laboratory investigation J Neurosurg 125:1080–1086, 2016

Convection-enhanced delivery improves MRI visualization of basal ganglia for stereotactic

Aaron E. Bond, MD, PhD,1 Robert F. Dallapiazza, MD, PhD,1 M. Beatriz Lopes, MD, PhD,2 and W. Jeffrey Elias, MD1

Departments of 1Neurosurgery and 2Neuropathology, University of Virginia Health Sciences Center, Charlottesville, Virginia

Objective Stereotactic deep stimulation surgery is most commonly performed while patients are awake. This allows for intraoperative clinical assessment and electrophysiological target verification, thereby promoting favorable outcomes with few side effects. Intraoperative CT and MRI have challenged this concept of clinical treatment validation. Image-guided surgery is capable of delivering electrodes precisely to a planned, stereotactic target; however, these methods can be limited by low anatomical resolution even with sophisticated MRI modalities. The authors are developing a novel method using convection-enhanced delivery to safely manipulate the extracellular space surrounding common anatomical targets for surgery. By altering the extracellular content of deep subcortical structures and their associated white matter tracts, the MRI visualization of the basal ganglia can be improved to better define the anatomy. This tech- nique could greatly improve the accuracy and success of stereotactic surgery, potentially eliminating the reliance on awake surgery. Methods Observations were made in the clinical setting where vasogenic and cytotoxic edema improved the MRI visualization of the basal ganglia. These findings were replicated in the experimental setting using an FDA-approved intracerebral catheter that was stereotactically inserted into the or basal ganglia of 7 swine. Five swine were infused with normal saline, and 2 were infused with autologous CSF. Flow rates varied between 1 μl/min to 6 μl/min to achieve convective distributions. Concurrent MRI was performed at 15-minute intervals to monitor the volume of infusion and observe the imaging changes of the deep subcortical structures. The animals were then clinically observed, and necropsy was performed within 48 hours, 1 week, or 1 month for histological analysis. Results In all animals, the white matter tracts became hyperintense on T2-weighted imaging as compared with basal ganglia nuclei, enabling better definition of the deep brain anatomy. The volume of distribution and infusion (Vd/Vi ratio) ranged from 2.5 to 4.5. There were no observed clinical effects from the infusions. Histological analysis demonstrated mild neuronal effects from saline infusions but no effects from CSF infusions. Conclusions This work provides the initial foundation for a novel approach to improve the visualization of deep brain anatomy during MRI-guided, stereotactic procedures. Convective infusions of CSF alter the extracellular fluid content of the brain for improved MRI without evidence of clinical or toxic effects. http://thejns.org/doi/abs/10.3171/2015.10.JNS151154 Key Words convection-enhanced delivery; basal ganglia; MRI; stereotactic ; image-guided; functional neurosurgery

tereotactic surgery is commonly used to treat Image-guided techniques have been developed to posi- chronic neurological disorders such as Parkinson’s tion a stimulating or lesioning electrode with a mean vec- disease, tremor, , and psychiatric diseases. tor error as low as 1.59 mm with intraoperative CT-guided Lesioning,S stimulation, and local delivery of drugs9,22 all techniques4 and 0.14–0.86 mm with intraoperative MRI.5 rely on the ability to treat specific brain regions that are Current methods for targeting these structures also of- often small, deep, and indistinct on conventional MRI. ten include referencing a stereotactic atlas, with potential Validation of their placement typically requires clinical errors due to patient-specific variations from the atlas.27 testing or electrophysiological mapping with microelec- Richter et al. reviewed 35 subthalamic nuclei (STN) in 18 trode recordings (MERs), or both of these.8,18,29,32–35 patients and compared their location based on T2-weight-

Abbreviations CED = convection-enhanced delivery; DBS = ; MER = microelectrode recording; MW = molecular weight; STN = subthalamic nuclei; UPDRS = Unified Parkinson’s Disease Rating Scale; Vd = volume of distribution; Vi = volume of infusion SUBMITTED May 20, 2015. ACCEPTED October 19, 2015. include when citing Published online February 5, 2016; DOI: 10.3171/2015.10.JNS151154.

