Introduction
Insular epilepsy (IE), or insular-opercular/peri-Sylvian epilepsy, is defined as epilepsy originating from the insula and extra-insular structures, typically the surrounding opercula with or without more extensive regional networks of the frontal, temporal/limbic, or parietal lobes ( Fig. 9.1 ). The first experience with surgery of the insular lobe for drug-resistant epilepsy (DRE) was carried out almost simultaneously by Penfield and Faulk at the Montreal Neurological Institute (MNI) in Canada and Guillaume & Mazars in France in the 1940s . Penfield and colleagues began sampling the inferior portion of the insula following removal of the temporal lobe in patients undergoing awake anterior temporal lobectomy (ATL) . In addition to studying the clinical symptoms elicited from cortical stimulation , they began removing varying degrees of the insular cortex when intraoperative electrocorticography (iECoG) showed insular epileptiform activity suggestive of possible insular lobe involvement as part of the epileptogenic zone (EZ) . While Guillaume and Mazars reported good seizure outcomes , the enthusiasm for insulectomy at the MNI declined over the course of a decade and they performed progressively fewer insulectomies due to suspected higher morbidity in the patient that underwent insulectomy. Silfvenius and Rasmussen eventually published the MNI results in 1962, which showed that the addition of insulectomy to ATL based on iECoG did not improve seizure freedom but came at the cost of increased neurological morbidity, including a 21% rate of motor impairment (of which 7% were permanent) as well as a 14% rate of language impairment, which they attributed to middle cerebral artery (MCA) vessel manipulation . Following this publication, there was little interest in investigating or surgically treating IE for almost half a century. With the introduction of the surgical microscope in the 1970s and a greater understanding of microsurgical anatomy, Prof. Yasargil demonstrated that the insular/peri-Sylvian region could be safely navigated surgically with a very low rate of permanent morbidity in patients with paralimbic/insular gliomas . Several other authors have since shown that insular gliomas can be safely resected, including using awake craniotomy with cortical and subcortical mapping techniques .
Over the course of the second half of the 20th century, advancements in technology and techniques in stereo-encephalography (SEEG) and subdural grid/strip placement allowed for European and Canadian groups to sample the insula more safely using these techniques, respectively. These advancements led to the modern description of IE semiology in 6 of 50 patients sampled for temporal lobe epilepsy (TLE) with atypical features . Over the course of the last 20 years, several groups around the world have sampled the insula with SEEG in patients with DRE and suspected insular involvement based on semiology or noninvasive workup . While some groups have shown that the addition of insular resection does not translate in improved seizure outcomes in patients with TLE and insular involvement , most reports have now shown IE surgery to be highly effective. Groups from around the world, including ours, have shown that resective surgery for IE is both feasible and effective, with a 60% seizure freedom at 5 years and a low rate of permanent impairment . However, 43% of patients harbor temporary impairment, including 30% motor deficits at the group level and 46% language deficits for language-dominant IE surgery . Over the last 10 years, many groups have reported outcomes surgically treating IE with an expanded armamentarium of less invasive stereotactic ablation [e.g., MR-guided laser thermal ablation (MRgLA), radiofrequency thermo-coagulation (RFTC)] procedures and neuromodulation (e.g., RNS) . It is thus imperative that in addition to selecting the right approach for the right patient, neurosurgeons master open and stereotactic ablation techniques, particularly in the context of emergent noninvasive neuromodulation options which harbor no risk of neurological morbidities, such as RNS ( Table 9.1 ).
| Open resection | SEEG-guided RFTC | Volumetric RFTC | MRgLA | RNS | |
|---|---|---|---|---|---|
| Technique | |||||
| Mean No. trajectories (range) | Not applicable | 2 (1–3) | 6 (2–8) | 2 (1–3) | Not applicable |
| Mean No. ablations (range) | Not applicable | 6 (2–15) | 20 (6–45) | Not available | Not applicable |
| Mean total ablation size, cm 3 (range) | Not applicable | Not available | 4 (0.2–10) | 13 (5–30) | Not applicable |
| Advantages |
|
|
|
|
|
| Disadvantages |
|
|
|
|
|
| Relative Indications |
|
|
|
|
|
| Seizure freedom b | 67% | 0%–20% | 53%–70% | 43%–50% | 0% |
| Responder rate c | 95% | 25%–60% | 89%–100% | 80%–100% | 50% |
| Neurological deficit d | 45% | 0% | 40%–42% | 35%–66% | 0% |
a Minimally invasive benefits: Smaller “stab” incision, avoidance of craniotomy, no ICU stay, less postoperative pain and reduced narcotic use, shorter hospitalization.
b Including data from larger case series and meta-analysis.
c “Responder” for open and LITT/RFA is defined as Engel 1–3 (or > 50% seizure reduction) and “responder” for RNS is defined as >50% reduction in seizures.
d Neurological deficit includes a combination of both temporary and permanent deficits.
