Presurgical Prediction guage, and other neurologic functions. In pediatric epilepsysurgery, this information about localization is essential, butoften difficult to acquire, because ambiguities often arise

lateral superior quadrantanopsia loss after temporal lobe

resections [5]. In other studies, tractography was used as

3901 Beaubien Boulevard; Detroit, MI 48201.E-mail: hchugani@pet.wayne.eduof Motor Functional

Loss Using

TractographyRajkumar Munian Govindan, MD*,Harry T. Chugani, MD*, Aimee F. Luat, MD*,and Sandeep Sood, MD

The usefulness of magnetic resonance imaging tractog-raphy is demonstrated in the presurgical planning of an8-year-old girl with intractable epilepsy. Imaging and in-tracranial electrode monitoring suggested a left hemi-spherectomy for complete control of her seizures.Although this child was hemiplegic, she retained consid-erable motor function in her right hand, and her parentsand the epilepsy team voiced significant concern that shewould lose right-hand function after a hemispherectomy.Tractography indicated near-complete absence of herleft corticospinal tract and a more robust than normalcorticospinal tract in the right hemisphere. This findingsuggested that her right motor function had reorganizedto the right hemisphere and the ipsilateral corticospinaltract. After surgery, her seizures were completely con-trolled, and no change in right motor activity was evidentcompared with her presurgical status. Tractographyhelped determine the extent of cortical resection and pre-dict the extent of motor functional loss. 2010 byElsevier Inc. All rights reserved.

Govindan RM, Chugani HT, Luat AF, Sood S. Presurgical

prediction of motor functional loss using tractography. Pe-

diatr Neurol 2010;43:70-72.

Introduction

The accurate localization of cerebral cortical regions serv-

ing crucial neurologic functions is important in neurosurgical

practice for the preservation of vital motor, sensory, lan-

From the *Department of Pediatrics and Neurology, and Department ofNeurosurgery, Childrens Hospital of Michigan, Wayne State University,Detroit, Michigan.70 PEDIATRIC NEUROLOGY Vol. 43 No. 1a navigational tool in neurosurgical procedures [6,7].

Here, we demonstrate the usefulness of tractography in

planning the extent of cortical resection and in predicting

motor functional loss in a child with intractable epilepsy.

Case Report

An 8-year-old girl, born at 38 weeks of gestation, was diagnosed with

hemiplegic cerebral palsy at age 1 year, and she had manifested complex

partial seizures since age 2 years. Her seizures were poorly controlled de-

spite the use of multiple antiepileptic medications. Her seizure semiology

generally consisted of an aura, staring episodes, and occasional jerking of

the right limbs, followed by vomiting. Her partial status epilepticus occa-

sionally lasted up to 2 hours. Scalp electroencephalography revealed slow

background activity in the left hemisphere, with spike-and-wave activity in

the left posterior quadrant. She had right hemiparesis, bilateral spasticity,

and cognitive delay. Right peripheral visual-field loss was suspected but

was difficult to evaluate formally during her examination. Facial move-

ments were symmetric, with no weakness. Her magnetic resonance imag-

ing scan at age 6 years suggested in utero vascular injury with decreased

white matter volume in the left centrum semiovale and peritrigonal areas,

and an increased signal consistent with gliosis. The corpus callosum was

thin in the posterior segment. A glucose metabolism positron emission to-

mography scan indicated decreased metabolism in the left parietal and tem-

poral cortices. The left basal ganglia and thalamus also exhibited severe

hypometabolism. The right hemisphere appeared normal.

Communications should be addressed to:Dr. Chugani; Department of Pediatrics and Neurology; Childrens Hospitalof Michigan; School of Medicine, Wayne State University;concerning the extent and localization of the epileptogenic

zone, in addition to the huge variation in the cortical location

of neurologic functions attributable to malformations or dys-

plasia, often leading to reorganization. For such localizations

of neurologic functions in the human brain, many investiga-

tive techniques, such as functionalmagnetic resonance imag-

ing [1,2] and magnetoencephography [3], have been used.

However, these methods contain some limitations because

they require active subject participation during the test,

which is especially difficult in children and neurologically

impaired patients.

Diffusion magnetic resonance imaging and tractography

provide an advantage in this respect, i.e., they can provide

valuable functional localization noninvasively, without ac-

tive patient participation. This technique was used to iden-

tify the normal course of the corticospinal tract in normal,

healthy children [4]. In one clinical application, this tech-

nique helped identify the optic radiation and predict contra-Received September 4, 2009; accepted February 22, 2010.

