Role of magnetoencephalography and stereo-electroencephalography in the presurgical evaluation in patients with drug-resistant epilepsy
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.201680
Source of Support: None, Conflict of Interest: None
In selected patients with drug-resistant focal epilepsies (DRFE), who otherwise are likely to be excluded from epilepsy surgery (ES) because of the absence of a magnetic resonance imaging (MRI)-demonstrable lesion or discordant anatomo-electro-clinical (AEC) data, magnetoencephalography (MEG) may help to generate an AEC hypothesis and stereo-electroencephalography (SEEG) may help to verify the hypothesis and proceed with ES. The sensitivity of MEG is much better in localizing the spiking zone in relation to lateral temporal and extratemporal cortical regions compared to the mesial temporal structures. MEG has a dominant role in the presurgical evaluation of patients with MRI-negative DRFEs, insular epilepsies, and recurrent seizures after failed epilepsy surgeries, and in guiding placement of invasive electrodes. Moreover, postoperative seizure freedom is better if MEG spike source localized cortical region is included in the resection. When compared to subdural grid electrode recording, SEEG is less invasive and safer. Those who are otherwise destined to suffer from uncontrolled seizures and their consequences, SEEG guided ES is a worthwhile and a cost-effective option. Depending on the substrate pathology, there is > 80-90% chance of undergoing ES and 60-80% chance of becoming seizure-free following SEEG. Recent noninvasive techniques aimed at better structural imaging, delineating brain connectivity and recording specific intracerebral EEG patterns such as high frequency oscillations might decrease the need for SEEG; but more importantly, make SEEG exploration more goal-directed and hypothesis-driven.
Keywords: Drug-resistant epilepsy, epilepsy surgery, invasive electroencephalography evaluation, magnetoencephalography, magnetic source imaging, presurgical evaluation, source localization, stereo-electroencephalography
In spite of therapy with appropriate antiepileptic drugs (AEDs), nearly one-third of newly diagnosed patients with epilepsy (PWE) continue to have chronic recurrent epileptic seizures., Availability of several new AEDs in recent years has not helped to reduce the burden of patients with drug-resistant epilepsy (DRE). In India, out of the nearly 6.5 million PWE, there would at least be one million with DRE. In selected patients with drug-resistant focal epilepsies (DRFE), surgical treatment has the potential to eliminate seizures and enhance the quality of life.
The objective of epilepsy surgery (ES) is to make the patient seizure-free without producing any disabling neurological or cognitive deficits. The presurgical evaluation, aimed to identify ideal PWE for ES, is a multimodality approach to generate an accurate anatomo-electro-clinical (AEC) hypothesis with respect to the cortical area responsible for generating the seizures (epileptogenic zone, EZ), its connections (epileptogenic networks, EN), and its relationship to eloquent cortical areas. The success of ES depends upon adequate and safe resection and/or disconnection of the EZ/EN.
In patients with an magnetic resonance imaging (MRI)-identified potentially epileptogenic lesion (such as hippocampal sclerosis, focal cortical dysplasia or low-grade neoplasm), in a majority, the non-invasive investigations such as seizure semiology, interictal and ictal EEG data obtained by long-term video-EEG (VEEG) monitoring, and neuropsychological assessment can provide an electro-clinical (EC) hypothesis concordant with the lesion sufficient enough to proceed with ES. Additional non-invasive investigations such as positron emission tomography (PET) and ictal single photon emission tomography (SPECT) may help to refine the hypothesis in those with EC discordance. However, nearly one-third to one-half of patients with DRFE do not have a MRI-demonstrable lesion to generate an AEC hypothesis., Others may have multiple lesions or EC, PET, and SPECT data that are discordant to the suspected epileptogenic substrate or overlapping with eloquent cortex. These patients often require invasive EEG evaluation.