Anatomy
Surgical anatomy
The insula is an inverted pyramid-shaped lobe located deep inside the Sylvian fissure. It is divided into seven gyri, with three short gyri anteriorly, two long gyri posteriorly, as well as an accessory gyrus and a transverse gyrus located at the most antero-inferior portion of the insula adjacent to the insular apex ( Fig. 9.2 ). The insula is divided into a larger anterior and smaller posterior portion by the central insular sulcus, which extends obliquely from the superior periinsular sulcus to the limen insulae at approximately the same level (up to 5 mm anterior) as the central sulcus. The insula is covered by the fronto-orbital, fronto-parietal and temporal opercula and it is separated from these structures by the anterior, superior and inferior peri-insular sulci, respectively. The peri-insular sulci surround the entire insula except for the limen insula, transverse gyrus, and accessory gyrus. The transverse gyrus is continuous with the posterior fronto-orbital cortex and the accessory gyrus with the orbital gyrus. The peri-insular sulcus represents an important surgical anatomical landmark during subpial operculo-insulectomy. The limen insula, located below the insular apex, is also an important surgical landmark, corresponding to the location of the genu of the MCA before its bifurcation. It is also a landmark for the underlying temporal stem which contains the inferior fronto-occipital fasciculus (IFOF) and uncinate fasciculus (UF) ( Fig. 9.3 ) and overlays the lenticulostriate perforators of M1 of the MCA. It is crucial to have a good working knowledge of the vascular anatomy of the peri-insular region. The M2 segments of the MCA branches cover and supply the insula until the periinsular sulcus, where they become the M3 segment of the MCA as they extend over the opercula towards the cerebral hemispheres ( Fig. 9.4 ). There are two types of perforators that arise from the M2 segments of the insula: (1) short M2 perforators irrigate the insular cortex, extreme capsule, claustrum, and external capsule, and (2) long insular perforator arteries (LIA) preferentially originate from the M2 segment of the MCA, usually in the postero-superior insula, and irrigate peri-insular white matter, particularly the cortico-spinal tract ( Fig. 9.4 ) . Long medullary arteries (LMAs) originate from M3 over the opercula and also irrigate the peri-insular white matter, notably the corticospinal tract (CST) and arcuate fasciculus (AF)-superior longitudinal fasciculus (SLF) complex ( Fig. 9.4 ) .
Functional anatomy
The insula harbors widespread reciprocal structural and functional connections with diverse subcortical and cortical brain regions including the dorsal thalamus, amygdala-hippocampal complex, hypothalamus, brainstem and various neocortices of the frontal, parietal and temporal lobes ( Fig. 9.2 ) . The insula has been structurally and functionally parcellated in vivo using diffusion tensor imaging and resting-state functional MRI (rs-fMRI) into three subdivisions with an antero-posterior gradient of connectivity. This tripartite “cognition-emotion-interoception” organization including the “cognitive” dorsal anterior insula (dAI) which harbors connections with pregenu anterior cingulate and frontal areas supporting higher-level cognitive processes, the ventral “socio-emotional” anterior insula (vAI) which harbors connections with limbic areas supporting affective processes, and the posterior insula (PI) which harbors connections with sensorimotor areas supporting interoceptive processes . There is also a fourth small olfacto-gustatory region located in the central insula and a transitional insular zone with overlapping connections ( Fig. 9.2 ) .
To understand seizure semiology, neuropsychological deficits and surgical risks for patients with IE, it is important to have a good working knowledge of the diverse functions that have been attributed to the insula. The insula plays a role in visceroscensory processing, autonomic control (blood pressure and heart rate), pain and thermosensory perception, olfaction, gustation , vestibular processing, but also higher-order functions such as self-awareness, introspection, empathy, risky decision making , language, and verbal memory. In addition, the dAI is central to the salience network and integrates external sensory information with internal emotional and bodily state signals to coordinate brain network dynamics and to initiate switches between the default mode network (DMN) and central executive network .
