- Email: [email protected]
Diffusion MR Tractography As a Tool for Surgical Planning
Diffusion MR Tractography As a Tool for Surgical Planning Jeffrey Berman, PhD KEYWORDS MR imaging Diffusion tensor imaging Tractography Surgical planning Brain Motor tract
tracking have driven the development of advanced HARDI techniques to provide more accurate presurgical tractography. This article includes a description of the most recent applications of q-ball tractography, one of these HARDI methods, to surgical planning. As DTI and HARDI tractography become more widely used for patient diagnosis and treatment, it is critical for users to understand the capabilities and limitations of these techniques.
SPECTRUM OF BRAIN MAPPING TECHNIQUES Tractography complements the information available from an array of brain mapping techniques available for surgical planning. Brain mapping techniques are differentiated by invasiveness, accuracy, and their physiologic basis. Diffusion MR is inherently noninvasive and does not, in most applications, directly detect brain function. Instead, tractography uses the diffusion of water to observe the microstructure of brain tissue, construct the gross anatomy of major white matter pathways, and then infer the trajectory of the delineated tract. Other noninvasive functional mapping techniques, such as functional MR imaging,7–10 magnetoencephalography,11–13 and electrocorticography,14 observe physiologic changes associated with brain function. However, these techniques are restricted to localizing activity in the gray matter. Cortical and subcortical electrical stimulation directly determines the function of neurons and remains the gold standard for functional mapping during neurosurgery.15,16 Stimulation is highly invasive, however, and subcortical stimulation
Department of Radiology and Biomedical Imaging, University of California, San Francisco, 185 Berry Street, Suite 350, San Francisco, CA 94107, USA E-mail address: [email protected] Magn Reson Imaging Clin N Am 17 (2009) 205–214 doi:10.1016/j.mric.2009.02.002 1064-9689/09/$ – see front matter ª 2009 Elsevier Inc. All rights reserved.
mri.theclinics.com
Diffusion MR tractography has rapidly become an important clinical tool that can delineate functionally important white matter tracts for surgical planning. One of the goals of brain surgery is to avoid damage to eloquent cortex and subcortical white matter. Diffusion tractography remains the only noninvasive method capable of segmenting the subcortical course of a white matter tract. This article reviews the technical and clinical issues surrounding presurgical diffusion tractography, including traditional diffusion tensor imaging (DTI) methods and more advanced high angular resolution diffusion imaging (HARDI) approaches, such as q-ball imaging. An overview of the presurgical DTI and q-ball tractography protocols used at our institution is also provided. Tractography based on diffusion MR exploits the correlation between water diffusion and brain structure to delineate the course of white matter pathways.1–5 The brain’s white matter is highly organized into fasciculi comprised of densely packed axons. Axonal membranes, myelin, and other structures affect the pattern of Brownian motion of water within white matter.6 Tractography follows a white matter tract from voxel to voxel in three dimensions by assuming that the direction of least restricted diffusion corresponds with the orientation of axons. The quantitative assessments of tissue microstructure central to the scientific applications of diffusion MR are not directly used for presurgical tractography. Instead, the goal of presurgical tractography is simply to identify the position of eloquent pathways, such as the motor, sensory, and language tracts. The shortcomings of DTI fiber
206
Berman cannot be performed until the resection cavity is close to the target pathway.17 Localization with stimulation also is subject to inaccuracies from the penetration of electrical current into surrounding tissues and variations in excitation thresholds.18,19 In the clinical setting, it is advantageous to use more than one brain mapping technique. Independent aspects of physiology are observed with each mapping technique, and their combined results can be used to validate each other and improve the confidence of surgical planning decisions.
ROUTINE PRESURGICAL DIFFUSION TENSOR IMAGING TRACTOGRAPHY PROTOCOL Translating DTI tractography into a useful clinical tool requires efficient integration of the technique into presurgical protocols. This section describes the presurgical DTI fiber-tracking protocol developed and routinely performed at University of California San Francisco. DTI tractography of the motor tract is prescribed for any brain tumor case with a lesion near the motor tract that involves awake or asleep intraoperative cortical stimulation mapping. The protocol involves image acquisition, post processing, quality assurance, and integration with a surgical navigation system. DTI acquisition sequences are readily available on all commercial MR scanner platforms. The following discussion provides several general guidelines for acquiring diffusion MR data suitable for tractography; more detailed specifications can be found in a recent review.20,21 A minimum of six diffusion gradient directions is required to calculate the diffusion tensor; however, significant improvements in data quality can be obtained by increasing the number of directions.22 For tractography, the slices must be contiguous and should not be thicker than 3 mm. At our institution, a typical acquisition at 1.5 T on a GE scanner has TR/TE 5 11 s/78 ms, 3-mm thick slices, matrix 5 128 128, field of view 5 260 260 mm, ASSET factor 5 2, and 15 diffusion directions with two averaged acquisitions at b 5 1000 s/mm2. Total diffusion imaging time is less than 8 minutes. In addition to diffusion MR, high-resolution three-dimensional T1- and T2-weighted anatomic images are acquired for use with the surgical navigation system. The postcontrast three-dimensional T1-weighted spoiled gradient recalled and the three-dimensional T2-weighted fast spin echo volumes have voxel resolution of 1 1 1.5 mm. DTI tractography is performed with multiple regions of interest using an algorithm based on the deterministic fiber assignment by continuous
tracking method.4 Fiber tracks are first generated from a region drawn in the cerebral peduncle with the aid of the b 5 0 s/mm2 echo planar images from the DTI sequence (Fig. 1A) and the postprocessed DTI anisotropy map. The starting region encompasses the entire cerebral peduncle, although the motor tract only occupies a fraction of this region. An overly large starting region reduces the potential for user-introduced error. Tractography algorithms such as the fiber assignment by continuous tracking method require a minimum anisotropy threshold to be specified for propagation of the ‘‘streamline’’ outlining the fiber trajectory. If the fiber track encounters a voxel with anisotropy below the threshold, which is usually caused by the presence of gray matter or cerebrospinal fluid, the streamline is terminated at that location. A relatively low fractional anisotropy threshold of 0.1 is used for presurgical tractography to assist in following white matter tracts through regions of edema or tumor infiltration, which exhibit low diffusion anisotropy.23 Each voxel in the starting region is densely seeded with 27 starting points evenly arranged in a 3 3 3 grid to improve fiber tracking performance.2 Extraneous fiber tracks are removed through the use of multiple target regions of interest (Fig. 1). The fiber tracks are first targeted to a region drawn in the posterior limb of the internal capsule. Then a second target region is drawn around fiber track bundles anterior to the central sulcus within the precentral gyrus. Fiber tracks passing through all three regions are retained as the delineated pyramidal tract (See Fig. 1). The echo planar images are unsuitable for surgical navigation because of poor image contrast and geometric warping artifacts that compromise anatomic fidelity, so the DTI tractography must be co-registered to the highresolution anatomic images. The b 5 0 s/mm2 echo planar images are first aligned to the anatomic images using automated image registration.24,25 A three-dimensional affine 12parameter model with a maximum of 50 iterations is used for registration. The registration’s performance is checked by comparing the position of sulci, ventricles, and cortex. The transformation matrix is applied to the DTI fiber track mask, and the registered mask is fused to the high-resolution anatomic images. Voxels that contain fiber tracks are given intensities 1.5 times higher than the brightest anatomic feature in the image (Fig. 1C, D). This ‘‘hot white’’ intensity ensures that fiber tracks can be distinguished easily from T2-hyperintense lesions and cerebrospinal fluid.
