35 Radiosurgical Treatment for Epilepsy
Radiosurgery is, by definition, a neurosurgical procedure that uses stereotactically focused, converging, narrow ionizing beams to induce a desired biological effect in a predetermined target, with minimal radiation to the surrounding tissues and without opening the skull. The increased worldwide use of gamma knife surgery (GKS) to treat various pathologies has rendered the side-effect profile of radio-surgery rare, generally transient, and quite easily predictable.1 Once it is established that resection of a small deeply seated lesion has a significant risk for surgical complications or functional worsening, GKS must be discussed as an alternative.
The first radiosurgical treatments for epilepsy surgery were performed by Talairach in the 1950s.2 As early as 1974, Talairach reported on the use of radioactive yttrium implants in patients with mesial temporal lobe epilepsy (MTLE) without space-occupying lesions and showed a high rate of seizure control in patients with epilepsies confined to the mesial structures of the temporal lobe.2 In 1980, Elomaa3 promoted the idea of the use of focal irradiation for the treatment of temporal lobe epilepsy based on the preliminary reports of Tracy, Von Wieser, and Baudouin.4,5 Furthermore, clinical experience of the use of GKS and linac-based radiosurgery in arteriovenous malformations and cortico-subcortical tumors revealed an anticonvulsive effect of radiosurgery in the absence of tissue necrosis.6–8 A series of experimental studies in small animals confirmed this effect9,10 and emphasized a relationship to the dose delivered.11,12
The Department of Functional Surgery in Marseille is a major referral center for epilepsy surgery and radiosurgery and has reported the first comprehensively evaluated series of MTLE successfully operated by GKS. The first case of GKS for MTLE was treated in 1993 and reported in 1994 by this group.13 Several prospective trials from this group have demonstrated
The safety efficacy of this approach,14,15
A very specific timetable of the clinical and radiological events,14,16
The importance of the anterior parahippocampal cortex for seizure cessation,17,18
The importance of the marginal dose (24 Gy) for efficacy,18
The feasibility of sparing verbal memory with GKS in dominant-side epilepsy,15 and
The nonlesional mechanism of action of radiosurgery.19
Recently, all these findings have been confirmed by a prospective trial in the United States.20 The Marseille group has treated 155 patients with epilepsy surgery using GKS radio-surgery among a total of 8,590 GKS procedures since 1993. The majority of these patients presented with MTLE (56 patients) and hypothalamic hamartoma (HH, 77 patients). The rest of the patients had severe epilepsy associated with small benign lesions such as cavernous malformations (42 patients) for which an epileptic zone was considered to be confined to the surrounding cortex.21
Seizure cessation may be generated by a specific neu-romodulatory effect of radiosurgery, without induction of a significant amount of histological necrosis.13,16,19,22,23 The selection of the appropriate technical parameters (e.g., dose, volume target) allowing us to accurately obtain the desired functional effect without histological damage remains an important challenge. A detailed review of these cases, as well as other clinical and experimental data, suggests that the use of radiosurgery is beneficial only to those patients in whom a strict preoperative definition of the extent of the epileptogenic network has been achieved,24 and where strict rules of dose planning have been followed.25 The strategy must be to define the patients in whom the safety/efficacy ratio makes radiosurgery advantageous or at least comparable to craniotomy and cortical resection.
