Clin Neuroradiol DOI 10.1007/s00062-016-0530-3

ORIGINAL ARTICLE

Visualization of CSF Flow with Time-resolved 3D MR Velocity Mapping in Aqueductal Stenosis Before and After Endoscopic Third Ventriculostomy A Feasibility Study Sebastian Brandner1 · Michael Buchfelder1 · Ilker Y. Eyuepoglu1 · Hannes Luecking2 · Arnd Doerfler2 · Andreas Stadlbauer1

Received: 4 May 2016 / Accepted: 13 July 2016 © Springer-Verlag Berlin Heidelberg 2016

Abstract Purpose The aim of this study was to evaluate timed-resolved three-dimensional (3D) magnetic resonance (MR) velocity mapping as a method for investigation of cerebrospinal fluid (CSF) flow changes in patients with aqueductal stenosis (AS) treated by endoscopic third ventriculostomy (ETV). Methods The MR velocity mapping was performed in 12 AS patients before and after ETV and in 10 healthy volunteers by using a 3-Tesla MR system. Time-resolved 3D MR velocity mapping data were acquired with a standard 3D phase contrast (PC) sequence with cardiac triggering. Values of mean (vmean) and maximum (vpeak) velocity were measured at several ventricular structures using dedicated software. Results Of the patients 11 showed a satisfactory clinical improvement after ETV, whereas one patient needed subsequent shunt implantation. All AS patients showed significant hypomotile CSF flow dynamics in the third ventricle when compared to healthy subjects before surgery (p < 0.05). In contrast, CSF flow velocity was increased within the Foramen of Monro in AS patients. After ETV, all AS patients showed a decrease of CSF flow dynamics within the third ventricle. Mean and peak CSF flow velocities Electronic supplementary material The online version of this article (doi:10.1007/s00062-016-0530-3) contains supplementary material, which is available to authorized users.  Sebastian Brandner

[email protected] 1

Department of Neurosurgery, Friedrich-Alexander University Erlangen-Nuremberg (FAU), Schwabachanlage 6, 91054 Erlangen, Germany

2

Department of Neuroradiology, Friedrich-Alexander University Erlangen-Nuremberg (FAU), Erlangen, Germany

through the ventriculostomy were 1.72 ± 0.59 cm/s (vmean) and 3.53 ± 0.79 cm/s (vpeak), respectively after ETV. The patient who needed shunt implantation after ETV had the lowest flow velocities through the ventriculostomy. Conclusion This study demonstrates that timed-resolved 3D MR velocity mapping is a useful imaging investigation for diagnostics and follow-up in patients with AS. This new technique provides an insight into the physiological CSF flow changes related with AS and its treatment. Keywords Hydrocephalus · Aqueductal stenosis · Cerebrospinal fluid · Endoscopic ventriculostomy · Magnetic resonance imaging · Velocity mapping · Phase contrast imaging

Introduction Hydrocephalus is a very common condition in neurosurgical practice. Development of hydrocephalus may either result from an impaired absorption of the cerebrospinal fluid (CSF) appearing as malresorptive communicating hydrocephalus or occur due to obstructive lesions within or around the ventricular system. The most common form of obstructive hydrocephalus in adults is aqueductal stenosis (AS) [1]. While idiopathic AS represents an isolated stenosis of the aqueduct, secondary AS is caused by aqueductal compression by space-occupying periaqueductal lesions, such as cysts or tumors. Treatment of obstructive hydrocephalus can be achieved by surgical removal of an obstructive lesion if possible, by shunt placement or using neuroendoscopic methods, notably endoscopic third ventriculostomy (ETV) [2]. To guide the surgical treatment of patients with obstructive hydrocephalus it is essential to accurately identify the nature of the obstruction to CSF flow.

