Technical Note Received: July 26, 2012 Accepted after revision: April 17, 2013 Published online: November 8, 2013
Stereotact Funct Neurosurg 2014;92:25–30 DOI: 10.1159/000351525
Direct Targeting of the Thalamic Anteroventral Nucleus for Deep Brain Stimulation by T1-Weighted Magnetic Resonance Imaging at 3 T Lars Buentjen a Klaus Kopitzki a, c Friedhelm C. Schmitt b Juergen Voges a, c Claus Tempelmann b Joern Kaufmann b Martin Kanowski b
Departments of a Stereotactic Neurosurgery and b Neurology, Otto von Guericke University, and c Leibniz Institute for Neurobiology, Magdeburg, Germany
Key Words Thalamus · Anteroventral nucleus · Deep brain stimulation · Epilepsy · Magnetic resonance imaging
Abstract Background: The thalamic anteroventral nucleus (AV) is a promising target structure for deep brain stimulation (DBS) in patients suffering from refractory epilepsy. Direct visualization of the AV would improve spatial accuracy in functional stereotactic neurosurgery for treatment of this disease. Methods: On 3-tesla magnetic resonance imaging (MRI), acquisition parameters were adjusted for optimal demarcation of the AV in 1 healthy subject. Reliability of AV visualization was then evaluated in 5 healthy individuals and 3 patients with refractory epilepsy. Results: In all individuals, an adjusted T1-weighted sequence allowed for demarcation of the AV. It was clearly distinguishable from hyperintense myelin-rich lamellae surrounding it ventrally and laterally and appeared hypo-intense compared to the adjacent thalamic nuclei. Image resolution and contrast facilitated direct stereotactic targeting of the AV prior to DBS surgery in all 3 patients. Conclusions: Direct targeting of the AV can be achieved, which has immediate implications for the accuracy of MRI-guided DBS in patients with refractory epilepsy. © 2013 S. Karger AG, Basel
© 2013 S. Karger AG, Basel 1011–6125/14/0921–0025$39.50/0 E-Mail [email protected] www.karger.com/sfn
Introduction
There are two distinctly different approaches for stereotactic targeting of cerebral nuclei in deep brain stimulation (DBS)—indirect targeting and visualization. Indirect targeting is based on salient landmarks like the anterior and posterior commissure. Distances of individual nuclei to specific points on the resulting intercommissural line are here derived from stereotactic atlases. Correction for individual brain measures such as length of the intercommissural line, width of the hemisphere and thalamus height can be carried out to increase accuracy. A major disadvantage of this approach results from the small number of brain specimens the reference atlases are based on. Interindividual variability can only partially be taken into account. Magnetic resonance imaging (MRI) employing dedicated imaging sequences has enabled direct visualization of the target region in the individual patient. The most prominent example is the subthalamic nucleus for which direct targeting has been established and validated [1]. Here the high iron content of the subthalamic nucleus results in a pronounced hypo-intensity on T2- and T2*weighted MR images. A growing body of evidence having emerged from neuro-anatomical considerations, various animal experiments and small non-randomized patient cohorts sugLars Buentjen Department for Stereotactic Neurosurgery Otto-von-Guericke Universität Magdeburg Leipziger Strasse 44, DE–39120 Magdeburg (Germany) E-Mail lars.buentjen @ med.ovgu.de
Downloaded by: Lund University Libraries 130.235.136.21 - 5/13/2015 2:46:07 PM
Materials and Methods MRI was conducted on a Siemens Verio (Siemens, Erlangen, Germany) equipped with a 32-channel head coil. The head was immobilized using a Tempur Original Pillow Junior (Tempur World, Inc.) to guarantee an optimal balance between tight fixation and patient comfort. After localization, T1-weighted MPRAGE [14] data sets with isometric voxel sizes of (0.8 mm)3 and/or (1.0 mm)3 were acquired [(0.8 mm)3: 320 × 240 × 224 matrix, repetition time TR = 2.91 s, inversion time TI = 1.1 s, echo time TE = 6.7 ms, 160 Hz/pixel bandwidth, 10: 11 min per average; (1 mm)3: 256 × 192 × 176 matrix, TR = 2.7 s, TI = 1.1 s, TE = 7.2 ms, 130 Hz/pixel bandwidth, 7:34 min per average; for both spatial resolutions: 7° flip angle, 7/8 kspace in phase direction, flow compensation in anterior-posterior direction, 2 averages]. The axial 3D block was oriented parallel to the anterior and posterior commissure plane. For estimation of longitudinal relaxation time T1, a series of turbo spin echo measurements (192 × 192 mm2 field of view, 320 × 320 matrix, 3 mm slice thickness, turbo factor 5, TE = 30 ms, 40 Hz/pixel bandwidth) with a TR of 1.75, 2.00, 2.25, 2.50, 2.75, 3.00, 3.25, and 3.50 s was acquired in 1 subject. For selected regions of interest, the signal intensity I was fitted to I = c {1 – exp[–(TR – TE,last)/T1]}, where TE,last denotes the echo time of the last echo in the echo train and c is a scaling factor.