1080 J Neurosurg Volume 125 • November 2016 ©AANS, 2016

Unauthenticated | Downloaded 09/29/21 06:53 PM UTC CED enhancement of basal ganglia on MRI ed MRI. They then compared the T2-weighted MRI loca- the right DBS electrode tract to the subthalamus (Fig. 1). tions to locations as depicted in the Talairach and Tourn- The remainder of the DBS system was explanted, and the oux, and Schaltenbrand and Wahren atlases and found the infection was cleared after several weeks of intravenous anterior border of the STN to vary from 4.1 mm to 3.7 mm aztreonam therapy. A second bilateral subthalamic DBS from the midcommissural point, posterior border 4.2 mm system was eventually reimplanted 1 year after the first. to 10 mm, medial border 7.9 mm to 12.1 mm, and lateral The presumed vasogenic edema around the infected border 12.3 mm to 15.4 mm.26,27,31 There have been ad- electrode highlighted the boundaries of the right STN and vances in the use of MRI including coregistering multiple adjacent deep brain anatomy in contrast to the unaffected, sequences,35 7-T imaging, 8,32,34 9.4-T imaging,18 fast gray contralateral, left side (Fig. 1). This observation led to the matter acquisition T1 inversion recovery,29 and others.33 To premise of this study, where we believe that the extracellu- date, a practical and widely adoptable solution to easily lar space could be manipulated experimentally to improve identify these deep subcortical nuclei remains elusive and MRI visualization. results in widespread reliance on awake neurosurgery. We first observed a patient with vasogenic edema Additional Clinical Observations around an STN deep brain stimulation (DBS) electrode After we recognized improved MRI visualization of that dramatically improved the MRI visualization of the the anatomy surrounding an infected DBS electrode, we deep brain anatomy. This observation was then confirmed noted similar incidences of edema in the deep brain re- by MRI in other patients with edema from a variety of gions from a variety of pathologies (Fig. 2). These shared a pathologies extending into the basal ganglia. Because va- common pathophysiology where vasogenic edema extend- sogenic edema represents an increase in the extracellular ed into surrounding tissues as evidenced by T2-weighted water content, we sought to replicate this experimentally MRI. These cases served as further clinical evidence in an animal model with convection-enhanced delivery that an alteration in extracellular fluid content changed (CED). The preclinical study presented here demonstrates the characteristics of T2-weighted MRI. The enhanced a technique for convecting an infusate to improve the definition of deep brain anatomy, and particularly the visualization of deep brain structures with conventional boundaries between nuclei and white matter tracts, was MRI. We envision that this technique potentially can be observed in patients with sarcoma , thalamic used in combination with MRI-guided surgery to improve glioblastoma, sarcoidosis of the basal ganglia, large B-cell stereotactic targeting and obviate the need for awake sur- lymphoma, clear cell , gliomatosis cerebri, gery with MER or stimulation mapping. stereotactic for tremor, central glioblastoma, and stereotactic for cerebral palsy athetosis Methods and dystonia. Case Observation Swine Model of CED-Enhanced MRI Visualization A 59-year-old retired psychologist presented for con- Following approval from the University of Virginia In- sideration of DBS surgery. He had been diagnosed 5 years stitutional Animal Care and Use Committee, Sus scrofa earlier with Parkinson’s disease when he presented with domestica (Yorkshire) female pigs weighing approxi- a stooped posture and diminished facial expressions. He mately 25 pounds were acquired and acclimated in the developed severe motor fluctuations, short duration re- vivarium facility for 3 days. Swine nuclei are smaller than sponses, and peak dose dyskinesia on a medication regi- men consisting of levodopa 25/100 every 2 hours, 5 daily doses of 200 mg of entacapone, and 1.5 mg of pramipexole 3 times a day. The unmedicated (off) motor examination was notable for mild hypophonia with hypomimia, mod- erately increased tone (especially in the patient’s neck), and significant bradykinesia in the absence of tremor. He was able to independently stand with effort to a stooped posture, and his gait was slowed with delayed initiation and a diminished arm swing. A formal neuropsychologi- cal assessment deemed that he was a good candidate for surgery, because he experienced no mood issues and a preservation of cognition. He underwent successful surgery consisting of bilat- eral STN DBS with microelectrode mapping and a staged pulse generator implantation. Unfortunately, he developed a Pseudomonas infection involving the pulse generator, which required explantation of the pulse generator and the Fig. 1. Axial T2-weighted MR images. Left: Patient who had a unilat- distal portion of the lead extensions. Despite ciprofloxa- eral infection of a DBS electrode that led to increased signal differentia- cin therapy, he presented again 2 months later with evi- tion between the surrounding white matter tracts and visualization of dence of infection at the cranial incision. Cerebral MRI the STN nucleus, which is not visible on the contralateral side. Right: with Gd revealed no intracranial abscess; however, there Enlarged view of infected side. GPi/GPe: internal and external globus was a hyperintense signal on T2-weighted images around pallidum.

