From resection to ablation: A review of resective surgical options for temporal lobe epilepsy and rationale for an ablation-based approach
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.201662
Source of Support: None, Conflict of Interest: None
Surgical intervention is of proven benefit in an appropriately selected subset of patients with medically refractory temporal lobe epilepsy. In these patients, a surgical cure both provides the quality of life improvement that comes from seizure freedom as well as a survival benefit. However, patients who undergo open surgical intervention may have a worsening in neurobehavioral outcomes. Laser interstitial thermal therapy (LITT) represents a minimally invasive surgical intervention that has shown promise in improving post-operative neurobehavioral outcomes. Further, the minimally invasive nature of this procedure holds the possibility to shift the significant under-penetration of surgical intervention that exists for eligible medically refractory patients. Herein, we review open surgical resection-based techniques and the clinical data to date for LITT.
Keywords: Anterior temporal lobectomy, disconnective procedures, laser interstitial thermal therapy, neuromodulation, selective amygdalohippocampectomy,radiosurgery, temporal lobe epilepsy
Despite an ever increasing number of available drugs over the past several decades, approximately 30% of adult epileptic patients remain medication resistant (>2 failed medications). For this group of patients without an effective medical treatment, surgical intervention is often of proven benefit. This is most strongly evidenced for epileptic foci localized to the temporal lobe; Class I evidence from 2001 supported 58% freedom from seizures impairing awareness at 1 year in patients who received an anterior temporal lobectomy (ATL). ATL, the gold standard surgical treatment for temporal lobe epilepsy, refers to the removal of the mesial (amygdala, hippocampus, entorhinal cortex) and lateral neocortical structures of the anterior temporal lobe either in stages or en bloc. In an ongoing effort to both optimize seizure freedom and minimize the neuropsychological deficits related to surgical intervention, alternative resection-, ablation-, disconnection, and neuromodulation-based approaches have been explored.
When compared to the ATL, all other surgical interventions attempt to preserve the lateral neocortical structures and adjacent white matter pathways. In the case of neuromodulation and disconnection-based approaches, an attempt is made to preserve elements of the hippocampus as well. The alternative resective approach, the selective amygdalohippocampectomy (selAH), utilizes either a transsylvian or transcortical corridor to the mesial structures with the aim of reducing white matter transgression. Disconnective procedures include hippocampal multiple subpial transection and hippocampal disconnection. Access, however, still requires some amount of white matter disruption. While early reports of hippocampal multiple subpial transection (MST) have been encouraging,,, more are needed; studies to date for hippocampal disconnection have indicated an elevated rate of visual and motor complications. Neuromodulation-based approaches have either been used as a palliative treatment to reduce seizure frequency and severity when a lesional target cannot be identified (e.g., deep brain stimulation of anterior thalamic nuclei [DBS ANT],,); or, alternatively when a lesional target cannot be safely resected (e.g., responsive neurostimulator). Targeted ablative approaches for treatment of pathology of the central nervous system have been explored in humans for over the past 70 years, enabled through the development of human stereotactic frames. In the recent past, stereotactic ablative approaches have benefitted enormously from improvements in imaging technology. This has allowed radiosurgical tissue ablation as well as stereotactic probe insertion to achieve thermocoagulation through either radiofrequency (RF) or laser energy [laser interstitital thermal therapy (LiTT)]. While both are minimally invasive approaches, the latter benefits from near-real time control of the ablation zone through incorporation of magnetic resonance (MR) thermometry, an algorithm that allows estimation of the ablation zone. Here, we will briefly review seizure and neurobehavioral outcomes from traditional resection-based treatment approaches to serve as a platform for a discussion of the LiTT procedural workflow and outcomes published to date.