Open microsurgical resective surgery for insular epilepsy
Open microsurgical resection, the mainstay and most widely available surgical treatment option for drug-resistant IE, can be roughly divided into two categories: selective insulectomies for cases with a strictly insular seizure focus and insulectomy with additional opercular resection or lobectomy in cases of insular/peri-Sylvian or “insular plus” epilepsy. Selective insulectomies can be carried out through a trans-Sylvian or trans-opercular corridor.
Open surgical technique
Rationale of trans-opercular versus trans-sylvian insulectomy for pure insular focus
When the EZ extends beyond the insula to the opercula, opercular resection can be leveraged as a corridor to reach the insula in a subpial fashion ( Fig. 9.5 ) . When the insular seizure focus is restricted to the insula, and (1) the EZ involves a large part of the insula and/or (2) MRgLA or vRFTC is not available, open resection is a very good surgical option. In these cases, one may select a trans-opercular or trans-Sylvian corridor which involves opening the Sylvian fissure. The rationale and advantage of the trans-Sylvian approach is anatomical preservation of the opercula, which harbor important functional zones, including the (1) precentral gyrus of the frontal operculum important for hemifacial motor function and swallowing, (2) pars triangularis/opercularis, posterior superior temporal gyrus (STG), supramarginal gyus (SMG) and angular gyrus (AG) of the inferior parietal lobule important for language in the dominant hemisphere ( Fig. 9.3 , Fig. 9.4 ) . Accordingly, resection of the frontal operculum is responsible for a high rate, albeit typically transient, of facial weakness and resection of the dominant temporal operculum or inferior parietal lobule is associated with language impairment . In the glioma surgery literature, some groups have advocated for a trans-Sylvian approach to avoid opercular injury and associated deficits . Our group has reported good long-term outcomes using this approach for IE . However, the trans-Sylvian approach is not without its pitfalls. Approaching the more posterosuperior insula through a trans-Sylvian corridor is particularly challenging as the Sylvian fissure becomes more difficult to dissect due to adherent opercula, the presence of large sylvian veins, and the ergonomic challenge of reaching a more deeply located insula below the Sylvian fissure compared to the more anterior and inferior insula ( Fig. 9.5 ). Therefore, some have argued that the trans-Sylvian approach of the posterosuperior insula increases the risk of opercular injury from retraction and venous sacrifice, which is not tolerated in a third of patients. Some groups have thus advocated for a trans-opercular approach (through nonepileptic opercular tissue) for IE, particularly in the more posterior and superior insula, to facilitate access to this challenging region, avoid fissure splitting and risk of vasospasm and opercular retraction . However, this approach results in a high rate of transient facial weakness and may put at risk other functional areas . Trans-opercular subpial insulectomy through nonepileptic opercular can be used if “safe corridors” are identified using functional mapping with invasive EEG, functional MRI and/or intra-operative neuromonitoring. In the glioma literature, applying intraoperative awake and/or cortical and subcortical direct electrical stimulation can make the trans-opercula approach safe with potentially greater likelihood of complete resection and reduced likelihood of permanent deficits . Because awake resection is not feasible in many pediatric patients, modern intra-operative methods such as cortico-cortical evoked potentials (CCEPs) can potentially avoid damage to peri-Sylvian language pathways . Tractography is also useful to avoid inadvertent injury to the CST/CBT (motor) and AF-SLF complex (language), particularly in resecting the more postero-superior insula ( Figs. 9.3 and 9.4 ).
Technique
Insular-opercular resective surgery is performed through a pterional-based fronto-temporal craniotomy. Frameless neuronavigation is a useful adjunct to correlate intraoperative anatomy with preoperative MRI. Co-registering the anatomical MRI with noninvasive workup, including fMRI, PET, SPECT, MEG cluster and tractography can also very helpful.
With the patient supine, the head is rotated 40–60 degrees to the contralateral side. A bolster pad may be placed under the ipsilateral shoulder. About 15 degrees of extension is placed on the neck and the head is elevated so the zygoma is the highest point. In patients in whom a prior craniotomy has not been performed, an inverse question mark incision is performed followed by pterional craniotomy tailored to the size of peri-sylvian corticectomy. After a fronto-temporal craniotomy, the lesser sphenoid wing is drilled down to the superior orbital fissure. This step is important to optimize exposure of the anterior insula and reduce brain retraction during insulectomy. After a c-shaped dural opening, the typical hemispheric anatomy can be seen. The subsequent steps depend on whether pure insulectomy or insular-opercular resection is to be performed.