Diffusion MR Tractography
Fig. 1. (A) Diffusion MR tractography of the motor tract (green) is shown overlaid on axial b 5 0 s/mm2 echo planar images. Starting and target regions of interest (red outlines) are shown in the cerebral peduncle, posterior limb of the internal capsule, and precentral gyrus. (B) The resultant fiber tracks (red) are shown adjacent to the lesion (green) in three dimensions. (C) Tractography (green) is registered to the T2-weighted fast spin echo anatomic volume and (D) T1-weighted spoiled gradient recalled volume.
Black and white anatomic images fused with the DTI tractography easily can be made Digital Imaging and Communications in Medicine (DICOM) compliant and recognized by multiple systems, including PACS and surgical navigation systems (Fig. 2). The DICOM standard and newer surgical navigation systems allow tractography to be viewed as a color overlay. The advantage of color overlays is that they can be removed easily if the surgeon desires to view the anatomic images beneath the tractography. Multiple white matter pathways can be represented by different colors and simultaneously viewed. Diffusion MR fiber tracks also can be viewed in relation to the tumor and skull as three-dimensional streamlines, streamtubes, glyphs, and solid volumes.26
TRACTOGRAPHY OF PATHWAYS OTHER THAN THE MOTOR TRACT Most presurgical tractography studies have focused on the motor tract because it is important for patient quality of life and is easy to track. However, tractography of functional pathways other than the cerebral corticospinal tract have
been integrated with a surgical navigation system or studied before surgery. A brainstem cavernous angioma was safely removed with the aid of sensory and motor pathway tractography.27 The optic radiation is a challenging pathway to delineate with tractography because it is a thin ribbon of white matter and turns at a sharp angle through Meyer’s loop (Fig. 3). Studies have shown the feasibility of using tractography to plan radiation therapy and guide surgical resection of arteriovenous malformations.28,29 The anatomy and operation of the language pathways are still under investigation. Tractography has been launched from mouth motor, speech arrest, and anomia stimulation sites on the frontal cortex.30,31 The subcortical language network was observed to include connections to the supplementary motor area, cerebral peduncle, and thalamus. Language tracts including the arcuate fasciculus and the uncinate fasciculus have been delineated with tractography and their position confirmed with stimulation.32,33 In patients undergoing anterior temporal lobectomy in the language dominant hemisphere, preoperative tractography has been used to predict language deficits.34
207
208
Berman
Fig. 2. Motor tractography as viewed during surgery with a surgical navigation system. Axial and coronal views are shown of the T2-weighted anatomic volume fused with the tractography. Voxels through which the fiber tracks pass are assigned bright white intensities. The red crosshairs indicate the location of a mouth motor subcortical stimulation site.
OTHER MAPPING TECHNIQUES USED WITH TRACTOGRAPHY
Fig. 3. Left and right optic radiations are identified with DTI tractography in an epileptic patient before surgery. Fiber tracks in three dimensions (purple) are projected on a b 5 0 s/mm2 axial echo planar slice. The anterior extent of Meyer’s loop is not fully delineated with DTI tractography because of the tract’s tight turn and thin dimensions.
Diffusion MR should never be solely relied on to guide a resection. Because diffusion MR measures a structural property of the brain’s tissue, functional information always must be added to infer the location of eloquent pathways. In most presurgical cases, our knowledge of brain anatomy enables us to draw starting and target regions in the cerebral peduncle and the precentral gyrus. In some brain tumor cases, however, the tumor has shifted or infiltrated the surrounding white matter, changing the geometry of the brain’s anatomy.35,36 These situations necessitate additional functional mapping to guide the placement of tractography starting and target regions. Motor functional MR imaging has been used to locate the motor and language cortex in patients who have brain tumor and define the starting region for diffusion tractography.37–39 The use of functional MR imaging allows detailed functional specificity to be assigned to the generated tractography.40 The subsection of the motor tract or language pathways specific to a particular muscle group or language function can be delineated. There is a disjoint between the functional MR imaging–activated volume in the gray matter and brain tissue with diffusion anisotropy in the white matter, however. The tractography starting region can be drawn manually in the white matter of the gyrus adjacent to the functional MR imaging activation.
Diffusion MR Tractography Validation of DTI tractography is necessary if the technique is to be trusted for surgical planning. Electrical stimulation of cortical and subcortical brain tissue has aided in the preservation of functional brain tissue for many decades.16 Comparison to electrical stimulation has been a necessary step to validate tractography’s use during surgery. Electrical stimulation studies test tractography and its integration with the surgical navigation system in a real clinical situation. Fig. 4 demonstrates the combination of DTI
tractography and cortical stimulation to delineate the portion of the motor tract that controls jaw motions. Diffusion tractography launched from stimulation sites on the primary motor cortex has been observed to follow the expected anatomic course and maintain somatotopic organization through the internal capsule.30 An initial study observed a gap between tractography and subcortical motor stimulation.41 The diffusion tractography was capable of revealing the relative position of
Fig. 4. DTI fiber tracks were launched from a jaw motor cortical stimulation site (blue circle). The DTI tractography (red) represents only the portion of the motor pathway descending from the cortical stimulation site. This case demonstrated the feasibility of using DTI fiber tracking to follow white matter tracts through some T2-hyperintense lesions.
209
210
Berman the tumor and corticospinal tract; however, the borders of the tract were underestimated. Additional subcortical motor stimulation studies have repeatedly observed an approximately 1-cm gap between the electrode tip and tractography.42–44 The gap between DTI tractography and bipolar stimulation represents the cumulative errors from diffusion tensor uncertainty, registration inaccuracies, and brain shift during surgery. Noise in the diffusion MR signal results in uncertainty in the orientation of the diffusion tensor and the position of fiber tracks.45,46 Geometric distortions are common on the echo planar images used to acquire diffusion MR. Care must be taken when
registering echo planar images to the high-resolution anatomic volumes. Numerous studies have examined the effect of brain shift during surgery.47–51 It is possible for the brain to shift either inward or outward, altering the position of tractography as perceived on the surgical navigation system. Bipolar stimulation currents can penetrate several millimeters into surrounding brain tissue.18 Stimulation of neurons distant from the tip of the bipolar electrode always contributes to the approximately 1-cm discrepancy between the stimulation coordinates reported by the navigation system and the tractography.
Fig. 5. Q-ball fiber tracks (green, top row) are compared with DTI fiber tracks (blue, bottom row). The tumor is in orange. The DTI fiber tracks are restricted to the portion of the motor tract originating in the superior-medial aspect of the motor cortex controlling lower extremity functions. The q-ball fiber tracks delineate a larger portion of the motor tract than the DTI fiber tracks. The q-ball fiber tracks show connectivity to the lateral portion of the motor cortex controlling facial motor functions.