The common indications for invasive evaluation are provided in [Table 1]. In summary, invasive evaluation becomes necessary due to the inability to generate a convincing AEC hypothesis either because of the absent/imprecise A (anatomical substrate) or multiple As, or discordance between A and EC. A majority of the MRI-negative patients with DRFE are denied ES in developing countries because of the absence of advanced technology and expertise to evolve an AEC hypothesis., Out of 285 patients with drug-resistant extratemporal focal epilepsies, who underwent presurgical evaluation in a leading ES center in India, surgery was deferred in 214 (75%) because of the inability to generate an AEC hypothesis. In such patients, magnetoencephalography (MEG) and stereo-electroencephalography (SEEG) are investigative methods that can be applied selectively to define the EZ and EN. We wish to describe, based on our experience from two leading ES centers in India, the indications, methodology, interpretation, utility and limitations of MEG and SEEG. Detailed technical descriptions of MEG and SEEG are beyond the scope of this review; interested readers may refer to recent comprehensive articles on MEG , and SEEG.,
The MEG noninvasively measures the magnetic field generated by intracellular neuronal currents.
As neuronal magnetic fields are considerably weaker (in the order of femto-Tesla to few pico-Tesla) than the earth's magnetic field (40–50 micro-Tesla), the measurement is carried inside a magnetically shielded room using highly unparalleled sensitive SQUID sensors (superconducting quantum interference device; is a very sensitive magnetometer used to measure extremely subtle magnetic fields) which are immersed in liquid helium., In [Table 2], we have compared the properties of MEG and EEG. Source modelling methods enable localisation of the measured MEG signal, including the traditionally used equivalent current dipole (ECD) models. Some of the more common algorithms employed for source localization include minimum-norm estimate (MNE), standardized low-resolution brain electromagnetic tomography (sLORETA), dynamic statistical parametric mapping (dSPM), and beam forming. Each method differs with the other in their assumptions about the cortical generators. The superimposition of sources localized by MEG on to MRI is referred to as magnetic source imaging (MSI). Although several studies have been attempted to compare the role of MEG/MSI with the existing standard presurgical noninvasive and invasive modalities, such comparisons are unrealistic as each of the investigative techniques provides a different set of information which varies from patient to patient, and patients are selected for each of these investigations based on the complexity of their epileptogenic substrate.
MEG in lesional and nonlesional focal epilepsies
In a patient with DRFE, MEG is usually undertaken when the AEC hypothesis has remained elusive after careful analysis and interpretation of the non-invasive presurgical evaluation findings obtained by MRI, VEEG, PET or SPECT., In such a situation, with respect to MEG, two questions are usually asked: (1) Can MEG delineate the irritative (spiking) zone better than the one provided by EEG?; and (2) Will the information obtained by MEG provide additional help in planning an invasive electrode implantation?
In patients with temporal lobe epilepsies (TLE), the value of MEG source localization depends on the location of EZ, i.e., mesial temporal (MTLE) or lateral (neocortical) temporal lobe epilepsy (LTLE) [Figure 1]. For MTLE, the sensitivity of MEG is low when compared to LTLE spike sources. The sensitivity of MEG in detecting spikes is high in extratemporal lobe epilepsies (ETLE) compared to MTLE [Figure 2]. MEG is especially useful in patients with MRI-negative focal neocortical epilepsies [Figure 3]. In such patients, MEG appears to be superior to fludeoxyglucose (FDG)-PET in hypothesizing the epileptogenic zone.
The low yield of MEG in MTLE patients may be due to the use of spherical head models, which although approximating well with neocortical temporal and extratemporal lobes, does not model the medial and inferior surfaces of the temporal lobe. Moreover, the intensity of the signal from the medial and inferior surfaces of the temporal lobe gets attenuated by the rule of the inverse square law (intensity is inversely proportional to the square of the distance from the source to the MEG sensor). It might be more advantageous to use a realistically shaped head model for more precise source localization. In recent years, several head models have been proposed such as the boundary element method (BEM), the finite difference method (FDM), and finite element method (FEM).,, Of these, the FEM is the most versatile and computationally comprehensive because it computes the anisotropic conductivity profiles of all tissues, including the white and grey matter distinction and the cerebrospinal fluid. Thus, it seems likely that source localization accuracy of MEG could be improved in future with the introduction and wider use of these effective versatile head models.