Trans-Sylvian selective insulectomy
In cases of pure insulectomy, a trans-Sylvian approach with splitting of the fissure followed by insulectomy can be performed. Sylvian fissure opening should be carried in a step-wise approach, starting from the superficial Sylvian cistern and extending down to the operculo-insular cistern (deep Sylvian fissure). Because the fissure is widest adjacent to the pars triangularis apex, Sylvian fissure dissection should be started from this point . Always using sharp dissection, the fissure can then be dissected in a stepwise manner using well-known techniques, in which MCA branches are followed down to open the operculo-insular cistern, all MCA branches and as many Sylvian veins as possible are preserved with the dissection coursing posteriorly, followed by anteriorly . Sylvian fissure dissection is tailored to the location of the insulectomy and a wide opening of the fissure is warranted when complete insulectomy is performed. However, when partial insulectomy is performed, a limited corresponding portion of the fissure may be split. Detailed anatomical knowledge of topographical correlation between the opercula and insular gyri is imperative .
Insular cortex resection is then performed by first identifying the superior and inferior sulcus. The cortex between the M2 branches of the MCA is coagulated and incised. The superficial cortex is resected in a subpial manner stopping just short of the white matter to avoid damage to the subcortical white matter pathways, basal ganglia or CST. The M2 perforators going to the insula can be difficult to preserve but the main arterial supply to the basal ganglia is from the lateral lenticulostriate vessels which are preserved by identifying and preserving the temporal stem.
Subpial operculo-insulectomy
Because almost all pediatric cases of IE involve the opercula, resection of the insular cortex can usually be carried out in a subpial manner following opercular corticectomy . Opercular and peri-Sylvian cortices are resected based on preoperative noninvasive and invasive workup mapping. Standard microsurgical techniques should be used, including preservation of “ en passage ” cortical M2 and M3 MCA branches and as many veins as possible. Functional areas should also be preserved, including Broca (pars opercularis and triangularis), Wernicke (posterior STG), and Geschwind’s area (supramarginal gyrus, angular gyrus of inferior parietal lobule) on the dominant side for language, and the sensory-motor area of the hand and face whenever possible ( Figs. 9.3 and 9.4 ).
Following opercular resection, the resection of the cortex can be carried down below the circular sulcus to the insular cortex in a subpial manner ( Fig. 9.5 ). In cases of insular-opercular epilepsy (IOE) in which a large insulectomy is warranted but only a small opercular resection is required, it may be necessary to still perform trans-Sylvian insulectomy following splitting of the fissure. In addition, in cases that have had an open trans-Sylvian placement of electrodes, insulectomy may be performed through this previously dissected corridor. The insular depth electrode can help guide the resection of involved insular cortex. Insular corticectomy can also be combined with temporal lobectomy which can be accomplished by splitting the Sylvian fissure, carrying out the temporal lobectomy, and then completing the insulectomy .
Outcomes
Efficacy
Open microsurgical resection results in very good seizure freedom rates that are comparable to those obtained with surgery for other forms of extra-temporal epilepsy (ETE) . In our experience at the University of Montreal, open microsurgical resection of IE results in seizure freedom (Engel I) in 70% of patients at 1 year and 50% at 5 years . In a recent meta-analysis of children and adults undergoing open resection for IE, the rate of seizure freedom (Engel I) was 82.1% at 1 year, 60.2% at 5 years, and 51.5% at 10 years . By comparison, the 1, 5- and 10-year seizure freedom rates following minimally invasive stereotactic ablation, including both MRgLITT and RFA, were 69%, 39%, and 31%, respectively. Thus, well-selected candidates with purely insular and “insular-plus” epilepsy can be counseled for a high likelihood of seizure freedom with open resection. The likelihood of success relates to the extent of resection of the epileptogenic network, and patients requiring invasive monitoring have a lower likelihood of seizure freedom and should thus be counseled accordingly. Pediatric patients also may have a lower likelihood of seizure control, likely as a result of a greater proportion of malformations of cortical development in this patient population and more extensive multilobar EZs that are difficult to localize, map and completely resect .