Diffusion MR Tractography HIGH ANGULAR RESOLUTION DIFFUSION IMAGING FOR FIBER TRACKING The diffusion tensor model is limited to a single fiber population per voxel, and DTI fiber tracking fails in regions with crossing white matter fibers or intravoxel partial volume averaging of adjacent tracts with different fiber orientations. This shortcoming of DTI can be overcome by sampling many more diffusion-weighted directions than the minimum of six required for DTI and acquiring the images with stronger diffusion weighting than the clinical standard of b 5 1000 s/mm2 for DTI. This technique is known as HARDI, and many different mathematical approaches have been devised recently to reconstruct fiber pathways based on HARDI information. One of the most popular of these is known as q-ball imaging.3,5 The q-ball reconstruction of HARDI data can distinguish distinct fiber populations in regions of complex white matter fiber architecture.3,5 Tractography based on these advanced diffusion MR methods has been shown to traverse
white matter regions with crossing fibers and reveal a greater portion of tracts.52–54 The major limitation of HARDI’s clinical application is the lengthy scan time required for acquiring many diffusion gradient directions. With only 55 gradient directions and parallel imaging at 3 T, however, scan time can be reduced to approximately 13 minutes or less.3,55 A typical acquisition at our institution on a 3 T GE Signa EXCITE scanner has parallel imaging reduction factor of 2, TR/TE 5 12.4 s/73 ms, 2.2 2.2 2.2 isotropic voxels, field of view 5 280 280 mm, and a 128 128 matrix zero-filled to a 256 256 matrix. The 55 diffusion gradient directions are acquired with b 5 2000 s/mm2. Although 13 minutes still may be a bit lengthy for many clinical situations, it can be anticipated that the required scan time will decrease with hardware and HARDI methodologic improvements. Our institution has begun using q-ball tractography for a limited number of patients who have brain tumors.55 As seen in Fig. 5, q-ball fiber tracking is able to delineate the motor tract from the midbrain to the
Fig. 6. The region of complex white matter in the centrum semiovale (pink rectangle) prevents DTI fiber tracks from traversing laterally to the motor cortex. The q-ball orientation density function distinguishes the motor pathway from the superior longitudinal fasciculus (SLF) and enables fiber tracks to course laterally through the centrum semiovale to the motor cortex.
211
212
Berman motor evoked potentials were combined with tractography to preserve motor function. This clinical outcome study and other subcortical stimulation validation studies suggested that stimulation be continually applied when resections are within 1 cm of motor tractography. Tractography can speed up radical resections by indicating which borders of a resection cavity deserve extensive stimulation. Future improvements in MR hardware and HARDI methods will further enable tractography’s clinical use for surgical planning. Protecting subcortical pathways and functional cortices is important during surgery. It is important to understand and account for the limitations of tractography and include multiple mapping techniques when planning surgeries.
ACKNOWLEDGMENTS
Fig. 7. Q-ball tractography (green) of the motor tract is shown in three dimensions in a patient who has a brain tumor. Cortical tongue motor (purple spheres) and subcortical mouth motor (orange sphere) stimulation sites are indicated. A cortical tongue sensory site (blue sphere) is also visible. The q-ball tractography generated presurgically showed connectivity from one cortical motor site, through the subcortical motor site, to the cerebral peduncle.
superior-medial and lateral portions of the motor cortex in our patients. The lateral sections of the motor tract control facial movements and are necessary for language functions. DTI fiber tracking revealed connectivity from the midbrain to only the medial-superior aspect of the motor cortex. The centrum semiovale is a region that contains multiple crossing white matter pathways, and DTI fails to accurately characterize this complex structure (Fig. 6). Initial studies have shown cortical and subcortical stimulations to be in good agreement with the course of the motor tract predicted by q-ball tractography (Fig. 7).55
IMPACT ON PATIENT CARE AND FUTURE DIRECTIONS Many studies have been conducted to improve tractography techniques and assess the accuracy of diffusion tractography. Recently a few studies were conducted to assess the technique’s impact on patient care. Mikuni examined postoperative motor function after tractography-guided surgery in patients who had brain tumors.56,57 Subcortical
I am grateful to Roland Henry, PhD, for his guidance on analyzing the data for many of the figures, and to Pratik Mukherjee, MD, PhD, for his help preparing this article.
REFERENCES 1. Basser PJ, Mattiello J, LeBihan D. Estimation of the effective self-diffusion tensor from the NMR spin echo. J Magn Reson B 1994;103(3):247–54. 2. Conturo TE, Lori NF, Cull TS, et al. Tracking neuronal fiber pathways in the living human brain. Proc Natl Acad Sci USA 1999;96(18):10422–7. 3. Hess CP, Mukherjee P, Han ET, et al. Q-ball reconstruction of multimodal fiber orientations using the spherical harmonic basis. Magn Reson Med 2006; 56(1):104–17. 4. Mori S, Crain BJ, Chacko VP, et al. Three-dimensional tracking of axonal projections in the brain by magnetic resonance imaging. Ann Neurol 1999; 45(2):265–9. 5. Tuch DS. Q-ball imaging. Magn Reson Med 2004; 52(6):1358–72. 6. Beaulieu C. The basis of anisotropic water diffusion in the nervous system – a technical review. NMR Biomed 2002;15(7–8):435–55. 7. Guye M, Parker GJ, Symms M, et al. Combined functional MRI and tractography to demonstrate the connectivity of the human primary motor cortex in vivo. Neuroimage 2003;19(4):1349–60. 8. Krishnan R, Raabe A, Hattingen E, et al. Functional magnetic resonance imaging-integrated neuronavigation: correlation between lesion-to-motor cortex distance and outcome. Neurosurgery 2004;55(4): 904–14 [discusssion 914–5]. 9. Ruff IM, Petrovich Brennan NM, Peck KK, et al. Assessment of the language laterality index in
Diffusion MR Tractography
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
patients with brain tumor using functional MR imaging: effects of thresholding, task selection, and prior surgery. AJNR Am J Neuroradiol 2008; 29(3):528–35. Schulder M, Maldjian JA, Liu WC, et al. Functional image-guided surgery of intracranial tumors located in or near the sensorimotor cortex. J Neurosurg 1998;89(3):412–8. Ganslandt O, Buchfelder M, Hastreiter P, et al. Magnetic source imaging supports clinical decision making in glioma patients. Clinical Neurology and Neurosurgery 2004;107(1):20–6. Nagarajan S, Kirsch H, Lin P, et al. Preoperative localization of hand motor cortex by adaptive spatial filtering of magnetoencephalography data. J Neurosurg 2008;109(2):228–37. Schiffbauer H, Berger MS, Ferrari P, et al. Preoperative magnetic source imaging for brain tumor surgery: a quantitative comparison with intraoperative sensory and motor mapping. J Neurosurg 2002;97(6):1333–42. Engel AK, Moll CK, Fried I, et al. Invasive recordings from the human brain: clinical insights and beyond. Nat Rev Neurosci 2005;6(1):35–47. Ojemann G. Brain organization for language from the perspective of electical stimulation mapping. The Behavioral and Brain Sciences 1983;6:189–230. Penfield W, Rasmussen T. The cerebral cortex of man: a clinical study of localization of function. New York: Macmillan; 1950. xv, 248. Keles GE, Lundin DA, Lamborn KR, et al. Intraoperative subcortical stimulation mapping for hemispherical perirolandic gliomas located within or adjacent to the descending motor pathways: evaluation of morbidity and assessment of functional outcome in 294 patients. J Neurosurg 2004;100(3):369–75. Haglund MM, Ojemann GA, Blasdel GG. Optical imaging of bipolar cortical stimulation. J Neurosurg 1993;78(5):785–93. Pouratian N, Cannestra AF, Bookheimer SY, et al. Variability of intraoperative electrocortical stimulation mapping parameters across and within individuals. J Neurosurg 2004;101(3):458–66. Mukherjee P, Berman JI, Chung SW, et al. Diffusion tensor MR imaging and fiber tractography: theoretic underpinnings. AJNR Am J Neuroradiol 2008;29(4): 632–41. Mukherjee P, Chung SW, Berman JI, et al. Diffusion tensor MR imaging and fiber tractography: technical considerations. AJNR Am J Neuroradiol 2008;29(5): 843–52. Jones DK. The effect of gradient sampling schemes on measures derived from diffusion tensor MRI: a Monte Carlo study. Magn Reson Med 2004;51(4): 807–15. Stadlbauer A, Nimsky C, Buslei R, et al. Diffusion tensor imaging and optimized fiber tracking in
