Seizure 52 (2017) 81–88
Contents lists available at ScienceDirect
Seizure journal homepage: www.elsevier.com/locate/yseiz
Temporal lobe surgery for intractable epilepsy in children: What to do with the hippocampus? Mony Beniflaa,* , Odeya Bennet-Backb , Zamir Shorerc , Iris Noymanc , Rima Bar-Yosefd , Dana Eksteind a
The Neurosurgical Pediatric Unit, Rambam Health Care Campus, Haifa, Israel Pediatric Neurology Department, Shaare-Zedek Medical Center, Jerusalem, Israel Pediatric Neurology Department, Soroka Medical Center, Beer-Sheva, Israel d Neurology Department, Agnes Ginges Center for Human Neurogenetics, Hadassah Medical Center, Jerusalem, Israel b c
A R T I C L E I N F O
A B S T R A C T
Article history: Received 25 February 2017 Received in revised form 7 July 2017 Accepted 28 September 2017 Available online xxx
Purpose: Resection of the hippocampus can cause verbal memory decline, especially in the pediatric population. Thus, preservation of the hippocampus can be crucial for the quality of life of children with intractable temporal lobe epilepsy (TLE) who are candidates for epilepsy surgery. We investigated techniques that determine whether the hippocampus is part of the epileptogenic zone and the outcomes of pediatric surgery aimed to spare the hippocampus. Methods: We accessed data of children with normal hippocampus on MRI, who underwent surgery for medically refractory TLE. To identify epileptogenic areas, electrocorticography was performed in patients with space occupying lesions adjacent to the hippocampus, and long term invasive monitoring in patients with nonlesional TLE. Postoperative seizure control was classified according to Engel I-IV; Class I indicates seizure-free. Results: Eleven females and 11 males met study inclusion criteria; the mean age at surgery was 11.3 years. Cortical and hippocampal electrocorticography was performed in 15 patients and long term invasive hippocampal monitoring in seven. The hippocampus was preserved in 16 patients (73%) while hippocampectomy was performed in 6 (27%). At the end of a mean follow-up of 3.5 years, 94% (15/16) of the patients who did not undergo hippocampectomy were classified as Engel I, compared to 50% (3/6) who underwent hippocampectomy. Conclusion: Sparing the hippocampus in temporal lobe epilepsy surgery is possible with excellent seizure outcome, while using the proper intraoperative technique. © 2017 British Epilepsy Association. Published by Elsevier Ltd. All rights reserved.
Keywords: Temporal lobe epilepsy Hippocampus Electrocorticography Invasive monitoring Pediatric neurosurgery
1. Introduction Temporal lobe epilepsy (TLE) is the cause of medically intractable seizures in 8 to 20% of pediatric epilepsy cases [1,2]. Good seizure outcome, 2–5 years after temporal lobe surgery, was reported in 72–86% of patients [3–6]; yet outcome has been shown not to remain stable over time and to decline to 41–67% in follow-
Abbreviations: TLE, temporal lobe epilepsy; GTR, gross-total resection; AH, amygdalohippocampectomy; ATL, anterior temporal lobectomy; DNET, dysembryoplastic neuroepithelial tumor. * Corresponding author at: The Department of Pediatric Neurosurgery, Rambam Health Care Campus, Haifa, Israel. E-mail addresses: benifl[email protected] (M. Benifla), [email protected] (O. Bennet-Back), [email protected] (Z. Shorer), [email protected] (I. Noyman), [email protected] (R. Bar-Yosef), [email protected] (D. Ekstein).
up of more than 10 years, both in children [6–8] and adults [9]. Three systematic reviews that analyzed mostly data of adult patients concluded that tailored surgery for lesional TLE that included the amygdala and hippocampus conferred better seizure outcomes than gross-total resection (GTR) alone [10–12]. Better epilepsy outcomes have been reported even when the hippocampus appeared normal in MRI studies [13] and when tumors did not invade the hippocampus [14]. In children, data are more sparse and less conclusive. Some pediatric studies have claimed better seizure outcomes following more extensive surgical resection [6,15–17]. However, no benefit in seizure outcomes was observed in a cohort of children who underwent tailored resection according to epileptogenicity compared to those who underwent GTR alone [18]. Moreover, unsatisfactory seizure results were reported in pediatric patients who underwent amygdalohippocampectomy
https://doi.org/10.1016/j.seizure.2017.09.020 1059-1311/© 2017 British Epilepsy Association. Published by Elsevier Ltd. All rights reserved.
82
M. Benifla et al. / Seizure 52 (2017) 81–88
(AH) and left-sided surgery, compared to patients who had anterior temporal lobectomies and lesionectomy alone [4]. The above studies were retrospective, and did not generally account for the presence and location of lesions in relation to the hippocampus. When a lesion is cortically based and is not near the mesial structure, lesionectomy alone can achieve good seizure outcome, also in children [16]. When the lesion involves the hippocampus or when the mesial structures look abnormal in the MRI there is a clear indication to resect the hippocampus and amygdala in pediatric patients [19]. On the other hand, it is unclear whether the mesial temporal structures should be resected when the lesion is in close proximity to a hippocampus that appears normal on MRI. The role of hippocampectomy in nonlesional temporal lobe epilepsy is also unclear. Moreover, the potential effects of resection of the medial temporal lobe structures and the risk factors for postoperative verbal memory decline in children have not been extensively investigated. Skirrow et al. [20] recently reported that better verbal memory was linked to greater post-surgical residual hippocampal volumes, mainly in the left hemisphere [20]. The goal of temporal lobe epilepsy surgery is to resect the epileptogenic zone while preserving the functional areas, specifically the hippocampus. In this paper we aimed to assess the decision making process regarding resection of the hippocampus in children with intractable temporal lobe epilepsy and without a lesion in the hippocampus as per MRI. We expected that patients who did not undergo hippocampectomy would have noninferior seizure outcomes compared to patients who underwent hippocampectomy.
2. Methods We reviewed the medical charts of all children ( < age 18 years) who underwent temporal lobe surgery for drug resistant epilepsy by one surgeon (MB) in two departments of neurosurgery between January 2008 and December 2015. We included in the study patients with temporal lesions involving the gyri adjacent to the hippocampus and with normal hippocampal appearance on MRI. We also included patients with nonlesional epilepsy and normal MRI. To state explicitly, children with cortical based lesions that were distant from the hippocampus by at least one gyrus gap and children with mesial temporal sclerosis and any hippocampal abnormality as per MRI were excluded from the analysis. MRI showing a normal hippocampus, on both sides, was a study inclusion criterion, as was the availability of follow-up data for at least one year. We reviewed the medical records for demographic data, age at seizure onset, age at surgery, and postsurgical seizure types and frequency. We also recorded the preoperative electroencephalogram (EEG) and video EEG (VEEG), neuroimaging studies and functional investigations. The operative notes were reviewed for intraoperative complications, before and after resection electrocorticography (ECoG) and pathology of the temporal resections. All children were with medically refractory epilepsy, had undergone a comprehensive presurgical evaluation and had been previously evaluated by EEG. All patients had undergone prolonged VEEG monitoring and MRI in a mesial temporal sclerosis protocol. For some patients, fMRI scanning was used to determine cerebral dominance for language.
Fig. 1. Illustrative case of a patient (#1) with nonlesional left temporal lobe epilepsy. Subdural grid electrodes with strips and depth electrodes were implanted (A). The depth electrodes aimed to the anterior (left panel) and posterior (right panel) hippocampus, as demonstrated in coregistration of post implantation CT and preoperative MRI (B).
M. Benifla et al. / Seizure 52 (2017) 81–88
2.1. Intraoperative ECoG For patients with space occupying lesions (SOL), ECoG was performed using a 20-electrode surface array (Ad-Tech Medical Instrument Co., Racine, WI). Another four-electrode strip array was placed over the hippocampus to record the pre-excisional epileptiform discharges. When the lesionectomy was completed, post-excisional ECoG was performed from both the cortex and the hippocampus, and according to the residual epileptiform discharges. Decisions were made regarding the requirement for further cortical resection or amygdalohippocampectomy.
83
decision as to the need for hippocampal resection was made on the basis of the invasive monitoring data alone. However, in some cases, intraoperative electrocorticography was conducted in addition, to guide the extent of neocortical resection. 2.3. Postoperative seizure outcome For the purposes of this study, postoperative seizure control was classified using the Engel scheme: Class I, seizure-free postoperatively; Class II, rare disabling seizures; Class III, worthwhile improvement; and Class IV, no worthwhile improvement.