Mesial Temporal Lobe Epilepsy
The first GKS operations for MTLE were performed in Marseille in March 1993. Because there was no similar experience available in the literature at that time, we were obliged to define our treatment criteria and technical choices based on present experience of radiosurgery for other pathological conditions at that time. We treated four patients with different technical choices in terms of dose, volume, and target definition. Then we observed some significant radiological changes in these patients several months after radiosurgery.16 This finding led us to stop such treatment, and we decided to follow the long-term results of these first four patients. Long-term results in these patients were reassuring with documented clinical safety of the procedure and gradual disappearance of the acute magnetic resonance imaging (MRI) changes several months later. Therefore, we again decided to treat a new series of patients under strict prospective controlled trial conditions. This classic planning (Fig. 35.1) was based on the use of two 18-mm shots, covering a volume of around 7 mL at the 50% isodose (24 Gy). The results were impressive with a high rate of seizure ces-sation.14,26 Then, we redefined our treatment target by focusing it on the parahippocampal cortex and sparing significant parts of the amygdaloidal complex and hippocampus to decrease morbidity. We further reduced the dose from 24 Gy to 20 and 18 Gy at the margin to find a dose that would create smaller amount of transient acute MRI changes to further refine our GKS technique. However, we observed that this reduction in the dose also caused a significant decrease in the rate of seizure cessation. We currently use 24 Gy for MTLE and recently reviewed our long-term follow-up data of our first 15 patients who were operated by GKS for MTLE at the state of the art (24 Gy). The mean follow-up in this group was 8 years and the seizure-free outcome rate was 73%, which is comparable with open surgery. We found no permanent neurological deficit except some visual field deficits in nine patients.27
We inform our patients before the procedure regarding the delayed efficacy of radiosurgery as its main drawback and define the typical course of the seizures after GKS treatment in three stages: no significant change in seizure frequency for the first few months, then a rapid and dramatic increase in auras for several days or weeks, and finally disappearance of the seizures. We observed that seizure cessation mostly occurred around the 8th to 18th month, although it may occur as late as 26 months after GKS, as we saw in one patient. We usually consider minimum duration after radiosurgery to assess the effect of GKS as 2 years. We observed the same pattern of MRI changes in all our patients unrelated to the amount of marginal dose (18–24 Gy) and volume of treatment (5–8.5 mL). However, the degree of these changes and the delay in their onset may vary according to the dose delivered to the margin, the volume treated, and the individual patient. If there is no or minimal initial radiological changes or clinical benefit, then the recommendation is to wait for the onset of the MRI changes and their subsequent disappearance. To allow an optimal evaluation, we recommend that subsequent microsurgery should not be considered before the third year after radiosurgery. If this is the case, then the reason for the treatment failure should be investigated carefully. We identified likely causes for GKS failure based on our file review as follows:
Poor patient selection (e.g., patients with epilepsy involving more than the MTL structures),
Too early surgical intervention with a diagnosis of “treatment failure” (< 3 years after radiosurgery),28
Targeting the amygdala and hippocampus instead of para-hippocampal cortex,29 and
Dose
Initially the targets in functional GKS radiosurgery (capsu-lotomy, thalamotomy of ventral intermediate (VIM) or the centromedian nuclei, pallidotomy) were treated using a high dose (300–150 Gy nuclei) delivered in very small volumes (3–5 mm in diameter).31 Then Barcia-Salorio et al documented a small and heterogeneous group of patients treated with different types of radiosurgical techniques and variable dosimetry according to the patient.32 Some of these patients were treated with very large volumes and very low dosage (approximately 10 Gy). These results led several other teams to consider using very low doses, as low as 10 Gy to 20 Gy at the margin, but to expect the same efficacy as the 24 Gy protocol (at the margin). This was the dose that we used for our first series of patients with MTLE.14 However the real rate of seizure cessation was only 36% (4/11) among the 11 patients reported by Barcia-Salorio et al, and this was much lower than what we would expect with resection in MTLE.32,33 In another study, Yang et al confirmed that only a very low rate of seizure control is achieved when low doses (from 9 to 13 Gy at the margin) are used based on the result of a heterogeneous group of 176 patients.30 Again, a recent de-escalation study has demonstrated poorer results in patients receiving doses of 18 or 20 Gy at the margin compared with 24 Gy.17,34 This finding is significant because any radiosurgical strategy associated with a much lower rate of seizure cessation in MTLE is unacceptable because of the high rate of seizure freedom achievable by surgical resection. Fractionated stereotactically guided radiotherapy was used by Grabenbauer et al in 12 patients; none of the patients become seizure free and only seizure reduction has been obtained in this series.35,36
Target