J Neurosurg Volume 125 • November 2016 1081

Unauthenticated | Downloaded 09/29/21 06:53 PM UTC A. E. Bond et al.

Fig. 2. Patients with various who developed perilesional edema in the basal ganglia. Patient A had sarcoid, and the inter- nal medullary lamina of the globus pallidum becomes visible delineating the interna clearly. A similar result is noted in patient B with a glioblastoma (GBM) and patient C with lymphoma. those in humans and more difficult to visualize, but they was opened, the outer cannula of the infusion catheter, are still within reasonable limits. For example, a swine with a stylet, was then connected to the stereotactic ma- STN is approximately 3 mm in maximal dimension.11 In nipulator and advanced to the target minus 1 cm in vector humans, the STN measures approximately 6 mm in maxi- length (the infusion cannula is 1 cm longer than the outer mal dimension.26 Most intraoperative MRI machines use cannula). The outer cannula was then affixed to the skull 1.5-T magnets; therefore, to compensate for the increased with dental acrylic. Then the animal was positioned in the difficulty of visualization of the smaller swine nuclei, the MRI machine. After removal of the stylet, the inner infu- CED swine experiments were performed on a 3.0-T MRI sion microcatheter was then advanced through the outer machine. The coil used for the swine was a standard head cannula as a final step to ensure there was no movement coil used for humans, and the swine were chosen to be the that could compromise the catheter/brain seal and lead to appropriate size to fit within the head coil with no detri- backflow. The PHD Ultra syringe pump was kept in the mental effects. MRI control room, and the tubing was passed through the Before surgery, the infusion apparatus was prepared. A wall. The swine was then advanced into the bore of the PHD Ultra syringe pump (Harvard Apparatus) was used magnet. Sagittal, axial, and coronal T2-weighted images with a 500-microliter, airtight glass syringe. The syringe were then obtained to verify the placement of the CED was connected to a series of four 60-inch extension sets catheter. In 2 instances, the CED catheter was misplaced; (Medex Micro Bore, Smiths Medical), each of which had therefore, the animal was returned to the large animal 0.4-ml priming volume. The extension sets were then con- operating room, and the outer cannula was repositioned nected to an EViTAR microcatheter (NexGen Medical without incident. Systems, Inc.). In experiments in which normal saline was Once the animal was in the MRI machine, the CED in- infused, the system was primed with normal saline via a fusion was started. Infusion rates varied from 1 ml/min to 3-way valve at an extension set connection point. Care was 6 ml/min, but imaging was constant at 15-minute intervals taken to ensure there were no air bubbles. When CSF was for a total of 90–120 minutes, depending on the rate of infused, the priming was performed in the same manner infusion. The measurement of volume of distribution (Vd) after CSF had been obtained from the lumbosacral region was performed using the average diameter of the infusion of the anesthetized animal. front, as measured from T2 changes, and calculating vol- On the day of surgery, anesthesia was induced with an ume of infusion (Vi) based on a sphere. intramuscular injection of tiletamine and xylazine and Following completion of CED infusion and MRI, the then maintained with inhaled isoflurane. External MRI -fi infusion catheter was then removed and the scalp was su- ducials were applied, and T2 images (TR 6000 msec, TE tured closed. Animals recovered under direct observation 93 msec, 2-mm slice width, no gap, resolution 320 × 320) for evidence of neurological, gait, or balance dysfunction. were acquired in all 3 orthogonal planes on a 3-T MRI Necropsy was performed at either 48 hours, 1 week, or system (Siemens Trio with Numaris 4 B17 software). The 1 month. The were then fixed in formalin and ex- midthalamic targets were identified on these presurgical amined for gross abnormalities before undergoing histo- MR images and then compared with a published stereotac- logical analysis with immunohistochemistry by a certified tic atlas.11,30 The coordinates for cannula placement were neuropathologist. A total of 7 animals were infused: 5 calculated and ranged in relation to the bregma as follows: with normal saline and 2 with CSF. A summary is shown 0–3 mm anterior, 6–8 mm lateral, and 33–36 mm deep. in Table 1. The animal was then positioned in a custom, MRI- compatible, stereotactic head frame such that the bregma and lambda points were in the same plane. The scalp was Results then shaved and prepared with betadine. A 4-cm linear in- Volume of Distribution cision was made centered at the bregma so that a small bur Once the model for CED was established, 7 animals hole could be made with a high-speed drill. After the dura underwent convection with varying rates of normal sa-