Outcome for resection-based approaches in systematic reviews
Comparisons on seizure outcome between ATL and selAH have been completed in multiple studies. Herein, we will focus on the data reviewed in two meta-analyses that were contemporaneously published., In 2013, Josephson et al., pooled 1203 patients across thirteen studies and five centers. This pooled analysis contained six prospective evaluations, one multicenter investigation, and no randomized controlled trials. The minimum follow-up was 12 months in 11 of these studies and was not specified in 2 studies. A modest seizure freedom improvement was observed for ATL over selAH, with an estimation that 13 patients would need to receive ATL for 1 patient to achieve an Engel Class I outcome, a result that otherwise would not be possible with selAH. The authors conclude by recommending the completion of a randomized controlled trial given that this benefit was of uncertain clinical worth when neurobehavioral outcomes were considered.
A similar conclusion was reached in a meta-analysis completed by Hu et al. These authors completed a separate systematic review encompassing a total of 13 studies. Nine studies purely evaluated seizure outcome, 2 studies explored the effect of the approach on intelligence quotient, and the remaining studies explored a combination of both. In this study, the authors noted a modest reduction in Engel Class I outcome with selAH as compared to ATL (66% vs 71%). They further failed to detect a difference in measures of intelligence quotient (IQ) between ATL and selAH. Given that many of the included studies were retrospective and nonrandomized, and also that the benefit of ATL over selAH was modest, the authors cautioned against making large inferences regarding the superiority of ATL over selAH. They also recommended completion of a prospective RCT to generate a direct comparison.
A final similar meta-analysis was completed by Kuang et al., in 2014. Echoing the findings of the above studies, these authors reviewed six studies that included a total of 626 patients,337 patients with SelAH and 289 patients with ATL H. The authors failed to detect a significant difference between groups in relation to postoperative seizure outcomes. Interestingly, two of the studies supported evidence for verbal memory improvement at 1 year postoperatively, and two studies indicated a decrement. In the pooled analysis, no association with improvement or worsening was attributable to the method of surgical intervention.
Neurobehavioral outcomes – Rationale for a minimally invasive ablative approach
The selAH was intended to achieve comparable seizure-free outcomes to the ATL while minimizing the neurobehavioral deficits seen in ATL. This was to be accomplished by minimizing trauma to the adjacent white matter pathways and neocortical structures. When variants of the selAH were first advocated, the strongest evidence for the role of the anterior temporal lobe in verbal function and object recognition was by way of deficits in these functions seen in a series of postsurgical ATL patients. Subsequently, the anatomic importance of lateral temporal neocortical structures and preserved functional connectivity for these functions has been supported by prospective surgical series with a neurobehavioral focus,, and in neuroimaging studies. In aggregate, these studies support a “functional adequacy” hypothesis indicating that postoperative decline in neurobehavioral outcomes results from iatrogenic injury to the neocortical structures or adjacent pathways. This is supported by imaging and histologic evaluations indicating injury to adjacent neocortical structures through both selAH approaches., In pooled analysis of a large number of patients, summarized in the prior section, the selAH demonstrates marginally worse seizure-free outcomes while showing a small improvement in neurobehavioral outcome. Patients have reliably been able to note the presence of subsequently quantified verbal deficits and have indicated attendant decrements in quality of life measures. It is in this setting that alternate innovative surgical approaches to treatment of mTLE are explored to realize the benefits of seizure freedom while mitigating the risk for neurobehavioral deficits. The recent clinical experience with LiTT indicates an alternate approach to the hippocampus and amygdala that, when coupled with modern imaging modalities, can be used to minimize injury to adjacent structures. Our institutional experience to date with LITT is provided herein.
A precedent – Radiofrequency ablation
Stereotactic access to the amygdala and hippocampus along the long axis of these structures has been considered as early as the 1970s. Small, updated radiofrequency ablation series were published in the late 1990s, reflecting improvements in targeting and imaging technologies., Subsequent expanded series both explored the outcomes of radiofrequency ablation alone  and as compared to the gold standard surgical treatment, ATL. These series provided important foundational work for the utilization of LITT by demonstrating (1) the reproducibility of stereotactic access to the long axis of the hippocampus, (2) the relative safety of adjacent structures (e.g., cerebral peduncles, cranial nerves) during mesial temporal thermocoagulation, and (3) the potential to generate postoperative outcomes (e.g. seizure, neurobehavioral) that were comparable to ATL. Compared to radiofrequency ablation, LITT has the critical advantage of therapy titration in the MR scanner based upon ablation estimates to the target structure and temperature estimates to surrounding tissue with MR thermometry.