26
Stereotact Funct Neurosurg 2014;92:25–30 DOI: 10.1159/000351525
Six healthy subjects (4 males, aged 22–27 years) and 3 patients (2 females, 25–42 years, contrast agent 0.1 mmol/kg Gadovist) with refractory epilepsy were scanned under approval of the ethics committee of the Medical Faculty of the University of Magdeburg. Written informed consent was obtained from all individuals prior to measurements. Trajectories were planned using the software Praezis Plus version 3.1 (Precisis AG, Heidelberg, Germany).
Results
MPRAGE images of the thalamus with (1.0 mm)3 and (0.8 mm)3 isometric resolution in axial, sagittal and coronal views are shown in figure 1. The AV (marked by arrows) is located at the anterior superior medial aspect of the thalamus and constitutes its anterodorsal border. It is partially enveloped by a myelin-rich sheath which belongs to the mamillothalamic tract (mtt) and the internal medullary lamina. The rostral medial branch of these laminae separates the AV from the medial dorsal nucleus (MD) which is situated ventrally, medially and posteriorly to it. The rostral lateral branch of the myelin-rich sheet separates the AV from the parvocellular ventral anterior nucleus (VApc) and the dorsal portion of the ventral lateral posterior nucleus (VLpd). The VApc abuts the AV laterally at its upper rostral aspect and the VLpd cushions the AV at its ventral posterior aspect. The AV can be distinguished from two of its embedding structures. First, it is enclosed by relatively thick myelin-rich layers, as described above. They exhibit the characteristic hyperintensity of white matter tracts like the mtt (arrowheads in fig. 1) and the corpus callosum as typical examples. Due to the hyperintense framing, nearly the complete border of the AV can be determined with high accuracy. Secondly, the AV appears more hypo-intense than the surrounding grey matter like, e.g., the VLpd, the VApc, and the MD. This hypo-intensity is best noticeable on axial slices, where VLpd/VApc and MD lie laterally and posteriorly to the AV, respectively. The reason for the hypo-intensity is the slightly longer T1 relaxation time of the AV. In the subject shown we determined T1 = 1.75 s in the AV and T1 = 1.61 s in VLp/VA. Images acquired with two averages are shown in figure 1 for a voxel size of (1.0 mm)3 and (0.8 mm)3. Overall image acquisition time was 15 and 20 min, respectively. The AV is clearly delineated at both resolutions, though definition of the AV seems slightly improved at (1.0 mm)3. Here signalto-noise ratio (SNR) seems to outweigh image resolution. To enable visual comparison of MPRAGE MR images of figure 1, histological stains of corresponding stereotacBuentjen/Kopitzki/Schmitt/Voges/ Tempelmann/Kaufmann/Kanowski
Downloaded by: Lund University Libraries 130.235.136.21 - 5/13/2015 2:46:07 PM
gested that the anteroventral nucleus (AV) might be a structure which could have an alleviating effect on focally initiated seizures when stimulated via an intracerebral electrode [2–6]. This publication uses the terminology first proposed by Hirai and Jones [7] and utilized in Morel’s Stereotactic Atlas of the Human Thalamus [8]. The corresponding term for the AV in the stereotactic atlas of Schaltenbrand and Wahren [9] is nucleus anteroprincipalis thalami (A.pr). The SANTE study (electrical stimulation of the anterior nucleus of the thalamus for treatment of refractory epilepsy) has demonstrated the efficacy of AV DBS to reduce epileptic seizures in a prospective randomized double-blind study [10]. As a result of this study, DBS surgery of the AV could become a more frequent procedure in the near future. Therefore, a dedicated MR protocol to delineate this nucleus seems desirable. Reports on MR-based localization of the AV are rare. They rely on complex acquisition strategies followed by sophisticated data post-processing like diffusion tensor imaging with subsequent clustering techniques [11], semi-quantitative magnetization transfer saturation mapping [12], or relaxometry [13]. Since these techniques are neither broadly available nor well standardized, we focused on the implementation of an MRI protocol suitable for routine application in DBS surgery as described in the present paper.
quired with a resolution of (1.0 mm)3 (top row) and (0.8 mm)3 (middle row) in axial (left), sagittal (centre) and coronal (right) views. Two averages were acquired in approximately 15 and 20 min for the (1.0 mm)3 and (0.8 mm)3 voxel size 3D data set, respectively. In the top row, arrows point to the AV and arrowheads mark the mtt. Due to its longer T1 relaxation time, the AV appears slightly darker than the remaining thalamic grey matter. This is most
discernible in the axial views. Further, the AV is enclosed by myelin-rich lamellae. They appear as thin hyperintense lines and significantly improve the visibility of the AV. Corresponding axial, sagittal and coronal slices (bottom row) were taken from the stereotaxic atlas of the human brain by Schaltenbrand and Wahren (plate 15, XXXVIII Hd +12; plate 39 LXXVIII S.l 5,5; plate 27, LXVIII F.a. 2,0) with friendly permission from the Thieme Verlag. Arrow markers are set analogous to the top row.