The MEG-guided refinement of the AEC hypothesis helps to decide the placement of invasive electrodes such as subdural, conventional depth and SEEG electrodes., MEG has also been shown to be especially useful in suspected insular epilepsy. Despite the inability to be localised by other noninvasive investigations, if these patients show a dense dipole cluster on MEG, they are likely to benefit from ES targeting the insula. In patients with skull defects due to previous surgery, unlike EEG electrical fields, magnetic fields are not distorted by skull and dural defects, and CSF collections. Therefore, MEG has a special place in investigating patients who have had a failure of their ES. In one study, MEG dipoles were localized adjacent to the margins of the previous resection, which agreed with the invasive EEG recordings. MEG has been shown to play a dominant role in accurately localizing the tangential activity from intrasylvian cortex in children with Landau–Kleffner syndrome, thereby facilitating the diagnosis and in selecting patients for multiple subpial transections.
Correlation between MEG spike source and surgical outcome
Among 132 patients with DRFE, it was found that if interictal MEG spikes sources were concordant to the surgical resection, 85% of the patients became seizure free; however, this outcome dropped to 37% in patients with discordant or nonspecific MEG spikes. Following MEG-guided cortical resection in patients with lesional ETLE and LTLE, 92% became seizure free. Patients with TLE, who underwent anterior temporal lobectomy and amygdalo-hippocampectomy (ATL-AH), became seizure free when there was anterior temporal MEG spike localization; however, seizure recurrence was associated with non-anterior temporal MEG localization. In another study, 88% (n = 23/26) of the TLE patients, who had anterior temporal horizontal and anterior temporal vertical MEG dipoles, became seizure free after ATL or selective AH. Patients with lesional frontal lobe epilepsy had a favorable surgical outcome when complete resection of MEG-defined EZ was undertaken. In pediatric patients with focal cortical dysplasia type Ia, which are typically not appreciated on MRI, resection of tight clustered MEG spike sources often results in a good post-operative seizure outcome.
Novel applications of MEG in epilepsy evaluation
Apart from studying the locations of cerebral sources, other electrophysiological characteristics of epileptiform spike and seizure rhythms can be studied by MEG. The neuronal oscillations that occur beyond gamma band (>80Hz) are known as high frequency oscillations (HFOs). The HFOs has been shown to be a biomarker of epileptogenesis, and the removal of cortical areas that generate pathological HFOs had been associated with better postsurgical seizure outcome. Most of the studies on HFOs are based on invasive EEG evaluation. MEG may be useful in noninvasively detecting HFOs.
The concept of epileptogenic zone is slowly being replaced by epileptogenic networks. Many MEG studies define the epileptic brain networks in the sensor space rather than in the brain space. Graph theory is being utilized in MEG for studying both focal epilepsies as well as generalized epilepsies., Although recording seizures during MEG is very challenging, ictal MEG data can be extremely useful for ictal onset zone localization. Although the electrophysiological characteristics of entire seizures cannot be determined because of significant head movements during seizure, the early ictal rhythms or changes can yield a considerable diagnostic value in patients with DRFE.
Functional localization of the eloquent cortex
Determining accurate location of the primary motor cortex, somatosensory cortex, and central sulcus is crucial in patients with proximity of the epileptogenic lesion to these areas. MEG localization of motor/sensory cortex had excellent agreement with direct cortical stimulations and mapping., MEG helps in determining the association between epileptogenic lesions and the visual cortex; thus, it guides the visual cortex resection. For language lateralization, various MEG paradigms have been utilized including a silent reading task, word recognition task, and verb generation with picture naming task., The overall concordance rate between MEG lateralization and Wada test ranged between 69–95%., MEG can also be used for intrahemispheric language distribution. MEG can provide the accurate location of both the receptive as well as the expressive language cortex; therefore, it can serve as a replacement for the Wada procedure.
