CSF flow imaging in Chiari 1 malformation Wende

CSF flow imaging in Chiari 1 malformation Wende Gibbs, MD, Department of Neuroradiology Gabriel Zada, MD, Department of Neurosurgery John Liu, MD, Department of Neurosurgery Patrick Hsieh, MD, Department of Neurosurgery Meng Law, MD, MBBS, Department of Neuroradiology University of Southern California, Keck School of Medicine Control # 1775 e. Ed. E-227

Disclosures Ø Wende Gibbs: none Ø Gabriel Zada: none Ø John Liu: none Ø Patrick Hsieh: none Ø Meng Law: Toshiba Grant Speakers Bureau; Bracco speaker and consultant; Guerbet Medical Advisory Board; Prism Stock; Fuji speaker and consultant

Purpose Ø Chiari 1 malformation (CM 1) has traditionally been defined by morphologic criteria: cerebellar tonsil herniation 3 -6 mm below the foramen magnum Ø It is increasingly clear that CM 1 is a complex disorder resulting from not only abnormal anatomy, but disordered CSF flow Ø This exhibit will review static and dynamic imaging tools applied to the study of CM 1, including new techniques that may improve diagnostic accuracy, patient management and surgical outcome

Approach Ø Review theories of pathogenesis of CM 1, focusing on the interplay of anatomy and CSF flow dynamics Ø Describe established and recently developed neuroimaging tools used to study CM 1: traditional MRI sequences and dynamic techniques including phase contrast MR and Time-spatial labeling inversion pulse technique (Time-SLIP) Ø Demonstrate the utility of Time-SLIP in providing a rapid, individualized assessment of CSF flow before and after surgical treatment

Introduction Ø CM 1 has historically been described as a change in hindbrain morphology characterized by cerebellar tonsil herniation 3 -6 mm* below the foramen magnum on sagittal MRI or CT images 1 -3 Ø However, 30 - 50% of individuals with tonsillar herniation greater than 5 mm are asymptomatic 4 Ø In those individuals with symptoms, the degree of herniation does not correlate well with symptom severity 5 Ø Further, a subset of patients with CM 1 symptoms have no tonsillar herniation, suggesting abnormal morphology alone cannot explain CM 1 symptomatology 6 *The degree of herniation is not universally agreed upon, and depends upon age

Introduction Both abnormal anatomy and aberrant CSF flow dynamics contribute to the pathophysiology of CM 1 Anatomy/Morphology Posterior fossa anatomy Tonsillar herniation Spinal canal geometry CSF dynamics Complex CSF flow patterns Resistance to flow Abnormal flow velocity Pressure gradients Craniospinal compliance Pathology/Symptoms Syrinx Suboccipital headache Motor or sensory dysfunction Fatigue Pain

Introduction Ø Suboccipital decompressive surgery is a standard treatment for CM 1 Ø 35 -45% of patients have minimal or no relief years after surgery 7, 8 Ø The search for a noninvasive method of selecting patients with CM 1 symptoms who will benefit from surgical intervention is an active area of research across many fields

CM 1 Imaging Ø The observation of cerebellar tonsillar ectopia in the absence of syrinx, hydrocephalus, or suggestive signs and symptoms has uncertain clinical significance Ø When signs and symptoms are present, neuroimaging is vital to diagnosis and management Symptoms Ø Headache (typically suboccipital) Ø Neck, back, face pain Ø “Cape” pain: neck, upper back, shoulders Ø Nonradicular limb pain Ø Weakness Ø Dizziness Ø Vertigo Ø Slurred speech Ø Syncope Ø Difficulty swallowing Ø Tinnitus Signs Cranial nerve: Ø Dysphagia Ø Dysarthria Ø Hoarseness Ø Cough Ø Nystagmus Cerebellar Ø Ataxia Ø Dysmetria Brainstem Ø Nystagmus Ø Sleep apnea Ø Sensorineural Ø Spastic gait hearing loss Ø Urinary Ø Hypertension incontinence, frequency Ø Sinus bradycardia/tachy Ø Extremity cardia weakness Ø Syncope Spinal cord Ø Hyperactive reflexes Ø Babinski, Hoffman reflex

CM 1 Imaging Ø Routine imaging sequences are obtained to evaluate for hydrocephalus, syrinx, or other craniocervical junction (CCJ) pathology Ø The midline sagittal image is used to quantify cerebellar tonsillar ectopia in relation to a line connecting the basion and opisthion (Mc. Rae’s line, foramen magnum)