tic atlas plates (Schaltenbrand/Wahren) are presented in the bottom row. The AV (A.pr.) is marked by white arrows and the mtt by white arrowheads. The images of 4 other subjects shown in figure 2 corroborate the reliability of the AV visualization. In an attempt to shorten the protocol, a scan time of only 8 min was used for the subject shown in the right column of figure 2. Even though this reduces image SNR, the AV and particularly its lower border are still definable. Figure 3 shows the thalamus of a patient. The demarcation of the AV is similar to the one demonstrated in figure 2. Unlike the previous examples, vessels are bright due to the contrast agent. Vessel visualization is necessary in our setting for image fusion with the intra-operative CT in the stereotactic frame. Adjacent bright vessels bear the risk to deteriorate image quality due to artefacts
caused by flow, partial voluming, or Gibbs ringing. A comparison with a second scan in one of the patients acquired without contrast agent confirmed that vessel artefacts are negligible (data not shown). To improve targeting of the AV, the MRI protocol was applied in 3 patients with refractory epilepsy scheduled for placement of DBS electrodes into the AV. Indirect planning was performed using atlas coordinates according to Hodaie et al. [4]. Entry points were defined close to the midline aiming for an angle between the intercommissural line and the trajectory of 60° in the sagittal plane. Final trajectories were adjusted to avoid intraventriclular vasculature and adapted to the individual shape of the AV as determined on 3-tesla MPRAGE images (direct targeting). Figure 4 shows two examples of trajectories for AV DBS. There is a notable asymmetry of the lateral ventri-
Direct Targeting of the Thalamic AV for DBS
Stereotact Funct Neurosurg 2014;92:25–30 DOI: 10.1159/000351525
27
Downloaded by: Lund University Libraries 130.235.136.21 - 5/13/2015 2:46:07 PM
Fig. 1. MPRAGE images of the thalamus of a healthy subject ac-
Fig. 2. Coronal (top) and sagittal (bottom) slices through the thal-
amus with (1.0 mm)3 resolution of 4 additional subjects at the level of the mtt. The images corroborate the robust delineation of the
AV. Columns 1–3 demonstrate images acquired with two averages, the right column shows images with only one average. Despite the lower SNR in the latter, the AV remains discernible.
Fig. 3. Preoperative axial, sagittal, and coronal slices through the thalamus of a patient with refractory epilepsy.
In comparison to brain tissue vessels appear hyperintense due to the administered contrast agent. Vessel visualization is a prerequisite for coregistration with intra-operative CT.
patients. Note the asymmetry of the left and the right AV in both individuals. Images in the bottom row are derived from the same patient as in figure 3.
28
Stereotact Funct Neurosurg 2014;92:25–30 DOI: 10.1159/000351525
Buentjen/Kopitzki/Schmitt/Voges/ Tempelmann/Kaufmann/Kanowski
Downloaded by: Lund University Libraries 130.235.136.21 - 5/13/2015 2:46:07 PM
Fig. 4. Preoperative trajectory planning for DBS electrode placement through the anterior thalamic nucleus in 2
Discussion
A major consideration in MRI protocol optimization is the trade-off between image resolution and SNR under the constraint of manageable scan duration. Here increased spatial resolution, i.e. a smaller voxel volume, reduces partial volume effects and hence might improve visibility of myelin-rich laminae between the AV and its surrounding nuclei. However, a concomitantly decreasing SNR might conceal the subtle differences in signal intensity between the AV and neighbouring thalamic nuclei. As shown in figure 1, partial volume effects are not detrimental to AV visualization at a resolution of (1.0 mm)3, even though the structural complexity of the thalamus would seem to require a higher resolution (0.8 mm)3. Furthermore, the combination of a tight fitting head coil with a field strength of 3 T is crucial for a high SNR. A multi-element phased array coil should not be mandatory as it is most effective near the coil elements, i.e. the cortex. For the centre of a head coil at best a marginal SNR gain is to be expected [15, 16]. Nevertheless, nowadays the smallest commercial head coils are array coils. Most of the presented images were measured with two averages. A scan time of about 15 min is recommended to ensure high image quality if involuntary head movements are neglectable. At lower resolution, images are less susceptible to motion artefacts. A tight and at the same time convenient fixation of the head reduces motion artefacts effectively. Even if the example given in figure 2 suggests that the image definition of a scan acquired with only one average can suffice, we recommend using two averages. The scan times can also be shortened by choosing the shortest possible repetition time TR. However, MPRAGE T1 contrast is significantly raised by the insertion of a delay (in this study ca. 400 ms) between the end of the echo train and the next inversion pulse. The MPRAGE sequence employed here is standard on all MR scanners and ubiquitous in clinical routine. Previous attempts to visualize thalamic nuclei required challenging data post-processing, or did at least rely on nonDirect Targeting of the Thalamic AV for DBS
commercial in-house software solutions [11–13]. A recently published work of Bender et al. [17] describes the identification of the four large thalamic nuclei groups (including the anterior group) by means of an optimized MPRAGE sequence. Their images show a good demarcation between the medial and the lateral group of thalamic nuclei whereas the image examples given here are supposed to allow a more precise targeting of the AV. Stereotactic access to the AV for the lack of delineable intrathalamic structures has so far relied upon indirect targeting. Typical coordinates as provided by Hodaie et al. [4] are 6 mm lateral to, 8 mm anterior to and 12 mm above the midcommisural point. However, asymmetries can be expected in a number of patients which have undergone epilepsy surgery. Also, atrophy of the entire thalamus has been described in patients with long-standing epilepsy [18]. The AV is a lengthy but narrow structure, and targeting is further complicated by the thalamostriate vein representing an extensive risk zone in its immediate vicinity. Therefore, knowledge about its boundaries is particularly important in choosing a stereotactic approach to this target. Here direct visualization of the AV resulted in a more anterior placement of electrodes compared to the approach described by Hodaie et al. [4]. However, it is important to stress that it would be difficult if not impossible to draw any reliable conclusions from only 3 patients. It is not our intention to clinically evaluate this procedure but to increase precision in targeting the AV, which in turn should facilitate such an evaluation in a more rigorous fashion.