2010 by Elsevier Inc. All rights reserved.doi:10.1016/j.pediatrneurol.2010.02.004 0887-8994/$see front matter

ractabickerg the b.Based on these clinical and imaging data, two-stage epilepsy surgery

was proposed, with the placement of intracranial cortical surface electrodes

for seizure localization and cortical mapping. Intracranial electrodes cov-

ered most of her left occipital, parietal, temporal, and frontal lobes, with

additional electrodes in the subtemporal regions and interhemisphere elec-

trodes in the parietal region. Intracranial monitoring indicated very fre-

quent interictal epileptic activity in the left occipital, frontal, and parietal

regions, including the precentral and postcentral gyri. These activities

were frequent enough to resemble electrographic seizures, suggesting

a multiple epileptogenesis involving the entire left hemisphere, including

the primary motor cortex.

Figure 1. (A, C) Right and left corticospinal tracts (CST) of child with inthemisphere, on visual evaluation, the isolated corticospinal tract (A) was thtrol subject (B). In the left hemisphere, no significant fiber bundle connectintractographic fiber bundle that failed to reach the sensorimotor cortex (C)Furthermore, using these electrodes, cortical motor and somatosensory

functional mapping was performed via repeated electrical stimulation of

the intracranial electrodes over the precentral sulcus and right tibial nerve, re-

spectively. Both these methods failed to elicit any motor or somatosensory

response in the left cortical hemisphere covered by the intracranial electrodes.

We performed tractography of the corticospinal tract, using diffusion

tensor imaging. The diffusion tensor imaging scan was acquired using

a General Electric (General Electric Company, Fairfield, CT) system

with a 3 T magnet. Diffusion tensor images were acquired in the axial

planes, using six diffusion-sensitized gradients in six noncollinear direc-

tions, each with a b-value of 1000 second/mm2, along with a T2-weighted

reference image with a b-value of zero. Each image volume was acquired

using six repetitions to increase image quality (i.e., to reduce geometric dis-

tortion and eddy current artifacts) and the signal-to-noise ratio in a matrix

(size, 128 128 41 axial slices), and the images were reconstructed intoa 256 256 matrix. Tensor calculation and tractography were performedusing DTI-Studio software, version 2.40 [8]. Tractography was performed

according to the fiber assignment by continuous tracking algorithm

(deterministic) [9], with fiber propagation starting at a fractional anisotropy

threshold value of 0.2, and with fiber propagation discontinued if the

fractional anisotropy value was 60.To isolate the corticospinal tract on the contralateral right side, an initial

wide-seed region-of-interest with an OR operator was drawn on the pos-

terior limb of the internal capsule on the transaxial slice. A second region of

interest with an AND operator was drawn on a transaxial slice at the

anterior aspect of the brainstem, just below the level of cerebellar efferent

fiber decussation. Afterward, multiple small regions of interest witha NOT function were drawn, to exclude fibers going to other nonsoma-

tosensory cortical regions. This procedure yielded a long corticospinal

tract, descending from the right precentral and postcentral cortex, and pass-

ing through the posterior limb of the internal capsule and anterior part of the

brainstem (Fig 1A). On visual evaluation, this tract was thicker compared

with its appearance in an age-matched, left-handed, normal control subject

(Fig 1B). On the ipsilateral (surgical) left side, an exhaustive search for the

corticospinal tract was performed by placing large-seed OR regions of

interest at several locations, i.e., the subcortical white matter underlying

the sensory motor cortex, the posterior limb of the internal capsule, and

the brainstem. Despite three different seed regions of interest placed along

le epilepsy. (B, D) An age-matched, left-handed healthy child. In the rightcompared with its appearance in an age-matched, left-handed normal con-rainstem to the sensorimotor cortex was evident, except for a single, smallthe expected course of the corticospinal tract, no significant fiber bundle

connecting the brainstem to the sensorimotor cortex was evident, except

for a single, small tractographic fiber bundle that failed to reach the senso-

rimotor cortex (Fig 1C).

Discussion

In this child, based on intracranial electrode recordings,

a left hemispherectomy was indicated to achieve seizure

control. Although this child was hemiplegic, she exhibited

considerable motor function in her right hand, and could

hold and drink from a cup. Therefore, her parents and the

epilepsy team expressed significant concern that she would

lose right-hand function after a hemispherectomy.

Because this child had manifested in utero injury, we

considered the possibility that right motor function had

reorganized in the opposite hemisphere. Such motor reorga-

nization from an ipsilateral corticospinal tract had been

demonstrated in a number of studies, using various tech-

niques [10-12]. If such reorganization had occurred, this

child had the potential to achieve a seizure-free outcome

after hemispherectomy, without any worsening of her right

hemiparesis.