K

S. Brandner et al.

Taking into consideration that shunt placement is a surgical procedure with a high risk for complications with a shunt complication rate up to 50 % within 5 years after implantation [3], it becomes evident that identifying patients suffering from specific forms of obstructive hydrocephalus, such as AS or obstruction of the fourth ventricle, which are suitable for endoscopic treatment options with significantly lower complication rates is of crucial importance [2]; however, an accurate diagnosis and a prolonged patency of the ventriculostomy are crucial for a favorable outcome after ETV. Routine magnetic resonance imaging (MRI) is not very sensitive for the detection of AS and confirmation of patency of the ventriculostomy and diagnostic criteria are often subjective [4]. Additionally, flow void has proven to be no guarantee for free communication between the ventricular system and the subarachnoidal space after ETV [5]. Moreover, reduction in the size of the ventricular system or CSF stroke volume within the ventriculostomy or the basal cisterns has failed to accurately detect the patency of ETV [6, 7]. To date, time-resolved two-dimensional phase contrast MRI (2D-PC-MRI) with velocity encoding in one spatial direction is an established method for the investigation of CSF flow [6, 8]. Time resolved 3D MR velocity mapping has particularly been used for investigation of blood flow patterns within the vascular system [9–12]. We previously performed 3D MR velocity mapping in healthy subjects to investigate CSF flow patterns within the ventricular system and described different CSF flow dynamics in hydrocephalus patients [13, 14]. The purpose of this study was to evaluate the suitability of timed-resolved 3D MR velocity mapping as a diagnostic method in patients with AS and to assess the success of ETV.

Methods Subjects

We performed 3D MR velocity mapping in a total of 12 patients (7 female and 5 male, mean age 37.7 ± 21.0 years, age range 1–76 years) with AS, 1–7 days before and 5–7 days after ETV. Of these patients, 2 had previously had a ventriculoperitoneal shunt before ETV for 30 years and 10 years, respectively, and the ETV was performed in these patients due to shunt insufficiency. Adult patients presented with signs of intracranial hypertension, such as headache and nausea or with postural and gait disturbances. Infant patients presented with macrocephaly. Additionally, we performed 3D MR velocity mapping in 10 healthy volunteers (5 male and 5 female, mean age 26.7 ± 2.75 years). The study protocol was approved by the local ethics committee and informed consent was obtained from each subject.

K

MRI Data Acquisition

All MRI examinations were performed on a 3 Tesla MR scanner (Magnetom TIM Trio, Siemens Healthcare Sector, Erlangen, Germany) equipped with the standard 8-channel head coil. The routine MRI protocol consisted of (a) coronal and sagittal T2-weighted turbo spin-echo (TSE) sequences, (b) a transverse fluid-attenuated inversion recovery (FLAIR) sequence, (c) a transverse T1-weighted gradient-echo sequence and (d) a transverse diffusion-weighted imaging (DWI) sequence. Time-resolved 3D MR velocity mapping data were acquired using a vector electrocardiogram (vector ECG) triggered gradient echo (GE) PC sequence. The volume of interest (VOI) was chosen in sagittal orientation to cover the whole ventricular system including both lateral ventricles except for the lateral parts and the inferior horn of the lateral ventricles especially in hydrocephalus patients. Retrospective vector ECG gating was used to cover the entire cardiac cycle. The measurement parameters were: Field of view (FOV) = 195 × 195 mm2 (FH×AP), slice thickness =1.5 mm, voxel size = 1.5 × 1.5 × 1.5 mm3, TR/TE = 53/11 ms, flip angle = 15°, velocity encoding in the feet-head direction only and velocity encoding factor (venc) = 3 cm/s. Velocity encoding was performed in anterior-posterior, left-right and feet-head directions. The sequence yielded at least 12 quantitative flow-encoded 3D data sets per cardiac cycle depending on the heart rate of the subject. The measurement time for the velocity mapping experiment was approximately 10 min and the total measurement time including routine MRI protocol ranged between 20 and 25 min depending on subjects’ heart rate. MRI Data Analysis

Magnetic resonance velocity mapping data were loaded into the commercially available GTFlow software tool (version 1.4.13, GyroTools, Zurich, Switzerland) for calculation of time-resolved 3D CSF flow patterns in the ventricular system. The boundary of the five ventricular structures, i. e. the lateral ventricles, the foramen of Monro, the third ventricle, the aqueduct of Sylvius and the fourth ventricle as well as the surgically created ventriculostoma were outlined using contours that were manually defined. Postprocessing of MR velocity mapping data required approximately 30 min. More details of the post-processing were published previously [14]. Interpretation of CSF flow patterns was performed by visual inspection. The CSF flow in patients was classified as hypermotile flow if it showed increased dynamics and as hypomotile flow if it showed attenuated dynamics compared with healthy volunteers and with findings of a previously published study that reported no variability with age [14]. The maximum peak value (vpeak) and the mean value (vmean) of the magnitude of the