2.2. Invasive monitoring 2.4. Data analysis For patients with nonlesional TLE, long term invasive monitoring was performed. After performing the craniotomy and the dural opening, the central sulcus was identified indirectly by determining the inversion polarity (phase reversals) on somatosensory evoked potential recording after median nerve electrical stimulation. Then, confirmation of the precentral (motor) gyrus was achieved using direct cortical stimulation. Sixty-four electrode surface array subdural grids (Ad-Tech Medical Instrument Co.) were placed to cover the lateral surface of the temporal and frontal lobes, using data from the semiology of the seizures, interictal and ictal scalp EEG recordings, interictal spike sources, and speech areas as detected by fMRI. In addition, ipsilateral subdural strips were placed to cover the anterior temporal tip and the inferior surface of the temporal lobe. Depth electrodes were placed to capture ictal data from the anterior greater regions of the temporal lobe, which could be provided by the subdural grid per se, specifically the anterior and posterior hippocampus, by frameless stereotaxic navigation (Fig. 1). The patients were monitored for 5 days for interictal epileptiform discharges and for ictal seizure onset. Mapping of motor, sensory and language functions was performed in 1–2 sessions on the 3rd or 4th day after grid implantation. After obtaining sufficient data regarding the functional cortex and regions of epileptogenesis, the subdural grid was removed and epileptic regions were resected. The
To determine factors that predict seizure outcome, comparison between groups, Pearson x2 and Fisher’s exact tests were used. To identify predicting factors of seizure outcomes, further statistical analyses were performed. These analyses included the following clinical variables: 1) age at seizure onset; 2) duration of seizures before surgery; 3) age at surgery; 4) whether or not the mesial structures were resected and 5) histological features (for example, neoplastic versus nonneoplastic status). Each of these variables was entered into a univariate logistic regression analysis in which the dependent variable was Engel Class I vs. Engel Class II–IV outcome. In this analysis, a p < 0.05 was used to indicate statistical significance. 3. Results 3.1. Patients Table 1 describes the patient demographics including gender, age of onset of seizures, age at time of surgery, developmental exam and hand dominance. Eleven females and 11 males met study inclusion criteria. The mean age at the time of surgery was 11.3 years.
Table 1 Demographic and presurgical characteristics. Patient
sex Age at seizure onset (months)
Seizure Age duration at (months) surgery (months)
f/u Hand (months) dominance
fMRIlanguage
Side of surgery
pre op NP
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
f f m F m f f m m m m f m f F
59 69 39 59 0 60 88 6 170 132 137 170 100 84 65
157 141 72 75 12 78 96 82 180 192 160 176 108 132 89
98 72 33 16 12 18 8 76 10 60 23 6 8 48 24
84 72 68 36 42 60 26 37 20 34 57 48 47 64 12
Rt Rt Rt Rt NA Rt Lt Rt Rt Rt Rt Rt Rt Rt Lt
Lt Lt Lt
Lt Lt Lt Lt Rt Lt Lt Lt Lt Rt Rt Rt Lt Rt Rt
16 17 18 19 20 21 22
F F F M M M M
108 190 166 158 166 151 131
120 205 198 183 212 205 141
12 15 32 25 46 54 10
12 30 32 29 19 16 55
Rt Rt Rt Rt Rt Rt Rt
Lt Bil
Weak verbal memory attention deficit Normal Normal ND ND ND Weak expressive language ND Below average verbal and non verbal performance Attention deficit Normal Normal Normal Normal Weak expressive language, weak visual and verbal memory Normal Weak expressive language, attention deficit Normal Low intellect, attention deficit low average intellect Normal Weak verbal skills; low working memory.
Lt Bil Lt
Lt Bil
Lt
Lt
Abbreviations: Bil– bilateral; Lt– left; Rt– right; NA– not applicable; NP– neuropsychology.
Lt Rt Rt Lt Rt Rt Lt
84
M. Benifla et al. / Seizure 52 (2017) 81–88
3.2. Seizure history The mean duration of seizures prior to surgery was 2.8 years. The mean age at the first seizure was 8.7 years. Semiology and preoperative interictal and ictal EEG recordings identified temporal lobe epilepsy in all patients; frontal epileptiform discharges were noted as well in 6 patients. The mean number of anticonvulsants tried before surgery per patient was 5.3. 3.3. Preoperative MRI Lesions were identified or suspected on MRI in 18 patients. For 4, the MRI scans were reported as normal (patients #1,2,3 and 9, Table 1). In all patients the hippocampus was reported as normal on both sides, this being a criterion for study inclusion. For 2 patients, the preoperative MRI suggested a diagnosis of focal cortical dysplasia. Other lesions identified by MRI included cavernous malformation, dysembryoplastic neuroepithelial tumor (DNET), ganglioglioma, low grade glioma and non-specific signal changes. Thirteen children underwent fMRI. Left language dominance was suggested in 9 patients and bilateral language activation in 3. For one additional patient, the fMRI results were unreliable due to the young age of the patient. 3.4. Temporal lobe surgery In 12 patients, a left temporal procedure was performed; and in 10, a right one. In 15 patients with suspected neoplastic lesion on MRI, the hippocampus was exposed and cortical and hippocampal ECoG was performed. A lesionectomy alone was performed in 7 patients, with significant improvement of the ECoG recording both from the cortex and from the hippocampus. In one patient (#5) an arachnoid cyst medial to the hippocampus was fenestrated after the first ECoG and was connected to the temporal horn. A second ECoG was performed about 10 min following the fenestration and no epileptiform discharges were noted (Fig. 2). In 4 patients lesionectomy and anterior temporal lobectomy (ATL) were performed with improvement of the ECoG recording. In another 3 patients (#16,17,20, Table 2) the hippocampal ECoG did not show any change before and after lesionectomy and ATL, so hippocampectomy was performed as well (Fig. 3). In 2 patients the hippocampus was resected on the right side and in the third, on the left. Long term invasive monitoring was performed in 7 children. Epileptic activity arising from the hippocampus, according to the depth electrode recording, was indication for AH. In four patients the preoperative MRI was reported as normal; of them, two (#3,9,
Table 2) underwent hippocampectomy. In 2 patients, cortical dysplasia was suspected: one of them (#14) underwent hippocampectomy. For one girl (#12) the suspected tumor was closely related to the Wernicke area; in a second surgery, she underwent hippocampectomy. For all 7 patients who underwent long term invasive monitoring, subdural grid and depth electrodes were inserted into the mesial temporal structures, or into a structural abnormality, as seen on frameless stereotactic MR imaging, as described. In three patients (#3,9,14, Table 2), contacts located in the hippocampus were part of the seizure onset zone and these children underwent AH. The hippocampus was spared in the remaining four children, in whom the epileptogenic zone was presumed to reside in the neocortex, according to invasive monitoring. 3.5. Neuropathology The neuropathological findings of the specimens removed at surgery for the 22 patients are described in Table 2. For all patients, a neuropathological diagnosis was achieved from the resected specimens. Four children had cortical dysplasia (3–Ib and 1–IIb), 2 cortical dysplasia (both type Ia) and gliosis, 2 had cavernous malformation; 5 a DNET; 4 ganglioglioma; 2 meningioangiomatosis, one pilocytic astrocytoma, one arachnoid cyst and one oligodendroglioma. All tumors were classified as WHO stage I, except the oligodendroglioma, which was WHO stage II. 3.6. Post-operative complications There were no postoperative mortalities. One child (# 18, Tables 1 and 2) underwent same day recraniotomy due to epidural hematoma in the surgical site. After evacuation of the hematoma the child recovered completely. For one child (#12, Tables 1 and 2) who underwent a GTR of the oligodendroglioma, the tumor recurred 4 years after the first procedure (she was seizure free during that period) and a second surgery was performed, this time with resection of the mesial structures. The pathology diagnosis was glioblastoma multiforme and the child passed away one year later. Only the outcome before tumor recurrence was taken into consideration for this report. Two children experienced homonymus superior quadrantopsia that was confirmed by a neuroophthalmologist. 3.7. Seizure outcome The mean follow-up of the 22 patients was 3.5 years. At the end of follow-up, 18 patients were classified as Engel I (82%), 2 as Engel
Fig. 2. A 15 month old boy (#15) with an arachnoid cyst (white arrow) compressing the hippocampus (A, B). The cyst was fenestrated using a transcortical approach through the medial temporal gyrus and opened into the right temporal horn. A post-operative MRI displays the decrease in the volume of the cyst and the release of pressure on the right hippocampus (C). The hippocampus remained intact and the patient is seizure free.
M. Benifla et al. / Seizure 52 (2017) 81–88
85
Table 2 Type of surgery, pathological diagnosis and seizure outcome. Patient
Surgery
Invasive monitoring
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
PH ATL + PH ATL + AH Lesionectomy Cyst fenestration Lesionectomy ATL + lesionectomy ATL + PH ATL + AH Lesionectomy Lesionectomy ATL + lesionectomy Lesionectomy ATL + AH ATL + lesionectomy ATL + lesionectomy + AH ATL + lesionectomy + AH Lesionectomy ATL + lesionectomy Lesionectomy + AH Lesionectomy
y y y
22
ATL + lesionectomy
ECoG
AH
Pathology
ENGEL
y Y y y Y y y
n n y n n n n n y n n n n y n y y n n y n
II I III I I I I I II I I I* I I I I I I I III I
y
n
CD Ia + gliosis CD Ib CD Ib DNET Arachnoid cyst DNET Ganglioglioma CD Ib CD Ia + gliosis Meningio- angiomatosis Ganglio OGD Cavernoma CD IIa Ganglioglioma Ganglioglioma JPA Cavernoma DNET DNET Meningioangiomatosis DNET
y y y y y y y y y y y
I
*This patient was seizure free for 4 years, then had recurrence of the glioma and underwent a second surgery that included AH; he died one year later. Abbreviations: ATL– Anterior temporal lobectomy; AH– Amygdalohippocampectomy; CD– Cortical dysplasia; DNET– Dysembryoplastic neuroepithelial tumor; JPA– Juvenile pilocytic astrocytoma; OGD– Oligodendroglioma; PH– Para hippocampectomy.