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
glioma patients: Histopathologic evaluation of tumor-invaded white matter structures. Neuroimage 2007;34(3):949–56. Woods RP, Grafton ST, Holmes CJ, et al. Automated image registration: I. General methods and intrasubject, intramodality validation. Journal of Computer Assisted Tomography 1998;22(1):139–52. Woods RP, Grafton ST, Watson JDG, et al. Automated image registration: II. Intersubject validation of linear and nonlinear models. Journal of Computer Assisted Tomography 1998;22(1):153–65. Nimsky C, Ganslandt O, Enders F, et al. Visualization Strategies for Major White Matter Tracts for Intraoperative Use. International Journal of Computer Assisted Radiology and Surgery 2006;1:13–22. Chen X, Weigel D, Ganslandt O, et al. Diffusion tensor-based fiber tracking and intraoperative neuronavigation for the resection of a brainstem cavernous angioma. Surg Neurol 2007;68(3): 285–91 [discussion 291]. Kikuta K, Takagi Y, Nozaki K, et al. Early experience with 3-T magnetic resonance tractography in the surgery of cerebral arteriovenous malformations in and around the visual pathway. Neurosurgery 2006;58(2):331–7 [discussion 331–7]. Maruyama K, Kamada K, Shin M, et al. Optic radiation tractography integrated into simulated treatment planning for Gamma Knife surgery. J Neurosurg 2007;107(4):721–6. Berman JI, Berger MS, Mukherjee P, et al. Diffusiontensor imaging-guided tracking of fibers of the pyramidal tract combined with intraoperative cortical stimulation mapping in patients with gliomas. J Neurosurg 2004;101(1):66–72. Henry RG, Berman JI, Nagarajan SS, et al. Subcortical pathways serving cortical language sites: initial experience with diffusion tensor imaging fiber tracking combined with intraoperative language mapping. Neuroimage 2004;21(2):616–22. Bello L, Gambini A, Castellano A, et al. Motor and language DTI Fiber Tracking combined with intraoperative subcortical mapping for surgical removal of gliomas. Neuroimage 2008;39(1):369–82. Kamada K, Todo T, Masutani Y, et al. Visualization of the frontotemporal language fibers by tractography combined with functional magnetic resonance imaging and magnetoencephalography. J Neurosurg 2007;106(1):90–8. Powell HW, Parker GJ, Alexander DC, et al. Imaging language pathways predicts postoperative naming deficits. J Neurol Neurosurg Psychiatry 2008;79(3): 327–30. Fernandez-Miranda JC, Rhoton AL Jr, AlvarezLinera J, et al. Three-dimensional microsurgical and tractographic anatomy of the white matter of the human brain. Neurosurgery 2008;62(6 Suppl. 3): 989–1026 [discussion 1026–8].
213
214
Berman 36. Wieshmann UC, Symms MR, Parker GJ, et al. Diffusion tensor imaging demonstrates deviation of fibres in normal appearing white matter adjacent to a brain tumour. J Neurol Neurosurg Psychiatry 2000;68(4): 501–3. 37. Hendler T, Pianka P, Sigal M, et al. Delineating gray and white matter involvement in brain lesions: threedimensional alignment of functional magnetic resonance and diffusion-tensor imaging. J Neurosurg 2003;99(6):1018–27. 38. Holodny AI, Ollenschleger MD, Liu WC, et al. Identification of the corticospinal tracts achieved using blood-oxygen-level-dependent and diffusion functional MR imaging in patients with brain tumors. AJNR Am J Neuroradiol 2001;22(1):83–8. 39. Schonberg T, Pianka P, Hendler T, et al. Characterization of displaced white matter by brain tumors using combined DTI and fMRI. Neuroimage 2006; 30(4):1100–11. 40. Smits M, Vernooij MW, Wielopolski PA, et al. Incorporating functional MR imaging into diffusion tensor tractography in the preoperative assessment of the corticospinal tract in patients with brain tumors. AJNR Am J Neuroradiol 2007;28(7):1354–61. 41. Kinoshita M, Yamada K, Hashimoto N, et al. Fibertracking does not accurately estimate size of fiber bundle in pathological condition: initial neurosurgical experience using neuronavigation and subcortical white matter stimulation. Neuroimage 2005; 25(2):424–9. 42. Berman JI, Berger MS, Chung SW, et al. Accuracy of diffusion tensor magnetic resonance imaging tractography assessed using intraoperative subcortical stimulation mapping and magnetic source imaging. J Neurosurg 2007;107(3):488–94. 43. Kamada K, Todo T, Masutani Y, et al. Combined use of tractography-integrated functional neuronavigation and direct fiber stimulation. J Neurosurg 2005; 102(4):664–72. 44. Mikuni N, Okada T, Nishida N, et al. Comparison between motor evoked potential recording and fiber tracking for estimating pyramidal tracts near brain tumors. J Neurosurg 2007;106(1):128–33. 45. Anderson AW. Theoretical analysis of the effects of noise on diffusion tensor imaging. Magn Reson Med 2001;46(6):1174–88. 46. Lazar M, Alexander AL. An error analysis of white matter tractography methods: synthetic diffusion tensor field simulations. Neuroimage 2003;20(2): 1140–53.
47. Coenen VA, Krings T, Weidemann J, et al. Sequential visualization of brain and fiber tract deformation during intracranial surgery with three-dimensional ultrasound: an approach to evaluate the effect of brain shift. Neurosurgery 2005;56(1 Suppl):133–41 [discussion 133–41]. 48. Keles GE, Lamborn KR, Berger MS. Coregistration accuracy and detection of brain shift using intraoperative sononavigation during resection of hemispheric tumors. Neurosurgery 2003;53(3):556–62 [discussion 562–4]. 49. Nimsky C, Ganslandt O, Hastreiter P, et al. Preoperative and intraoperative diffusion tensor imagingbased fiber tracking in glioma surgery. Neurosurgery 2005;56(1):130–7. 50. Reinges MHT, Nguyen HH, Krings T, et al. Course of brain shift during microsurgical resection of supratentorial cerebral lesions: limits of conventional neuronavigation. Acta Neurochirurgica 2004;146(4): 369–77. 51. Roberts DW, Hartov A, Kennedy FE, et al. Intraoperative brain shift and deformation: a quantitative analysis of cortical displacement in 28 cases. Neurosurgery 1998;43(4):749–58 [discussion 758–60]. 52. Berman JI, Chung S, Mukherjee P, et al. Probabilistic streamline Q-Ball tractography using the residual bootstrap. Neuroimage 2008;39:215–22. 53. Campbell JS, Siddiqi K, Rymar VV, et al. Flow-based fiber tracking with diffusion tensor and q-ball data: validation and comparison to principal diffusion direction techniques. Neuroimage 2005;27(4): 725–36. 54. Wedeen VJ, Wang RP, Schmahmann JD, et al. Diffusion spectrum magnetic resonance imaging (DSI) tractography of crossing fibers. Neuroimage 2008; 41(4):1267–77. 55. Berman JI, Clark DJ, Berger MS, et al. Improved Diffusion MR Fiber Tracking for Neurosurgical Application. ISMRM 2008; Toronto. 56. Mikuni N, Okada T, Enatsu R, et al. Clinical impact of integrated functional neuronavigation and subcortical electrical stimulation to preserve motor function during resection of brain tumors. J Neurosurg 2007; 106(4):593–8. 57. Mikuni N, Okada T, Enatsu R, et al. Clinical significance of preoperative fibre-tracking to preserve the affected pyramidal tracts during resection of brain tumours in patients with preoperative motor weakness. J Neurol Neurosurg Psychiatry 2007; 78(7):716–21.
tracking have driven the development of advanced HARDI techniques to provide more accurate presurgical tractography. This article includes a description of the most recent applications of q-ball tractography, one of these HARDI methods, to surgical planning. As DTI and HARDI tractography become more widely used for patient diagnosis and treatment, it is critical for users to understand the capabilities and limitations of these techniques.