1082 J Neurosurg Volume 125 • November 2016

Unauthenticated | Downloaded 09/29/21 06:53 PM UTC CED enhancement of basal ganglia on MRI

TABLE 1. Summary of swine CED experiments Discussion Time Alive We observed that vasogenic edema associated with an No Description Infusate Rate (days) infected DBS electrode enhanced the T2-weighted MRI 1 Trial 1 NS 1 μl/min × 45 mins, then 2 characteristics for improved delineation of the deep brain 3 μl/min for 45 mins, then anatomy (Fig. 1). This phenomenon was then similarly 6 μl/min for 30 mins recognized with other CNS diseases that caused cerebral edema in the basal ganglia (Fig. 2). These consistent ob- 2 Trial 2 NS 3 μl/min × 90 mins 28 servations led us to design a laboratory experiment in a 3 Trial 3 NS 5 μl/min × 90 mins 7 large-brain animal model to simulate vasogenic edema 4 Trial 4 NS 5 μl/min × 90 mins 28 by manipulating the extracellular water content in a con- 5 Trial 5 NS 5 μl/min × 90 mins 7 trolled fashion with CED (Fig. 3). In swine, mild neuronal 6 Trial 1 CSF 5 μl/min × 90 mins 28 effects were noted in the thalamus by histological analysis 7 Trial 2 CSF 5 μl/min × 90 mins 7 when normal saline was convected (Fig. 4) but not with infusions of autologous CSF. NS = normal saline. Convection-enhanced infusion, or convection, is distin- guished from diffusion because the infusate is driven by a pressure wave through the extracellular space. More op- line (n = 5) or autologous CSF (n = 2). The Vd/Vi, mea- timal convection with minimal backflow along the cath- sured from T2 weighted MRI sequences, varied from eter occurs with a smaller catheter diameter, infusates of smaller molecular weight (MW), stepped catheter designs, 2.5 to 4.5 regardless of the infusate. The rate for CED 1,3,6,7,16,20,21,23 is generally accepted as greater than 0.05 ml per minute. and continuous slow infusion rates. These ex- These experimental rates varied between 1 and 6 ml per periments involved convection of normal saline or autolo- minute without noticeable differences in the Vd/Vi. Total gous CSF, achieving Vd/Vi ratios of 2.5–4.5. Infusions volume of infusion was bounded by either MRI time for into white matter tracts distribute to larger distances than low flow rates (limited to 90 minutes), or by observing in nuclear structures. Our Vd/Vi ratios of between 2.5 and no further increase in the volume of distribution when a 4.5 are consistent with those of other studies using low- molecular-weight infusates with Vd/Vi ratios of 3.8 and 6 pial boundary was encountered, and this ranged from 90 3,16 to 360 ml. in gray and white matter, respectively. The distribution of infusates is bounded by the pia, thus decreasing the Vd/ 14 MR Imaging Vi ratio. It has been demonstrated in previous models that in- All MRI assessments were made on T2-weighted se- fusates preferentially distribute along white matter tracts quences. The CED infusion of normal saline or CSF compared with denser gray matter. When infusing in the became noticeable with increased T2 hyperintensity at spinal cord, the infusate will appear as cylinders as it infus- roughly 45 ml. By 90 ml, a clearer definition of nuclear es along the direction of axons.17 Furthermore, Lieberman structures began to develop. MRI changes in the infusion et al. infused the striatum of rodents and rhesus monkeys, ceased once infusate reached the pial boundaries. CED and they showed that high-molecular-weight compounds infusions into the thalamus delineated the dorsal medi- like phaseolus vulgaris-leukoagglutinin (MW 126kD) had al, ventral anterior, reticular, and central median nuclei, a gray matter Vd/Vi ratio of 1.3, whereas low-molecular- which were not apparent before the infusion (Fig. 3). Sim- weight molecules like biotinylated dextran (MW 10kD) ilar results were obtained with all animals; qualitatively were higher at 3.8. Interestingly, they compared the bio- the enhanced appearance of the nuclear structures was tinylated dextran Vd/Vi ratio to their previous experience present to the same degree in all animals. The ventral an- using transferrin (MW 80kD) in white matter infusions terior nucleus provided the most distinctive improvement (Vd/Vi = 6) and found Vd/Vi for the smaller biotinylated in appearance, with the dorsal medial nucleus being the dextran molecule to be 63% of the larger transferrin mol- second most distinctive. None of these nuclear structures ecule, highlighting the enhanced infusion in white mat- were visible prior to the infusion in any animal. Animal ter relative to gray matter.3,16 In all of our infusions, the 7 had a postinfusion scan performed 30 minutes after the T2 images of the infused side demonstrated significant completion of the infusion, and the infusion effects per- signal differentiation between gray and white structures sisted. Long-term imaging was not performed. relative to the noninfused side. Figure 3A and B are rep- resentative samples of 2 animals, infused with saline, in Histological Analysis which the ventral anterior nucleus is clearly demarcated Histological analysis was examined at either acute from the surrounding white matter tracts and highlights (within 2 days), subacute (7 days), or chronic (28 days) the dorsomedial, centromedian, and reticular nucleus; Fig. time points for evidence of early or latent neuronal toxic- 3C shows a CSF-infused animal with similar delineation. ity due to the infusions. All specimens were analyzed with Figure 3D is a time-lapse series of a third animal. Our H & E stains. Animals infused with normal saline dem- infusions were consistent in demonstrating a reproducible onstrated mild neuronal effects with pyknotic neurons and technique that highlights deep nuclear structures. Our hy- axonal retractions, possibly attributed to the tonicity of the pothesis is that the improved visualization derives from 2 infusate (Fig. 4). Those infused with autologous CSF had major contributions. First, the convective infusion of sa- no evidence of neuronal change at any time point. line or CSF increases the water content of the extracellular