If the radiosurgical target is a lesion, then it can be precisely defined radiologically and the question of the selection of the marginal dose can be quite easily addressed by correlating safety and efficacy with individual outcome to the marginal dose and can be refined based on stratification according to volume, location, age, etc. However, this is not the case in patients presenting with MTLE for two reasons: there is no consensus regarding the required extent of mesial temporal lobe resection for good seizure control, and the concept of MTLE syndrome with a stable extent of the epileptogenic zone that can be defined as surgical target is controversial.37,38 The volume, in association with marginal dose, is well known to be a major determinant of the tissue effect in radiosurgery, as shown in integrated risk/dose volume formulas.39 Therefore, target determination is critical in the effectiveness of GKS in MTLE patients. In the first series of patients we treated, our marginal isodose volume was ~7mL (range 5–8.5 mL). In a recent published study, authors tried to correlate dose/volume, degree of MRI changes, and seizure control rates.34 In this study, it was shown that the higher the dose and the volume were, the higher the risk was of having more severe MRI changes, as well as the higher the chance was of achieving seizure cessation. It is clear that efficacious dose-planning strategies with smaller prescription isodose volumes need more precise definition of the essential targets in mesial temporal lobe. However this is a difficult task. There is growing evidence in the current literature that defines the organization of the epileptogenic zone as a network. According to this hypothesis, epileptogenic zone includes several different and possibly distant structures that discharge simultaneously at the onset of the electro-clinical seizures. This perspective helps to explain the high failure risk in a simple lesionectomy without preoperative investigations in the management of severe drug-resistant epilepsies associated with a benign lesion.40 This has been also reported in MTLE cases.37,38 Therefore precise definition of target in radiosurgery is very critical. Wieser et al analyzed the postoperative MRIs of patients who were operated by Yasargil and had amygdalohippocampectomy.41 In this study, they were able to correlate the degree of the resection of each substructure of the mesial temporal lobe and correlate the result with the seizure outcome.41 They reported that only the resection amount of the anterior parahippo-campal cortex was correlated strongly with a higher chance of seizure cessation.41 We also tried to perform a similar study in patients treated with GKS radiosurgery.34
Patient Selection
In a previously published study, Whang and Kwon reported seizure cessation in only 38% (12/31) of the patients who were treated for epilepsy associated with slowly growing lesions.42 This observation as well as the arguments we summarized previously emphasizes the importance of pre-operative definition of the extent of the epileptic zone and its relationship with the lesion.40,43 Therefore, the philosophy in our institution is to choose appropriate investigation techniques and management strategies in each case individually. In some patients, the electroclinical data, structural and functional imaging, and neuropsychological examination may be sufficiently concordant, and surgery of the temporal lobe is proposed without depth electrode recording. In other cases, preoperative assessment results may be discordant to define MTLE reliably, and a stereoelectroencephalographic (stereo-EEG) study is performed. Stereo-EEG implantation is used to assess the reliability of the primary hypothesis (mesial epileptogenic zone) or alternative hypothesis (early involvement of the temporal pole, lateral cortex, basal cortex, insular cortex, or other cortical areas). The purpose is to record the patient’s habitual seizures to define the temporo-spatial pattern of cortical involvement during these seizures. In these patients, depth electrode recording allows us precise tailoring of the extent of surgical resection according to the temporo-spatial course of the seizures. Furthermore, if depth electrode investigation enables us to define a particular subtype of MTLE, then further tailoring of the treatment volume, even reducing it, becomes feasible. Because the main limitation of radiosurgery is the size of the target (prescription isodose volume), the requirement for precision and accuracy to define the epileptogenic zone is higher in this technique. Although this requirement makes the GKS planning for MTLE very challenging, it also makes it the most selective surgical therapy modality for this patient group.
Complications
It is well known that the radiotherapy in young patients has been associated with a significant rate of cognitive decline44,45 and tumorogenesis,46 including some carcinogen-esis.47 However, such reported cases48–50 are extremely rare and frequently do not meet the classic criteria to define the tumors as radiation induced.51 Even if this risk exists, it is most likely approximately 1/10 000, which is far lower than the mortality risk associated with temporal lobectomy.52–56 Epilepsy may cause sudden unexplained death in epileptic patients, and the death rate among epilepsy patients is higher than in the general population.57,58 This risk is especially higher in patients treated with more than two antiepileptic drugs and with a lower IQ (<70). The surgical treatment of epilepsy provides a possibility of immediate seizure cessation and reduces the mortality risk to that of the general population.58 Conversely, this risk persists for a longer period of time in the patients who received GKS treatment because of the delayed benefits of radiosurgical treatment in the epilepsy patient. Therefore, we systematically inform our patients about this disadvantage of radio-surgery.