SPECTRUM OF BRAIN MAPPING TECHNIQUES Tractography complements the information available from an array of brain mapping techniques available for surgical planning. Brain mapping techniques are differentiated by invasiveness, accuracy, and their physiologic basis. Diffusion MR is inherently noninvasive and does not, in most applications, directly detect brain function. Instead, tractography uses the diffusion of water to observe the microstructure of brain tissue, construct the gross anatomy of major white matter pathways, and then infer the trajectory of the delineated tract. Other noninvasive functional mapping techniques, such as functional MR imaging,7–10 magnetoencephalography,11–13 and electrocorticography,14 observe physiologic changes associated with brain function. However, these techniques are restricted to localizing activity in the gray matter. Cortical and subcortical electrical stimulation directly determines the function of neurons and remains the gold standard for functional mapping during neurosurgery.15,16 Stimulation is highly invasive, however, and subcortical stimulation
Department of Radiology and Biomedical Imaging, University of California, San Francisco, 185 Berry Street, Suite 350, San Francisco, CA 94107, USA E-mail address: [email protected] Magn Reson Imaging Clin N Am 17 (2009) 205–214 doi:10.1016/j.mric.2009.02.002 1064-9689/09/$ – see front matter ª 2009 Elsevier Inc. All rights reserved.
mri.theclinics.com
Diffusion MR tractography has rapidly become an important clinical tool that can delineate functionally important white matter tracts for surgical planning. One of the goals of brain surgery is to avoid damage to eloquent cortex and subcortical white matter. Diffusion tractography remains the only noninvasive method capable of segmenting the subcortical course of a white matter tract. This article reviews the technical and clinical issues surrounding presurgical diffusion tractography, including traditional diffusion tensor imaging (DTI) methods and more advanced high angular resolution diffusion imaging (HARDI) approaches, such as q-ball imaging. An overview of the presurgical DTI and q-ball tractography protocols used at our institution is also provided. Tractography based on diffusion MR exploits the correlation between water diffusion and brain structure to delineate the course of white matter pathways.1–5 The brain’s white matter is highly organized into fasciculi comprised of densely packed axons. Axonal membranes, myelin, and other structures affect the pattern of Brownian motion of water within white matter.6 Tractography follows a white matter tract from voxel to voxel in three dimensions by assuming that the direction of least restricted diffusion corresponds with the orientation of axons. The quantitative assessments of tissue microstructure central to the scientific applications of diffusion MR are not directly used for presurgical tractography. Instead, the goal of presurgical tractography is simply to identify the position of eloquent pathways, such as the motor, sensory, and language tracts. The shortcomings of DTI fiber
206
Berman cannot be performed until the resection cavity is close to the target pathway.17 Localization with stimulation also is subject to inaccuracies from the penetration of electrical current into surrounding tissues and variations in excitation thresholds.18,19 In the clinical setting, it is advantageous to use more than one brain mapping technique. Independent aspects of physiology are observed with each mapping technique, and their combined results can be used to validate each other and improve the confidence of surgical planning decisions.
ROUTINE PRESURGICAL DIFFUSION TENSOR IMAGING TRACTOGRAPHY PROTOCOL Translating DTI tractography into a useful clinical tool requires efficient integration of the technique into presurgical protocols. This section describes the presurgical DTI fiber-tracking protocol developed and routinely performed at University of California San Francisco. DTI tractography of the motor tract is prescribed for any brain tumor case with a lesion near the motor tract that involves awake or asleep intraoperative cortical stimulation mapping. The protocol involves image acquisition, post processing, quality assurance, and integration with a surgical navigation system. DTI acquisition sequences are readily available on all commercial MR scanner platforms. The following discussion provides several general guidelines for acquiring diffusion MR data suitable for tractography; more detailed specifications can be found in a recent review.20,21 A minimum of six diffusion gradient directions is required to calculate the diffusion tensor; however, significant improvements in data quality can be obtained by increasing the number of directions.22 For tractography, the slices must be contiguous and should not be thicker than 3 mm. At our institution, a typical acquisition at 1.5 T on a GE scanner has TR/TE 5 11 s/78 ms, 3-mm thick slices, matrix 5 128 128, field of view 5 260 260 mm, ASSET factor 5 2, and 15 diffusion directions with two averaged acquisitions at b 5 1000 s/mm2. Total diffusion imaging time is less than 8 minutes. In addition to diffusion MR, high-resolution three-dimensional T1- and T2-weighted anatomic images are acquired for use with the surgical navigation system. The postcontrast three-dimensional T1-weighted spoiled gradient recalled and the three-dimensional T2-weighted fast spin echo volumes have voxel resolution of 1 1 1.5 mm. DTI tractography is performed with multiple regions of interest using an algorithm based on the deterministic fiber assignment by continuous
tracking method.4 Fiber tracks are first generated from a region drawn in the cerebral peduncle with the aid of the b 5 0 s/mm2 echo planar images from the DTI sequence (Fig. 1A) and the postprocessed DTI anisotropy map. The starting region encompasses the entire cerebral peduncle, although the motor tract only occupies a fraction of this region. An overly large starting region reduces the potential for user-introduced error. Tractography algorithms such as the fiber assignment by continuous tracking method require a minimum anisotropy threshold to be specified for propagation of the ‘‘streamline’’ outlining the fiber trajectory. If the fiber track encounters a voxel with anisotropy below the threshold, which is usually caused by the presence of gray matter or cerebrospinal fluid, the streamline is terminated at that location. A relatively low fractional anisotropy threshold of 0.1 is used for presurgical tractography to assist in following white matter tracts through regions of edema or tumor infiltration, which exhibit low diffusion anisotropy.23 Each voxel in the starting region is densely seeded with 27 starting points evenly arranged in a 3 3 3 grid to improve fiber tracking performance.2 Extraneous fiber tracks are removed through the use of multiple target regions of interest (Fig. 1). The fiber tracks are first targeted to a region drawn in the posterior limb of the internal capsule. Then a second target region is drawn around fiber track bundles anterior to the central sulcus within the precentral gyrus. Fiber tracks passing through all three regions are retained as the delineated pyramidal tract (See Fig. 1). The echo planar images are unsuitable for surgical navigation because of poor image contrast and geometric warping artifacts that compromise anatomic fidelity, so the DTI tractography must be co-registered to the highresolution anatomic images. The b 5 0 s/mm2 echo planar images are first aligned to the anatomic images using automated image registration.24,25 A three-dimensional affine 12parameter model with a maximum of 50 iterations is used for registration. The registration’s performance is checked by comparing the position of sulci, ventricles, and cortex. The transformation matrix is applied to the DTI fiber track mask, and the registered mask is fused to the high-resolution anatomic images. Voxels that contain fiber tracks are given intensities 1.5 times higher than the brightest anatomic feature in the image (Fig. 1C, D). This ‘‘hot white’’ intensity ensures that fiber tracks can be distinguished easily from T2-hyperintense lesions and cerebrospinal fluid.