J Neurosurg Volume 125 • November 2016 1083

Unauthenticated | Downloaded 09/29/21 06:53 PM UTC A. E. Bond et al.

Fig. 3. A: Axial view of animal 1 after a 360-μl infusion of saline. B: Axial view of animal 4 after a 450-μl infusion of saline. C: Coronal view of animal 6 after a 450-μl infusion of CSF. D: Axial time-lapsed infusion of animal 3 with time and infusion volume labeled (infusion rate = 5 μl/min). The white matter tract medial to the ventral anterior nucleus, lateral to dorsomedial and anterior to centromedian, is readily apparent after infusion. In panel A, the reticular nucleus is noted lateral to the ventral anterior nucleus. Images A, B, and D are axial MR images. space and leads to an increased T2 signal analogous to MRI-guided surgery was initially developed in the that seen in vasogenic edema. Second, the infusate pref- 1990s for a more precise brain .2,12 This may have a erentially distributes into white matter, increasing the T2 more attractive allure to functional neurosurgery for ste- signal in these structures relative to gray matter. reotactically implanting DBS electrodes into visualized Safety is paramount in determining whether CED in- deep brain targets. Clinically, general anesthesia could fusions are feasible to enhance MRI during stereotactic procedures. There were no neurological or behavioral ef- fects observed after any of our CED infusions in swine. Histological examination, however, revealed mild neu- ronal changes from normal saline infusions with axonal retractions and pyknotic neurons. Because the volume infused was much less than 0.5 ml, this likely resulted from an osmotic mismatch, because normal saline is more hypertonic. There were no histological effects when the same volumes of autologous CSF was used as the infusate. These preliminary results indicate that the infusion of CSF does not demonstrate neuronal toxicity and has equivalent effects of improved MRI delineation between white and gray matter. This work was pursued on the assumption that im- proved resolution of these deep brain structures will aid stereotactic targeting, ultimately improving patient out- comes. Unfortunately, the actual percentage of DBS elec- trodes that are suboptimally positioned within a target nucleus is difficult to determine. The assessment of DBS electrode placement relative to a selected target location in x, y, z space is relatively straightforward and requires rou- tine postoperative imaging, but whether this is the correct clinical target cannot be ultimately determined until final clinical outcome. McClelland et al. analyzed electrodes placed using frame-based stereotaxy and MER guidance, Fig. 4. Photomicrographs of formalin-fixed brain after infusion with and found that 40 of 52 electrodes required repositioning normal saline identifying pyknotic neuron (dashed arrow) and axonal based on MER localization.19 Unfortunately, one cannot retraction figures solid( arrow) in a saline-infused thalamus. Infusion of determine whether this is from frame placement error or CSF did not demonstrate any neuronal or neuropil damage (not shown). from MRI target definition error. H & E, original magnification ×200.