Visualization of CSF Flow with Time-resolved 3D MR Velocity Mapping in Aqueductal Stenosis Before and After Endoscopic Third...

p velocity vector, i. e. |v| D (vx2 + vy2 + vz2), where vx2, vy2 and vz2 represent velocity vector components, respectively, were determined for the five ventricular structures and the ventriculostoma. Surgical Procedure

The ETV was performed with the subjects under general anesthesia. In each case we used a rigid endoscope (Karl Storz, Tuttlingen, Germany) with a working length of 21.6 cm, an outer diameter of 4.0 mm and a 2.3 mm optic. Patients were placed in a supine position. A frontal burr hole was placed at typical location at Kocher’s point approximately 2.5 cm lateral from the midline and just anterior to the coronal suture. After incision of the dura mater, the endoscope was introduced together with the operating sheath into the lateral ventricle and maneuvered into the third ventricle through the foramen of Monro. After identification of the mammillary bodies and the tuber cinereum, the site of perforation was determined. Blunt forceps and a Fogarty balloon catheter were used in all cases for perforation and widening of the floor of the third ventricle thus achieving a diameter of the ventriculostomy of approximately 5 mm. The margin of the ventriculostomy was coagulated using monopolar coagulation. Finally, the

operating sheath was withdrawn together with the endoscope. External ventricular drains were not inserted. The burr hole was covered with a collagen sponge coated with fibrinogen and thrombin (TachoSil, Nycomed, Switzerland) to prevent a CSF fistula. The skin was closed using an atraumatic suture. Statistical Analysis

All values are presented as mean ± SD. Statistical analysis was performed using the SPSS® statistics software (IBM, Version 21). The Mann-Whitney U-test was used for analyzing group differences between AS patients and healthy volunteers. The Wilcoxon test was used for statistical evaluation of CSF flow changes in AS patients after ETV. Statistical significance was accepted for p < 0.05.

Results The MR velocity mapping was successfully performed in all 12 patients before and after ETV and in the 10 healthy volunteers. Fig. 1 illustrates the preoperative and postoperative velocity mapping in a representative patient. The dynamics of preoperative and postoperative CSF flow in the ventric-

Fig. 1 Preoperative and postoperative velocity mapping in a representative patient (45-year-old female patient who presented with headache, vertigo, and slight gait disturbance). The colour codes for the CSF velocity encoding of the velocity mapping data are shown on the right-hand side. The dynamics of preoperative and postoperative CSF flow in the ventricular system of this patient over the whole cardiac cycle in sagittal direction are given in Online Resource 1 and 2

K

S. Brandner et al.

vmean-volunteers = 0.45 ± 0.05 cm/s, p < 0.001 and vpeak-AS = 1.41 ± 0.70 cm/s, vpeak-volunteers = 0.79 ± 0.13 cm/s, p < 0.001). Mean and peak CSF flow velocities within the lateral ventricles were not significantly different between AS patients before ETV and healthy volunteers. Within the fourth ventricle we found a significantly lower peak velocity in AS patients before ETV compared to healthy volunteers (vpeak-AS = 1.20 ± 0.19 cm/s and vpeak-volunteers = 2.33 ± 0.74 cm/s, p < 0.005), whereas the vmean was not significantly different. Postoperative Changes of CSF flow Velocities After ETV

Fig. 2 a Mean (vmean) and b peak (vpeak) CSF flow velocities within the ventricular system in AS patients before treatment (black boxes), after ETV (light grey boxes) and in healthy volunteers (dark grey boxes). Significant differences (p < 0.005) between preoperative and postoperative CSF flow velocities are marked with asterisks (*), AS aqueductal stenosis, ETV endoscopic third ventriculostomy

ular system of this patient over the whole cardiac cycle in sagittal direction are presented in Online Resource 1 and 2. Mean and peak velocities of preoperative and postoperative CSF flow within the ventricular system of AS patients as well as of CSF flow in healthy volunteers are illustrated in Fig. 2. Differences of CSF Flow Velocities Between AS Patients Before ETV and Healthy Volunteers