II (9%), and 2 as Engel III (9%). Fifteen of the 16 (94%) patients who did not undergo AH were classified as Engel I, and the remaining one as Engel II. Of the 6 patients who underwent AH, only 3 were classified as Engel I; one as Engel II and 2 as Engel III. Of the 2 patients classified as Engel III, one (#3) had nonlesional epilepsy and the mesial structures were resected on the left side, after long term invasive monitoring. The other patient (#20) had right-sided DNET with intraoperative abnormal hippocampal ECoG. He underwent gross total resection with AH, and was seizure free for 3 months. A second surgery with long term invasive monitoring was suggested, yet the family refused. Left-sided surgery was performed in the 2 patients classified as Engel II (#1, #9); both were right-hand dominant with left fMRI-language. A favorable surgical outcome (Engel I and II) was obtained in 3 of the four patients with non-lesional MRI (#1, 2, 3 and 9) and in 17 of the 18 lesional cases. However, only 25% (1/4) of the non-lesional cases achieved Engel I outcome, compared with 94% of the lesional ones. None of the clinical variables tested for prediction of seizure outcome: age at seizure onset, seizure duration, type of surgery, pathological features and age at surgery reached statistical significance. Thus, we could not identify predictors for favorable seizure outcome. 4. Discussion In this study, we reviewed the medical records of 22 children who had surgery for intractable temporal lobe epilepsy and normal hippocampus. The surgeries were intended to resect epileptogenic zones, while attempting to preserve the mesial structures. Overall, favorable outcomes, according to Engel class I, were found in 82% (18/22) of the patients. While this rate appears consistent with the pediatric medical literature, no other study selected patients according to the criteria applied here, specifically a normal appearing hippocampus on MRI; and either a temporal lesion adjacent to the hippocampus or no lesion at all. For pediatric patients with these characteristics, we confirmed our hypothesis of noninferior seizure outcomes following surgery without AH, compared to surgery with AH; Engel I in 94% (15/16) vs. 50% (3/6).
In the procedures documented in this study, ECoG in children with lesions, and invasive monitoring in children without lesions, were used to identify epileptic foci and to determine the extent of resection, including the possibility of hippocampectomy. In their pediatric study, Clusmann et al. [4] suggested that difficulty in precisely locating the epileptogenic focus in children could explain the lack of improvement they observed in seizure outcomes following AH compared to ATL. Indeed, good seizure control was achieved for 83% of their patients with hippocampal sclerosis; and poorer results for children who underwent AH due to a less defined abnormality in the lateral temporal lobe, or due to a tumor, or who had normal histological features or dysplasia. Other studies have reported good seizure outcomes in pediatric patients with hippocampal sclerosis [6,19]; outcomes in children were assessed as comparable to those in adults [21]. Disagreement as to the performance of hippocampectomy arises in cases of no apparent lesion in the hippocampus. Dual pathology [22] has become well recognized as a driver of epileptogenesis in TLE, in both children and adults. The implication is that seizures continue after surgery, even following GTR. Thus, procedures that identify extra-tumoral epileptogenic tissue and tailor resection accordingly have become the mainstay of treatment for TLE, and improving means for identification of epileptogenic zones is a priority. In the current study, all 15 patients with SOL underwent cortical and hippocampal ECoG, and all 12 children in whom the hippocampus was spared based on the ECoG results had Engel I outcome. While the benefit of this procedure in achieving better seizure outcomes is not conclusive, its intraoperative use in identifying epileptogenic areas outside tumors has persisted over many years. Two systematic reviews of observational studies [10,11] reported similar rates of seizure freedom between patients for whom ECoG was and was not used. However, they warranted caution in interpretation of their findings, since ECoG may be used in more severe cases of epilepsy, as was demonstrated in a study of adults that addressed this matter [23]. Pediatric studies have shown positive seizure outcomes following ECoG [16,24], which, in long term follow up, were shown to surpass those of patients who underwent lesionectomy without ECoG [24]. In an earlier study, we
86
M. Benifla et al. / Seizure 52 (2017) 81–88
Fig. 3. Preoperative MRI of patient #17 with temporal lobe lesion (white arrow) compressing the hippocampus (A). The explored hippocampus (H) and strip electrodes over it are shown in B. After lesionectomy, electrocorticography showed significant epileptiform discharges from the hippocampus and the patient underwent hippocampectomy. Post-excisional MRI (C) demonstrates the area of resection and the posterior end of the hippocampal resection.
M. Benifla et al. / Seizure 52 (2017) 81–88
showed that the grading of postexcision ECoG according to epileptiform discharges did not predict seizure outcome [3]. Nonetheless, ECoG continues to be used in our department to guide the extent of resection in pediatric epilepsy surgery. We report seizure free results in 93% (14/15) of patients with SOL compared to 57% (4/7) with nonlesional and dysplastic lesions. In retrospect, some of the pathological lesions of the four MRInegative cases may have been detected on higher Tesla MR's. Epileptogenic zones in the hippocampus were more often detected in nonlesional than lesional patients: 57% (4/7) vs. 20% (3/15). As described, the performance of AH did not necessarily contribute to seizure freedom. Studies in adults show poorer seizure and neuropsychological outcomes in patients without lesions. One study showed better seizure outcomes following lobectomy in TLE patients with lesions than without; outcomes of the latter were similar to patients with lesions outside the excision area [9]. In another study, postoperative memory loss was greater for TLE patients without lesions, according to preoperative MRI and histopathological findings, than for matched TLE patients with abnormal MRI, despite better preoperative memory in the former [25] Overall, according to a meta-analysis that included 697 patients with nonlesional epilepsy and 2860 patients with lesional epilepsy, the odds for becoming seizure free were 2.7 higher for the lesional cases, in both children and adults [26]. Thus, the better surgical outcome in our lesional cases is in line with the literature and is not directly related to whether the hippocampus was resected or spared. The procedure for AH in the current study included resection of the TLE. Selective AH as described by Yasargil et al. [27] was the procedure documented in several studies, including Clusmann et al's pediatric study [4]. A systematic review that analyzed 3 prospective and 10 retrospective studies concluded that seizure free outcomes were better following ATL than selective AH [28] Interestingly, a review of 53 studies did not find any certain surgical approach for TLE to demonstrate an advantage in seizure outcomes [29]. This highlights the importance of clearly defining preoperative characteristics for any comparison between surgical procedures, as was done in the current study. We did not find the side of surgery to correlate with seizure outcome. In contrast, left-sided surgery was found to predict poor seizure outcome among pediatric patients who underwent ATL, and particularly among those who underwent AH [4]. Laterality of the lesion (dominant or nondominant hemisphere) [19,30], neuropsychological assessment [19], intraoperative language mapping [31], and the timing (early or late) of seizure onset [32,33] are factors that have been considered in determining extent of resection. As with most studies on epileptic surgery, the main outcome investigated in this study was freedom from seizures. Postoperative neuropsychology testing was not performed, this being a limitation of the study. A recent retrospective study showed a decrease in verbal memory in children following left temporal lobe surgery that includes mesial structures, particularly when language representation is in the left hemisphere [34]. Importantly, that study did not find a correlation between seizure freedom and postsurgical decline in verbal memory. This highlights the importance of preserving the hippocampus, when possible, especially in children; and calls for more attention to neuropsychology evaluation, which is evidently a distinct outcome from seizure freedom. We emphasize that the findings of this study are relevant to the pediatric population only. Apparently, studies on epilepsy surgery should distinguish clearly between patient and adult populations. One systematic review of TLE surgery concluded that children became seizure-free more often than adults after tumor resection [10] and another reported no difference in seizure outcomes
87
between adults and children [11]. In two studies that compared seizure outcomes between children and adults following similar surgeries, the adults fared better [35,36]. Differences in outcomes between children and adults may stem from differences in preoperative evaluation, such as greater difficulty to evaluate pediatric MRI scans;[4] in clinical characteristics such as less clear ictal EEG pattern [1] or more multifocal discharges in children; [36] in pediatric pathology, such as lower frequency of hippocampal sclerosis and higher frequency of dual pathology and etiologies such as temporal sclerosis, cortical dysplasia, trauma and vascular malformations;[37] and in a broader impact on neurocognitive function in children due to brain development. In children, the plasticity of the brain may promote better recovery, yet the impact of cognitive decline on development may have serious consequences. We did not find duration of epilepsy to correlate with seizure outcomes; however, our cohort was small. Long epilepsy duration has been identified as an important risk factor for surgically refractory seizures [37]. Thus, the consideration of early surgical treatment for medically intractable TLE is recommended for children. Seizure outcome following epilepsy surgery during childhood does not remain stable over time [9,38]. A review of 42 children following temporal lobe surgery revealed a seizure freedom rate of 77% at 1 year, which decreased to 67% at 10 years [8]. A similar decline, from 61% at 1 year to 41% at 10 years was noted in a large retrospective review of 325 adult and pediatric patients who underwent anterior temporal lobectomy [9]. A cohort study of patients who underwent temporal lobectomy during childhood found that 58% of the patients who were seizure free at 5 years follow-up experienced seizure recurrence within the subsequent 10 years [7]. Second surgery may be effective in patients with seizure relapse. At that point, re-evaluation of the epileptogenic zone may lead to hippocampal resection if it was spared in the first operation. The retrospective design is a limitation of this, as well as the vast majority of studies on epilepsy surgery. In particular, analysis of clinical decision making is generally problematic in retrospective studies. However, the fact that the same surgeon performed all the operations during the study period mitigates for such. Other limitations of this study are the relatively small number of patients and the heterogenicity of the lesions. The analysis of seizure outcomes in pediatric patients with similar clinical characteristics and the inclusion of children without lesions are strengths of this study. 5. Conclusion In this paper, we presented seizure outcomes of intraoperative techniques to determine whether the hippocampus is part of the epileptogenic zone in pediatric patients with lesions located close to the hippocampus (but separated from it by at least one gyrus gap) or with nonlesional TLE. Such patients would be subjected to AH according to the literature regarding adults [14]. In the current pediatric cohort, the hippocampus was preserved in 16 patients (73%) while hippocampectomy was performed in 6 (27%). Excellent seizure outcomes were achieved when the hippocampus was spared, implying that hippocampal sparing does not negatively influence epilepsy surgery results in children. Since verbal memory decline can significantly influence a child's quality of life, every effort should be exerted to preserve the mesial temporal structures, mainly the hippocampus. Conflict of interest None of the authors has any conflict of interest to disclose.