Safety
The effectiveness of open microsurgical resection of IE comes at the cost of neurological impairment in about half of patients, although deficits are transient in the vast majority of these cases and permanent in about 8% . Neurological deficits are significantly more likely to occur in patients operated by surgeons with less experience, which underscores the complex nature of this surgery and the nonnegligible learning curve .
Motor outcome
The most common neurological deficit following IE surgery is motor impairment, which occurs in a third of patients . While most patients recover completely, 5% harbor permanent motor impairment, which is nonnegligible and can have a significant impact on quality of life . There is robust evidence from lesion and stimulation studies of motor function within the insula proper that removal of the insular cortex could theoretically explain postoperative motor impairments, particularly those that are transient . However, there is no clear evidence attributing surgical removal of the insula as a direct cause of postoperative motor deficits, especially permanent deficits.
In terms of facial/limb motor deficits, 20% of patients undergoing IE surgery develop hemiparesis and 10% develop brachio-facial or facial weakness with most recovering ad integrum within 3 months . Facial and limb motor impairments are principally due to damage to the corticobulbar (CBT) and CST, respectively, which can be either directly injured from resection or surgical manipulation or indirectly injured through ischemic stroke. Most patients undergoing open resection (60%) develop subcortical ischemia involving the CST/CBT, and while many are asymptomatic, this likely represents the major pathophysiological explanation of most motor impairments following open resection for IE . While strokes can rarely result from injury to the MCA branches themselves or lenticulostriate perforator injury at M1 adjacent to the temporal stem, almost all strokes occur as a result of subcortical ischemic lesions of the corona radiata. These strokes arise from the inadvertent sacrifice of the LIAs that originate from M2, typically injured during removal of the more postero-superior insula, as well as LMAs that arise from M3 following removal of the frontal or parietal opercula ( Fig. 9.4 ). Because these small perforators do not harbor anastomoses, their sacrifice is not well tolerated . More extensive opercular resections and postero-superior insular resections thus lead to a higher risk of stroke and motor deficits, including hemiparesis, through these long perforator injuries . This complication is specific to open microsurgical resection and represents a disadvantage of this approach compared to stereotactic ablation techniques. Sparing resection of nonepileptic cortical structures (fronto-parietal opercula, postero-superior insula) may in theory reduce the risk of stroke and deficits, although this may come at the cost of reduced likelihood of seizure freedom in patients .
Direct injury to cortical and subcortical structures subserving motor function is also a significant cause of motor deficits, and patients are particularly at risk when resection extends beyond the insula. Resection of the precentral gyrus (M1) at the level of the frontal operculum results in hemi-facial weakness, which is typically transient, as has been shown in both epilepsy and glioma literature . When resection involves more extensive M1 resection, one can expect brachio-facial weakness. The CST white matter, which is located adjacent to the posterosuperior margin of the insula ( Fig. 9.4 ), can be directly injured by inadvertently extending resection too deep beyond the superior periinsular sulcus at this location, although this is very rare.
Dysarthria or apraxia of speech (AOS) is not uncommon following resective IE surgery, occurring in 3% of patients, possibly related to damage to peri-Sylvian nodes of the cortico-subcortical motor speech network , including the insular cortex, frontal operculum/inferior primary sensorimotor cortex, STG and/or AF . A small subset of patients (0.5%) experience dysphagia following surgery, likely related to damage to the cortical or subcortical network responsible for voluntary swallowing, namely the sensorimotor integration areas of the insula and peri-Sylvian cortical structures (e.g., frontal opercula, precentral/postcentral gyrus, SMG/AG, STG) and their corresponding white matter tracts (e.g., CBT) .
Higher-order motor deficits may occur following IE surgery and likely originate from resection of the insula or peri-sylvian intra-hemispheric associative pathways, such as the frontal aslant tract (FAT). The FAT is a track that originates from the pars opercularis/triangularis and inferior precentral gyrus of the frontal operculum and anterior insula extending to the supplementary motor area (SMA) complex [e.g., pre-SMA/SMA of the superior frontal gyrus (SFG) and dorsolateral prefrontal cortex (DLPFC)] ( Fig. 9.4 ). From a motor standpoint, the FAT contributes to the negative motor network to offer inhibitory regulation of motor actions . The FAT is also involved in other higher-order motor function, such as visual–motor activities, construction praxis, orofacial movements, as well as executive function abilities. Injury to the FAT during resective surgery has been linked to transient higher-order motor impairment involving these domains , such as an SMA-like syndrome and rarely Foix-Chavany-Marie Syndrome .