The methods have been previously described in a recent report by Kang et al. All patients receiving LITT have been prospectively tracked at Thomas Jefferson University Hospital, including those who received a standard presurgical evaluation. The patients in this review included patients with complex partial seizures, mesial temporal sclerosis present on MRI on fluid-attenuated inversion recovery (FLAIR) imaging, hippocampal atrophy on thin-section T1 coronal MR imaging, and hypometabolism on 18-flourodeoxyglucose positron emission tomography (FDG-PET). These patients also had a preoperative electroencephalographic evaluation that was in agreement. Concordant with prior publication of our series, all surgical candidates with isolated hippocampal sclerosis were offered either LITT or anterior temporal lobectomy. Given the possibility of improved neurocognitive outcome, patients considered for intervention on the dominant side were encouraged to consider LITT. If there was clinical evidence (e.g. multiple seizure semiologies) or radiographic evidence (bilateral hippocampal sclerosis) of dual pathologies, patients received invasive evaluation (e.g. stereoencephalography) prior to offering a resective or ablative procedure. If there was concomitant evidence of lateral temporal pathology, patients were offered a temporal lobectomy. Surgical failures were considered for re-intervention with LITT, progression to ATL, or invasive monitoring. LITT was favored if there was clear evidence of significant remaining amygdala or hippocampus and there was no change in the seizure semiology. ATL was favored if post-failure semiology remained stable and the ablation was deemed adequate. Invasive monitoring was favored if there was no clear imaging evidence supporting the reason for surgical failure or if there was a semiologic change in the postoperative seizures. According to our management practice, any temporal lobe surgical failure that demonstrated evidence of dual pathology and postoperative seizures from the contralateral temporal lobe were not candidates for further resection or ablation. These patients would be considered for an alternative neuromodulatory procedure (e.g., responsive neurostimulation (RN); Neuropace (Mountainview, CA). Neuropsychological testing and electroencephalography were performed on every patient preoperatively; however, this data was not present on all cases postoperatively. All patients underwent LITT, and some patients who persisted with seizures post-procedure were offered and underwent ATL subsequently. Follow-up evaluation was performed by patient visit, or in some cases, by phone.
Gross et al.,have previously provided a detailed description and flow chart-based algorithm to describe the patient selection process. Separate articles have been published exploring the procedural workflow for LITT., Please find a pictorial review of the workflow in the accompanying [Figure 1]. To briefly review, successful LITT completion comprises two primary steps (1) stereotactic trajectory planning to achieve satisfactory ablation of the desired structures (e.g., amygdala, hippocampus, and entorhinal cortex); and, (2) thermocoagulation of the desired target structures while ensuring the relative protection of nearby critical structures. The considerations for both these steps are discussed.
Stereotactic placement of the laser fiber was achieved with a Cosman-Roberts-Wells (CRW) stereotactic frame (Integra Neurosciences, Plainsboro, NJ, USA) Trajectories were selected to maximize the cannulation of the hippocampus. Early in the experience, the target aimed more inferiorly towards the entorhinal cortex and the latter group had targets aiming more towards the amygdala complex. This was not systematically studied with the exception of conducting post-operative analysis of the volume of resection. We acknowledge that the change in targeting could have affected the seizure results.