1084 J Neurosurg Volume 125 • November 2016

Unauthenticated | Downloaded 09/29/21 06:53 PM UTC CED enhancement of basal ganglia on MRI be used to decrease patient anxiety and discomfort and Conclusions the need for unmedicated (off-state) intraoperative assess- This preliminary work demonstrates that CED of au- ments. Technically, true image-guided surgery adjusts to tologous CSF enhances the T2-weighted MRI visualiza- brain shifting during the procedure and the inherent reg- tion of deep subcortical structures without neuronal injury. istration errors that can be associated with image fusion 10,13 This technique could potentially be used to better delin- techniques. eate image-guided targets during stereotactic neurosur- True, real-time image-guided DBS placement has only gery. We envision improved stereotactic targeting with recently been realized with compatible skull-mounted the incorporation of CED-enhanced MRI and higher field frames and high-field MRI. Radial vector error during magnets during image-guided .25,28 DBS for disease progression has been measured at 1.2 ± 0.65 mm and 0.8 ± 0.4 mm.24,28 Clinical outcomes note improved off-motor Unified Parkinson’s Disease Rating Acknowledgments Scale (UPDRS) scores by 60% and 49%, respectively, This work was supported by the Wallace H. Coulter which compares similarly to a recent meta-analysis of 34 Translational Partnership ROPE and Translational Re- studies estimating 52% in disease progression with STN search Grants, and NexGen Medical Systems, Inc. DBS.15 Even with highly precise MRI-guided placement, the need for occasional electrode repositioning exists.28 References CT has been used as a modality to guide asleep DBS sur- geries, although it is primarily used for image registration 1. Asthagiri AR, Walbridge S, Heiss JD, Lonser RR: Effect of concentration on the accuracy of convective imaging distri- and confirmation of the placement as opposed to true im- bution of a gadolinium-based surrogate tracer. J Neurosurg age-guided localization during the procedure. Burchiel et 115:467–473, 2011 al. reported similar accuracies, but clinical outcomes are 2. Black PM, Moriarty T, Alexander E III, Stieg P, Woodard EJ, not yet available.4 Gleason PL, et al: Development and implementation of intra- We are interested in finding out if the combined use operative magnetic resonance imaging and its neurosurgical of MRI-guided electrode placement with improved tar- applications. Neurosurgery 41:831–845, 1997 get selection based on the use of CED would lead to even 3. Bobo RH, Laske DW, Akbasak A, Morrison PF, Dedrick RL, Oldfield EH: Convection-enhanced delivery of macromole- greater improvements in UPDRS scores on average, and cules in the brain. Proc Natl Acad Sci U S A 91:2076–2080, just as importantly, more uniform results among patients. 1994 The infusion procedure adds time and a new step to the 4. Burchiel KJ, McCartney S, Lee A, Raslan AM: Accuracy of insertion of DBS electrodes, establishing that the improve- deep brain stimulation electrode placement using intraopera- ments gained are central to evaluating the risk-benefit ratio tive computed without microelectrode recording. of such a procedure. J Neurosurg 119:301–306, 2013 The potential applicability of this technique to humans 5. Chabardes S, Isnard S, Castrioto A, Oddoux M, Fraix V, has many issues. First and foremost, the risk-benefit ra- Carlucci L, et al: Surgical implantation of STN-DBS leads using intraoperative MRI guidance: technique, accuracy, and tio must be favorable to ultimately consider infusion-en- clinical benefit at 1-year follow-up. Acta Neurochir (Wien) hanced, MRI-guided surgery as a valid surgical option. 157:729–737, 2015 We perceive the addition of a small-volume infusion to be 6. Chen MY, Lonser RR, Morrison PF, Governale LS, Oldfield associated with low morbidity. The increased risk of mi- EH: Variables affecting convection-enhanced delivery to the crocatheter insertion should be similar to or less than that striatum: a systematic examination of rate of infusion, can- of the current electrophysiological mapping process that nula size, infusate concentration, and tissue-cannula sealing requires sharp microelectrode penetrations. It is impos- time. J Neurosurg 90:315–320, 1999 7. Chittiboina P, Heiss JD, Warren KE, Lonser RR: Magnetic sible to assess or predict the clinical effects of a CED infu- resonance imaging properties of convective delivery in dif- sion, anticipated as less than 0.5–1.0 ml, into the human fuse intrinsic pontine . J Neurosurg Pediatr 13:276– basal ganglia without a clinical trial. The safest infusate to 282, 2014 test intuitively would be autologous CSF, which is easily 8. Cho ZH, Min HK, Oh SH, Han JY, Park CW, Chi JG, et and safely obtained by lumbar puncture from patients who al: Direct visualization of deep brain stimulation targets in are under general anesthesia. Parkinson disease with the use of 7-tesla magnetic resonance Finally, this study describes a proof-of-concept tech- imaging. J Neurosurg 113:639– 647, 2010 nique that mainly relies on the improved visualization 9. Dallapiazza R, McKisic MS, Shah B, Elias WJ: Neuro- modulation for movement disorders. Neurosurg Clin N Am of structures like the STN from T2 signal manipulation 25:47–58, 2014 in the laboratory and clinic. Other structures like the 10. Elias WJ, Fu KM, Frysinger RC: Cortical and subcortical internal segment of the globus pallidum or the ventral brain shift during stereotactic procedures. J Neurosurg intermediate nucleus of the thalamus are better visual- 107:983–988, 2007 ized with other MR sequences such as inversion recov- 11. Félix B, Léger ME, Albe-Fessard D, Marcilloux JC, Rampin ery or susceptibility-weighted images, respectively. We O, Laplace JP: Stereotaxic atlas of the pig brain. Brain Res envision that infusion-enhanced MRI would be tailored Bull 49:1–137, 1999 12. Hall WA, Martin AJ, Liu H, Nussbaum ES, Maxwell RE, to the desired brain target with target-specific sequences Truwit CL: using high-field strength interven- and infusates. Imaging of all modalities will continue to tional magnetic resonance imaging. Neurosurgery 44:807– improve with technology, but tailored intracerebral infu- 814, 1999 sions may serve as an adjunct for improved visualiza- 13. Ivan ME, Yarlagadda J, Saxena AP, Martin AJ, Starr PA, tion, not as a replacement for current imaging. Sootsman WK, et al: Brain shift during bur hole–based