Limitations of magnetoencephalography studies
MEG cannot be utilized for chronic recordings, and only interictal MEG recordings can be obtained in the clinical setting, whereas ictal recordings are not usually feasible. Few other limitations are described in [Table 2]. MEG studies have reported to be nondiagnostic or inconclusive in one-third of patients. In patients with infrequent or no MEG spikes, a benign variant of MEG spikes are recorded in bilateral fronto-parietal MEG sensors localizing to sensorimotor areas. These benign MEG spike variants are often misinterpreted as pathological spikes. Widely distributed or multifocal spike sources in MEG need to be interpreted in conjunction with the information from other presurgical modalities. The maintenance of MEG imaging centres and the cost of MEG acquisition are considerably expensive, which limits their availability even in developed countries. Furthermore, well-trained personnel and neurophysiologists are required for optimal deployment of the facility.
The SEEG was born out of the pain-staking efforts of the epileptologist - neurosurgeon duo, Jean Bancaud and Jean Talairach, to advance intracranial EEG by allowing more prolonged extraoperative interictal and ictal recordings in people with DRFE, stimulation and mapping of eloquent cortical areas, and real time correlation of semiological manifestations with specific cortical structures. In recent years, there has a been a revival of interest in SEEG, due to the following significant developments: (1) Sophistication in electrode design and signal interpretation with modern computing technology, (2) major developments in brain imaging technology and improved neurosurgical techniques, and (3) better understanding of the surgical management of epilepsy, its pathological substrates and expected outcomes. The more recent acceptance that epilepsy is caused not by structural focus, but due to abnormal neural networks, albeit in relation to a circumscribed lesion, represents a paradigm shift in the understanding of epilepsy. In this particular context, SEEG is capable of mapping the entire epileptic network and spatiotemporal relationships across even distant structures [Figure 4].
Indications and contraindications of SEEG
The common indications for an invasive evaluation are provided in [Table 1]. Among the patients who require an invasive EEG monitoring, indications for SEEG in preference to subdural electrode recordings are listed in [Table 3]. In summary, SEEG scores over the conventional invasive recording when the potential EZ is deep seated, bilateral and extensive, and when specific functional networks need to be explored. Patients in whom previous the subdual recording has failed to generate an AEC hypothesis, SEEG would be the choice for a repeat evaluation. In adults and children who were otherwise deemed nonsurgical candidates, SEEG exploration is sometimes undertaken based on the relative risk-to-benefit ratio (30% chance of seizure-freedom combined with a low complication rate).,
The relative contraindications to the usage of SEEG include absence of a strong AEC hypothesis, coexisting bleeding diathesis or severe medical comorbidities. A relative contraindication is DRFE with infrequent seizures, since a monitoring period in excess of 2-3 weeks with few seizures may offset the risk-benefit ratio. During the preoperative counseling for SEEG exploration, it is important for the patient and family to understand that the diagnostic nature of the procedure implies one of two possibilities: (1) Localization of the epileptogenic focus and subsequent epilepsy surgery (reported to occur 80-90%); or, (2) unresectability and continuation of modified medical management, ketogenic diet, or palliative surgical management, such as vagus nerve or deep brain stimulations.
The steps of SEEG methodology are summarized in [Figure 5] and the steps are illustrated through a patient scenario [Figure 6].
The single most important step in SEEG implantation remains the formulation of an elaborate preimplantation AEC hypothesis. This is accomplished through an epilepsy management conference in which the neurologists, neurosurgeons, neuroradiologists, and neuropsychologists, who are involved regularly with the epilepsy surgery program participate. The data gathered by the non-invasive presurgical evaluation are meticulously reviewed in the following order: (1) Seizure semiology obtained by history and neurological deficits such as hemiparesis or visual field defects; (2) semiology of the recorded seizures by VEEG monitoring; (3) scalp recorded interictal and ictal EEG abnormalities; and, (4) structural MRI and functional imaging (such as PET, SPECT, fMRI and MEG, if undertaken) findings. If the clinical, electrical and imaging data are concordant, and EZ/EN can be reliably localized and its non-encroachment to eloquent cortical areas can be reliably concluded, patient could proceed with surgery; otherwise, an invasive evaluation is planned.