An adequate ablation can often be achieved with the use of a single laser fiber passed in a single trajectory down the long axis of the amygdala and hippocampus. The laser fiber is encased within a cooling jacket, and once at the target depth, can be activated at different points along the cooling jacket. Therefore, with long axis cannulation, a comparable amount of amygdala and hippocampus can be ablated, as is commonly achieved with an ATL or selAH. Commonly, ablations can include the hippocampal destruction back to either the lateral mesencephalic sulcus or the collicular plate. This is facilitated by a medial trajectory that allows maximal cannulation of the mesial structures. Anatomic features that generate steric constraints on the ablation trajectory include the presence of cortical vasculature, ventricular size, and the location of the collateral sulcus. The latter two features create a relative bottleneck to the posterior entry to the hippocampus at the coronal level of the quadrigeminal cistern. Prohibitive features of these structures may result in a need to explore increasingly flat and lateral trajectories. Occasionally, more than one orthogonal (or near orthogonal) trajectory may be required if an appropriate long axis approach cannot be found. In our institutional experience, only one trajectory was used in each case. Of importance, transventricular trajectories should be considered carefully or avoided as ventricular transgression may make parenchymal re-entry difficult. This may result in attendant targeting inaccuracy with parenchymal re-entry at the level of the amygdala. While the ventricle is directly adjacent to target structures, the cerebrospinal fluid acts as a heatsink, preventing effective ablation of the target structures if the laser fiber is in the cerebrospinal fluid. One of the most pronounced advantages of the LITT technology over radiofrequency ablation is the ability to complete the procedure in the MRI scanner. This provides the following critical features: (1) near-real time estimate of the ablation territory through use of MR thermometry; and, (2) an automatic shutoff if predefined critical structures pass preset temperature thresholds.
Patients presented here represent an extension of those discussed in the article by Kang et al. These represent patients who received LITT for temporal lobe epilepsy. Patient demographics are provided in [Table 1]. Lengthened follow-up is provided for the initial twenty patients in their study. A total of 26 patients (12 males, 14 females) have undergone the LITT procedure at our institution for completing a selective laser amygdalohippocampectomy between December 2011 and August 2016. [Table 2] summarizes the clinical outcomes for these patients. Fourteen of the 26 patients (53.8%) are currently seizure-free (Engel 1A, B outcome) with follow-up ranging from 0.2 to 4.2 years. Six of the 13 patients followed for 2 years (46%), 9 of the 19 patients followed for 1 year (47%), and 13 of the 23 patients followed for 6 months (56.5%) were seizure free. In patients who had a seizure following LITT, the average time to seizure relapse was 2.6 months; 10 of the 12 patients had a seizure relapse in less than 6 months; the remaining 2 patients relapsed in less than 1 year.
Early published series exploring the clinical outcomes of LITT procedure have focused on seizure freedom rates, neurobehavioral outcomes, impact on healthcare utilization, and have attempted the comparison of the procedure with the resective options (e.g. ATL, selAH). Willie et al., prospectively evaluated results in an early series of patients with mesial temporal lobe epilepsy (n = 13, 9 with MTS). At a median follow-up of 14 months (range 5 to 26), 7/13 (54%) and 6/9 MTS (67%) had an Engel Class I outcome. The median hospitalization was 1 day and the median ablation volume was 60%. The only neurologic complication was visual field deficit related to deviation of a stereotactically aligned rod. Subsequently, our group, Kang et al., have published a prospective evaluation (n = 20, 17 with MTS). At the time of publication, the median follow-up was 13.4 months (range 1.3 months to 3.2 years). Published Engel I seizure freedom rates were 8/15 (53%) at a 6 month follow-up, 4/11 (36.4%) at a 1 year follow-up, and 3/5 (60%) at a 2 year follow-up. A few patients had intermediate Engel classification, indicating an immediate improvement or no improvement. Three of the 4 patients who subsequently underwent ATL had an Engel I classification (follow up 2.7 to 25.4 months). In both the studies, patients who went on to seize postoperatively did so early. In the study by Willie et al., all observed seizures occurred within the first six months. Kang et al., found that 4/6 (67%) of failures occurred within the first 6 months. The seizure outcome presented above represents an extension of the same findings provided in the experience presented by Kang et al. Seizure freedom rates have been stable in a greater proportion of patients at more than a two-year follow-up. In ten of twelve patients who have gone on to seize in our institutional dataset, seizures occurred within 6 months. If this holds true as patient series grow larger, it will have implications for patient counseling. Of note, in patients in our institution who had a subsequent failure of LITT and had received a revision ATL, it was found that having had a previous LITT procedure complicates the revision surgery. This occurs both due to a change in the tissue consistency and because of a loss of surgical anatomic landmarks in the mesial temporal lobe.
