Pre-Surgical Planning: Multimodality Imaging to Optimize Outcomes in Pediatric Epilepsy Surgery

Abstract

Neuroimaging is an important component of the pre-surgical planning for pediatric epilepsy. High-resolution structural magnetic resonance images are combined with advanced structural and functional imaging techniques to better define the surgical lesion and decrease morbidity postoperatively. The combination of neuroimaging, electroencephalography (EEG), and neuropsychiatric testing in a multidisciplinary epilepsy conference setting is essential for determining a plan for surgical management.

Magnetic Resonance Imaging

The foundational imaging modality for the diagnostic workup of epilepsy is magnetic resonance imaging (MRI). A dedicated epilepsy protocol is recommended which is generally tailored towards a higher-resolution assessment of the cerebral cortex and associated signal abnormalities. Imaging resolution of the brain can be optimized with the use both a higher field strength 3T magnet and a multichannel head coil such as 32 or 64 channel.¹ A whole brain isotropic T1 weighted sequence in generally the mainstay for providing structural or anatomic detail of the brain (Figure 1A). The detection of signal abnormalities in the brain may also be increased with the use of a thin section T2 axial sequence (Figure 1B) or an isotropic fluid attenuated inversion recovery (FLAIR) sequence. Isotropic sequences are an essential part of the imaging protocol because they are commonly used for surgical navigation. The sequences are also useful for co-registration with other structural and functional imaging techniques. Because of the high propensity of mesial temporal sclerosis in childhood epilepsy, dedicated thin section coronal T2 and T2 FLAIR sequences angled perpendicular to the hippocampi are very useful sequences as this orientation provides improved visualization of the hippocampal morphology. Susceptibility weighted imaging (SWI) is a useful sequence since it is sensitive for the detection of hemosiderin and calcification. Diffusion tensor imaging (DTI) is also useful for both the detection of diffusion signal abnormalities and for the post processing of white matter fiber tracts.

Figure 1.

Axial isotropic T1 (A) sequence demonstrating abnormal thickened cortex in the left frontoparietal region (arrow). Axial isotropic T2 (B) sequence showing abnormal T2 signal extending from the thickened cortex to deep white matter (arrow).

Diffusion Tensor Imaging

Diffusion Tensor Imaging (DTI) is an MRI sequence that assesses both the magnitude and direction of the microscopic motion of water molecules. The directional information of white matter tracts obtained through DTI allows for the three-dimensional (3-D) reconstruction of fiber tracts. Through post-processing, 3-D fiber tracts can be reconstructed and fused to the structural isotropic T1 sequence for more precise anatomic localization (Figure 2). White matter structural information can be obtained that is not visible on traditional sequences. The proximity of an epileptogenic focus to the fiber tract can be delineated for surgical planning and to help predict morbidity. For example, fiber tracking of the optic radiations in patients with anterior lobe epilepsy has been show to assist with the prediction of postoperative visual field deficits. ² Additional parameters which may be evaluated with DTI include white matter fractional anisotropy (FA) and mean diffusivity (MD). FA and MD measurements have been shown to correlate with memory and language impairments in temporal lobe epilepsy which may help predict neuropsychological status in patients with epilepsy. ³

Figure 2.

Tractography fused to axial isotropic T2 sequence demonstrating proximity of the corticospinal tracts (color) to the thickened cortex and deep white matter lesion (arrow).

Functional MRI

Functional MRI (fMRI) is a noninvasive means to help identify eloquent cortex for pre-surgical planning. The most commonly used technique for fMRI is blood oxygen level dependent (BOLD) imaging. The principle of this technique is that increased cerebral blood flow corresponds to areas of increased neuronal activity. Areas of increased neuronal activity are generated during fMRI through a stimulus delivery or paradigm. Commonly used paradigms for pre-surgical planning include motor, sensory, speech, and visual (Figure 3A, 3B. The implementation of the paradigms in the pediatric setting can be challenging since the patient must not only be awake and cooperative during the exam but also relatively motionless. Stimulus delivery may be accomplished through specialized equipment including DVD goggles or flat screen television with a mirror mounted on the MRI head coil. Depending on the paradigm, patient responses to the stimuli may also be recorded with MRI compatible response pads. The BOLD imaging sequence is very susceptible to patient motion and dental hardware so proper patient screening prior to the exam is essential. One of the more common uses for fMRI for epilepsy pre-surgical planning is language lateralization in which fMRI has be shown to help predict post-operative language deficits.⁴

Figure 3.

fMRI language paradigm fused to axial isotropic T2 sequence demonstrating left hemisphere language dominance with activations corresponding to Broca’s (A) and Wernicke’s (B) areas.