With respect to SEEG, the most important component is seizure semiology (phenomenology), which results from coactivation and/ or deactivation between multiple hubs in the epileptic network. A detailed discussion of ictal symptoms and signs and their anatomical correlation is beyond the scope of this review, but few important points are worth mentioning. For example, in an analysis of SEEG-video recorded EC manifestations of 54 patients with frontal lobe epilepsy, Bonini et al., identified four groups – a finding that significantly aids in pre-implantation SEEG planning. They described a hierarchy of semiological manifestations following a posterior-anterior axis based on the temporal manifestation in isolation or in combination of: (1) Elementary motor signs, (2) nonintegrated gestural motor behavior, (3) distal stereotypies and (4) fearful behavior resulting from involvement of the paralimbic system. Proximity to primary visual, auditory and sensory cortices allows for a simple topographic localization using auras. Elementary visual hallucinations like seeing colored circles in the left superior visual quadrant localizes the epileptic focus to the right inferior calcarine cortex, but when epileptic discharges involve areas away from the primary visual cortex, like the occipito-temporal region, they may present with complex visual phenomena like hallucinations and illusions. A similar phenomenon also applies to the motor system. Primary motor area stimulation results in cortical myoclonus. As we move towards the supplementary motor and premotor areas, the motor phenomenon becomes more complicated. Finally, upon reaching the rostral end of the supplementary motor area (pre-SMA), cessation of motor activity (motor arrest) is more commonly seen due to the inhibitory nature of these motor areas.
Apart from being determined by the anatomical substrate, semiology has also been shown to depend upon the discharge frequency. For example, a phasic discharge while stimulating the primary motor cortex leads to a focal myoclonic jerk, whereas a high frequency oscillatory discharge of the same area results in motor arrest. The semiological expression by the same area may also vary depending on whether it is initially involved in the pathological discharge or it is driven by a distant focus. For example, a seizure discharge starting in the primary motor cortex results in clonic or tonic-clonic movements, whereas a discharge originating from a distant focus like the orbitofrontal region and subsequently spreading to the motor cortex may manifest a different semiology of hypermotor seizures. Hence, it is important to integrate primary and secondary networks into the SEEG sampling plan since multiple hubs of epileptogenicity are described which may contribute to surgical failures.,, These hubs often express significant interlobar connectivity, for example, fronto-parietal, orbitofrontal-temporopolar, insulo-temporal networks. In contrast to subdural electrode implantations, a well-planned SEEG exploration conveniently allows sampling of distant hubs that may be intimately connected – an understanding of which may be important when finally planning resection or lesioning.