Current Indications
Radiosurgery for MTLE is still an investigational treatment. The advantages of radiosurgery are the comfort and non-invasiveness of the procedure; avoiding general anesthesia and surgical complications, including mortality; the very short hospital stay; and, finally, the immediate return to the previous function level and employment. Whether or not radiosurgery provides a better result in sparing memory function is still a matter of debate and needs to be confirmed with further comparative studies. Long-term efficacy and safety of radiosurgery also needs to be documented. Micro-surgical management of MTLE provides very satisfactory results because of the rarity of surgical complications and the high rate of seizure freedom. Therefore, the most appropriate treatment modality should be chosen carefully, and the patient should clearly understand the advantages, disadvantages, and limitations of both modalities. The patient should be able to understand the limits and constraints of radiosurgery very well. In our opinion, the most important selection parameter for GKS is the demonstration of the purely mesial location of the epileptogenic zone. Another good candidate for GKS is the patient with proven MTLE who has had a surgical failure because of insufficient extent of resection. Overall, the best candidates are young patients with moderately severe epilepsy, socially well-adapted people with a high functioning level and quite a high risk of memory deficit with open surgery (such as MTLE on the dominant side with a subtle hippocampal atrophy and slight preoperative deficit in verbal memory). Postoperative memory deficit exposes these patients to potentially huge social and professional consequences; therefore, GKS constitutes a very good alternative treatment modality for this patient group.
Hypothalamic Hamartomas
HHs are rare, congenital heterotopic lesions that are intrinsically epileptogenic when closely connected to the mammillary bodies.59,60 Patients classically present with gelastic seizures during the first years of life.61 In the more severe forms of the disease, affected patients develop an epileptic encephalopathy during the following years that is characterized by drug resistance, various types of seizures including generalized seizures and drop attacks,61 cognitive decline,62–64 and severe psychiatric co-morbidity.65 Usually the seizures begin early in life and are often particularly drug resistant from the onset. Commonly, seizure semiology suggests the involvement of temporal or frontal lobe region and even secondary epileptogenesis. HHs may also be asymptomatic or associated with precocious puberty, associated with neurological disorders (including epilepsy, behavior disturbances, and cognitive impairment), or both.
The natural history is unfavorable in the majority of the patients because of behavioral symptoms (particularly aggressive behavior) and mental decline, which occur as a direct effect of the seizures.66 Interestingly, in our experience, the reversal of these behavioral symptoms after radiosurgery appears to begin even before complete cessation of the seizures and appears to be correlated to the improvement in background EEG activity. It is the authors’ speculation that these continuous discharges lead to the disorganization of several systems, including the limbic system, and that their disappearance accounts for the improvement seen in attention, memory, cognitive performance, and impulsive behavior. In these cases, the role of radiosurgery in the reversal of the behavioral symptoms may be as or more important than its effect on decreasing seizure occurrence. Consequently, we consider that it is essential to operate on these young patients as early as possible, whatever the surgical approach considered (resection or radiosurgery).
Even though the first successful and safe removal of an HH was reported by Paillas et al in 1969,59 the interest for the surgical cure of this specific group of patients developed only in the 1990s. According to Valdueza et al, epilepsy in HH is observed only in medium/large sessile HH broadly attached to tuber cinereum or mammillary body.67 The microsurgical resection in this critical area is related to a significant risk of oculomotor palsy, hemiparesis, and visual field deficit.68,69 The first clinical series evaluating microsurgical resection using pterional and midline frontal approaches did not emphasize complications.67,70–73 However, in 2002, Palmini and coworkers analyzed the patients with HH who were treated in several of the best centers for epilepsy surgery around the world, and they reported severe complications after micro-surgical resection in 7 of 13 patients.74 These complications included four thalamocapsular infarcts with contralateral hemiplegia, transient third nerve paresis in four subjects, a central diabetes insipidus, and a nonreversible hyperphagia.74 Conversely, they also confirmed the efficacy of surgery in this pathological condition. More specifically, 3 patients showed complete seizure cessation and the remaining 10 subjects had more than 90% reduction in the frequency of their seizures.