Conclusion
This study has shown that it is possible to delineate the anteroventral thalamic nucleus on T1-weighted MR images to a degree which enables a stereotactic neurosurgeon to perform direct targeting. The 3-tesla MRI protocol presented can easily be integrated into clinical routine and should improve preciseness of DBS in this target.
References
1 Brunenberg EJL, Platel B, Hofman PAM, Ter Haar Romeny BM, Visser-Vandewalle V: Magnetic resonance imaging techniques for visualization of the subthalamic nucleus. J Neurosurg 2011;115:971–984. 2 Mirski MA, McKeon AC, Ferrendelli JA: Anterior thalamus and substantia nigra: two distinct structures mediating experimental generalized seizures. Brain Res 1986; 397: 377– 380.
Stereotact Funct Neurosurg 2014;92:25–30 DOI: 10.1159/000351525
29
Downloaded by: Lund University Libraries 130.235.136.21 - 5/13/2015 2:46:07 PM
cles and also of the AV. Nonetheless penetration of the AV could be achieved in the given examples. Average coordinates of electrode contacts in the AV were derived from intra-operative stereotactic X-ray images registered with the preoperative MR images. The target coordinates ranged from 4.5 to 6 mm lateral to the midcommissural line, 12 to 16 mm anterior and 12 to 14.5 mm above the posterior commissure.
30
8 Morel A: Stereotactic Atlas of the Human Thalamus and Basal Ganglia. New York, Informa Healthcare, 2007. 9 Schaltenbrand G, Wahren W: Atlas for Stereotaxy of the Human Brain, ed 2. Stuttgart, Thieme, 1977. 10 Fisher R, Salanova V, Witt T, Worth R, Henry T, Gross R, et al: Electrical stimulation of the anterior nucleus of thalamus for treatment of refractory epilepsy. Epilepsia 2010; 51: 899– 908. 11 Wiegell MR, Tuch DS, Larsson HBW, Wedeen VJ: Automatic segmentation of thalamic nuclei from diffusion tensor magnetic resonance imaging. Neuroimage 2003; 19: 391– 401. 12 Gringel T, Schulz-Schaeffer W, Elolf E, Frölich A, Dechent P, Helms G: Optimized high-resolution mapping of magnetization transfer (MT) at 3 Tesla for direct visualization of substructures of the human thalamus in clinically feasible measurement time. J Magn Reson Imaging 2009;29:1285–1292. 13 Traynor CR, Barker GJ, Crum WR, Williams SCR, Richardson MP: Segmentation of the thalamus in MRI based on T1 and T2. Neuroimage 2011;56:939–950.
Stereotact Funct Neurosurg 2014;92:25–30 DOI: 10.1159/000351525
14 Mugler JP, Brookeman JR: Three-dimensional magnetization-prepared rapid gradientecho imaging (3D MP RAGE). Magn Reson Med 1990;15:152–157. 15 Wright SM, Wald LL: Theory and application of array coils in MR spectroscopy. NMR Biomed 1997;10:394–410. 16 Wiggins GC, Polimeni JR, Potthast A, Schmitt M, Alagappan V, Wald LL: 96-Channel receive-only head coil for 3 Tesla: design optimization and evaluation. Magn Reson Med 2009;62:754–762. 17 Bender B, Mänz C, Korn A, Nägele T, Klose U: Optimized 3D magnetization-prepared rapid acquisition of gradient echo: identification of thalamus substructures at 3T. AJNR Am J Neuroradiol 2011;32:2110–2115. 18 Margerison JH, Corsellis JA: Epilepsy and the temporal lobes. A clinical, electroencephalographic and neuropathological study of the brain in epilepsy, with particular reference to the temporal lobes. Brain 1966;89:499–530.
Buentjen/Kopitzki/Schmitt/Voges/ Tempelmann/Kaufmann/Kanowski
Downloaded by: Lund University Libraries 130.235.136.21 - 5/13/2015 2:46:07 PM
3 Mirski MA, Rossell LA, Terry JB, Fisher RS: Anticonvulsant effect of anterior thalamic high frequency electrical stimulation in the rat. Epilepsy Res 1997;28:89–100. 4 Hodaie M, Wennberg RA, Dostrovsky JO, Lozano AM: Chronic anterior thalamus stimulation for intractable epilepsy. Epilepsia 2002; 43:603–608. 5 Hamani C, Ewerton FIS, Bonilha SM, Ballester G, Mello LEAM, Lozano AM: Bilateral anterior thalamic nucleus lesions and high-frequency stimulation are protective against pilocarpineinduced seizures and status epilepticus. Neurosurgery 2004;54:191–195, discussion 195–197. 6 Kerrigan JF, Litt B, Fisher RS, Cranstoun S, French JA, Blum DE, et al: Electrical stimulation of the anterior nucleus of the thalamus for the treatment of intractable epilepsy. Epilepsia 2004;45:346–354. 7 Hirai T, Jones EG: A new parcellation of the human thalamus on the basis of histochemical staining. Brain Res Brain Res Rev 1989;14:1–34.