Historically, SEEG evolved around the cerebral tele-angiogram for two important reasons: (1) The need to avoid intracranial hemorrhage resulting from a vascular collision with the stereotactic insertion; and, (2) this being the prevailing brain imaging at that time prior to the CT era. The former concern has continued through contemporary case series and remains a common reason for the continued use of pre-implantation digital subtraction angiography (DSA) of the cerebral vessels at many centers., All present frame or frameless stereotactic systems function use thin-cut DICOM (digital images and communications in medicine) images of the head - hence a volumetric navigation series imaging of the brain with adequate pial and sulcal vascular delineation is the minimum imaging required for planning SEEG cortical trajectories. In this process, a heavily contrast-enhanced T1-weighted navigational sequence has become the routine primary imaging sequence used to plan SEEG implantation at many centers. Adequate visualization of pial and sulcal vessels to avoid an entry-point vascular collision is most important. The remaining trajectory should also be inspected carefully; however, most deep sulcal and parenchymal vessels yield more readily to the electrode passage and hence represent less of a hemorrhage risk. Against the navigation-sequence MR image, other additional presurgical images such as PET or SPECT data could be co-registered or visually correlated to enhance the targeting of suspected foci depending on the capabilities of the planning software. For most stereotaxy systems (frame-based, frameless, or robotic), a thin-cut navigational sequence CT is also required to give either stereotactic frame coordinate localization or more accurate surface registration (e.g., ROSA robotic system, Medtech, Montpellier, France). MRI surface registration is inaccurate and MRI-based frame localization is both tedious and time-consuming in planning multiple stereotactic trajectories. This thin-cut CT can be completed as a cerebral CT angiogram with emphasis on the cerebral venous phase to further enhance the surface vascular markings during stereotactic trajectory planning. A detailed description of the implantation strategies are beyond the scope of this review and readers may consult other recent articles on this subject.,
Recording and interpretation
Judicious and individualized tapering of AEDs is advised as in any chronic monitoring to capture the ictal events. The minimum number of events needed to arrive at a conclusion of the EZ is variable and must be tailored to each patient. In a recent meta-analysis of the safety of SEEG technique reviewing 30 published studies with 2,624 patients, the average duration of monitoring was 11 days with a range from 2 to 33 days. Most image guidance softwares allow rapid co-registration of thin-cut CT data with the preoperative MRI, allowing precise three dimensional visualization of the electrode contacts in relation to the cortical anatomy. This should be done almost immediately following implantation to ascertain the final position of electrode contacts, and to assist the electrophysiologist in developing a spatiotemporal interpretation of the onset and spread of ictal rhythms.
After recording sufficient number of electro-clinical events, the next step is a spectral and temporal correlation of the ictal discharge pattern with anatomic structures. During the initial decades of SEEG practice, emphasis had been on the identification of low voltage fast activity in the beta and gamma frequency ranges to define the seizure-onset zone., Alarcon et al., demonstrated that resection of the cortex that was generating early fast activity recorded on intracranial EEG was a good predictor of postoperative seizure-freedom. The next important factor in determining the epileptogenic significance of a discharge is its timing with respect to the clinical seizure onset - the earlier the appearance of the discharge, the more it is likely to be epileptogenic. This forms the basis for a qualitative visual analysis of the SEEG recordings. Since the turn of the millennium, improved signal acquisition technology has allowed sampling at much higher frequencies, which revealed the presence of oscillations >400 Hz at the seizure-onset. A calculated epileptogenicity index was proposed by Bartolomei et al., to define the seizure-onset zone in a more quantitative, controlled manner, which mixes both spectral and temporal information of each electrode.
Akin to subdural electrodes, SEEG electrodes can be used for the precise electrocortical stimulation mapping of cerebral functions. For the same purpose, a large subdural grid naturally gives a wider surface coverage of the lobes and would seem to be more appropriate for mapping studies when compared to depth electrodes. However, it should be noted that only 10% of grey matter is present on the surface. The main difference between using subdural grids versus SEEG-depth electrodes for stimulation mapping is that depth electrodes need to be positioned specifically for cortical mapping if required, whereas a large surface grid allows for wider mapping when in doubt. Mapping involves current stimulation between 1-6 mA between contacts and careful observation of positive sensorimotor, auditory-visual or other experiential phenomena as well as inhibition of continuous function (e.g, language, motor tasks). Stimulation mapping can also be used to recreate specific seizure auras to correlate with EZ mapping, and to recreate the habitual seizure patterns from electrode contacts in the suspected EZ/EN.