CM 1 Imaging: Classification Ø Ø Ø Chiari 0: Ø Tonsils descend 3 mm or less below the foramen magnum Ø Syrinx Ø +/- crowding at foramen magnum Chiari 1: Ø Greater than 5 mm tonsillar descent in age >15 year Ø Greater than 6 mm tonsillar descent in age < 15 years Ø 3 -5 mm is borderline, and abnormal if syrinx or symptoms Ø 4 th ventricle remains in normal position Chiari 1. 5: Ø Herniation of tonsils Ø Elongation and displacement of 4 th ventricle and brainstem Chiari 0 and Chiari 1. 5 are controversial classifications

CM 1 Imaging Static features of CM 1 - anatomy and morphology – are studied with traditional MRI and CT sequences Degree of tonsillar herniation Tonsillar shape Posterior fossa volume Posterior fossa crowdedness Linear Posterior Fossa Measures: Ø Clivus Ø Supraocciut Ø Twining line Ø Mc. Rae Line Dynamic aspects of CM 1 are evaluated with CSF flow techniques and computational fluid dynamic simulations CSF velocity CSF stroke volume Tonsil and cord movement Pressure Resistance to flow Craniospinal compliance

CM 1 Imaging: Morphology Ø The morphologic abnormality in CM 1 is diverse Ø In general, CM 1 is characterized by: Ø Ø Pointed configuration of the tonsils Ø More vertically oriented cerebellar folia Ø Crowded foramen magnum Ø Narrowed retrocerebellar and premedullary subarachnoid space Ø Lower limits of normal or small posterior fossa Ø Short clivus Ø Inferior elongation of the 4 th ventricle with mildly low-lying nucleus gracilis (the demarcation of obex and central canal) 40 -80% of symptomatic CM 1 have a syrinx 4

CM 1 Imaging: Morphology Ø Posterior cranial fossa (PCF) volumetry is a potential predictor of surgical outcome 9, 10 Ø Alpern et. al. studied 20 morphologic and physiologic measures, of which 10 were found to discriminate CM 1 from healthy controls better than tonsillar herniation alone Ø The three parameters that best characerized CM 1: Ø Cord displacement Ø Posterior cranial fossa crowdedness Ø Normalized posterior cranial fossa volume Ø Using these three parameters, 37 healthy subjects and 35/36 CM 1 subjects were correctly classified 10

CM 1 Imaging: Morphology Ø Complex CMI (c. CMI) has recently been described in the neurosurgical literature as a CMI variant with more severe clinical phenotype Ø Recognition by the radiologist is useful as c. CMI may require more extensive or repeat neurosurgical procedures Ø Moore and Moore evaluated a number of morphologic measures and found that obex level was the most important differentiating factor between CMI and c. CMI 11 Ø Inferior herniation of the obex below the foramen magnum (FM) and a prominent dorsal bump was observed in all patients with c. CMI in their study Ø Typical CMI is characterized by obex above or at the FM Complex CMI: The obex lies just below the FM (arrow).

CM 1 Imaging: Dynamic Ø CSF dynamics in the cranial and spinal subarachnoid space may be equally important to morphology in the pathophysiology of CM 1 Ø CSF velocity, resistance to flow, pressure, and craniospinal compliance cannot be measured with static MR techniques Ø Dynamic evaluation of CSF flow has primarily been studied with 2 D phase contrast MRI Ø New techniques developed to study flow include 4 D PC MRI and Time inversion recovery pulse (Time. SLIP)

Morphology Abnormal morphology of cerebellar tonsils at the FM: -crowding of neural structures -narrowed subarachnoid space Obstruction of CSF pulsations Cranial arterial driving pressure forces same volume of CSF against obstruction Increased pressure: may further displace or damage neural structures Hydrodynamics Increased pressure gradient Elevated CSF velocity Altered craniospinal compliance Increased resistance to CSF flow Increased pressure: over time may alter neural elasticity, permeability, water content Surgical decompression alleviates crowding, results in decreased peak CSF velocity, and alters craniocervical CSF compliance

CM 1 Imaging: Phase Contrast Ø 2 D PC MRI in axial and/or sagittal orientation has been used to quantitatively and qualitatively evaluate dynamic CSF features such as: Direction of flow Ø Peak CSF velocity Ø Pulse wave velocity in the subarachnoid space Ø Relative timing of CSF and arterial pulsations Ø Ø Before PC MR images are acquired, maximum CSF velocity must be anticipated in order to set the Venc (velocity encoding) To optimize signal, CSF velocity should be the same or slightly below the venc Ø Velocities above the Venc produce aliasing artifact Ø Velocities significantly below the Venc have weak signal Ø