Stereotact Funct Neurosurg 2014;92:25–30 DOI: 10.1159/000351525
Direct Targeting of the Thalamic Anteroventral Nucleus for Deep Brain Stimulation by T1-Weighted Magnetic Resonance Imaging at 3 T Lars Buentjen a Klaus Kopitzki a, c Friedhelm C. Schmitt b Juergen Voges a, c Claus Tempelmann b Joern Kaufmann b Martin Kanowski b
Departments of a Stereotactic Neurosurgery and b Neurology, Otto von Guericke University, and c Leibniz Institute for Neurobiology, Magdeburg, Germany
Key Words Thalamus · Anteroventral nucleus · Deep brain stimulation · Epilepsy · Magnetic resonance imaging
Abstract Background: The thalamic anteroventral nucleus (AV) is a promising target structure for deep brain stimulation (DBS) in patients suffering from refractory epilepsy. Direct visualization of the AV would improve spatial accuracy in functional stereotactic neurosurgery for treatment of this disease. Methods: On 3-tesla magnetic resonance imaging (MRI), acquisition parameters were adjusted for optimal demarcation of the AV in 1 healthy subject. Reliability of AV visualization was then evaluated in 5 healthy individuals and 3 patients with refractory epilepsy. Results: In all individuals, an adjusted T1-weighted sequence allowed for demarcation of the AV. It was clearly distinguishable from hyperintense myelin-rich lamellae surrounding it ventrally and laterally and appeared hypo-intense compared to the adjacent thalamic nuclei. Image resolution and contrast facilitated direct stereotactic targeting of the AV prior to DBS surgery in all 3 patients. Conclusions: Direct targeting of the AV can be achieved, which has immediate implications for the accuracy of MRI-guided DBS in patients with refractory epilepsy. © 2013 S. Karger AG, Basel
© 2013 S. Karger AG, Basel 1011–6125/14/0921–0025$39.50/0 E-Mail [email protected] www.karger.com/sfn
Introduction
There are two distinctly different approaches for stereotactic targeting of cerebral nuclei in deep brain stimulation (DBS)—indirect targeting and visualization. Indirect targeting is based on salient landmarks like the anterior and posterior commissure. Distances of individual nuclei to specific points on the resulting intercommissural line are here derived from stereotactic atlases. Correction for individual brain measures such as length of the intercommissural line, width of the hemisphere and thalamus height can be carried out to increase accuracy. A major disadvantage of this approach results from the small number of brain specimens the reference atlases are based on. Interindividual variability can only partially be taken into account. Magnetic resonance imaging (MRI) employing dedicated imaging sequences has enabled direct visualization of the target region in the individual patient. The most prominent example is the subthalamic nucleus for which direct targeting has been established and validated [1]. Here the high iron content of the subthalamic nucleus results in a pronounced hypo-intensity on T2- and T2*weighted MR images. A growing body of evidence having emerged from neuro-anatomical considerations, various animal experiments and small non-randomized patient cohorts sugLars Buentjen Department for Stereotactic Neurosurgery Otto-von-Guericke Universität Magdeburg Leipziger Strasse 44, DE–39120 Magdeburg (Germany) E-Mail lars.buentjen @ med.ovgu.de
Downloaded by: Lund University Libraries 130.235.136.21 - 5/13/2015 2:46:07 PM
Materials and Methods MRI was conducted on a Siemens Verio (Siemens, Erlangen, Germany) equipped with a 32-channel head coil. The head was immobilized using a Tempur Original Pillow Junior (Tempur World, Inc.) to guarantee an optimal balance between tight fixation and patient comfort. After localization, T1-weighted MPRAGE [14] data sets with isometric voxel sizes of (0.8 mm)3 and/or (1.0 mm)3 were acquired [(0.8 mm)3: 320 × 240 × 224 matrix, repetition time TR = 2.91 s, inversion time TI = 1.1 s, echo time TE = 6.7 ms, 160 Hz/pixel bandwidth, 10: 11 min per average; (1 mm)3: 256 × 192 × 176 matrix, TR = 2.7 s, TI = 1.1 s, TE = 7.2 ms, 130 Hz/pixel bandwidth, 7:34 min per average; for both spatial resolutions: 7° flip angle, 7/8 kspace in phase direction, flow compensation in anterior-posterior direction, 2 averages]. The axial 3D block was oriented parallel to the anterior and posterior commissure plane. For estimation of longitudinal relaxation time T1, a series of turbo spin echo measurements (192 × 192 mm2 field of view, 320 × 320 matrix, 3 mm slice thickness, turbo factor 5, TE = 30 ms, 40 Hz/pixel bandwidth) with a TR of 1.75, 2.00, 2.25, 2.50, 2.75, 3.00, 3.25, and 3.50 s was acquired in 1 subject. For selected regions of interest, the signal intensity I was fitted to I = c {1 – exp[–(TR – TE,last)/T1]}, where TE,last denotes the echo time of the last echo in the echo train and c is a scaling factor.
26
Stereotact Funct Neurosurg 2014;92:25–30 DOI: 10.1159/000351525
Six healthy subjects (4 males, aged 22–27 years) and 3 patients (2 females, 25–42 years, contrast agent 0.1 mmol/kg Gadovist) with refractory epilepsy were scanned under approval of the ethics committee of the Medical Faculty of the University of Magdeburg. Written informed consent was obtained from all individuals prior to measurements. Trajectories were planned using the software Praezis Plus version 3.1 (Precisis AG, Heidelberg, Germany).