J Neurosurg Volume 125 • November 2016 1085

Unauthenticated | Downloaded 09/29/21 06:53 PM UTC A. E. Bond et al.

procedures using interventional MRI. J Neurosurg 121:149– using high-field interventional magnetic resonance imaging 160, 2014 and a skull-mounted aiming device: technique and applica- 14. Jagannathan J, Walbridge S, Butman JA, Oldfield EH, Lonser tion accuracy. J Neurosurg 112:479–490, 2010 RR: Effect of ependymal and pial surfaces on convection- 29. Sudhyadhom A, Haq IU, Foote KD, Okun MS, Bova FJ: A enhanced delivery. J Neurosurg 109:547–552, 2008 high resolution and high contrast MRI for differentiation of 15. Kleiner-Fisman G, Herzog J, Fisman DN, Tamma F, Lyons subcortical structures for DBS targeting: the Fast Gray Mat- KE, Pahwa R, et al: Subthalamic nucleus deep brain stimula- ter Acquisition T1 Inversion Recovery (FGATIR). Neuroim- tion: summary and meta-analysis of outcomes. Mov Disord age 47 (Suppl 2):T44–T52, 2009 21 (Suppl 14):S290–S304, 2006 30. Sudhyadhom A, Okun MS, Foote KD, Rahman M, Bova FJ: 16. Lieberman DM, Laske DW, Morrison PF, Bankiewicz KS, A Three-dimensional deformable brain atlas for DBS target- Oldfield EH: Convection-enhanced distribution of large mol- ing. I. Methodology for atlas creation and artifact reduction. ecules in gray matter during interstitial drug infusion. J Neu- Open J 6:92–98, 2012 rosurg 82:1021–1029, 1995 31. Talairach J, Tournoux P: Co-Planar Stereotaxic Atlas of the 17. Lonser RR, Gogate N, Morrison PF, Wood JD, Oldfield EH: Human Brain. Stuttgart: Thieme Medical Publishers, 1988 Direct convective delivery of macromolecules to the spinal 32. Tani N, Joly O, Iwamuro H, Uhrig L, Wiggins CJ, Poupon C, cord. J Neurosurg 89:616–622, 1998 et al: Direct visualization of non-human primate subcortical 18. Massey LA, Miranda MA, Zrinzo L, Al-Helli O, Parkes HG, nuclei with contrast-enhanced high field MRI. Neuroimage Thornton JS, et al: High resolution MR anatomy of the sub- 58:60–68, 2011 thalamic nucleus: imaging at 9.4 T with histological valida- 33. Tanner M, Gambarota G, Kober T, Krueger G, Erritzoe D, tion. Neuroimage 59:2035–2044, 2012 Marques JP, et al: Fluid and white matter suppression with 19. McClelland S III, Ford B, Senatus PB, Winfield LM, Du YE, the MP2RAGE sequence. J Magn Reson Imaging 35:1063– Pullman SL, et al: Subthalamic stimulation for Parkinson 1070, 2012 disease: determination of electrode location necessary for 34. Tourdias T, Saranathan M, Levesque IR, Su J, Rutt BK: Visu- clinical efficacy. Neurosurg Focus 19(5):E12, 2005 alization of intra-thalamic nuclei with optimized white-mat- 20. Morrison PF, Chen MY, Chadwick RS, Lonser RR, Oldfield ter-nulled MPRAGE at 7T. Neuroimage 84:534–545, 2014 EH: Focal delivery during direct infusion to brain: role of 35. Xiao Y, Beriault S, Pike GB, Collins DL: Multicontrast