A focal resection, disconnection or lesioning of the proposed epileptogenic cortex should naturally result from mapping of the ictal onset pattern, epileptogenic network and adjacent or intervening functional cortical networks. At most SEEG centers today, this is undertaken several days to even months after the removal of SEEG electrodes. This is a major benefit of using SEEG over subdural grids implanted through a craniotomy for invasive evaluation performed in that the diagnostic phase is clearly separated from the therapeutic phase. This allows for: (1) An adequate time for careful analysis of the SEEG and mapping data, (2) resuming adequate AED levels which would have been otherwise tapered during the ictal recordings, (3) a refreshing break for both the patient and the surgical team, and (4) helping reduce any bacterial colonization of the scalp and cranial implantation tracts. During ES planning, the SEEG data should be integrated again with all the preimplantation modalities to plan the best resection, disconnection or lesioning that will lead to seizure freedom with the least neurological morbidity. During a focal resection, the scalp and bone entry point scars should be matched with the original electrode implantation trajectories to navigate the surgeon. Many experienced SEEG centers now also consider thermoablation when the epileptogenic zone is observed to be very focal, especially in the insula where a craniotomy and large microsurgical exposure might otherwise be required to reach the intended small target. This can be achieved with conventional radiofrequency (RF) or more modern laser induced thermal therapy system., In one of the largest modern clinical series reviewing the accuracy, safety and outcome of 500 SEEG implantations in 482 patients over a 5-year period, 18.5% did not proceed to resection due to epileptological/functional concerns or patient refusal. Of the remaining patients, 73% underwent resection, 4.1% were awaiting surgery and 3.1% underwent SEEG-guided RF thermoablation.
Depending on technique, mean accuracies vary between 1-2 mm. Gonzalez-Martinez et al., recently reported an accuracy in the range of 0.78-1.83 mm with maximum range of 0.3 to 5.1 mm at entry, and similarly, 1.2 to 2.3 mm with a maximum range of 0.4 to 7.1 mm at target for robotic implantation. Therefore, it is imperative that the operating surgeon periodically assesses the post-implantation images with the pre-implantation trajectory planning and standardize the stereotactic technique to achieve an implantation accuracy of ~ 1 mm. In their extensive review of 419 patients implanted with SEEG, Cardinale et al., reported a median entry point localization error of 1.43 mm (interquartile range, 0.91-2.21 mm) with the traditional Talairach frame plus digital subtraction angiogram (DSA) method and 0.78 mm (interquartile range, 0.49-1.08 mm) with a new workflow pattern consisting of three-dimensional volumetric angiogram-brain surface image reconstruction with robotic stereotaxy (Neuromate, Renishaw-Mayfield SA, Nyon, Switzerland). Their median target point accuracies were 2.69 mm (interquartile range, 1.89-3.67 mm) for the traditional, and 1.77 mm (interquartile range, 1.25-2.51 mm) for their newer method.
The most serious complication of SEEG electrode implantation is hemorrhage, with major hemorrhages resulting in re-exploration or focal deficits occurring in about 0.4-1% of patients., Controllable factors related to avoidance of hemorrhage include: (1) Adequate preoperative vascular imaging and trajectory planning, (2) stereotactic accuracy, (3) careful dural penetration, and (4) careful probe insertion through the pia-arachnoid planes. A recent systematic meta-analysis of the incidence of complications reported during SEEG electrode implantation evaluating 22,085 electrodes implanted in 2,624 patients revealed a remarkably low overall complication rate of 1.3%. The pooled prevalence of hemorrhagic complications was 1% with SEEG compared to 3.2-4% observed with subdural grid placements., The complications and their frequencies from this meta-analysis are summarized in [Table 4].