The rationale for the application of disconnective surgeries in the treatment of HH is that the lesion is not a neoplasm, and its removal is therefore not mandatory. Another factor favoring disconnection technique is the possibility of avoiding the complications that may occur during the dissection in the cisterns, a maneuver necessary for the micro-surgical resection. Delalande et al68 actually stressed this point as favoring the simple disconnection of an HH instead of its complete excision because of the occurrence of severe complications in their first patient. When the clinical result is not satisfactory and the upper part of the lesion is mainly in the third ventricle, Delalande proposed a second step via an endoscopic approach to the third ventricle. In 2003, the same author published a series of 17 patients with a follow-up between 1 month and 5.4 years.68 A second intervention (usually endoscopic) was necessary in 8 patients. In this excellent series, 47% of the patients (8 of 17) were seizure free, including three patients who were operated on twice. The author reported some permanent severe complications, namely, one case of hemiplegia, one case of hemiparesis, two cases of hyperphagia, one case of panhypopituitarism, one case of hypothyroidism, and another case with growth hormone deficiency. Transient morbidity included one case of meningitis and two cases of diabetes insipidus. In addition, the author reported a postoperative frontal lobe ische-mic complication, which was apparently asymptomatic. In conclusion, only six patients (35%) were seizure free with no permanent deficit. Contrary to the other reports describing the results with transcallosal approach, Delalande did not observe any memory deficit. Finally, the author observed a correlation between completeness of disconnection and control of the seizures.
We retrospectively analyzed the results of radiosurgical management in a series of 10 HH patients collected from centers around the world.75 The excellent safety–efficacy ratio (all improved, 50% cured, and no adverse effects except one case of poikilothermia) in this series led us to organize a prospective multicenter trial. In this trial, we prospectively evaluated 55 patients and published the results in a preliminary report.60 In this study, preoperative cognitive deficits and behavioral disturbances were investigated, and the relationship of seizure severity and anatomical type as well as cognitive abilities were characterized.62 The goal of the preoperative workup was to adequately select the candidates for inclusion and to evaluate the baseline neurological and endocrinological functions. All radiosurgical procedures were performed using the Leksell 201-source Cobalt 60 Gamma Knife (Elekta Instrument, Stockholm, Sweden). We consistently used multi-isocentric complex dose planning of high conformity and selectivity. We also used low peripheral doses to take into account the close relationship with optic pathways and hypothalamus (median 17 Gy; range 13 to 26 Gy). We paid special attention to the dose delivered to the mamillary body and to the fornix, and we tried to tailor the dose plan for each patient on the basis of the use of a single run of shots with the 4-mm collimator. The lesions were generally small (median 9.5 mm; range 5–26). Patients were evaluated with respect to seizures, cognition, behavior, and endocrine status at 6, 12, 18, 24, and 36 months after radiosurgery and then every year. There was satisfactory follow-up for 27 patients. Among these patients, 10 were seizure free (37%) and 6 were very much improved (22.2%) with a significant seizure reduction (usually only rare residual gelastic seizures) associated with a dramatic behavioral and cognitive improvement. Overall, an excellent result was obtained in 60% of the patients. According to our policy, the patient and the family are offered a second radiosurgery in case of partial benefit when the lesion is anatomically small and well defined. Five patients (18.5%) with small hamarto-mas were only modestly improved and are being considered for a second session of radiosurgery. Two of them reported no significant improvement until now. A microsurgical approach has been performed in 4 patients (14.8%) with quite large HH and poor efficacy of radiosurgery. Of these patients, 2 are cured and 2 failed to respond. The radiosurgical treatment has been performed twice in 9 patients.