Diffusion MR Tractography
Fig. 1. (A) Diffusion MR tractography of the motor tract (green) is shown overlaid on axial b 5 0 s/mm2 echo planar images. Starting and target regions of interest (red outlines) are shown in the cerebral peduncle, posterior limb of the internal capsule, and precentral gyrus. (B) The resultant fiber tracks (red) are shown adjacent to the lesion (green) in three dimensions. (C) Tractography (green) is registered to the T2-weighted fast spin echo anatomic volume and (D) T1-weighted spoiled gradient recalled volume.
Black and white anatomic images fused with the DTI tractography easily can be made Digital Imaging and Communications in Medicine (DICOM) compliant and recognized by multiple systems, including PACS and surgical navigation systems (Fig. 2). The DICOM standard and newer surgical navigation systems allow tractography to be viewed as a color overlay. The advantage of color overlays is that they can be removed easily if the surgeon desires to view the anatomic images beneath the tractography. Multiple white matter pathways can be represented by different colors and simultaneously viewed. Diffusion MR fiber tracks also can be viewed in relation to the tumor and skull as three-dimensional streamlines, streamtubes, glyphs, and solid volumes.26
TRACTOGRAPHY OF PATHWAYS OTHER THAN THE MOTOR TRACT Most presurgical tractography studies have focused on the motor tract because it is important for patient quality of life and is easy to track. However, tractography of functional pathways other than the cerebral corticospinal tract have
been integrated with a surgical navigation system or studied before surgery. A brainstem cavernous angioma was safely removed with the aid of sensory and motor pathway tractography.27 The optic radiation is a challenging pathway to delineate with tractography because it is a thin ribbon of white matter and turns at a sharp angle through Meyer’s loop (Fig. 3). Studies have shown the feasibility of using tractography to plan radiation therapy and guide surgical resection of arteriovenous malformations.28,29 The anatomy and operation of the language pathways are still under investigation. Tractography has been launched from mouth motor, speech arrest, and anomia stimulation sites on the frontal cortex.30,31 The subcortical language network was observed to include connections to the supplementary motor area, cerebral peduncle, and thalamus. Language tracts including the arcuate fasciculus and the uncinate fasciculus have been delineated with tractography and their position confirmed with stimulation.32,33 In patients undergoing anterior temporal lobectomy in the language dominant hemisphere, preoperative tractography has been used to predict language deficits.34
207
208
Berman
Fig. 2. Motor tractography as viewed during surgery with a surgical navigation system. Axial and coronal views are shown of the T2-weighted anatomic volume fused with the tractography. Voxels through which the fiber tracks pass are assigned bright white intensities. The red crosshairs indicate the location of a mouth motor subcortical stimulation site.
OTHER MAPPING TECHNIQUES USED WITH TRACTOGRAPHY
Fig. 3. Left and right optic radiations are identified with DTI tractography in an epileptic patient before surgery. Fiber tracks in three dimensions (purple) are projected on a b 5 0 s/mm2 axial echo planar slice. The anterior extent of Meyer’s loop is not fully delineated with DTI tractography because of the tract’s tight turn and thin dimensions.
Diffusion MR should never be solely relied on to guide a resection. Because diffusion MR measures a structural property of the brain’s tissue, functional information always must be added to infer the location of eloquent pathways. In most presurgical cases, our knowledge of brain anatomy enables us to draw starting and target regions in the cerebral peduncle and the precentral gyrus. In some brain tumor cases, however, the tumor has shifted or infiltrated the surrounding white matter, changing the geometry of the brain’s anatomy.35,36 These situations necessitate additional functional mapping to guide the placement of tractography starting and target regions. Motor functional MR imaging has been used to locate the motor and language cortex in patients who have brain tumor and define the starting region for diffusion tractography.37–39 The use of functional MR imaging allows detailed functional specificity to be assigned to the generated tractography.40 The subsection of the motor tract or language pathways specific to a particular muscle group or language function can be delineated. There is a disjoint between the functional MR imaging–activated volume in the gray matter and brain tissue with diffusion anisotropy in the white matter, however. The tractography starting region can be drawn manually in the white matter of the gyrus adjacent to the functional MR imaging activation.
Diffusion MR Tractography Validation of DTI tractography is necessary if the technique is to be trusted for surgical planning. Electrical stimulation of cortical and subcortical brain tissue has aided in the preservation of functional brain tissue for many decades.16 Comparison to electrical stimulation has been a necessary step to validate tractography’s use during surgery. Electrical stimulation studies test tractography and its integration with the surgical navigation system in a real clinical situation. Fig. 4 demonstrates the combination of DTI
tractography and cortical stimulation to delineate the portion of the motor tract that controls jaw motions. Diffusion tractography launched from stimulation sites on the primary motor cortex has been observed to follow the expected anatomic course and maintain somatotopic organization through the internal capsule.30 An initial study observed a gap between tractography and subcortical motor stimulation.41 The diffusion tractography was capable of revealing the relative position of
Fig. 4. DTI fiber tracks were launched from a jaw motor cortical stimulation site (blue circle). The DTI tractography (red) represents only the portion of the motor pathway descending from the cortical stimulation site. This case demonstrated the feasibility of using DTI fiber tracking to follow white matter tracts through some T2-hyperintense lesions.
209
210
Berman the tumor and corticospinal tract; however, the borders of the tract were underestimated. Additional subcortical motor stimulation studies have repeatedly observed an approximately 1-cm gap between the electrode tip and tractography.42–44 The gap between DTI tractography and bipolar stimulation represents the cumulative errors from diffusion tensor uncertainty, registration inaccuracies, and brain shift during surgery. Noise in the diffusion MR signal results in uncertainty in the orientation of the diffusion tensor and the position of fiber tracks.45,46 Geometric distortions are common on the echo planar images used to acquire diffusion MR. Care must be taken when
registering echo planar images to the high-resolution anatomic volumes. Numerous studies have examined the effect of brain shift during surgery.47–51 It is possible for the brain to shift either inward or outward, altering the position of tractography as perceived on the surgical navigation system. Bipolar stimulation currents can penetrate several millimeters into surrounding brain tissue.18 Stimulation of neurons distant from the tip of the bipolar electrode always contributes to the approximately 1-cm discrepancy between the stimulation coordinates reported by the navigation system and the tractography.
Fig. 5. Q-ball fiber tracks (green, top row) are compared with DTI fiber tracks (blue, bottom row). The tumor is in orange. The DTI fiber tracks are restricted to the portion of the motor tract originating in the superior-medial aspect of the motor cortex controlling lower extremity functions. The q-ball fiber tracks delineate a larger portion of the motor tract than the DTI fiber tracks. The q-ball fiber tracks show connectivity to the lateral portion of the motor cortex controlling facial motor functions.