Results
MPRAGE images of the thalamus with (1.0 mm)3 and (0.8 mm)3 isometric resolution in axial, sagittal and coronal views are shown in figure 1. The AV (marked by arrows) is located at the anterior superior medial aspect of the thalamus and constitutes its anterodorsal border. It is partially enveloped by a myelin-rich sheath which belongs to the mamillothalamic tract (mtt) and the internal medullary lamina. The rostral medial branch of these laminae separates the AV from the medial dorsal nucleus (MD) which is situated ventrally, medially and posteriorly to it. The rostral lateral branch of the myelin-rich sheet separates the AV from the parvocellular ventral anterior nucleus (VApc) and the dorsal portion of the ventral lateral posterior nucleus (VLpd). The VApc abuts the AV laterally at its upper rostral aspect and the VLpd cushions the AV at its ventral posterior aspect. The AV can be distinguished from two of its embedding structures. First, it is enclosed by relatively thick myelin-rich layers, as described above. They exhibit the characteristic hyperintensity of white matter tracts like the mtt (arrowheads in fig. 1) and the corpus callosum as typical examples. Due to the hyperintense framing, nearly the complete border of the AV can be determined with high accuracy. Secondly, the AV appears more hypo-intense than the surrounding grey matter like, e.g., the VLpd, the VApc, and the MD. This hypo-intensity is best noticeable on axial slices, where VLpd/VApc and MD lie laterally and posteriorly to the AV, respectively. The reason for the hypo-intensity is the slightly longer T1 relaxation time of the AV. In the subject shown we determined T1 = 1.75 s in the AV and T1 = 1.61 s in VLp/VA. Images acquired with two averages are shown in figure 1 for a voxel size of (1.0 mm)3 and (0.8 mm)3. Overall image acquisition time was 15 and 20 min, respectively. The AV is clearly delineated at both resolutions, though definition of the AV seems slightly improved at (1.0 mm)3. Here signalto-noise ratio (SNR) seems to outweigh image resolution. To enable visual comparison of MPRAGE MR images of figure 1, histological stains of corresponding stereotacBuentjen/Kopitzki/Schmitt/Voges/ Tempelmann/Kaufmann/Kanowski
Downloaded by: Lund University Libraries 130.235.136.21 - 5/13/2015 2:46:07 PM
gested that the anteroventral nucleus (AV) might be a structure which could have an alleviating effect on focally initiated seizures when stimulated via an intracerebral electrode [2–6]. This publication uses the terminology first proposed by Hirai and Jones [7] and utilized in Morel’s Stereotactic Atlas of the Human Thalamus [8]. The corresponding term for the AV in the stereotactic atlas of Schaltenbrand and Wahren [9] is nucleus anteroprincipalis thalami (A.pr). The SANTE study (electrical stimulation of the anterior nucleus of the thalamus for treatment of refractory epilepsy) has demonstrated the efficacy of AV DBS to reduce epileptic seizures in a prospective randomized double-blind study [10]. As a result of this study, DBS surgery of the AV could become a more frequent procedure in the near future. Therefore, a dedicated MR protocol to delineate this nucleus seems desirable. Reports on MR-based localization of the AV are rare. They rely on complex acquisition strategies followed by sophisticated data post-processing like diffusion tensor imaging with subsequent clustering techniques [11], semi-quantitative magnetization transfer saturation mapping [12], or relaxometry [13]. Since these techniques are neither broadly available nor well standardized, we focused on the implementation of an MRI protocol suitable for routine application in DBS surgery as described in the present paper.
quired with a resolution of (1.0 mm)3 (top row) and (0.8 mm)3 (middle row) in axial (left), sagittal (centre) and coronal (right) views. Two averages were acquired in approximately 15 and 20 min for the (1.0 mm)3 and (0.8 mm)3 voxel size 3D data set, respectively. In the top row, arrows point to the AV and arrowheads mark the mtt. Due to its longer T1 relaxation time, the AV appears slightly darker than the remaining thalamic grey matter. This is most
discernible in the axial views. Further, the AV is enclosed by myelin-rich lamellae. They appear as thin hyperintense lines and significantly improve the visibility of the AV. Corresponding axial, sagittal and coronal slices (bottom row) were taken from the stereotaxic atlas of the human brain by Schaltenbrand and Wahren (plate 15, XXXVIII Hd +12; plate 39 LXXVIII S.l 5,5; plate 27, LXVIII F.a. 2,0) with friendly permission from the Thieme Verlag. Arrow markers are set analogous to the top row.