Language outcome
Patients with language-dominant IE often have poor baseline language, particularly involving expressive functions such as naming and verbal fluency. While baseline impairment could be related to long-standing epilepsy and structural abnormalities of the insula itself, it is likely predominantly attributed to the extensive atrophy involving a broader network of language-related structures such as the mesial and lateral temporal lobe . New or worsening language impairment occurs in half of patients operated for IE in the dominant hemisphere, though the likelihood of permanent deficit is very rare in well-selected candidates . Although there is converging evidence from lesion, electrostimulation and neuroimaging studies that the insula plays a role in language function, it is unclear if direct resection/injury to the insula itself contributes to language deficits following IE surgery . Because most patients have extra-insular and often temporal resections, it is difficult to ascertain whether postoperative language deficits can be attributed to resection of the insula proper or extra-insular resections/damage. A recent meta-analysis showed that resection of portions of the peri-insular language network in—specifically the STG-is a risk factor for postoperative language impairment in resective surgery for IE, a finding that is not surprising given its essential role in semantic integration as part of the ventral pathway of language . Other subcortical language pathways, such as temporal branch of AF-SLF complex of the dorsal “phonological” stream near the postero-superior insula and the IFOF-UF complex of the ventral “semantic” stream that bottle-necks in the temporal stem near the limen insula (antero- inferior insula) are at risk during peri-insular dissection and resections ( Fig. 9.3 ) . The above-mentioned FAT is also involved in diverse speech and language functions in the dominant hemisphere (e.g., verbal fluency/initiation and inhibition of speech, control of the articulation apparatus, lexical and semantic word selection) and surgical damage to the dominant hemisphere FAT can lead to transient motor language impairment involving these domains that typically recovers within 8–12 weeks .
Neurological outcome
The insula harbors diverse functions and a range of postoperative manifestations similar to those seen following ischemic stroke of the insula are not uncommon and patients should be counseled to the possibility of these symptoms . Neurological manifestations constitute a high proportion of sensory disturbances and including but not limited to: somatosensory deficits (4%), olfactory-gustatory alterations (2%), auditory impairment (1%), and hemineglect (0.5%) .
Cognitive outcome
Because of the insula’s central role in diverse neurological (motor, gustatory), higher-order cognitive and socio-affective functions – through its connections within the DMN and salience networks – a primary concern with surgical removal of the insula and neighboring structures is the impact on these important functions . While there is no evidence that surgical removal of the insula proper results in major cognitive deficits, studies in adults comparing IE to TLE cohorts and normal controls have shown that insular resections are associated with mild cognitive and social impairments that are typically only revealed on detailed neuropsychological testing but not necessarily spontaneously reported by patients. For example, most adult patients (60%) undergoing IE surgery experience a change in appetite which may be related to behavioral signs of dysfunctional interoception and gustatory functions, likely related to the insula’s involvement in the DMN and gustatory centers, respectively . Patients are also more likely to harbor mild to moderate alterations in sensory processing, such as impairments in the “active behavioral responses” for gustatory/olfactory sensory modalities . These patients are also at greater risk of developing altered emotional processing, such as emotional recognition of happiness and surprise as well as poor emotional judgment for neutral memories (oromotor speed) . There is also evidence of impairment for emotional words on color naming speed . However, these subtle neuropsychological impairments do not impact quality of life and while individual patients may have mood complaints following surgery, there is no significant difference in psychological status, depression, anxiety or quality of life following IE surgery compared to their TLE counterparts at the group level .
In children, the postoperative trajectory in cognitive function remains stable or improves following surgery, with no evidence of cognitive decline at the group level . In a cohort of 17 children undergoing resective IE surgery, neurocognitive outcome was unchanged in 62% and improved in 38% . In a study of pediatric patients undergoing multilobar insular resections, 70% had an improvement in neuropsychological outcome . In a recent study of 15 children with IE related to FCD characterized by impaired cognition and several poor prognostic factors, such as early seizure onset, frequent seizures, and widely distributed interictal epileptiform activity, resective surgery was not associated with further decline in cognitive function .