Diffusion MR Tractography HIGH ANGULAR RESOLUTION DIFFUSION IMAGING FOR FIBER TRACKING The diffusion tensor model is limited to a single fiber population per voxel, and DTI fiber tracking fails in regions with crossing white matter fibers or intravoxel partial volume averaging of adjacent tracts with different fiber orientations. This shortcoming of DTI can be overcome by sampling many more diffusion-weighted directions than the minimum of six required for DTI and acquiring the images with stronger diffusion weighting than the clinical standard of b 5 1000 s/mm2 for DTI. This technique is known as HARDI, and many different mathematical approaches have been devised recently to reconstruct fiber pathways based on HARDI information. One of the most popular of these is known as q-ball imaging.3,5 The q-ball reconstruction of HARDI data can distinguish distinct fiber populations in regions of complex white matter fiber architecture.3,5 Tractography based on these advanced diffusion MR methods has been shown to traverse
white matter regions with crossing fibers and reveal a greater portion of tracts.52–54 The major limitation of HARDI’s clinical application is the lengthy scan time required for acquiring many diffusion gradient directions. With only 55 gradient directions and parallel imaging at 3 T, however, scan time can be reduced to approximately 13 minutes or less.3,55 A typical acquisition at our institution on a 3 T GE Signa EXCITE scanner has parallel imaging reduction factor of 2, TR/TE 5 12.4 s/73 ms, 2.2 2.2 2.2 isotropic voxels, field of view 5 280 280 mm, and a 128 128 matrix zero-filled to a 256 256 matrix. The 55 diffusion gradient directions are acquired with b 5 2000 s/mm2. Although 13 minutes still may be a bit lengthy for many clinical situations, it can be anticipated that the required scan time will decrease with hardware and HARDI methodologic improvements. Our institution has begun using q-ball tractography for a limited number of patients who have brain tumors.55 As seen in Fig. 5, q-ball fiber tracking is able to delineate the motor tract from the midbrain to the
Fig. 6. The region of complex white matter in the centrum semiovale (pink rectangle) prevents DTI fiber tracks from traversing laterally to the motor cortex. The q-ball orientation density function distinguishes the motor pathway from the superior longitudinal fasciculus (SLF) and enables fiber tracks to course laterally through the centrum semiovale to the motor cortex.
211
212
Berman motor evoked potentials were combined with tractography to preserve motor function. This clinical outcome study and other subcortical stimulation validation studies suggested that stimulation be continually applied when resections are within 1 cm of motor tractography. Tractography can speed up radical resections by indicating which borders of a resection cavity deserve extensive stimulation. Future improvements in MR hardware and HARDI methods will further enable tractography’s clinical use for surgical planning. Protecting subcortical pathways and functional cortices is important during surgery. It is important to understand and account for the limitations of tractography and include multiple mapping techniques when planning surgeries.
ACKNOWLEDGMENTS
Fig. 7. Q-ball tractography (green) of the motor tract is shown in three dimensions in a patient who has a brain tumor. Cortical tongue motor (purple spheres) and subcortical mouth motor (orange sphere) stimulation sites are indicated. A cortical tongue sensory site (blue sphere) is also visible. The q-ball tractography generated presurgically showed connectivity from one cortical motor site, through the subcortical motor site, to the cerebral peduncle.
superior-medial and lateral portions of the motor cortex in our patients. The lateral sections of the motor tract control facial movements and are necessary for language functions. DTI fiber tracking revealed connectivity from the midbrain to only the medial-superior aspect of the motor cortex. The centrum semiovale is a region that contains multiple crossing white matter pathways, and DTI fails to accurately characterize this complex structure (Fig. 6). Initial studies have shown cortical and subcortical stimulations to be in good agreement with the course of the motor tract predicted by q-ball tractography (Fig. 7).55
IMPACT ON PATIENT CARE AND FUTURE DIRECTIONS Many studies have been conducted to improve tractography techniques and assess the accuracy of diffusion tractography. Recently a few studies were conducted to assess the technique’s impact on patient care. Mikuni examined postoperative motor function after tractography-guided surgery in patients who had brain tumors.56,57 Subcortical
I am grateful to Roland Henry, PhD, for his guidance on analyzing the data for many of the figures, and to Pratik Mukherjee, MD, PhD, for his help preparing this article.
REFERENCES 1. Basser PJ, Mattiello J, LeBihan D. Estimation of the effective self-diffusion tensor from the NMR spin echo. J Magn Reson B 1994;103(3):247–54. 2. Conturo TE, Lori NF, Cull TS, et al. Tracking neuronal fiber pathways in the living human brain. Proc Natl Acad Sci USA 1999;96(18):10422–7. 3. Hess CP, Mukherjee P, Han ET, et al. Q-ball reconstruction of multimodal fiber orientations using the spherical harmonic basis. Magn Reson Med 2006; 56(1):104–17. 4. Mori S, Crain BJ, Chacko VP, et al. Three-dimensional tracking of axonal projections in the brain by magnetic resonance imaging. Ann Neurol 1999; 45(2):265–9. 5. Tuch DS. Q-ball imaging. Magn Reson Med 2004; 52(6):1358–72. 6. Beaulieu C. The basis of anisotropic water diffusion in the nervous system – a technical review. NMR Biomed 2002;15(7–8):435–55. 7. Guye M, Parker GJ, Symms M, et al. Combined functional MRI and tractography to demonstrate the connectivity of the human primary motor cortex in vivo. Neuroimage 2003;19(4):1349–60. 8. Krishnan R, Raabe A, Hattingen E, et al. Functional magnetic resonance imaging-integrated neuronavigation: correlation between lesion-to-motor cortex distance and outcome. Neurosurgery 2004;55(4): 904–14 [discusssion 914–5]. 9. Ruff IM, Petrovich Brennan NM, Peck KK, et al. Assessment of the language laterality index in
Diffusion MR Tractography
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
patients with brain tumor using functional MR imaging: effects of thresholding, task selection, and prior surgery. AJNR Am J Neuroradiol 2008; 29(3):528–35. Schulder M, Maldjian JA, Liu WC, et al. Functional image-guided surgery of intracranial tumors located in or near the sensorimotor cortex. J Neurosurg 1998;89(3):412–8. Ganslandt O, Buchfelder M, Hastreiter P, et al. Magnetic source imaging supports clinical decision making in glioma patients. Clinical Neurology and Neurosurgery 2004;107(1):20–6. Nagarajan S, Kirsch H, Lin P, et al. Preoperative localization of hand motor cortex by adaptive spatial filtering of magnetoencephalography data. J Neurosurg 2008;109(2):228–37. Schiffbauer H, Berger MS, Ferrari P, et al. Preoperative magnetic source imaging for brain tumor surgery: a quantitative comparison with intraoperative sensory and motor mapping. J Neurosurg 2002;97(6):1333–42. Engel AK, Moll CK, Fried I, et al. Invasive recordings from the human brain: clinical insights and beyond. Nat Rev Neurosci 2005;6(1):35–47. Ojemann G. Brain organization for language from the perspective of electical stimulation mapping. The Behavioral and Brain Sciences 1983;6:189–230. Penfield W, Rasmussen T. The cerebral cortex of man: a clinical study of localization of function. New York: Macmillan; 1950. xv, 248. Keles GE, Lundin DA, Lamborn KR, et al. Intraoperative subcortical stimulation mapping for hemispherical perirolandic gliomas located within or adjacent to the descending motor pathways: evaluation of morbidity and assessment of functional outcome in 294 patients. J Neurosurg 2004;100(3):369–75. Haglund MM, Ojemann GA, Blasdel GG. Optical imaging of bipolar cortical stimulation. J Neurosurg 1993;78(5):785–93. Pouratian N, Cannestra AF, Bookheimer SY, et al. Variability of intraoperative electrocortical stimulation mapping parameters across and within individuals. J Neurosurg 2004;101(3):458–66. Mukherjee P, Berman JI, Chung SW, et al. Diffusion tensor MR imaging and fiber tractography: theoretic underpinnings. AJNR Am J Neuroradiol 2008;29(4): 632–41. Mukherjee P, Chung SW, Berman JI, et al. Diffusion tensor MR imaging and fiber tractography: technical considerations. AJNR Am J Neuroradiol 2008;29(5): 843–52. Jones DK. The effect of gradient sampling schemes on measures derived from diffusion tensor MRI: a Monte Carlo study. Magn Reson Med 2004;51(4): 807–15. Stadlbauer A, Nimsky C, Buslei R, et al. Diffusion tensor imaging and optimized fiber tracking in