88
M. Benifla et al. / Seizure 52 (2017) 81–88
Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. References [1] Nickels KC, Wong-Kisiel LC, Moseley BD, et al. Temporal lobe epilepsy in children. Epilepsy Res Treat 2012;2012:849540. [2] Blume WT. Temporal lobe epilepsy surgery in childhood: rationale for greater use. Can J Neurol Sci 1997;24:95–8. [3] Benifla M, Otsubo H, Ochi A, et al. Temporal lobe surgery for intractable epilepsy in children: an analysis of outcomes in 126 children. Neurosurgery 2006;59:1203–13 discussion 1213-4. [4] Clusmann H, Kral T, Gleissner U, et al. Analysis of different types of resection for pediatric patients with temporal lobe epilepsy. Neurosurgery 2004;54:847–59 discussion 859–60. [5] Mittal S, Montes JL, Farmer JP, et al. Long-term outcome after surgical treatment of temporal lobe epilepsy in children. J Neurosurg 2005;103:401– 12. [6] Lopez-Gonzalez MA, Gonzalez-Martinez JA, Jehi L, et al. Epilepsy surgery of the temporal lobe in pediatric population: a retrospective analysis. Neurosurgery 2012;70:684–92. [7] Jarrar RG, Buchhalter JR, Meyer FB, et al. Long-term follow-up of temporal lobectomy in children. Neurology 2002;59:1635–7. [8] Benifla M, Rutka JT, Otsubo H, et al. Long-term seizure and social outcomes following temporal lobe surgery for intractable epilepsy during childhood. Epilepsy Res 2008;82:133–8. [9] McIntosh AM, Kalnins RM, Mitchell LA, et al. Temporal lobectomy: long-term seizure outcome, late recurrence and risks for seizure recurrence. Brain 2004;127:2018–30. [10] Englot DJ, Han SJ, Berger MS, et al. Extent of surgical resection predicts seizure freedom in low-grade temporal lobe brain tumors. Neurosurgery 2012;70:921–8 discussion 928. [11] Englot DJ, Berger MS, Barbaro NM, et al. Factors associated with seizure freedom in the surgical resection of glioneuronal tumors. Epilepsia 2012;53:51–7. [12] Bonney PA, Glenn CA, Ebeling PA, et al. Seizure freedom rates and prognostic indicators after resection of gangliogliomas: a review. World Neurosurg 2015;84:1988–96. [13] Bell ML, Rao S, So EL, et al. Epilepsy surgery outcomes in temporal lobe epilepsy with a normal MRI. Epilepsia 2009;50:2053–60. [14] Ghareeb F, Duffau H. Intractable epilepsy in paralimbic word health organization grade II gliomas: should the hippocampus be resected when not invaded by the tumor? J Neurosurg 2012;116:1226–34. [15] Bilginer B, Yalnizoglu D, Soylemezoglu F, et al. Surgery for epilepsy in children with dysembryoplastic neuroepithelial tumor: clinical spectrum, seizure outcome, neuroradiology, and pathology. Childs Nerv Syst 2009;25:485–91. [16] Ogiwara H, Nordli DR, DiPatri AJ, et al. Pediatric epileptogenic gangliogliomas: seizure outcome and surgical results. J Neurosurg Pediatr 2010;5:271–6. [17] Babini M, Giulioni M, Galassi E, et al. Seizure outcome of surgical treatment of focal epilepsy associated with low-grade tumors in children. J Neurosurg Pediatr 2013;11:214–23.
[18] Brahimaj B, Greiner HM, Leach JL, et al. The surgical management of pediatric brain tumors causing epilepsy: consideration of the epileptogenic zone. Childs Nerv Syst 2014;30:1383–91. [19] Cataltepe O, Turanli G, Yalnizoglu D, et al. Surgical management of temporal lobe tumor-related epilepsy in children. J Neurosurg 2005;102:280–7. [20] Skirrow C, Cross JH, Harrison S, et al. Temporal lobe surgery in childhood and neuroanatomical predictors of long-term declarative memory outcome. Brain 2015;138:80–93. [21] Mohamed A, Wyllie E, Ruggieri P, et al. Temporal lobe epilepsy due to hippocampal sclerosis in pediatric candidates for epilepsy surgery. Neurology 2001;56:1643–9. [22] Spencer S, Huh L. Outcomes of epilepsy surgery in adults and children. Lancet Neurol 2008;7:525–37. [23] Southwell DG, Garcia PA, Berger MS, et al. Long-term seizure control outcomes after resection of gangliogliomas. Neurosurgery 2012;70:1406–13; discussion 1413-4. [24] Gelinas JN, Battison AW, Smith S, et al. Electrocorticography and seizure outcomes in children with lesional epilepsy. Childs Nerv Syst 2011;27:381–90. [25] Helmstaedter C, Petzold I, Bien CG. The cognitive consequence of resecting nonlesional tissues in epilepsy surgery–results from MRI- and histopathologynegative patients with temporal lobe epilepsy. Epilepsia 2011;52:1402–8. [26] Tellez-Zenteno JF, Hernandez Ronquillo L, Moien-Afshari F, et al. Surgical outcomes in lesional and non-lesional epilepsy: a systematic review and metaanalysis. Epilepsy Res 2010;89:310–8. [27] Yasargil MG, Antic J, Laciga R, et al. Microsurgical pterional approach to aneurysms of the basilar bifurcation. Surg Neurol 1976;6:83–91. [28] Hu WH, Zhang C, Zhang K, et al. Selective amygdalohippocampectomy versus anterior temporal lobectomy in the management of mesial temporal lobe epilepsy: a meta-analysis of comparative studies. J Neurosurg 2013;119:1089– 97. [29] Schramm J. Temporal lobe epilepsy surgery and the quest for optimal extent of resection: a review. Epilepsia 2008;49:1296–307. [30] Phi JH, Kim SK, Cho BK, et al. Long-term surgical outcomes of temporal lobe epilepsy associated with low-grade brain tumors. Cancer 2009;115:5771–9. [31] Benifla M, Otsubo H, Ochi A, et al. Temporal lobe surgery for intractable epilepsy in children: an analysis of outcomes in 126 children. Neurosurgery 2006;59:1203–13 discussion 1213-4. [32] Goldring S, Rich KM, Picker S. Experience with gliomas in patients presenting with a chronic seizure disorder. Clin Neurosurg 1986;33:15–42. [33] Fried I, Kim JH, Spencer DD. Limbic and neocortical gliomas associated with intractable seizures: a distinct clinicopathological group. Neurosurgery 1994;34:815–23 discussion 823-4. [34] Law N, Benifla M, Rutka J, et al. Verbal memory after temporal lobe epilepsy surgery in children: do only mesial structures matter? Epilepsia 2016. [35] Datta A, Sinclair DB, Wheatley M, et al. Selective amygdalohippocampectomy: surgical outcome in children versus adults. Can J Neurol Sci 2009;36:187–91. [36] Lee YJ, Kang HC, Bae SJ, et al. Comparison of temporal lobectomies of children and adults with intractable temporal lobe epilepsy. Childs Nerv Syst 2010;26:177–83. [37] Lee YJ, Lee JS. Temporal lobe epilepsy surgery in children versus adults: from etiologies to outcomes. Korean J Pediatr 2013;56:275–81. [38] Hauptman JS, Pedram K, Sison CA, et al. Pediatric epilepsy surgery: long-term 5-year seizure remission and medication use. Neurosurgery 2012;71:985–93.