tic atlas plates (Schaltenbrand/Wahren) are presented in the bottom row. The AV (A.pr.) is marked by white arrows and the mtt by white arrowheads. The images of 4 other subjects shown in figure 2 corroborate the reliability of the AV visualization. In an attempt to shorten the protocol, a scan time of only 8 min was used for the subject shown in the right column of figure 2. Even though this reduces image SNR, the AV and particularly its lower border are still definable. Figure 3 shows the thalamus of a patient. The demarcation of the AV is similar to the one demonstrated in figure 2. Unlike the previous examples, vessels are bright due to the contrast agent. Vessel visualization is necessary in our setting for image fusion with the intra-operative CT in the stereotactic frame. Adjacent bright vessels bear the risk to deteriorate image quality due to artefacts
caused by flow, partial voluming, or Gibbs ringing. A comparison with a second scan in one of the patients acquired without contrast agent confirmed that vessel artefacts are negligible (data not shown). To improve targeting of the AV, the MRI protocol was applied in 3 patients with refractory epilepsy scheduled for placement of DBS electrodes into the AV. Indirect planning was performed using atlas coordinates according to Hodaie et al. [4]. Entry points were defined close to the midline aiming for an angle between the intercommissural line and the trajectory of 60° in the sagittal plane. Final trajectories were adjusted to avoid intraventriclular vasculature and adapted to the individual shape of the AV as determined on 3-tesla MPRAGE images (direct targeting). Figure 4 shows two examples of trajectories for AV DBS. There is a notable asymmetry of the lateral ventri-
Direct Targeting of the Thalamic AV for DBS
Stereotact Funct Neurosurg 2014;92:25–30 DOI: 10.1159/000351525
27
Downloaded by: Lund University Libraries 130.235.136.21 - 5/13/2015 2:46:07 PM
Fig. 1. MPRAGE images of the thalamus of a healthy subject ac-
Fig. 2. Coronal (top) and sagittal (bottom) slices through the thal-
amus with (1.0 mm)3 resolution of 4 additional subjects at the level of the mtt. The images corroborate the robust delineation of the
AV. Columns 1–3 demonstrate images acquired with two averages, the right column shows images with only one average. Despite the lower SNR in the latter, the AV remains discernible.
Fig. 3. Preoperative axial, sagittal, and coronal slices through the thalamus of a patient with refractory epilepsy.
In comparison to brain tissue vessels appear hyperintense due to the administered contrast agent. Vessel visualization is a prerequisite for coregistration with intra-operative CT.
patients. Note the asymmetry of the left and the right AV in both individuals. Images in the bottom row are derived from the same patient as in figure 3.
28
Stereotact Funct Neurosurg 2014;92:25–30 DOI: 10.1159/000351525
Buentjen/Kopitzki/Schmitt/Voges/ Tempelmann/Kaufmann/Kanowski
Downloaded by: Lund University Libraries 130.235.136.21 - 5/13/2015 2:46:07 PM
Fig. 4. Preoperative trajectory planning for DBS electrode placement through the anterior thalamic nucleus in 2
Discussion
A major consideration in MRI protocol optimization is the trade-off between image resolution and SNR under the constraint of manageable scan duration. Here increased spatial resolution, i.e. a smaller voxel volume, reduces partial volume effects and hence might improve visibility of myelin-rich laminae between the AV and its surrounding nuclei. However, a concomitantly decreasing SNR might conceal the subtle differences in signal intensity between the AV and neighbouring thalamic nuclei. As shown in figure 1, partial volume effects are not detrimental to AV visualization at a resolution of (1.0 mm)3, even though the structural complexity of the thalamus would seem to require a higher resolution (0.8 mm)3. Furthermore, the combination of a tight fitting head coil with a field strength of 3 T is crucial for a high SNR. A multi-element phased array coil should not be mandatory as it is most effective near the coil elements, i.e. the cortex. For the centre of a head coil at best a marginal SNR gain is to be expected [15, 16]. Nevertheless, nowadays the smallest commercial head coils are array coils. Most of the presented images were measured with two averages. A scan time of about 15 min is recommended to ensure high image quality if involuntary head movements are neglectable. At lower resolution, images are less susceptible to motion artefacts. A tight and at the same time convenient fixation of the head reduces motion artefacts effectively. Even if the example given in figure 2 suggests that the image definition of a scan acquired with only one average can suffice, we recommend using two averages. The scan times can also be shortened by choosing the shortest possible repetition time TR. However, MPRAGE T1 contrast is significantly raised by the insertion of a delay (in this study ca. 400 ms) between the end of the echo train and the next inversion pulse. The MPRAGE sequence employed here is standard on all MR scanners and ubiquitous in clinical routine. Previous attempts to visualize thalamic nuclei required challenging data post-processing, or did at least rely on nonDirect Targeting of the Thalamic AV for DBS
commercial in-house software solutions [11–13]. A recently published work of Bender et al. [17] describes the identification of the four large thalamic nuclei groups (including the anterior group) by means of an optimized MPRAGE sequence. Their images show a good demarcation between the medial and the lateral group of thalamic nuclei whereas the image examples given here are supposed to allow a more precise targeting of the AV. Stereotactic access to the AV for the lack of delineable intrathalamic structures has so far relied upon indirect targeting. Typical coordinates as provided by Hodaie et al. [4] are 6 mm lateral to, 8 mm anterior to and 12 mm above the midcommisural point. However, asymmetries can be expected in a number of patients which have undergone epilepsy surgery. Also, atrophy of the entire thalamus has been described in patients with long-standing epilepsy [18]. The AV is a lengthy but narrow structure, and targeting is further complicated by the thalamostriate vein representing an extensive risk zone in its immediate vicinity. Therefore, knowledge about its boundaries is particularly important in choosing a stereotactic approach to this target. Here direct visualization of the AV resulted in a more anterior placement of electrodes compared to the approach described by Hodaie et al. [4]. However, it is important to stress that it would be difficult if not impossible to draw any reliable conclusions from only 3 patients. It is not our intention to clinically evaluate this procedure but to increase precision in targeting the AV, which in turn should facilitate such an evaluation in a more rigorous fashion.