Minimally invasive stereotactic ablative procedure
Minimally invasive stereotactic ablation procedures can be divided into two categories: MRgLA and RFTC. In both procedures, the epileptic focus is thermally ablated by the delivery of current into the insular/peri-Sylvian tissue using either light-diffusing laser probe (e.g., MRgLA) or radio-frequency waves (e.g., RFTC) through SEEG electrode contacts or mono/bipolar tip RF probe . These procedures remain excellent options and are well suited as curative options for patients with IE related to small lesional foci or palliative options in larger networks, particularly when conventional surgery harbors greater surgical risks (e.g., reoperations) or when patients prefer a minimally invasive alternative .
Radio-frequency thermo-coagulation
Brief history
The implementation of stereotactic techniques in the second half of the 20th century paved the way for the first stereotactic ablation procedures in the 1960s that involved lesioning of deep-seated targets, initially of the amygdala to treat behavior disorders and soon thereafter of the thalamus to treat epilepsy as an alternative to open surgical resection . The advancements and eventual widespread adoption of SEEG have concurrently enabled the accurate mapping of the EZ and epileptic networks (ENs) . SEEG-guided RFTC, which involves SEEG exploration with RFTC stereotactic lesioning directly through the recording electrodes, represents the combination of both these technologies . Since the first modern series using SEEG-guided RFTC for DRE in 2004 , SEEG-guided RFTC has been increasingly used for various types of DRE, including IE in children and adults alike .
Technology and technique
From a technical standpoint, SEEG-guided RFTC involves connecting the SEEG electrode to a radiofrequency (RF) generator. The RF generator sends current through a single electrode to a dipole between adjacent electrode contacts on an electrode or between contacts of two adjacent electrodes ( Fig. 9.6 ) . The first step is thus the placement of SEEG electrodes, nuances of which are covered in depth in Chapter 3. It is important to highlight that electrode placement is crucial both in terms of RFTC efficacy, ensuring precise and optimal coverage of the target, and safety, as a minimum of 2 mm distance from MCA vessels to avoid vascular injury and hemorrhage. While the main goal of SEEG is mapping the EZ and function, if there is a high degree of suspicion for an epileptic focus in the insula or opercula, a higher density electrode coverage could be planned with the forethought of potentially ablating and reducing seizure burden or even curing seizures in selected candidates. This approach is particularly useful for relatively small (1–3 gyri, ≤2 cm 3 ) seizure focus restricted to the insula.
A variant of the RFTC technique involves a “high-density focal stereo-array” or “volumetric” RFTC (vRFTC), in which electrodes or monopolar probes are placed in a separate setting to create a convergent, high-density conformational ablation of the EZ ( Fig. 9.6 ) . This technique is usually performed in a second procedure after a first SEEG exploration or after a first failed “standard” SEEG-guided RFTC.
Different electrodes can be used for RFTC, and several are approved for RFTC procedures (Microdeep electrodes, Dixi Medical, Besançon, France; Alcis electrodes, Alcis—Temis Santé, Besançon, France). Examples of RF generators include Radionics (Medical Products Inc., 22 TerryAv., Burlington, MA 01803, United States) and Cosman G3 or G (Cosman Medical, Burlington, MA, United States). The size of the “dipole” ablation depends on the (1) dipole configuration and distance (e.g., between adjacent contacts on 1 electrode vs. between contacts on two different electrodes) and (2) radiofrequency parameters (e.g., current power) . Once the SEEG electrode is connected to the RF generator, the RF current is delivered and increased until the impedance suddenly increases (e.g., power delivered collapses), which typically occurs within seconds as a result of protein coagulation. Current settings increase local temperature to about 80°C within a few seconds, thus producing a lesion around the electrode contact in 10–30 seconds . Each dipole between adjacent contacts on a SEEG electrode creates an ablation size of about 6×3 mm or a volume of around 30–60 mm 2 . Each electrode can thus create a “bead” shaped ablation along its length ( Fig. 9.6 ). The ability to coalesce between ablations depends on current power (W) and distance between contacts, with the greatest likelihood of coalescing between contacts at a lower power of 3 W (compared to 5 or 7 W, for example) for up to 10 mm intercontact distance ( Fig. 9.6 ). Ablation coalescence does not occur when intercontact distance is greater than 15 mm. Greater ablation volume can be created by using dipoles between contacts on adjacent electrodes rather than contacts on the same electrode .