Contents lists available at ScienceDirect
Seizure journal homepage: www.elsevier.com/locate/yseiz
Temporal lobe surgery for intractable epilepsy in children: What to do with the hippocampus? Mony Beniflaa,* , Odeya Bennet-Backb , Zamir Shorerc , Iris Noymanc , Rima Bar-Yosefd , Dana Eksteind a
The Neurosurgical Pediatric Unit, Rambam Health Care Campus, Haifa, Israel Pediatric Neurology Department, Shaare-Zedek Medical Center, Jerusalem, Israel Pediatric Neurology Department, Soroka Medical Center, Beer-Sheva, Israel d Neurology Department, Agnes Ginges Center for Human Neurogenetics, Hadassah Medical Center, Jerusalem, Israel b c
A R T I C L E I N F O
A B S T R A C T
Article history: Received 25 February 2017 Received in revised form 7 July 2017 Accepted 28 September 2017 Available online xxx
Purpose: Resection of the hippocampus can cause verbal memory decline, especially in the pediatric population. Thus, preservation of the hippocampus can be crucial for the quality of life of children with intractable temporal lobe epilepsy (TLE) who are candidates for epilepsy surgery. We investigated techniques that determine whether the hippocampus is part of the epileptogenic zone and the outcomes of pediatric surgery aimed to spare the hippocampus. Methods: We accessed data of children with normal hippocampus on MRI, who underwent surgery for medically refractory TLE. To identify epileptogenic areas, electrocorticography was performed in patients with space occupying lesions adjacent to the hippocampus, and long term invasive monitoring in patients with nonlesional TLE. Postoperative seizure control was classified according to Engel I-IV; Class I indicates seizure-free. Results: Eleven females and 11 males met study inclusion criteria; the mean age at surgery was 11.3 years. Cortical and hippocampal electrocorticography was performed in 15 patients and long term invasive hippocampal monitoring in seven. The hippocampus was preserved in 16 patients (73%) while hippocampectomy was performed in 6 (27%). At the end of a mean follow-up of 3.5 years, 94% (15/16) of the patients who did not undergo hippocampectomy were classified as Engel I, compared to 50% (3/6) who underwent hippocampectomy. Conclusion: Sparing the hippocampus in temporal lobe epilepsy surgery is possible with excellent seizure outcome, while using the proper intraoperative technique. © 2017 British Epilepsy Association. Published by Elsevier Ltd. All rights reserved.
Keywords: Temporal lobe epilepsy Hippocampus Electrocorticography Invasive monitoring Pediatric neurosurgery
1. Introduction Temporal lobe epilepsy (TLE) is the cause of medically intractable seizures in 8 to 20% of pediatric epilepsy cases [1,2]. Good seizure outcome, 2–5 years after temporal lobe surgery, was reported in 72–86% of patients [3–6]; yet outcome has been shown not to remain stable over time and to decline to 41–67% in follow-
Abbreviations: TLE, temporal lobe epilepsy; GTR, gross-total resection; AH, amygdalohippocampectomy; ATL, anterior temporal lobectomy; DNET, dysembryoplastic neuroepithelial tumor. * Corresponding author at: The Department of Pediatric Neurosurgery, Rambam Health Care Campus, Haifa, Israel. E-mail addresses: benifl[email protected] (M. Benifla), [email protected] (O. Bennet-Back), [email protected] (Z. Shorer), [email protected] (I. Noyman), [email protected] (R. Bar-Yosef), [email protected] (D. Ekstein).
up of more than 10 years, both in children [6–8] and adults [9]. Three systematic reviews that analyzed mostly data of adult patients concluded that tailored surgery for lesional TLE that included the amygdala and hippocampus conferred better seizure outcomes than gross-total resection (GTR) alone [10–12]. Better epilepsy outcomes have been reported even when the hippocampus appeared normal in MRI studies [13] and when tumors did not invade the hippocampus [14]. In children, data are more sparse and less conclusive. Some pediatric studies have claimed better seizure outcomes following more extensive surgical resection [6,15–17]. However, no benefit in seizure outcomes was observed in a cohort of children who underwent tailored resection according to epileptogenicity compared to those who underwent GTR alone [18]. Moreover, unsatisfactory seizure results were reported in pediatric patients who underwent amygdalohippocampectomy
https://doi.org/10.1016/j.seizure.2017.09.020 1059-1311/© 2017 British Epilepsy Association. Published by Elsevier Ltd. All rights reserved.
82
M. Benifla et al. / Seizure 52 (2017) 81–88
(AH) and left-sided surgery, compared to patients who had anterior temporal lobectomies and lesionectomy alone [4]. The above studies were retrospective, and did not generally account for the presence and location of lesions in relation to the hippocampus. When a lesion is cortically based and is not near the mesial structure, lesionectomy alone can achieve good seizure outcome, also in children [16]. When the lesion involves the hippocampus or when the mesial structures look abnormal in the MRI there is a clear indication to resect the hippocampus and amygdala in pediatric patients [19]. On the other hand, it is unclear whether the mesial temporal structures should be resected when the lesion is in close proximity to a hippocampus that appears normal on MRI. The role of hippocampectomy in nonlesional temporal lobe epilepsy is also unclear. Moreover, the potential effects of resection of the medial temporal lobe structures and the risk factors for postoperative verbal memory decline in children have not been extensively investigated. Skirrow et al. [20] recently reported that better verbal memory was linked to greater post-surgical residual hippocampal volumes, mainly in the left hemisphere [20]. The goal of temporal lobe epilepsy surgery is to resect the epileptogenic zone while preserving the functional areas, specifically the hippocampus. In this paper we aimed to assess the decision making process regarding resection of the hippocampus in children with intractable temporal lobe epilepsy and without a lesion in the hippocampus as per MRI. We expected that patients who did not undergo hippocampectomy would have noninferior seizure outcomes compared to patients who underwent hippocampectomy.
2. Methods We reviewed the medical charts of all children ( < age 18 years) who underwent temporal lobe surgery for drug resistant epilepsy by one surgeon (MB) in two departments of neurosurgery between January 2008 and December 2015. We included in the study patients with temporal lesions involving the gyri adjacent to the hippocampus and with normal hippocampal appearance on MRI. We also included patients with nonlesional epilepsy and normal MRI. To state explicitly, children with cortical based lesions that were distant from the hippocampus by at least one gyrus gap and children with mesial temporal sclerosis and any hippocampal abnormality as per MRI were excluded from the analysis. MRI showing a normal hippocampus, on both sides, was a study inclusion criterion, as was the availability of follow-up data for at least one year. We reviewed the medical records for demographic data, age at seizure onset, age at surgery, and postsurgical seizure types and frequency. We also recorded the preoperative electroencephalogram (EEG) and video EEG (VEEG), neuroimaging studies and functional investigations. The operative notes were reviewed for intraoperative complications, before and after resection electrocorticography (ECoG) and pathology of the temporal resections. All children were with medically refractory epilepsy, had undergone a comprehensive presurgical evaluation and had been previously evaluated by EEG. All patients had undergone prolonged VEEG monitoring and MRI in a mesial temporal sclerosis protocol. For some patients, fMRI scanning was used to determine cerebral dominance for language.
Fig. 1. Illustrative case of a patient (#1) with nonlesional left temporal lobe epilepsy. Subdural grid electrodes with strips and depth electrodes were implanted (A). The depth electrodes aimed to the anterior (left panel) and posterior (right panel) hippocampus, as demonstrated in coregistration of post implantation CT and preoperative MRI (B).
M. Benifla et al. / Seizure 52 (2017) 81–88
2.1. Intraoperative ECoG For patients with space occupying lesions (SOL), ECoG was performed using a 20-electrode surface array (Ad-Tech Medical Instrument Co., Racine, WI). Another four-electrode strip array was placed over the hippocampus to record the pre-excisional epileptiform discharges. When the lesionectomy was completed, post-excisional ECoG was performed from both the cortex and the hippocampus, and according to the residual epileptiform discharges. Decisions were made regarding the requirement for further cortical resection or amygdalohippocampectomy.
83
decision as to the need for hippocampal resection was made on the basis of the invasive monitoring data alone. However, in some cases, intraoperative electrocorticography was conducted in addition, to guide the extent of neocortical resection. 2.3. Postoperative seizure outcome For the purposes of this study, postoperative seizure control was classified using the Engel scheme: Class I, seizure-free postoperatively; Class II, rare disabling seizures; Class III, worthwhile improvement; and Class IV, no worthwhile improvement.