Conclusion
This study has shown that it is possible to delineate the anteroventral thalamic nucleus on T1-weighted MR images to a degree which enables a stereotactic neurosurgeon to perform direct targeting. The 3-tesla MRI protocol presented can easily be integrated into clinical routine and should improve preciseness of DBS in this target.
References
1 Brunenberg EJL, Platel B, Hofman PAM, Ter Haar Romeny BM, Visser-Vandewalle V: Magnetic resonance imaging techniques for visualization of the subthalamic nucleus. J Neurosurg 2011;115:971–984. 2 Mirski MA, McKeon AC, Ferrendelli JA: Anterior thalamus and substantia nigra: two distinct structures mediating experimental generalized seizures. Brain Res 1986; 397: 377– 380.
Stereotact Funct Neurosurg 2014;92:25–30 DOI: 10.1159/000351525
29
Downloaded by: Lund University Libraries 130.235.136.21 - 5/13/2015 2:46:07 PM
cles and also of the AV. Nonetheless penetration of the AV could be achieved in the given examples. Average coordinates of electrode contacts in the AV were derived from intra-operative stereotactic X-ray images registered with the preoperative MR images. The target coordinates ranged from 4.5 to 6 mm lateral to the midcommissural line, 12 to 16 mm anterior and 12 to 14.5 mm above the posterior commissure.
30
8 Morel A: Stereotactic Atlas of the Human Thalamus and Basal Ganglia. New York, Informa Healthcare, 2007. 9 Schaltenbrand G, Wahren W: Atlas for Stereotaxy of the Human Brain, ed 2. Stuttgart, Thieme, 1977. 10 Fisher R, Salanova V, Witt T, Worth R, Henry T, Gross R, et al: Electrical stimulation of the anterior nucleus of thalamus for treatment of refractory epilepsy. Epilepsia 2010; 51: 899– 908. 11 Wiegell MR, Tuch DS, Larsson HBW, Wedeen VJ: Automatic segmentation of thalamic nuclei from diffusion tensor magnetic resonance imaging. Neuroimage 2003; 19: 391– 401. 12 Gringel T, Schulz-Schaeffer W, Elolf E, Frölich A, Dechent P, Helms G: Optimized high-resolution mapping of magnetization transfer (MT) at 3 Tesla for direct visualization of substructures of the human thalamus in clinically feasible measurement time. J Magn Reson Imaging 2009;29:1285–1292. 13 Traynor CR, Barker GJ, Crum WR, Williams SCR, Richardson MP: Segmentation of the thalamus in MRI based on T1 and T2. Neuroimage 2011;56:939–950.
Stereotact Funct Neurosurg 2014;92:25–30 DOI: 10.1159/000351525
14 Mugler JP, Brookeman JR: Three-dimensional magnetization-prepared rapid gradientecho imaging (3D MP RAGE). Magn Reson Med 1990;15:152–157. 15 Wright SM, Wald LL: Theory and application of array coils in MR spectroscopy. NMR Biomed 1997;10:394–410. 16 Wiggins GC, Polimeni JR, Potthast A, Schmitt M, Alagappan V, Wald LL: 96-Channel receive-only head coil for 3 Tesla: design optimization and evaluation. Magn Reson Med 2009;62:754–762. 17 Bender B, Mänz C, Korn A, Nägele T, Klose U: Optimized 3D magnetization-prepared rapid acquisition of gradient echo: identification of thalamus substructures at 3T. AJNR Am J Neuroradiol 2011;32:2110–2115. 18 Margerison JH, Corsellis JA: Epilepsy and the temporal lobes. A clinical, electroencephalographic and neuropathological study of the brain in epilepsy, with particular reference to the temporal lobes. Brain 1966;89:499–530.
Buentjen/Kopitzki/Schmitt/Voges/ Tempelmann/Kaufmann/Kanowski
Downloaded by: Lund University Libraries 130.235.136.21 - 5/13/2015 2:46:07 PM
3 Mirski MA, Rossell LA, Terry JB, Fisher RS: Anticonvulsant effect of anterior thalamic high frequency electrical stimulation in the rat. Epilepsy Res 1997;28:89–100. 4 Hodaie M, Wennberg RA, Dostrovsky JO, Lozano AM: Chronic anterior thalamus stimulation for intractable epilepsy. Epilepsia 2002; 43:603–608. 5 Hamani C, Ewerton FIS, Bonilha SM, Ballester G, Mello LEAM, Lozano AM: Bilateral anterior thalamic nucleus lesions and high-frequency stimulation are protective against pilocarpineinduced seizures and status epilepticus. Neurosurgery 2004;54:191–195, discussion 195–197. 6 Kerrigan JF, Litt B, Fisher RS, Cranstoun S, French JA, Blum DE, et al: Electrical stimulation of the anterior nucleus of the thalamus for the treatment of intractable epilepsy. Epilepsia 2004;45:346–354. 7 Hirai T, Jones EG: A new parcellation of the human thalamus on the basis of histochemical staining. Brain Res Brain Res Rev 1989;14:1–34.