HH Classification and Treatment Strategy
Topological classification of the lesion is a key feature in the decision-making process60 (Fig. 35.2). This should be done based on a good, high-resolution MRI study. Previous classifications have been based on anatomical76–78 or surgical76 considerations. These classifications failed to describe the large diversity of these lesions and their therapeutic consequences. As underlined by Palmini and coworkers, the exact location of the lesion and its relation with interpeduncular fossa and third ventricle walls is critical. These topographical characteristics correlate with the extent of excision, seizure control, and complication rate.12 Because of this, we classify the HH according to its topology based on our original classification.45 In our experience, this classification correlates very well with the clinical semiology and its severity. It is especially critical for determining the appropriate surgical strategy. According to this classification, six types of HH have been described.
Type I lesions correspond to small HHs located inside the hypothalamus that also extend more or less in the third ventricle. This group includes the best candidates for GKS, because the morbidity of microsurgical removal in this group of patient is potentially high.
In type II, the lesions are small and mainly located in the third ventricle. Radiosurgery is again certainly a safer alternative in this group of patients (Fig. 35.3). Even though the endoscopic and transcallosal interforniceal approaches have been well described, the risks of short-term memory worsening, endocrinological disturbance (hyperphagia with obesity, low thyroxin, sodium metabolism disturbance), and thalamic or thalamocapsular infarcts have been reported even in the hands of highly skilled and experience neurosur-geons. However, in exceptional cases with a very severe recurrent status epilepticus, we still recommend open surgery through either a transcallosal interforniceal approach or an endoscopic approach for these patients. If the lesion is small and the third ventricle is large, then an endoscopic approach is a reasonable option.
In type III, the lesions are located essentially in the floor, and we prefer GKS for these patients because of the extremely close relationship between the mammillary body the fornix and the lesion. We believe that sessile HHs always more or less have an extension into the hypothalamus close to the mammillary body. Thus, when a lesion is classified as a type II or III, that means that the lesion appears on the MRI as being located in the third ventricle and is likely to have a root in the hypothalamus.
In type IV, the lesion is generally sessile in the cistern, and surgical disconnection can be preferred using pteri-onal approach with or without orbitozygomatic osteotomy. However, if the lesion is small, GKS can be recommended because of its safety and its ability to simultaneously treat the small associated part of the lesion in the hypothalamus itself that is frequently visible on the high-resolution MRI. In Delalande’s series, only two patients among 14 are seizure free after a single disconnection through a pterional ap-proach.76 Consequently, we only recommend this approach in the cases with too large lesions for GKS and as a first step of a staged approach. In most circumstances, the patient is improved but stays still not seizure free after the first surgical step and GKS is planned at 3 months as a second step in the treatment of these cases.
Type V constitutes a group of lesions with pedicle. These lesions are rarely epileptic and can be easily cured by radiosurgery or disconnection through a pterional approach. In cases of severe epilepsy, surgical disconnection will certainly allow a faster seizure cessation than GKS. However, a distant extension of the lesion in the hypothalamus must be cautiously searched on high-resolution MRI, because the presence of an extension close to the mammillary bodies will eventually lead to recommending GKS that would provide treatment for both parts of the lesion. This is especially true in cases in which the cisternal component is small.
Type VI includes giant HHs. These patients are not good candidates for up-front radiosurgery, and a combination of several therapeutic modalities should be used in nearly all cases. Even if GKS does not appear to be suitable when the lesion is large, radiosurgical disconnection can be considered. The result of radiosurgical targeting of only the superior part of the lesion located in the hypothalamus or the third ventricle and leaving the lower part of the lesion below the floor untreated has been uniformly disappointing. In our opinion, this strategy may result in a loss of a precious developmental stage of a child, and we do not recommending this approach. Again a staged approach can be considered if microsurgical resection leaves a small remnant in the third ventricle, and the patient is still not seizure free. We do recommend GKS in these patients as well.
Related posts:
15 Implantation of Strip, Grid, and Depth Electrodes for Invasive Electrophysiological Monitoring
18 Transsylvian Selective Amygdalohippocampectomy
22 Surgical Approaches in Cortical Dysplasia
34 Cortical and Deep Brain Stimulation
28 Transsylvian Hemispheric Deafferentation
32 Hippocampal Transection
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