Because vRFTC involves the placement of three times more trajectories than SEEG-guided RFTC (average six trajectories for 20 ablations with vRFTC versus three trajectories for six ablations with SEEG-guided RFTC), the total ablation volume obtainable with vRFTC is at least twice that of standard SEEG-guided RFTC (e.g., up to 10 cm 3 with vRFTC). However, the total ablation size is half that obtained with MRgLA, which has an average ablation volume of about 13 cm 3 and maximal ablation of up to 30 cm 3 ( Table 9.1 ) . The insular cortex has a volume of 5–8 cm 3 and a complete insulectomy with vRFTC requires approximately up to 30 coagulations and over 5–10 trajectories . The resulting lesion after multiple ablations is often larger than the addition of individual spheres due to peri-lesional edema and secondary degeneration of the tissues located in the gap between thermal .
Indications
The use of SEEG-guided RFTC begins with the decision to proceed with a SEEG exploration, the indications of which are covered in Chapter 3. While SEEG exploration is performed in 70% of ILE cases to map function and the EN, in rare instances it can be performed with the main goal of carrying out RFTC . It is important to highlight that SEEG-guided RFTC distinguished itself from other modalities as it can be classified as a diagnostic/prognostic in addition to therapeutic tool.
The goal of SEEG-guided RFTC is to ablate the insular EN without incurring functional deficits. Thus, the first step is to perform functional mapping of potential RFTC dipoles by bipolar electrical stimulations during the video-SEEG recording session . Ictal patterns necessary to reach the threshold for warranting therapeutic targeting with RFTC (e.g., with higher specificity and likelihood to respond), include low voltage fast activity (LVFA) or spike-wave discharges at seizure onset . Highly specific interictal repetitive rhythmic spiking in cases of Type II FCD can also be used .
For patients in whom an insular/peri-Sylvian focus has been confirmed with SEEG, RFTC can help prognosticate the likelihood of seizure freedom following an eventual resection. The temporary abolition or reduction in seizures shows a high positive predictive value (PPV, 93%) of seizure improvement with a more extensive open resection. The absence of improvement, however, does not preclude the benefit of a broader open resection (positive negative value, PNV, 43%). This “indication” is thus extremely useful when considering resective surgery of the insula, which is—as discussed above-associated with a significant rate of temporary neurological impairment and a nonnegligible rate of permanent deficits. Providing the patient with the experience of seizure reduction or freedom can help inform the decision-making process and counsel and outweigh the benefits of seizure reduction against the risks of deficits.
SEEG-guided RFTC can also be a therapeutic alternative in patients who are candidates for open resection but with very focal EZ (e.g., <2 cm³) and favorable predictive factors. A cure can sometimes be obtained in these cases with a focal, constrained EZ in which the volume of ablation covers the full extent of the EZ such as in patients with small tumors or a purely insular focus restricted to 1–2 gyri.
While SEEG-guided RFTC is generally not considered a therapeutic indication in IE with extensive EZ, as is often seen in pediatric settings with extensive FCD, it can be considered as a palliative treatment in patients who are not candidates for open surgery such as nonlesional patients with diffuse insular involvement or large FCD involving insula and peri-Sylvian structures of eloquent language or motor pathways. In these cases, targeting nodes in the network can reduce the seizure burden .
Outcome of SEEG-guided RFTC
The data on SEEG-guided RFTC for IE is limited to a handful of single-center studies totaling about 70 patients ( Table 9.2 ). Compared to open surgery and MRgLA, RFTC is less effective but likely safer, associated with reduced risk of temporary and permanent deficits. It is readily applicable given the high proportion of IE patients who require SEEG investigation ( Table 9.2 ). Most groups have carried out SEEG-guided RFTC using the conventional method (e.g., RFTC through already placed SEEG contacts). The effective accuracy and safety of SEEG are now well documented . Although the hemorrhagic risk of SEEG placement is low, including a 0.2% risk of asymptomatic hemorrhage and a 0.05% risk of permanent neurologic deficit or death, the risk is greater with a higher number of electrodes (particularly over 13) and this should also be considered when planning more dense coverage for RFTC.
Related posts:
Evidence in pediatric epilepsy surgery
Controversies in the timing of pediatric epilepsy surgery: is earlier better?
Epilepsy in eloquent cortex: resection versus responsive neurostimulation
Lesional epilepsy: lesionectomy versus ECoG-guided resection
Temporal lobe epilepsy with preserved function: multiple hippocampal transection versus neuromodulation (deep brain stimulation, responsive neurostimulation)
Corpus callosotomy: anterior two-thirds (two-stage) versus complete (one-stage)
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