CM 1 Imaging: Phase Contrast Ø Magnitude and phase images provide information about anatomy and velocity Ø The phase image, reflecting spin phase shifts, is the most sensitive to flow Ø Quantitative information is acquired with images in the axial plane with through-plane velocity encoding in the craniocaudal direction Ø Qualitative features of flow are observed in the sagittal plane with in -plane velocity encoding Ø Peripheral cardiac gating allows for collection of 12 -24 phases during the repetition interval, depending on HR Ø By convention, bright signal reflects caudal motion during systole and dark signal represents cranial motion during diastole

CM 1 Imaging: Phase Contrast Phase images in sagittal orientation in cine mode “White” flow is moving caudally during systole “Black” flow is moving cranially during diastole In this patient with CMI, there is craniocaudal flow ventral to the brainstem and upper cervical cord Craniocaudal flow dorsal to the tonsils and cord is minimal. Click to play cine clip.

CM 1 Imaging: Phase Contrast Ø The majority of studies utilizing 2 D PC MRI show that CM 1 is characterized by elevated peak CSF flow velocity at the foramen magnum, and that peak velocity decreases after decompression Ø However, there is not an established correlation between change in velocity and the degree of clinical improvement Ø CM 1 is characterized by inhomogeneous flow patterns and simultaneous bidirectional flow: important findings confirmed in subsequent studies using different techniques, including 4 D PC MRI and computation flow models 5, 12, 13

CM 1 Imaging: Phase Contrast Ø Mc. Girt et al. found that pediatric CM 1 patients with normal CSF flow at the FM as assessed by PC MR were 4. 8 -fold more likely to experience symptom recurrence following surgery regardless of the degree of tonsillar herniation or presence of syrinx 8 Ø Abnormal ventral and dorsal flow was associated with a 2. 6 -fold reduction in risk of symptom recurrence after surgery Ø These findings support the role of inhomogeneous flow patterns in CM 1 pathophysiology

CM 1 Imaging: Phase Contrast Ø Time-resolved three directional velocity encoded phase contrast MRI (4 D PC MRI) is a recent advance that can better assess the three dimensional complexities of the CSF flow field Ø Using 4 D PC MRI, Bunck et al. showed that in CM 1, the anterior subarachnoid space (SAS) is markedly narrowed, with CSF flow diversion to the anterolateral SAS Ø This results in flow jets with elevated velocities and flow vortices 14 A B Coronal 4 D PC MRI images in control (A) and CM 1 (B). Compare uniform, homogeneous flow in A, with lateral flow diversion and left sided flow jet in B. Bunck et al, Eur Radiology (2012) 22: 1860 -1870.

CM 1 Imaging: Phase Contrast Ø Peak CSF velocities were significantly greater at the craniocervical junction in CM 1 patients than in controls, a finding in most, but not all prior studies using the 2 D PC MR technique Ø The volumetric measurement facilitated by the 4 D technique demonstrated variability among patients as to the level where peak systolic flow was found Ø Inconsistent results in prior studies may relate to the inability of the 2 D technique to capture the correct level for peak flow measurement

CM 1 Imaging: Time-SLIP Ø Another recently developed MR technique applied to the study of CSF flow dynamics is Time-spatial labeling inversion pulse (Time-SLIP) Ø Time-SLIP is based upon the arterial spin labeling concept Ø In this case, instead of flowing blood, CSF is used as an endogenous tracer Ø Advantages over phase contrast include: Ø Superior anatomic detail Ø Shorter acquisition time Ø Improved evaluation of non periodic or turbulent flow

CM 1 Imaging: Time-SLIP Ø First the background signal is suppressed with a non selective inversion recovery pulse Ø This is followed by a second, spatially selective pulse perpendicular to the direction of flow Ø When images are acquired, the labeled CSF flows into regions of suppressed background with high conspicuity Ø CSF bulk flow can be observed for up to 5 seconds before contrast between tagged and non tagged CSF is lost

Time-SLIP A nonselective IR pulse is applied, inverting all signal in the field of view A second, spatially selective inversion pulse is applied to the region of interest After a short period of time, tagged CSF is seen moving into the nontagged background (red arrow)

CM 1 Imaging: Time-SLIP A B Time-SLIP in CM 1 after surgical decompression. The initial image (A) shows the location of the selective pulse (gold lines). Tagged CSF is bright in this slice. All other CSF is suppressed (dark). After several seconds (B) tagged CSF is seen above and below the slice (gold lines), ventral to the brainstem and cord and dorsal to the cerebellum and cord (orange arrows). *Note the exquisite anatomic detail of the images allowing precise localization of CSF flow, a significant advantage over PC MR images.