Govindan et al: Tractography in Presurgical Planning 71

Although functional motor cortical mapping, performed

using repetitive electrical stimulation of the intracranial

electrodes on the left precentral gyrus, failed to elicit any

right motor activity, this test was not sufficiently conclusive

to exclude an absence of motor activity in the precentral

cortex (and other electrode-covered areas) because of

a lack of confidence about the accurate placement of intra-

cranial electrodes on the corresponding cortex, and also

because of variations in the stimulus threshold required to

elicit a motor response. Similarly, somatosensory cortical

mapping failed to produce a conclusive sensory functional

localization in the left hemisphere. Under these circum-

medial precentral regions were not visualized. This limita-

tion was based on the presence of a large corona radiata

and corpus callosum in the lateral and medial aspects of

the corticospinal tract, acting as a huge diffusion barrier im-

peding fiber propagation. Complex multiple-fiber model

tractographic methods may overcome this limitation, but

these methods are still in developmental stages. Further-

more, in many disease conditions, functionally active axons

often traverse the edematous and gliotic tissue regions, and

these axons may be difficult to isolate even with the most

robust imaging or tractographic techniques. Therefore, for

present clinical use, tractography must be applied with

of motor function in patients with pontine infarct. Neurorehabilitation

2006;21:233-7.stances, a robust motor mapping method such as functional

magnetic resonance imaging would have been useful for

a precise localization of the right sensory motor cortex.

However, this procedure requires a level of patient cooper-

ation that the child could not provide. Therefore, the near-

complete absence of a tractographic corticospinal tract in

the left hemisphere and a normal, if not more robust, corti-

cospinal tract in the right hemisphere reinforced our confi-

dence that motor function in the left hemisphere was

probably absent, suggesting that right motor function had

reorganized to the right hemisphere. Thus, tractography

exerted an impact on our clinical decision, and allowed us

to extend the surgical resection to involve the entire left

hemisphere, including the primary motor and sensory cor-

tex, with the presumption that the child would not undergo

any further right motor functional loss. Indeed, after sur-

gery, the seizures were completely controlled, and no

change in the strength and tone of the right upper and lower

extremities were evident compared with her presurgical

status.

The limitations of diffusion tensor tractography should

be mentioned. Tractography alone is unreliable in terms

of suggesting the presence or absence of the corticospinal

or any other white matter tract. This unreliability is inevita-

ble because diffusion magnetic resonance imaging is inher-

ently low in spatial resolution. Diffusion values were

measured in a range of millimeters, whereas the actual

size of white matter axons are in a range of micrometers

or even less. To overcome this issue, at least partly,

higher-resolution prolonged image-acquisition protocols

are needed, but these protocols are very time-consuming

and difficult to apply in clinical practice. Moreover, our

tractographic method (deterministic) only involved cortico-

spinal tract segments mostly arising from the superior part

of the precentral gyrus. Fibers arising from the lateral and72 PEDIATRIC NEUROLOGY Vol. 43 No. 1[12] Vandermeeren Y, Davare M, Duque J, Olivier E. Reorganiza-

tion of cortical hand representation in congenital hemiplegia. Eur J Neuro-

sci 2009;29:845-54.some caution. Nevertheless, in selected cases, it can provide

important clinical information that affects patient manage-

ment.

References

[1] Bookheimer S. Pre-surgical language mapping with functional

magnetic resonance imaging. Neuropsychol Rev 2007;17:145-55.

[2] Toma K, Nakai T. Functional MRI in human motor control studies

and clinical applications. Magn Reson Med Sci 2002;1:109-20.

[3] Tovar-Spinoza ZS, Ochi A, Rutka JT, Go C, Otsubo H. The role

of magnetoencephalography in epilepsy surgery. Neurosurg Focus 2008;

25. E16.

[4] Kumar A, Juhasz C, Asano E, et al. Diffusion tensor imaging

study of the cortical origin and course of the corticospinal tract in healthy

children. AJNR 2009;30:1963-70.

[5] 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:

331-7.

[6] Romano A, DAndrea G, Minniti G, et al. Pre-surgical planning

and MR-tractography utility in brain tumour resection. Eur Radiol 2009

Jun 16 (Epub ahead of print).

[7] Berman J. Diffusion MR tractography as a tool for surgical plan-

ning. Magn Reson Imag Clin North Am 2009;17:205-14.

[8] Jiang H, van Zijl PC, Kim J, Pearlson GD, Mori S. Dtistudio: Re-

source program for diffusion tensor computation and fiber bundle tracking.

Comput Methods Programs Biomed 2006;81:106-16.

[9] Mori S, Crain BJ, Chacko VP, van Zijl PC. Three-dimensional

tracking of axonal projections in the brain by magnetic resonance imaging.

Ann Neurol 1999;45:265-9.

[10] Kim YH, Jang SH, Han BS, et al. Ipsilateral motor pathway con-

firmed by diffusion tensor tractography in a patient with schizencephaly.

Neuroreport 2004;15:1899-902.

[11] Ahn YH, You SH, Randolph M, et al. Peri-infarct reorganization

Presurgical Prediction of Motor Functional Loss Using TractographyIntroductionCase ReportDiscussionReferences