‐4.00
‐3.00
‐2.00
‐1.00
0.00
1.00
2.00
3.00
0.00 0.20 0.40 0.60 0.80
Flow
rate (cm3/s)
Time (s)
4D mri flow rate (cm3/s)
C1
C2P
C2M
C3
C4
C5
C6
C7
FM
th
Results: velocity profile comparison
14
A
PLR
Healthy
Data not yet processed
Mrs sHealthy 8 Healthy 3 Healthy 4 Mrs F Mr V Mr G
**
*
*
**
+
+
++
++
+
+
+
+
+
+++
+
+
↑
↑
↑
↑
↑
↑
Patients3D CFD 4D MRI
Patient (Mrs S)Healthy 8
15
3D CFD 4D MRI
A
P
LR
C7
C6
C5
C4
C3
C2P
3D CFD 4D MRI
C7
C6
C5
C4
C3
C2P1 cm
Velocity(cm/s)3.0
0.0 0.0
Velocity(cm/s)3.0
0.0 0.0
1 cm
Velocity(cm/s)6.0
Velocity(cm/s)9.0
↑
↑
↑
↑
↑
+
+
+
+
+
+ +
+
↑
↑
↑
Results: comparison of peak velocities• 4D MRI was consistently higher than 3D CFD• Differences were pronounced in patients near brain base (FM)
16
17
Data not yet processed
Mrs S Mrs F Mr V Mr G
* *
*
*
**
+
+
++
+ +
+
+
Patients
Interesting observation: For Mrs S. and Mrs F, CSF flow was present in the mid/lower cervical spine but not at the foramen magnum!
No flow?
18
Mr. V Mr. G Mrs. F Mrs. S
+
‐5
‐4
‐3
‐2
‐1
0
1
2
0 0.2 0.4 0.6
‐6
‐4
‐2
0
2
4
0.00 0.20 0.40 0.60
‐4
‐3
‐2
‐1
0
1
0 0.2 0.4 0.6 0.8
‐3
‐2
‐1
0
1
2
0 0.2 0.4 0.6
FM
C2MCSF flo
w (m
l/s)
Increasing tonsillar herniation
TissueMotion
Tonsillar herniation and CSF flow jetsMild herniation at FM
– Flow stenosis– Elevated pressure gradients at FM & stress on tissue– Flow Jets– Mild reduction in CSF pulse amplitude at FM– Mild tissue displacement– Tissue remodeling?
Severe herniation at FM– Flow blockage– No flow jets– CSF pulse zero at FM – Severe tissue displacement – Lower cervical CSF pulse from bulk tissue displacement?
19
Increasin
g tonsillar herniation
Why are there differences in the flow field obtained by subject specific 3D CFD
compared to 4D MRI?
20
Tissue motion• Alters flow field• Quantified in literature
(Cousins et al., Greitz)
21
Chiari patient
*Unpublished imaging video.
Cousins, J. & Haughton, V. Motion of the cerebellar tonsils in the foramen magnum during the cardiac cycle. AJNR. American journal of neuroradiology 30, 1587‐1588, doi:10.3174/ajnr.A1507 (2009).1991‐Greitz‐Neuroradiology‐CSF Piston
4D velocity (03/10/2010)
22
Error/noise from MRI
4D MRI in 3rd ventricle of brain4D MRI in aqueduct of Sylvius
Inconsistent / inaccurate CSF mesh
05/31/2010 04/27/2010 03/22/2010
04/22/2010 03/10/2010
All geometries from same personSegmentation performed by same personScans obtained on two different scanners
Error from neglect of fine structure• Arachnoid trebeculae within the spinal subarachnoid space were neglected. Is the distribution isotropic?
24(2) Cloyd, M. W. and F. N. Low (1974). "Scanning electron microscopy of the subarachnoid space in the dog. I. Spinal cord levels." The Journal of comparative neurology 153(4): 325‐368.
Error in CFD from neglect of other structures
• Nerve roots • Denticulate ligaments
25
• CFD in cervical spine with and w/o fine anatomy
7T
3T
Future directions
26Sigmund, E. E. et al. High‐resolution human cervical spinal cord imaging at 7 T. NMR in biomedicine, doi:10.1002/nbm.1809 (2011).
Our new model in development
Test 4D MRI measurements with in vitro model
27
L=
12.5
mm
L=
12.5
mm
Images courtesy of Dr. Navid Borhani, EPFL thermodynamics laboratory
Model 1 Model 2
Conclusion• Fine anatomy and tissue motion in the cervical spine appear
to have an important impact on CSF hydrodynamics.• To accurately model the cervical spine hydrodynamics in
disease we likely need to include fine anatomy and tissue motion.
• 4D MRI is useful to understand complex flow in craniospinal disorders.
28
29
Theresia Yiallourou
Prof. Nikos Stergiopulos
Work funded by: Swiss National Science FoundationAmerican Syringomyelia Alliance ProjectConquer ChiariEngineering Collaborators• Francis Loth, Ph.D. U. Akron• Francesco Santini, Ph.D., Uni. Basel• Matthias Stuber, Ph.D., CHUV CIBM• John N. Oshinski, Ph.D., Emory Univ.• Rick Labuda, Conquer Chiari• Jean‐Marie Drezet, Ph.D., EPFL• Jelena Bock, Ph.D., Uni. Freiburg • Eleonora Fonari, Ph.D., CHUV CIBM• Novak Elliott, Ph.D. Univ. Warwick• Chris Bertram, Ph.D. Project advise• Carlo Marcucci, M.D. CHUV • John D. Heiss, M.D. NIH USA• Christopher Bertram, Ph.D.• Lynne Bilston, Ph.D.• Vartan Kurtcuoglu, Ph.D., ETZ, CH
Thanks!
EPFL neurohydrodynamics team• Thiresia Yiallourou, EPFL, Ph.D. student• Benedict Danny, Royal Inst., M.S. student• Emelie Farine, M.S. student• Anass Chiki, M.S. student• Arthur Hirsch, EPFL, M.S. student• Shahim Kamal, EPFL LSMX, Ph.D. student• Serge Metarailler, EPFL, B.S. student• Leonie Asboth, EPFL, M.S. student• Christian Meuli, EPFL M.S. student• Aurélie Picquot, EPFL masters student• Adrien DeMuralt, EPFL masters student• Navid Borhani, Ph.D.Medical collaborators• Jan Novy, M.D. CHUV neurology• Lorenz Hirt, M.D., CHUV neurology• Celine Odier, M.D. CHUV neurology• Raphael Heinzer, M.D. CHUV sleep disorders• Jose Haba‐Rubio, M.D. CHUV sleep disorders• Andreas Gottaschalk, M.D., Germany • Alexander Bunck, M.D. Muenster, G.• Shahan Momjain, M.D., HUG, Switzerland• Bermans Iskandar, M.D. Univ. Wisc.
30
31
Questions?