CM 1 imaging: Time-SLIP Ø In this cine clip, we watch the movement of CSF over 5 seconds. Gold lines mark the selective pulse with tagged (bright) CSF Ø Notice the movement of CSF above and below the slice with time Ø We observe features of flow not possible by any other technique: Ø CSF moves from the 4 th ventricle superiorly into the aqueduct Ø Turbulent flow is seen in the 4 th ventricle (moving dark lines in the midst of bright CSF) Ø CSF moves within the cervical syrinx Ø At the end of 5 seconds, contrast between Click to play cine clip. tagged and untagged CSF is lost

CM 1 Imaging: Time-SLIP Click to play cine clip. 2 D Phase Contrast Click to play cine clip. Time-SLIP 2 D PC demonstrates the presence and direction of flow. Time. SLIP allows better visualization of location of flow, as well as periodic and turbulent flow.

Case 1: 35 year-old man with 2 years of worsening headache, facial pain, and developing slurred speech. MRI demonstrated tonsillar herniation 16 mm below the foramen magnum. After decompression with C 1 and partial C 2 laminectomies and duraplasty, the patient had marked improvement of his headaches and resolution of his facial pain. A. Initial imaging. Sagittal T 1 -W image shows ectopic, pointed cerebellar tonsils and crowding at the foramen magnum (FM). A B B. The preoperative Time-SLIP image demonstrates flow ventral the cervical cord (blue arrow). No flow is seen dorsal to the cord below the FM (gold arrow). Yellow lines indicate the tagged slice. All bright CSF above and below the lines has flowed from the tagged slice. C. Post decompression sagittal T 2 -W image reveals relief of crowding at the FM. More CSF is seen dorsal and inferior to the tonsils. A small amount of fluid is seen posterior to the duraplasty. C D D. Post operative Time-SLIP image again shows CSF flow ventral to cervical cord (blue arrow). There is now flow dorsal to the cord at this level (gold arrow).

Case 2: 43 -year-old man with 10 year history of upper and lower extremity weakness and numbness, ataxia, increasing difficulty with fine motor control. Click to play cine clip. A The pre-operative Time-SLIP image demonstrates flow ventral to the cord, but only trace flow dorsal to the cord at the tagged level (level marked by gold dots). A small amount of flow is also seen within the syrinx. Click to play cine clip. B The post-decompression study reveals increased flow dorsal to the cord. Craniocaudal flow within the syrinx has also increased. Interestingly, there is a small amount of cranial flow of fluid within the pseudomeningocele. This may relate to respiration. The high intrinsic signal to noise and temporal resolution of Time-SLIP in comparison to 2 D PC MRI allows visualization of CSF movement in response to respiration. Respiratory motion may have a greater effect on CSF flow than cardiac pulsation. 15

Case 3: 58 -year-old man with complicated history of cervical stenosis with myelopathy post C 6 -T 1 laminectomy and fusion one year ago. Increasing difficulty with gait, balance, and left leg weakness prompted imaging. He was found to have an increase in his preexisting tonsillar ectopia. He underwent posterior fossa decompression, C 1 laminectomy, and duraplasty. A B 5/2014. The preoperative image shows crowding at the FM, increased from prior studies. The focus of abnormal T 2 signal in the cord at C 7/T 1 reflects myelomalacia. The post decompression image shows relief of crowding. There is a large pseudomeningocele. Click to play cine clip. C Click to play cine clip. D 8/2014 and 12/2014. The patient had only partial improvement in symptoms. Time-SLIP shows that there is no flow dorsally at the FM. The brisk ventral flow is already apparent on the first image. 4 months later, the pseudomeningocele is noted to be larger. There is no change in CSF flow pattern. There is no cranial flow in the fourth ventricle, an abnormal finding.

Case 4. 49 -year-old woman with history of suboccipital headaches. Click to play cine clip. Note the short clivus, small posterior fossa, and superiorly oriented straight sinus, typical of CM 1. Crowding at the FM is minimal but there is a large cervical syrinx (Chiari 0? ) Click to play cine clip. Note the change in shape of the syrinx in relation to the flow in the ventral and dorsal SAS at that level. No flow is seen dorsally at the FM. Note the cranially-directed flow through the aquaduct into the third ventricle, a normal finding. Selective tagging pulses can be performed at multiple levels and in different orientations, as long as the slice is perpendicular to the flow direction of interest. A coronal orientation can evaluate flow between the lateral and third ventricles.