Both mean and peak flow velocities through the aqueduct were significantly lower in AS patients before ETV than in healthy volunteers (vmean-AS = 0.59 ± 0.06 cm/s, vmean-volunteers = 0.68 ± 0.19 cm/s, vpeak-AS = 0.91 ± 0.22 cm/s and vpeak-volunteers = 1.90 ± 1.00 cm/s, p < 0.005). We also found a hypomotile CSF flow in AS patients before ETV within the third ventricle in comparison to healthy volunteers with significantly lower flow velocities (vmean-AS = 0.56 ± 0.10 cm/s, vmean-volunteers = 0.69 ± 0.14 cm/s, p < 0.05 and vpeak-AS = 1.31 ± 0.47 cm/s, vpeak-volunteers = 1.93 ± 0.65 cm/s, p < 0.05). In contrast, mean and peak CSF flow velocities through the foramen of Monro were significantly higher in AS patients before ETV in comparison to the group of healthy volunteers (vmean-AS = 0.78 ± 0.51 cm/s,

K

Of the patients 11 showed prolonged clinical improvement after ETV. In these patients MR velocity mapping performed 5–7 days after ETV showed a laminar flow through the surgically created ventriculostomy in all patients. Quantitative analysis of CSF flow revealed a mean CSF flow velocity through the ventriculostomy (vmean) of 1.72 ± 0.59 cm/s and a peak CSF flow velocity (vpeak) of 3.53 ± 0.79 cm/s. Additionally, we found a significantly increased CSF flow within the third ventricle compared to the preoperative findings (vmean-preop = 0.56 ± 0.10 cm/s, vmean-postop = 1.01 ± 0.51 cm/s, p < 0.005 and vpeak-preop = 1.31 ± 0.47 cm/s, vpeak-postop = 2.59 ± 1.04 cm/s, p < 0.005). In contrast, the hyperdynamic CSF flow velocity through the foramen of Monro increased after ETV even more (vmean-preop = 0.78 ± 0.51 cm/s, vmean-postop = 1.47 ± 0.63 cm/s, p < 0.005 and vpeak-preop = 1.41 ± 0.70 cm/s, vpeak-postop = 2.64 ± 1.19 cm/s, p < 0.005). Within the lateral ventricles, the fourth ventricle and the aqueduct we did not find significant changes of CSF flow velocities. The patient who had had a ventriculoperitoneal shunt 30 years prior to ETV showed recurrent clinical signs of elevated intracranial pressure few days after ETV. In this patient, the postoperative MR velocity mapping elucidated a persistent hypomotile CSF flow within the third ventricle and the CSF flow through the ventriculostomy was the lowest of all patients (vmean = 0.61 cm/s and vpeak = 1.54 cm/s). In contrast, in the patients with successful ETV (n = 11) the lowest vmean and vpeak values within the ventriculostomy were 1.01 and 2.37 cm/s, respectively. Due to progressive clinical deterioration the patient received a new ventriculoperitoneal shunt.

Discussion Due to the low morbidity associated with this technique, ETV has become the treatment of choice in AS [6]. Although success rates are assumed to be high, the determination of the effectiveness of the new internal CSF pathway is of crucial importance in predicting a favorable outcome.

Visualization of CSF Flow with Time-resolved 3D MR Velocity Mapping in Aqueductal Stenosis Before and After Endoscopic Third...