2.2. Invasive monitoring 2.4. Data analysis For patients with nonlesional TLE, long term invasive monitoring was performed. After performing the craniotomy and the dural opening, the central sulcus was identified indirectly by determining the inversion polarity (phase reversals) on somatosensory evoked potential recording after median nerve electrical stimulation. Then, confirmation of the precentral (motor) gyrus was achieved using direct cortical stimulation. Sixty-four electrode surface array subdural grids (Ad-Tech Medical Instrument Co.) were placed to cover the lateral surface of the temporal and frontal lobes, using data from the semiology of the seizures, interictal and ictal scalp EEG recordings, interictal spike sources, and speech areas as detected by fMRI. In addition, ipsilateral subdural strips were placed to cover the anterior temporal tip and the inferior surface of the temporal lobe. Depth electrodes were placed to capture ictal data from the anterior greater regions of the temporal lobe, which could be provided by the subdural grid per se, specifically the anterior and posterior hippocampus, by frameless stereotaxic navigation (Fig. 1). The patients were monitored for 5 days for interictal epileptiform discharges and for ictal seizure onset. Mapping of motor, sensory and language functions was performed in 1–2 sessions on the 3rd or 4th day after grid implantation. After obtaining sufficient data regarding the functional cortex and regions of epileptogenesis, the subdural grid was removed and epileptic regions were resected. The
To determine factors that predict seizure outcome, comparison between groups, Pearson x2 and Fisher’s exact tests were used. To identify predicting factors of seizure outcomes, further statistical analyses were performed. These analyses included the following clinical variables: 1) age at seizure onset; 2) duration of seizures before surgery; 3) age at surgery; 4) whether or not the mesial structures were resected and 5) histological features (for example, neoplastic versus nonneoplastic status). Each of these variables was entered into a univariate logistic regression analysis in which the dependent variable was Engel Class I vs. Engel Class II–IV outcome. In this analysis, a p < 0.05 was used to indicate statistical significance. 3. Results 3.1. Patients Table 1 describes the patient demographics including gender, age of onset of seizures, age at time of surgery, developmental exam and hand dominance. Eleven females and 11 males met study inclusion criteria. The mean age at the time of surgery was 11.3 years.
Table 1 Demographic and presurgical characteristics. Patient
sex Age at seizure onset (months)
Seizure Age duration at (months) surgery (months)
f/u Hand (months) dominance
fMRIlanguage
Side of surgery
pre op NP
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
f f m F m f f m m m m f m f F
59 69 39 59 0 60 88 6 170 132 137 170 100 84 65
157 141 72 75 12 78 96 82 180 192 160 176 108 132 89
98 72 33 16 12 18 8 76 10 60 23 6 8 48 24
84 72 68 36 42 60 26 37 20 34 57 48 47 64 12
Rt Rt Rt Rt NA Rt Lt Rt Rt Rt Rt Rt Rt Rt Lt
Lt Lt Lt
Lt Lt Lt Lt Rt Lt Lt Lt Lt Rt Rt Rt Lt Rt Rt
16 17 18 19 20 21 22
F F F M M M M
108 190 166 158 166 151 131
120 205 198 183 212 205 141
12 15 32 25 46 54 10
12 30 32 29 19 16 55
Rt Rt Rt Rt Rt Rt Rt
Lt Bil
Weak verbal memory attention deficit Normal Normal ND ND ND Weak expressive language ND Below average verbal and non verbal performance Attention deficit Normal Normal Normal Normal Weak expressive language, weak visual and verbal memory Normal Weak expressive language, attention deficit Normal Low intellect, attention deficit low average intellect Normal Weak verbal skills; low working memory.
Lt Bil Lt
Lt Bil
Lt
Lt
Abbreviations: Bil– bilateral; Lt– left; Rt– right; NA– not applicable; NP– neuropsychology.
Lt Rt Rt Lt Rt Rt Lt
84
M. Benifla et al. / Seizure 52 (2017) 81–88
3.2. Seizure history The mean duration of seizures prior to surgery was 2.8 years. The mean age at the first seizure was 8.7 years. Semiology and preoperative interictal and ictal EEG recordings identified temporal lobe epilepsy in all patients; frontal epileptiform discharges were noted as well in 6 patients. The mean number of anticonvulsants tried before surgery per patient was 5.3. 3.3. Preoperative MRI Lesions were identified or suspected on MRI in 18 patients. For 4, the MRI scans were reported as normal (patients #1,2,3 and 9, Table 1). In all patients the hippocampus was reported as normal on both sides, this being a criterion for study inclusion. For 2 patients, the preoperative MRI suggested a diagnosis of focal cortical dysplasia. Other lesions identified by MRI included cavernous malformation, dysembryoplastic neuroepithelial tumor (DNET), ganglioglioma, low grade glioma and non-specific signal changes. Thirteen children underwent fMRI. Left language dominance was suggested in 9 patients and bilateral language activation in 3. For one additional patient, the fMRI results were unreliable due to the young age of the patient. 3.4. Temporal lobe surgery In 12 patients, a left temporal procedure was performed; and in 10, a right one. In 15 patients with suspected neoplastic lesion on MRI, the hippocampus was exposed and cortical and hippocampal ECoG was performed. A lesionectomy alone was performed in 7 patients, with significant improvement of the ECoG recording both from the cortex and from the hippocampus. In one patient (#5) an arachnoid cyst medial to the hippocampus was fenestrated after the first ECoG and was connected to the temporal horn. A second ECoG was performed about 10 min following the fenestration and no epileptiform discharges were noted (Fig. 2). In 4 patients lesionectomy and anterior temporal lobectomy (ATL) were performed with improvement of the ECoG recording. In another 3 patients (#16,17,20, Table 2) the hippocampal ECoG did not show any change before and after lesionectomy and ATL, so hippocampectomy was performed as well (Fig. 3). In 2 patients the hippocampus was resected on the right side and in the third, on the left. Long term invasive monitoring was performed in 7 children. Epileptic activity arising from the hippocampus, according to the depth electrode recording, was indication for AH. In four patients the preoperative MRI was reported as normal; of them, two (#3,9,
Table 2) underwent hippocampectomy. In 2 patients, cortical dysplasia was suspected: one of them (#14) underwent hippocampectomy. For one girl (#12) the suspected tumor was closely related to the Wernicke area; in a second surgery, she underwent hippocampectomy. For all 7 patients who underwent long term invasive monitoring, subdural grid and depth electrodes were inserted into the mesial temporal structures, or into a structural abnormality, as seen on frameless stereotactic MR imaging, as described. In three patients (#3,9,14, Table 2), contacts located in the hippocampus were part of the seizure onset zone and these children underwent AH. The hippocampus was spared in the remaining four children, in whom the epileptogenic zone was presumed to reside in the neocortex, according to invasive monitoring. 3.5. Neuropathology The neuropathological findings of the specimens removed at surgery for the 22 patients are described in Table 2. For all patients, a neuropathological diagnosis was achieved from the resected specimens. Four children had cortical dysplasia (3–Ib and 1–IIb), 2 cortical dysplasia (both type Ia) and gliosis, 2 had cavernous malformation; 5 a DNET; 4 ganglioglioma; 2 meningioangiomatosis, one pilocytic astrocytoma, one arachnoid cyst and one oligodendroglioma. All tumors were classified as WHO stage I, except the oligodendroglioma, which was WHO stage II. 3.6. Post-operative complications There were no postoperative mortalities. One child (# 18, Tables 1 and 2) underwent same day recraniotomy due to epidural hematoma in the surgical site. After evacuation of the hematoma the child recovered completely. For one child (#12, Tables 1 and 2) who underwent a GTR of the oligodendroglioma, the tumor recurred 4 years after the first procedure (she was seizure free during that period) and a second surgery was performed, this time with resection of the mesial structures. The pathology diagnosis was glioblastoma multiforme and the child passed away one year later. Only the outcome before tumor recurrence was taken into consideration for this report. Two children experienced homonymus superior quadrantopsia that was confirmed by a neuroophthalmologist. 3.7. Seizure outcome The mean follow-up of the 22 patients was 3.5 years. At the end of follow-up, 18 patients were classified as Engel I (82%), 2 as Engel
Fig. 2. A 15 month old boy (#15) with an arachnoid cyst (white arrow) compressing the hippocampus (A, B). The cyst was fenestrated using a transcortical approach through the medial temporal gyrus and opened into the right temporal horn. A post-operative MRI displays the decrease in the volume of the cyst and the release of pressure on the right hippocampus (C). The hippocampus remained intact and the patient is seizure free.