mul- flow rate, catheter diameter, and tissue mechanics. Am J tiecho FLASH MRI for targeting the subthalamic nucleus. Physiol 277:R1218–R1229, 1999 Magn Reson Imaging 30:627–640, 2012 21. Nduom EK, Walbridge S, Lonser RR: Comparison of pulsed versus continuous convective flow for central nervous sys- tem tissue perfusion: laboratory investigation. J Neurosurg 117:1150–1154, 2012 22. Nikkhah G, Pinsker M: Stereotactic and Functional Neu- Disclosures rosurgery. Acta Neurochirurgica Supplementum 117. US Patent Application Serial No. 14/429,592. Filed March 19, Vienna: Springer, 2013 2015, “Method and System for Enhanced Imaging Visualization 23. Olson JJ, Zhang Z, Dillehay D, Stubbs J: Assessment of a of Deep Brain Anatomy Using Infusion.” balloon-tipped catheter modified for intracerebral convec- tion-enhanced delivery. J Neurooncol 89:159–168, 2008 Author Contributions 24. Ostrem JL, Galifianakis NB, Markun LC, Grace JK, Mar- Conception and design: Elias, Bond. Acquisition of data: Bond, tin AJ, Starr PA, et al: Clinical outcomes of PD patients Dallapiazza, Lopes. Analysis and interpretation of data: Elias, having bilateral STN DBS using high-field interventional Bond, Dallapiazza. Drafting the article: Bond. Critically revising MR-imaging for lead placement. Clin Neurol Neurosurg the article: all authors. Reviewed submitted version of manu- 115:708–712, 2013 script: all authors. Statistical analysis: Bond. Administrative/tech- 25. Richardson RM, Kells AP, Martin AJ, Larson PS, Starr PA, nical/material support: Elias, Bond. Study supervision: Elias. Piferi PG, et al: Novel platform for MRI-guided convection- enhanced delivery of therapeutics: preclinical validation Supplemental Information in nonhuman primate brain. Stereotact Funct Neurosurg Previous Presentations 89:141–151, 2011 26. Richter EO, Hoque T, Halliday W, Lozano AM, Saint-Cyr This work was presented at the 82nd AANS Annual Scientific JA: Determining the position and size of the subthalamic Meeting in San Francisco, California, April 5–9, 2014; the 2014 nucleus based on magnetic resonance imaging results in ASSFN Scientific Meeting in Washington, DC, May 31–June 3, patients with advanced Parkinson disease. J Neurosurg 2014; and the NSA Annual Scientific Meeting in Perth, Western 100:541–546, 2004 Australia, on October 2–4, 2014. 27. Schaltenbrand G, Bailey P: [Introduction to Stereotaxis, with an Atlas of the Human Brain.] Stuttgart: Georg Correspondence Thieme, 1959 (Ger) W. Jeff Elias, Department of Neurosurgery, University of 28. Starr PA, Martin AJ, Ostrem JL, Talke P, Levesque N, Lar- Virginia Health Science Center, Box 800212, Charlottesville, VA son PS: Subthalamic nucleus deep brain stimulator placement 22908. email: [email protected].

1086 J Neurosurg Volume 125 • November 2016

Unauthenticated | Downloaded 09/29/21 06:53 PM UTC