Routine MRI is limited by some restrictions for preoperative and postoperative evaluation in AS. Although numerous imaging parameters have been evaluated to assess success of ETV none of them, including reduction of the size of the ventricular system or CSF stroke volume within the ventriculostomy or the basal cisterns, have proven to be accurate in differentiating between successful and failed ETV [2, 6, 7, 15]. Radionuclide cisternography was found to be more sensitive for detecting AS and the patency of the ventriculostomy after ETV [2]; however, cisternographic and ventriculographic studies are very invasive and cumbersome to perform in routine practice [4]. In the present study, we assessed time-resolved 3D MR velocity mapping for preoperative and postoperative evaluation in AS patients before and after ETV. The advantages of time-resolved 3D MR velocity mapping can be described as follows: (a) velocity encoding in all three dimensions additionally allows for detection of CSF flow in the left-right and anterior-posterior directions, (b) calculation of velocity vectors for the whole ventricular system enables assessment of flow patterns and vortices in the whole ventricular system and not only at the suspected location of abnormality (e. g. aqueduct) and (c) color coding of the magnitude of the velocity vector allows for a more direct assessment of CSF flow [13]. Before ETV we found a hypomotile CSF flow within the aqueduct and the third ventricle with significantly decreased CSF flow velocities when compared to the CSF flow of healthy volunteers. In contrast, we found significantly increased CSF flow velocities through the foramen of Monro in AS patients. In healthy individuals CSF passes from the lateral and third ventricles through the aqueduct into the fourth ventricle and finally into the basal cisterns and the spinal canal during the systolic increase of the cerebral blood volume. During the diastolic decrease of the cerebral blood volume the CSF flow returns in the opposite direction. Thus, in physiological conditions, the cardiac cycle-related CSF flow is pulsatile and bidirectional [16]. In AS only a pulsatile flow between the lateral ventricles and the third ventricle is possible, while the physiological CSF outflow into the extracranial space is impaired [7]. The increased CSF flow velocities through the foramen of Monro in AS patients before ETV might reflect this condition in which the increased intraventricular pressure leads to higher pulsatile velocities through the foramen of Monro. Interestingly, flow velocities through the foramen of Monro even increased after ETV. This might be explained by the formation of a new CSF flow pathway which is characterized by a lower flow resistance compared to the physiological pathway via the aqueduct. This may be associated with an increase in flow velocity in the foramen of Munro. After ETV, we additionally observed significantly increased CSF flow velocities within the third ventricle compared to

the preoperative velocities with a laminar flow through the ventriculostomy. We determined a mean CSF flow velocity through the ventriculostomy of 1.72 ± 0.59 cm/s and a peak CSF flow velocity of 3.53 ± 0.79 cm/s in the patients who clinically improved after the procedure. We found the lowest mean and peak velocities through the ventriculostomy in the single patient with a subsequent deterioration after ETV requiring insertion of a ventriculoperitoneal shunt. This finding clearly indicates a prognostic role of the determined flow velocities through the ventriculostomy for the outcome after ETV, as a favorable outcome after ETV depends on an adequate flow through the new surgically created pathway. On another note, the adequate flow is not only determined by the patency of the ventriculostomy but also by the adequate capacity of reabsorption of the CSF [17, 18]. As a consequence, the low flow velocities in the patient who subsequently needed a ventriculoperitoneal shunt placement after ETV might be caused by an impaired capacity of CSF reabsorption with a shunt history of 30 years. In summary, our data provide new insights into CSF flow patterns in AS patients and indicate that the determination of flow velocities within the ventriculostomy using 3D timeresolved MR velocity mapping can be used for prediction of effectiveness and clinical outcome after EVT. Although performing the postoperative MR velocity mapping relatively early after EVT could be considered a limiting factor in this study, it still provided a baseline postoperative evaluation of the surgery as well as a foundation for the prediction of effectiveness of the ventriculostomy. Nevertheless, the ability of 3D time-resolved velocity mapping for longterm outcome prediction and possible changes of CSF need further refinement in future studies. Acknowledgements This study was supported by a grant from the ELAN program (Erlanger Leistungsbezogene Anschubfinanzierung und Nachwuchsförderung, grant number 14-03-04-1-Brandner)

Compliance with Ethical Guidelines Conflict of Interests S. Brandner, M. Buchfelder, I.Y. Eyuepoglu, H. Luecking, A. Doerfler and A. Stadlbauer state that there are no conflicts of interest. Ethical standards All studies on humans described in this manuscript were carried out with the approval by the local ethics committee and performed in accordance with national law and the Declaration of Helsinki from 1975 (in its current revised form). Informed consent was obtained from all subjects included in the study.