Prospect of epilepsy surgery and seizure outcome
Postoperative seizure-free outcomes rates ranging from 50% to 88% have been reported for ES following SEEG.,, The outcome is influenced by several factors such as complexity of patient selection, accuracy of the preimplantation AEC hypothesis, and experience in SEEG implantation technique and interpretation. In a retrospective series of 215 patients (age range 2-56 years) implanted with SEEG, 87% of patients underwent ES. Amongst 165 patients with a minimum of 1-year follow-up, 56.4% had Engel Class 1 outcome. In a series of 100 patients reported by Guénot et al., SEEG was helpful in 84% of patients to confirm or annul ES and to plan the extent of resection, and allowed resections that would not have otherwise been considered possible in 14%. Chassoux et al., reporting seizure outcome following SEEG-based resection specific for type II focal cortical dysplasias (FCD) which were largely already detected in imaging, noted Engel Class 1 outcomes for 88% of MRI-negative and 94% of MRI-positive patients with a mean follow-up of 4.6 years and 3.9 years, respectively. Even in MRI negative cases, SEEG recordings, when utilized with MRI-PET fusion findings, allowed restriction of the resection to a single gyrus in more than half of the patients. A prospective study of 200 patients with DRFE from the Cleveland Clinic, Cleveland, OH, who underwent 2663 electrode implantations, concluded that a confirmation of the EZ was possible in 77%. In their study, which included patients who previously had failed subdural invasive monitoring and ES, 134 (67%) patients subsequently underwent SEEG-guided resection and 61 of 90 (68%) patients with a minimum follow-up of 12 months (mean 2.4 years) achieved Engel Class 1 outcome. Resected tissue revealed 61 type 1 FCD in 61%, type 2 FCD in 10% and no specific pathology in 9%. In summary, the available data indicate that, following SEEG, there is at least a two-third chance that a patient will undergo ES and he/she a similar chance of becoming seizure-free following ES.
AIMS, Kochi experience
The first SEEG implantation in AIMS, Kochi, was performed on January 2, 2014. Since then, until September 30, 2016, we have undertaken SEEG implantations in 45 patients. Their age ranged from 5 – 49 (mean 26.6) years. Eleven of them were <18 years. The number of SEEG electrodes implanted ranged from 2 – 14 (mean 9), each electrode with 8 or 16 contacts. After implantation, the patients were monitored for an average of 5.0 + 2.7 days, recording a mean of 10.2 electro-clinical and 8.1 electrographic seizures. The indications for SEEG in our patients are depicted in [Figure 7]. The most common indication was no demonstrable lesion in 3T MRI images in 32 (71%) patients. Following SEEG, all except 2 patients underwent ES. No serious complications such as hemorrhage or intracranial infections occurred. Out of 33 patients, who had completed ≥6-month postoperative follow-up, 23 (69.7%) were seizure-free.
Research and technology developments, aimed at increasing the yield of presurgical evaluation and in making it more precise and noninvasive, are progressing in three directions. First, advances in structural neuroimaging such as a 7T MRI and a variety of post-processing techniques may help us to ascertain the lesions that are invisible in 3T MRI images or make them more distinctly demarcated. Second, the Human Connectome Project is engaged in noninvasively investigating brain connectivity in normal subjects and in various brain diseases including epilepsy, utilizing techniques such as diffusion tensor imaging tractography, resting state fMRI, EEG-fMRI, electrical source imaging (ESI) and magnetic source imaging (MSI).,, Third, specific intracerebral EEG patterns such as HFOs might help us to more objectively assess the epileptogenic brain structures., These developments may obviate the need for SEEG in a proportion of patients in the future who require it now. More importantly, they may assist in making SEEG exploration more goal-directed and hypothesis-driven, and may decrease the number of electrodes required to be inserted in an individual patient, thereby, decreasing the cost and complication rate. Indigenous SEEG electrode development can bring down the cost of SEEG and make it affordable to a larger segment of the population in developing countries.
The Editor, Neurology India as well as the Guest editor of the Neurology India Supplement are grateful to Drs. V. Jayabal, N. Mariyappa, S. Sinha, and P. Satishchandra for contributing the part of this chapter on 'magnetoencephalography;' and, to Drs. A. Pillai, S. Gopinath, and K. Radhakrishnan for contributing the part of this chapter on 'stereo-electroencephalography.'
The Editor, Neurology India as well as the Guest editor of the Neurology India Supplement also wish to thank Dr. K. Radhakrishnan for his help in extensively editing 'magnetoencephalography' and 'stereo-electroencephalography' manuscripts, and merging them.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5], [Figure 6], [Figure 7]
[Table 1], [Table 2], [Table 3], [Table 4]