Time-SLIP Ø Investigators are currently devising methods to quantify flow velocity on Time-SLIP 16 Ø PC MR and Time-SLIP provide complementary information for the characterization of pathologies with aberrant CSF flow dynamics Ø CSF dynamics visualized with Time-SLIP differ from classic CSF circulation theories, and the development of this method has advanced knowledge of CSF physiology Ø A better understanding of CSF dynamics in health and disease may lead to increased diagnostic accuracy and better patient selection for surgical interventions

Summary Ø Historically and currently, neuroimaging is vital to diagnosis and management of CM 1 Ø The search for a noninvasive method of selecting patients with CM 1 symptoms who will benefit from surgical intervention is an active area of research across many fields Ø Emerging techniques such as 4 D PC MRI and Time. SLIP are providing unique insights into CSF flow dynamics in CM 1 and other pathologies resulting from disordered CSF flow dynamics

References 1. Barkovich AJ, Wippold FJ, Sherman JL, Citrin CM. Significance of cerebellar tonsillar position on MR. AJNR Am J Neuroradiol. 1986; 7(5): 795 -9. 2. Elster AD, Chen MY. Chiari I malformations: clinical and radiologic reappraisal. Radiology. 1992; 183(2): 347 -53. 3. Milhorat TH, Chou MW, Trinidad EM, et al. Chiari I malformation redefined: clinical and radiographic findings for 364 symptomatic patients. Neurosurgery. 1999; 44(5): 1005 -17. 4. Osborn AG. Osborn's Brain, Imaging, Pathology, and Anatomy. Lippincott Williams & Wilkins; 2012. 5. Iskandar BJ, Quigley M, Haughton VM. Foramen magnum cerebrospinal fluid flow characteristics in children with Chiari I malformation before and after craniocervical decompression. J Neurosurg. 2004; 101(2 Suppl): 169 -78. 6. Sekula RF, Jannetta PJ, Casey KF, Marchan EM, Sekula LK, Mccrady CS. Dimensions of the posterior fossa in patients symptomatic for Chiari I malformation but without cerebellar tonsillar descent. Cerebrospinal Fluid Res. 2005; 2: 11. 7. Aliaga L, Hekman KE, Yassari R, et al. A novel scoring system for assessing Chiari malformation type I treatment outcomes. Neurosurgery. 2012; 70(3): 656 -64. 8. Mcgirt MJ, Nimjee SM, Fuchs HE, George TM. Relationship of cine phase-contrast magnetic resonance imaging with outcome after decompression for Chiari I malformations. Neurosurgery. 2006; 59(1): 140 -6. 9. Badie B, Mendoza D, Batzdorf U. Posterior fossa volume and response to suboccipital decompression in patients with Chiari I malformation. Neurosurgery. 1995; 37(2): 214 -8. 10. Alperin N, Loftus JR, Oliu CJ, et al. Magnetic resonance imaging measures of posterior cranial fossa morphology and cerebrospinal fluid physiology in Chiari malformation type I. Neurosurgery. 2014; 75(5): 515 -22. 11. Moore HE, Moore KR. Magnetic resonance imaging features of complex Chiari malformation variant of Chiari 1 malformation. Pediatr Radiol. 2014; 44(11): 1403 -11. 12. Haughton VM, Korosec FR, Medow JE, Dolar MT, Iskandar BJ. Peak systolic and diastolic CSF velocity in the foramen magnum in adult patients with Chiari I malformations and in normal control participants. AJNR Am J Neuroradiol. 2003; 24(2): 169 -76. 13. Haughton V, Mardal KA. Spinal fluid biomechanics and imaging: an update for neuroradiologists. AJNR Am J Neuroradiol. 2014; 35(10): 1864 -9. 14. Bunck AC, Kroeger JR, Juettner A, et al. Magnetic resonance 4 D flow analysis of cerebrospinal fluid dynamics in Chiari I malformation with and without syringomyelia. Eur Radiol. 2012; 22(9): 1860 -70. 15. Yamada S, Miyazaki M, Yamashita Y, et al. Influence of respiration on cerebrospinal fluid movement using magnetic resonance spin labeling. Fluids Barriers CNS. 2013; 10: 36. 16. Shiodera T, Nitta S, Takeguchi T, et al. Automated flow quantification for spin labeling MR imaging. MAGMA. 2014; 27(5): 425 -33.