K

S. Brandner et al.

References 1. Höglund M, Tisell M, Wikkelsø C. Incidence of surgery for hydrocephalus in adults surveyed: same number afflicted by hydrocephalus as by multiple sclerosis. Lakartidningen. 2001;98:1681–5. 2. Tisell M. How should primary aqueductal stenosis in adults be treated? A review. Acta Neurol Scand. 2005;111:145–53. 3. Borgbjerg BM, Gjerris F, Albeck MJ, Hauerberg J, Børgesen SE. Frequency and causes of shunt revisions in different cerebrospinal fluid shunt types. Acta Neurochir (Wien). 1995;136:189–94. 4. Algin O, Hakyemez B, Parlak M. Phase-contrast MRI and 3D-CISS versus contrast-enhanced MR cisternography on the evaluation of the aqueductal stenosis. Neuroradiology. 2010;52:99–108. 5. Hopf NJ, Grunert P, Fries G, Resch KD, Perneczky A. Endoscopic third ventriculostomy: outcome analysis of 100 consecutive procedures. Neurosurgery. 1999;44:795–804, discussion 804–6. 6. Bargalló N, Olondo L, Garcia AI, Capurro S, Caral L, Rumia J. Functional analysis of third ventriculostomy patency by quantification of CSF stroke volume by using cine phase-contrast MR imaging. AJNR Am J Neuroradiol. 2005;26:2514–21. 7. Lev S, Bhadelia RA, Estin D, Heilman CB, Wolpert SM. Functional analysis of third ventriculostomy patency with phase-contrast MRI velocity measurements. Neuroradiology. 1997;39:175–9. 8. Abbey P, Singh P, Khandelwal N, Mukherjee KK. Shunt surgery effects on cerebrospinal fluid flow across the aqueduct of Sylvius in patients with communicating hydrocephalus. J Clin Neurosci. 2009;16:514–8. 9. Markl M, Draney MT, Hope MD, Levin JM, Chan FP, Alley MT, Pelc NJ, Herfkens RJ. Time-resolved 3-dimensional velocity mapping in the thoracic aorta: visualization of 3-directional blood flow patterns in healthy volunteers and patients. J Comput Assist Tomogr. 2004;28:459–68.

K

10. Markl M, Kilner PJ, Ebbers T. Comprehensive 4D velocity mapping of the heart and great vessels by cardiovascular magnetic resonance. J Cardiovasc Magn Reson. 2011;13:7. 11. Bammer R, Hope TA, Aksoy M, Alley MT. Time-resolved 3D quantitative flow MRI of the major intracranial vessels: initial experience and comparative evaluation at 1.5 and 3.0 T in combination with parallel imaging. Magn Reson Med. 2007;57:127–40. 12. Wetzel S, Meckel S, Frydrychowicz A, Bonati L, Radue EW, Scheffler K, Hennig J, Markl M. In vivo assessment and visualization of intracranial arterial hemodynamics with flow-sensitized 4D MR imaging at 3 T. AJNR Am J Neuroradiol. 2007;28:433–8. 13. Stadlbauer A, Salomonowitz E, Brenneis C, Ungersböck K, van der Riet W, Buchfelder M, Ganslandt O. Magnetic resonance velocity mapping of 3D cerebrospinal fluid flow dynamics in hydrocephalus: preliminary results. Eur Radiol. 2012;22:232–42. 14. Stadlbauer A, Salomonowitz E, van der Riet W, Buchfelder M, Ganslandt O. Insight into the patterns of cerebrospinal fluid flow in the human ventricular system using MR velocity mapping. Neuroimage. 2010;51:42–52. 15. Algin O. Role of complex hydrocephalus in unsuccessful endoscopic third ventriculostomy. Childs Nerv Syst. 2010;26:3–4, author reply 5–6. 16. Enzmann DR, Pelc NJ. Normal flow patterns of intracranial and spinal cerebrospinal fluid defined with phase-contrast cine MR imaging. Radiology. 1991;178:467–74. 17. Drake JM. Ventriculostomy for treatment of hydrocephalus. Neurosurg Clin N Am. 1993;4:657–66. 18. Jones RF, Stening WA, Kwok BC, Sands TM. Third ventriculostomy for shunt infections in children. Neurosurgery. 1993;32:855–9, discussion 860.