M. Benifla et al. / Seizure 52 (2017) 81–88
85
Table 2 Type of surgery, pathological diagnosis and seizure outcome. Patient
Surgery
Invasive monitoring
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
PH ATL + PH ATL + AH Lesionectomy Cyst fenestration Lesionectomy ATL + lesionectomy ATL + PH ATL + AH Lesionectomy Lesionectomy ATL + lesionectomy Lesionectomy ATL + AH ATL + lesionectomy ATL + lesionectomy + AH ATL + lesionectomy + AH Lesionectomy ATL + lesionectomy Lesionectomy + AH Lesionectomy
y y y
22
ATL + lesionectomy
ECoG
AH
Pathology
ENGEL
y Y y y Y y y
n n y n n n n n y n n n n y n y y n n y n
II I III I I I I I II I I I* I I I I I I I III I
y
n
CD Ia + gliosis CD Ib CD Ib DNET Arachnoid cyst DNET Ganglioglioma CD Ib CD Ia + gliosis Meningio- angiomatosis Ganglio OGD Cavernoma CD IIa Ganglioglioma Ganglioglioma JPA Cavernoma DNET DNET Meningioangiomatosis DNET
y y y y y y y y y y y
I
*This patient was seizure free for 4 years, then had recurrence of the glioma and underwent a second surgery that included AH; he died one year later. Abbreviations: ATL– Anterior temporal lobectomy; AH– Amygdalohippocampectomy; CD– Cortical dysplasia; DNET– Dysembryoplastic neuroepithelial tumor; JPA– Juvenile pilocytic astrocytoma; OGD– Oligodendroglioma; PH– Para hippocampectomy.
II (9%), and 2 as Engel III (9%). Fifteen of the 16 (94%) patients who did not undergo AH were classified as Engel I, and the remaining one as Engel II. Of the 6 patients who underwent AH, only 3 were classified as Engel I; one as Engel II and 2 as Engel III. Of the 2 patients classified as Engel III, one (#3) had nonlesional epilepsy and the mesial structures were resected on the left side, after long term invasive monitoring. The other patient (#20) had right-sided DNET with intraoperative abnormal hippocampal ECoG. He underwent gross total resection with AH, and was seizure free for 3 months. A second surgery with long term invasive monitoring was suggested, yet the family refused. Left-sided surgery was performed in the 2 patients classified as Engel II (#1, #9); both were right-hand dominant with left fMRI-language. A favorable surgical outcome (Engel I and II) was obtained in 3 of the four patients with non-lesional MRI (#1, 2, 3 and 9) and in 17 of the 18 lesional cases. However, only 25% (1/4) of the non-lesional cases achieved Engel I outcome, compared with 94% of the lesional ones. None of the clinical variables tested for prediction of seizure outcome: age at seizure onset, seizure duration, type of surgery, pathological features and age at surgery reached statistical significance. Thus, we could not identify predictors for favorable seizure outcome. 4. Discussion In this study, we reviewed the medical records of 22 children who had surgery for intractable temporal lobe epilepsy and normal hippocampus. The surgeries were intended to resect epileptogenic zones, while attempting to preserve the mesial structures. Overall, favorable outcomes, according to Engel class I, were found in 82% (18/22) of the patients. While this rate appears consistent with the pediatric medical literature, no other study selected patients according to the criteria applied here, specifically a normal appearing hippocampus on MRI; and either a temporal lesion adjacent to the hippocampus or no lesion at all. For pediatric patients with these characteristics, we confirmed our hypothesis of noninferior seizure outcomes following surgery without AH, compared to surgery with AH; Engel I in 94% (15/16) vs. 50% (3/6).
In the procedures documented in this study, ECoG in children with lesions, and invasive monitoring in children without lesions, were used to identify epileptic foci and to determine the extent of resection, including the possibility of hippocampectomy. In their pediatric study, Clusmann et al. [4] suggested that difficulty in precisely locating the epileptogenic focus in children could explain the lack of improvement they observed in seizure outcomes following AH compared to ATL. Indeed, good seizure control was achieved for 83% of their patients with hippocampal sclerosis; and poorer results for children who underwent AH due to a less defined abnormality in the lateral temporal lobe, or due to a tumor, or who had normal histological features or dysplasia. Other studies have reported good seizure outcomes in pediatric patients with hippocampal sclerosis [6,19]; outcomes in children were assessed as comparable to those in adults [21]. Disagreement as to the performance of hippocampectomy arises in cases of no apparent lesion in the hippocampus. Dual pathology [22] has become well recognized as a driver of epileptogenesis in TLE, in both children and adults. The implication is that seizures continue after surgery, even following GTR. Thus, procedures that identify extra-tumoral epileptogenic tissue and tailor resection accordingly have become the mainstay of treatment for TLE, and improving means for identification of epileptogenic zones is a priority. In the current study, all 15 patients with SOL underwent cortical and hippocampal ECoG, and all 12 children in whom the hippocampus was spared based on the ECoG results had Engel I outcome. While the benefit of this procedure in achieving better seizure outcomes is not conclusive, its intraoperative use in identifying epileptogenic areas outside tumors has persisted over many years. Two systematic reviews of observational studies [10,11] reported similar rates of seizure freedom between patients for whom ECoG was and was not used. However, they warranted caution in interpretation of their findings, since ECoG may be used in more severe cases of epilepsy, as was demonstrated in a study of adults that addressed this matter [23]. Pediatric studies have shown positive seizure outcomes following ECoG [16,24], which, in long term follow up, were shown to surpass those of patients who underwent lesionectomy without ECoG [24]. In an earlier study, we
86
M. Benifla et al. / Seizure 52 (2017) 81–88
Fig. 3. Preoperative MRI of patient #17 with temporal lobe lesion (white arrow) compressing the hippocampus (A). The explored hippocampus (H) and strip electrodes over it are shown in B. After lesionectomy, electrocorticography showed significant epileptiform discharges from the hippocampus and the patient underwent hippocampectomy. Post-excisional MRI (C) demonstrates the area of resection and the posterior end of the hippocampal resection.
M. Benifla et al. / Seizure 52 (2017) 81–88
showed that the grading of postexcision ECoG according to epileptiform discharges did not predict seizure outcome [3]. Nonetheless, ECoG continues to be used in our department to guide the extent of resection in pediatric epilepsy surgery. We report seizure free results in 93% (14/15) of patients with SOL compared to 57% (4/7) with nonlesional and dysplastic lesions. In retrospect, some of the pathological lesions of the four MRInegative cases may have been detected on higher Tesla MR's. Epileptogenic zones in the hippocampus were more often detected in nonlesional than lesional patients: 57% (4/7) vs. 20% (3/15). As described, the performance of AH did not necessarily contribute to seizure freedom. Studies in adults show poorer seizure and neuropsychological outcomes in patients without lesions. One study showed better seizure outcomes following lobectomy in TLE patients with lesions than without; outcomes of the latter were similar to patients with lesions outside the excision area [9]. In another study, postoperative memory loss was greater for TLE patients without lesions, according to preoperative MRI and histopathological findings, than for matched TLE patients with abnormal MRI, despite better preoperative memory in the former [25] Overall, according to a meta-analysis that included 697 patients with nonlesional epilepsy and 2860 patients with lesional epilepsy, the odds for becoming seizure free were 2.7 higher for the lesional cases, in both children and adults [26]. Thus, the better surgical outcome in our lesional cases is in line with the literature and is not directly related to whether the hippocampus was resected or spared. The procedure for AH in the current study included resection of the TLE. Selective AH as described by Yasargil et al. [27] was the procedure documented in several studies, including Clusmann et al's pediatric study [4]. A systematic review that analyzed 3 prospective and 10 retrospective studies concluded that seizure free outcomes were better following ATL than selective AH [28] Interestingly, a review of 53 studies did not find any certain surgical approach for TLE to demonstrate an advantage in seizure outcomes [29]. This highlights the importance of clearly defining preoperative characteristics for any comparison between surgical procedures, as was done in the current study. We did not find the side of surgery to correlate with seizure outcome. In contrast, left-sided surgery was found to predict poor seizure outcome among pediatric patients who underwent ATL, and particularly among those who underwent AH [4]. Laterality of the lesion (dominant or nondominant hemisphere) [19,30], neuropsychological assessment [19], intraoperative language mapping [31], and the timing (early or late) of seizure onset [32,33] are factors that have been considered in determining extent of resection. As with most studies on epileptic surgery, the main outcome investigated in this study was freedom from seizures. Postoperative neuropsychology testing was not performed, this being a limitation of the study. A recent retrospective study showed a decrease in verbal memory in children following left temporal lobe surgery that includes mesial structures, particularly when language representation is in the left hemisphere [34]. Importantly, that study did not find a correlation between seizure freedom and postsurgical decline in verbal memory. This highlights the importance of preserving the hippocampus, when possible, especially in children; and calls for more attention to neuropsychology evaluation, which is evidently a distinct outcome from seizure freedom. We emphasize that the findings of this study are relevant to the pediatric population only. Apparently, studies on epilepsy surgery should distinguish clearly between patient and adult populations. One systematic review of TLE surgery concluded that children became seizure-free more often than adults after tumor resection [10] and another reported no difference in seizure outcomes
87
between adults and children [11]. In two studies that compared seizure outcomes between children and adults following similar surgeries, the adults fared better [35,36]. Differences in outcomes between children and adults may stem from differences in preoperative evaluation, such as greater difficulty to evaluate pediatric MRI scans;[4] in clinical characteristics such as less clear ictal EEG pattern [1] or more multifocal discharges in children; [36] in pediatric pathology, such as lower frequency of hippocampal sclerosis and higher frequency of dual pathology and etiologies such as temporal sclerosis, cortical dysplasia, trauma and vascular malformations;[37] and in a broader impact on neurocognitive function in children due to brain development. In children, the plasticity of the brain may promote better recovery, yet the impact of cognitive decline on development may have serious consequences. We did not find duration of epilepsy to correlate with seizure outcomes; however, our cohort was small. Long epilepsy duration has been identified as an important risk factor for surgically refractory seizures [37]. Thus, the consideration of early surgical treatment for medically intractable TLE is recommended for children. Seizure outcome following epilepsy surgery during childhood does not remain stable over time [9,38]. A review of 42 children following temporal lobe surgery revealed a seizure freedom rate of 77% at 1 year, which decreased to 67% at 10 years [8]. A similar decline, from 61% at 1 year to 41% at 10 years was noted in a large retrospective review of 325 adult and pediatric patients who underwent anterior temporal lobectomy [9]. A cohort study of patients who underwent temporal lobectomy during childhood found that 58% of the patients who were seizure free at 5 years follow-up experienced seizure recurrence within the subsequent 10 years [7]. Second surgery may be effective in patients with seizure relapse. At that point, re-evaluation of the epileptogenic zone may lead to hippocampal resection if it was spared in the first operation. The retrospective design is a limitation of this, as well as the vast majority of studies on epilepsy surgery. In particular, analysis of clinical decision making is generally problematic in retrospective studies. However, the fact that the same surgeon performed all the operations during the study period mitigates for such. Other limitations of this study are the relatively small number of patients and the heterogenicity of the lesions. The analysis of seizure outcomes in pediatric patients with similar clinical characteristics and the inclusion of children without lesions are strengths of this study. 5. Conclusion In this paper, we presented seizure outcomes of intraoperative techniques to determine whether the hippocampus is part of the epileptogenic zone in pediatric patients with lesions located close to the hippocampus (but separated from it by at least one gyrus gap) or with nonlesional TLE. Such patients would be subjected to AH according to the literature regarding adults [14]. In the current pediatric cohort, the hippocampus was preserved in 16 patients (73%) while hippocampectomy was performed in 6 (27%). Excellent seizure outcomes were achieved when the hippocampus was spared, implying that hippocampal sparing does not negatively influence epilepsy surgery results in children. Since verbal memory decline can significantly influence a child's quality of life, every effort should be exerted to preserve the mesial temporal structures, mainly the hippocampus. Conflict of interest None of the authors has any conflict of interest to disclose.