A more traditional radiologic examination for language lateralization is the Intracarotid Sodium Amybarbital Procedure or Wada Test. During a Wada test, a femoral arterial puncture is performed and after an abbreviated cerebral angiogram, intra-arterial sodium amybarbital is administered via each internal carotid artery. Both language and memory testing is usually performed by a neurologist and/or neuropsychologist during the Wada test in which a cooperative patient is given various tasks. In the pediatric setting, it may be difficult to perform a Wada test considering the more invasive nature of the exam and the need for a cooperative patient. Functional MRI may be the best option in this context considering how various studies have shown good correlation for language lateralization between the more invasive Wada test and fMRI.⁵ Functional MRI has also been shown to reliably delineate the proximity of eloquent motor cortex to a surgical lesion and therefore improve surgical outcome. ⁶ This may mitigate the need for more invasive intraoperative cortical stimulation.

Fluorodeoxyglucose Positron Emission Tomography

Fluorodeoxyglucose positron emission tomography (FDG-PET) is utilized to assess glucose uptake in the brain. The brain normally demonstrates avid FDG uptake as normal metabolically active neuronal cells readily utilize glucose. The FDG isotope is injected intravenously and after an approximate thirty-minute uptake time, imaging is performed on the PET scanner. If sedation is necessary for the examination, it is preferentially administered after the isotope uptake is completed as sedation can affect glucose uptake in the brain. The patient is monitored with EEG during the isotope uptake and scan to monitor for the presence or absence of seizure activity. In the absence of seizure activity, a seizure focus tends to be hypometabolic which manifests as decreased FDG uptake. One context in which FDG-PET has been proven useful is the diagnostic evaluation of seizures in the setting of a normal brain MRI. An FDG-PET directed “second read” of an MRI has been proven helpful in the detection of seizure foci such as subtle cortical malformations. ⁷ Detection is only further improved by post-processing techniques whereby the structural MRI is fused to the FDG-PET scan, providing better correlation of the abnormal structure on MRI to the abnormal function on the FDG-PET scan (Figure 4A).

Figure 4.

FDG-PET (A) fused to axial isotropic T1 sequence demonstrating an area of hypometabolic activity in the left frontoparietal region corresponding to abnormal cortical thickening. *

SPECT images fused to axial isotropic T1 sequence: interictal (B), ictal (C), and subtracted (D) images highlight an area of increased perfusion in the left frontoparietal region. *

* Postprocessed PET and SPECT images are courtesy of Dr. Lalit Bansal

Single Photon Emission Computed Tomography

Single Photon Emission Computed Tomography (SPECT) is an imaging technique utilized to assess brain perfusion. A radioisotope which can cross the blood brain barrier is injected intravenously. The distribution of this isotope in the brain represents brain perfusion and correspondingly neuronal activity. Both ictal and interictal imaging is typically performed with this technique. For ictal imaging, the patient is admitted to the epilepsy monitoring unit and taken off seizure medications. Over a three day admission, the patient is actively monitored by the bedside and with EEG. If a seizure event occurs the radioisotope, which is ready for injection at the bedside, is injected within as close proximity to the seizure event as possible. During ictal isotope uptake, a seizure focus tends to be hyperperfused. Interictal imaging is also performed during this admission. Ictal (Figure 4B) and interictal (Figure 4C) images may be post processed to produce a subtracted image series (Figure 4D) which may further highlight an area of abnormal brain perfusion. The use of this ictalinterictal subtraction technique with fusion to MRI has been proven to assist in the detection of a seizure focus.⁸

Multidisciplinary Epilepsy Conference

A multidisciplinary epilepsy conference is the ideal setting to integrate information from the patient’s diagnostic radiology workup with information obtained from the remainder of the patient’s clinical assessment. ⁹ The information gained from the structural and functional diagnostic imaging techniques discussed above are of even greater value when considered in the context of the patent’s clinical evaluation. Clinical information presented from the epileptologists regarding seizure semiology and from neuropsychiatric testing may be effectively integrated with diagnostic studies through this multidisciplinary team setting into a cohesive management plan. The presentation of the information in this cohesive manner assists with further neurosurgical preoperative planning with the overall goal of optimizing the patient’s postoperative outcome.

Biography

Timothy P. Zinkus, MD, is Assistant Professor, Pediatric Radiology, Children’s Mercy, University of Missouri, Kansas City, Kansas City, Mo.Contact: tpzinkus@cmh.edu