Neurobehavioral outcomes following LITT have also been explored. In the aforementioned study by our group, Kang et al., a subset of 6 patients with a 6-month follow-up, completed contextualized and noncontextualized memory tasks and their results were comparison with that of a baseline study. While the patients had a decline in noncontextualized tasks, the addition of a contextualized language component improved the testing results. This indicated a preserved role for lateral neocortical language function. Separately, Drane et al., compared LITT patients with a control ATL cohort, and failed to find neuropsychological worsening in LITT patients in the assessed measures (object recognition in the nondominant lobe, and naming in the dominant lobe), which was at variance with the finding that was commonly observed in ATL patients. Given the above results, the findings obtained in our study were perhaps related to the contextualized nature of this data. Therefore, while previously discussed studies have failed to detect a significantly improved seizure outcome or neurobehavioral advantage when comparing between the resective surgical options, neurobehavioral investigations when assessing patients who have undergone the LITT procedure indicate that a LITT-tailored ablation does preserve white matter pathways to lateral temporal structures. This allows for improvements in memory scores to occur and accounts for contextualization of information, using the functions of adjacent white matter tissue that help with recall of information.
The publication of Class I evidence supporting ATL in patients with medically refractory temporal lobe epilepsy was truly a watershed moment. This evidence simultaneously established a substantial opportunity for patients suffering from medically refractory seizures, raised in them the possibility of a life that provided complete freedom from seizures and set a high bar for future interventions to overcome. The selAH was subsequently developed and evaluated to mitigate the potential for neurobehavioral decline which was thought to be related to collateral damage to adjacent neocortical structures and their white matter pathways. Comparison in multiple meta-analyses have failed to support the superiority of selAH on the grounds of improved seizure outcome or neurobehavioral outcomes. LITT has been rapidly adopted within the epilepsy surgery community as a targeted ablative approach to achieve destruction of the mesial temporal structures (e.g., hippocampus, amygdala, entorhinal cortex) while sparing the adjacent tissue. This has been made possible by a combination of technological advances in imaging algorithms (e.g., MR thermometry) and energy delivery (e.g., highly controllable laser-based heating). The early published outcome data, including those studies published at our centres, are encouraging with a possible modest reduction in an Engel I outcome but with evidence of improvement in the neurobehavioral outcome. This appears related to sparing of adjacent white matter pathways and neocortical function. The efficacy of LITT for mesial temporal sclerosis is currently under investigation in a multicenter randomized controlled study. While primary outcomes pertain to seizure freedom, secondary outcomes focus prominently on neuropsychological outcomes. This Class I evidence will both serve as a direct comparator to the previously generated Class I evidence supporting the anterior temporal lobectomy and should begin to provide more definite guidance regarding the role of lateral neocortical preservation in neurobehavioral outcomes.
Unequivocally, LITT-based treatment is better tolerated than open surgical treatment. With an incision that can be closed with a stitch, many patients could conceivably be treated on an outpatient basis. They are currently kept in the hospital overnight. This represents a significant reduction in healthcare resource utilization. Finally, while anecdotal, it is clear that the introduction of a minimally-invasive first-line treatment approach has resulted in an influx of patients who have not previously considered surgical intervention. This is of critical importance because, despite the strong data supporting surgery for medically refractory temporal lobe epilepsy, temporal trends suggest that there is an enormous under-penetrance of surgical treatments for this patient population. In this setting, and because the ability to later consider ATL is not sacrificed as a revision alternative, threshold analysis studies suggest that LITT is an appropriate first line therapy even if more definitive ongoing investigations demonstrate seizure freedom rates to be less than ATL or selAH. Considering that epilepsy surgery can be curative and that it remains severely underutilized, the successful outreach of this minimally invasive procedure to a new subset of patients (medically refractory but unwilling to consider surgery) is perhaps the greatest success of all. For these patients, the only true alternative can be medical management.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest
[Table 1], [Table 2]