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
glioma patients: Histopathologic evaluation of tumor-invaded white matter structures. Neuroimage 2007;34(3):949–56. Woods RP, Grafton ST, Holmes CJ, et al. Automated image registration: I. General methods and intrasubject, intramodality validation. Journal of Computer Assisted Tomography 1998;22(1):139–52. Woods RP, Grafton ST, Watson JDG, et al. Automated image registration: II. Intersubject validation of linear and nonlinear models. Journal of Computer Assisted Tomography 1998;22(1):153–65. Nimsky C, Ganslandt O, Enders F, et al. Visualization Strategies for Major White Matter Tracts for Intraoperative Use. International Journal of Computer Assisted Radiology and Surgery 2006;1:13–22. Chen X, Weigel D, Ganslandt O, et al. Diffusion tensor-based fiber tracking and intraoperative neuronavigation for the resection of a brainstem cavernous angioma. Surg Neurol 2007;68(3): 285–91 [discussion 291]. Kikuta K, Takagi Y, Nozaki K, et al. Early experience with 3-T magnetic resonance tractography in the surgery of cerebral arteriovenous malformations in and around the visual pathway. Neurosurgery 2006;58(2):331–7 [discussion 331–7]. Maruyama K, Kamada K, Shin M, et al. Optic radiation tractography integrated into simulated treatment planning for Gamma Knife surgery. J Neurosurg 2007;107(4):721–6. Berman JI, Berger MS, Mukherjee P, et al. Diffusiontensor imaging-guided tracking of fibers of the pyramidal tract combined with intraoperative cortical stimulation mapping in patients with gliomas. J Neurosurg 2004;101(1):66–72. Henry RG, Berman JI, Nagarajan SS, et al. Subcortical pathways serving cortical language sites: initial experience with diffusion tensor imaging fiber tracking combined with intraoperative language mapping. Neuroimage 2004;21(2):616–22. Bello L, Gambini A, Castellano A, et al. Motor and language DTI Fiber Tracking combined with intraoperative subcortical mapping for surgical removal of gliomas. Neuroimage 2008;39(1):369–82. Kamada K, Todo T, Masutani Y, et al. Visualization of the frontotemporal language fibers by tractography combined with functional magnetic resonance imaging and magnetoencephalography. J Neurosurg 2007;106(1):90–8. Powell HW, Parker GJ, Alexander DC, et al. Imaging language pathways predicts postoperative naming deficits. J Neurol Neurosurg Psychiatry 2008;79(3): 327–30. Fernandez-Miranda JC, Rhoton AL Jr, AlvarezLinera J, et al. Three-dimensional microsurgical and tractographic anatomy of the white matter of the human brain. Neurosurgery 2008;62(6 Suppl. 3): 989–1026 [discussion 1026–8].
213
214
Berman 36. Wieshmann UC, Symms MR, Parker GJ, et al. Diffusion tensor imaging demonstrates deviation of fibres in normal appearing white matter adjacent to a brain tumour. J Neurol Neurosurg Psychiatry 2000;68(4): 501–3. 37. Hendler T, Pianka P, Sigal M, et al. Delineating gray and white matter involvement in brain lesions: threedimensional alignment of functional magnetic resonance and diffusion-tensor imaging. J Neurosurg 2003;99(6):1018–27. 38. Holodny AI, Ollenschleger MD, Liu WC, et al. Identification of the corticospinal tracts achieved using blood-oxygen-level-dependent and diffusion functional MR imaging in patients with brain tumors. AJNR Am J Neuroradiol 2001;22(1):83–8. 39. Schonberg T, Pianka P, Hendler T, et al. Characterization of displaced white matter by brain tumors using combined DTI and fMRI. Neuroimage 2006; 30(4):1100–11. 40. Smits M, Vernooij MW, Wielopolski PA, et al. Incorporating functional MR imaging into diffusion tensor tractography in the preoperative assessment of the corticospinal tract in patients with brain tumors. AJNR Am J Neuroradiol 2007;28(7):1354–61. 41. Kinoshita M, Yamada K, Hashimoto N, et al. Fibertracking does not accurately estimate size of fiber bundle in pathological condition: initial neurosurgical experience using neuronavigation and subcortical white matter stimulation. Neuroimage 2005; 25(2):424–9. 42. Berman JI, Berger MS, Chung SW, et al. Accuracy of diffusion tensor magnetic resonance imaging tractography assessed using intraoperative subcortical stimulation mapping and magnetic source imaging. J Neurosurg 2007;107(3):488–94. 43. Kamada K, Todo T, Masutani Y, et al. Combined use of tractography-integrated functional neuronavigation and direct fiber stimulation. J Neurosurg 2005; 102(4):664–72. 44. Mikuni N, Okada T, Nishida N, et al. Comparison between motor evoked potential recording and fiber tracking for estimating pyramidal tracts near brain tumors. J Neurosurg 2007;106(1):128–33. 45. Anderson AW. Theoretical analysis of the effects of noise on diffusion tensor imaging. Magn Reson Med 2001;46(6):1174–88. 46. Lazar M, Alexander AL. An error analysis of white matter tractography methods: synthetic diffusion tensor field simulations. Neuroimage 2003;20(2): 1140–53.
47. Coenen VA, Krings T, Weidemann J, et al. Sequential visualization of brain and fiber tract deformation during intracranial surgery with three-dimensional ultrasound: an approach to evaluate the effect of brain shift. Neurosurgery 2005;56(1 Suppl):133–41 [discussion 133–41]. 48. Keles GE, Lamborn KR, Berger MS. Coregistration accuracy and detection of brain shift using intraoperative sononavigation during resection of hemispheric tumors. Neurosurgery 2003;53(3):556–62 [discussion 562–4]. 49. Nimsky C, Ganslandt O, Hastreiter P, et al. Preoperative and intraoperative diffusion tensor imagingbased fiber tracking in glioma surgery. Neurosurgery 2005;56(1):130–7. 50. Reinges MHT, Nguyen HH, Krings T, et al. Course of brain shift during microsurgical resection of supratentorial cerebral lesions: limits of conventional neuronavigation. Acta Neurochirurgica 2004;146(4): 369–77. 51. Roberts DW, Hartov A, Kennedy FE, et al. Intraoperative brain shift and deformation: a quantitative analysis of cortical displacement in 28 cases. Neurosurgery 1998;43(4):749–58 [discussion 758–60]. 52. Berman JI, Chung S, Mukherjee P, et al. Probabilistic streamline Q-Ball tractography using the residual bootstrap. Neuroimage 2008;39:215–22. 53. Campbell JS, Siddiqi K, Rymar VV, et al. Flow-based fiber tracking with diffusion tensor and q-ball data: validation and comparison to principal diffusion direction techniques. Neuroimage 2005;27(4): 725–36. 54. Wedeen VJ, Wang RP, Schmahmann JD, et al. Diffusion spectrum magnetic resonance imaging (DSI) tractography of crossing fibers. Neuroimage 2008; 41(4):1267–77. 55. Berman JI, Clark DJ, Berger MS, et al. Improved Diffusion MR Fiber Tracking for Neurosurgical Application. ISMRM 2008; Toronto. 56. Mikuni N, Okada T, Enatsu R, et al. Clinical impact of integrated functional neuronavigation and subcortical electrical stimulation to preserve motor function during resection of brain tumors. J Neurosurg 2007; 106(4):593–8. 57. Mikuni N, Okada T, Enatsu R, et al. Clinical significance of preoperative fibre-tracking to preserve the affected pyramidal tracts during resection of brain tumours in patients with preoperative motor weakness. J Neurol Neurosurg Psychiatry 2007; 78(7):716–21.