88
M. Benifla et al. / Seizure 52 (2017) 81–88
Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. References [1] Nickels KC, Wong-Kisiel LC, Moseley BD, et al. Temporal lobe epilepsy in children. Epilepsy Res Treat 2012;2012:849540. [2] Blume WT. Temporal lobe epilepsy surgery in childhood: rationale for greater use. Can J Neurol Sci 1997;24:95–8. [3] Benifla M, Otsubo H, Ochi A, et al. Temporal lobe surgery for intractable epilepsy in children: an analysis of outcomes in 126 children. Neurosurgery 2006;59:1203–13 discussion 1213-4. [4] Clusmann H, Kral T, Gleissner U, et al. Analysis of different types of resection for pediatric patients with temporal lobe epilepsy. Neurosurgery 2004;54:847–59 discussion 859–60. [5] Mittal S, Montes JL, Farmer JP, et al. Long-term outcome after surgical treatment of temporal lobe epilepsy in children. J Neurosurg 2005;103:401– 12. [6] Lopez-Gonzalez MA, Gonzalez-Martinez JA, Jehi L, et al. Epilepsy surgery of the temporal lobe in pediatric population: a retrospective analysis. Neurosurgery 2012;70:684–92. [7] Jarrar RG, Buchhalter JR, Meyer FB, et al. Long-term follow-up of temporal lobectomy in children. Neurology 2002;59:1635–7. [8] Benifla M, Rutka JT, Otsubo H, et al. Long-term seizure and social outcomes following temporal lobe surgery for intractable epilepsy during childhood. Epilepsy Res 2008;82:133–8. [9] McIntosh AM, Kalnins RM, Mitchell LA, et al. Temporal lobectomy: long-term seizure outcome, late recurrence and risks for seizure recurrence. Brain 2004;127:2018–30. [10] Englot DJ, Han SJ, Berger MS, et al. Extent of surgical resection predicts seizure freedom in low-grade temporal lobe brain tumors. Neurosurgery 2012;70:921–8 discussion 928. [11] Englot DJ, Berger MS, Barbaro NM, et al. Factors associated with seizure freedom in the surgical resection of glioneuronal tumors. Epilepsia 2012;53:51–7. [12] Bonney PA, Glenn CA, Ebeling PA, et al. Seizure freedom rates and prognostic indicators after resection of gangliogliomas: a review. World Neurosurg 2015;84:1988–96. [13] Bell ML, Rao S, So EL, et al. Epilepsy surgery outcomes in temporal lobe epilepsy with a normal MRI. Epilepsia 2009;50:2053–60. [14] Ghareeb F, Duffau H. Intractable epilepsy in paralimbic word health organization grade II gliomas: should the hippocampus be resected when not invaded by the tumor? J Neurosurg 2012;116:1226–34. [15] Bilginer B, Yalnizoglu D, Soylemezoglu F, et al. Surgery for epilepsy in children with dysembryoplastic neuroepithelial tumor: clinical spectrum, seizure outcome, neuroradiology, and pathology. Childs Nerv Syst 2009;25:485–91. [16] Ogiwara H, Nordli DR, DiPatri AJ, et al. Pediatric epileptogenic gangliogliomas: seizure outcome and surgical results. J Neurosurg Pediatr 2010;5:271–6. [17] Babini M, Giulioni M, Galassi E, et al. Seizure outcome of surgical treatment of focal epilepsy associated with low-grade tumors in children. J Neurosurg Pediatr 2013;11:214–23.
[18] Brahimaj B, Greiner HM, Leach JL, et al. The surgical management of pediatric brain tumors causing epilepsy: consideration of the epileptogenic zone. Childs Nerv Syst 2014;30:1383–91. [19] Cataltepe O, Turanli G, Yalnizoglu D, et al. Surgical management of temporal lobe tumor-related epilepsy in children. J Neurosurg 2005;102:280–7. [20] Skirrow C, Cross JH, Harrison S, et al. Temporal lobe surgery in childhood and neuroanatomical predictors of long-term declarative memory outcome. Brain 2015;138:80–93. [21] Mohamed A, Wyllie E, Ruggieri P, et al. Temporal lobe epilepsy due to hippocampal sclerosis in pediatric candidates for epilepsy surgery. Neurology 2001;56:1643–9. [22] Spencer S, Huh L. Outcomes of epilepsy surgery in adults and children. Lancet Neurol 2008;7:525–37. [23] Southwell DG, Garcia PA, Berger MS, et al. Long-term seizure control outcomes after resection of gangliogliomas. Neurosurgery 2012;70:1406–13; discussion 1413-4. [24] Gelinas JN, Battison AW, Smith S, et al. Electrocorticography and seizure outcomes in children with lesional epilepsy. Childs Nerv Syst 2011;27:381–90. [25] Helmstaedter C, Petzold I, Bien CG. The cognitive consequence of resecting nonlesional tissues in epilepsy surgery–results from MRI- and histopathologynegative patients with temporal lobe epilepsy. Epilepsia 2011;52:1402–8. [26] Tellez-Zenteno JF, Hernandez Ronquillo L, Moien-Afshari F, et al. Surgical outcomes in lesional and non-lesional epilepsy: a systematic review and metaanalysis. Epilepsy Res 2010;89:310–8. [27] Yasargil MG, Antic J, Laciga R, et al. Microsurgical pterional approach to aneurysms of the basilar bifurcation. Surg Neurol 1976;6:83–91. [28] Hu WH, Zhang C, Zhang K, et al. Selective amygdalohippocampectomy versus anterior temporal lobectomy in the management of mesial temporal lobe epilepsy: a meta-analysis of comparative studies. J Neurosurg 2013;119:1089– 97. [29] Schramm J. Temporal lobe epilepsy surgery and the quest for optimal extent of resection: a review. Epilepsia 2008;49:1296–307. [30] Phi JH, Kim SK, Cho BK, et al. Long-term surgical outcomes of temporal lobe epilepsy associated with low-grade brain tumors. Cancer 2009;115:5771–9. [31] Benifla M, Otsubo H, Ochi A, et al. Temporal lobe surgery for intractable epilepsy in children: an analysis of outcomes in 126 children. Neurosurgery 2006;59:1203–13 discussion 1213-4. [32] Goldring S, Rich KM, Picker S. Experience with gliomas in patients presenting with a chronic seizure disorder. Clin Neurosurg 1986;33:15–42. [33] Fried I, Kim JH, Spencer DD. Limbic and neocortical gliomas associated with intractable seizures: a distinct clinicopathological group. Neurosurgery 1994;34:815–23 discussion 823-4. [34] Law N, Benifla M, Rutka J, et al. Verbal memory after temporal lobe epilepsy surgery in children: do only mesial structures matter? Epilepsia 2016. [35] Datta A, Sinclair DB, Wheatley M, et al. Selective amygdalohippocampectomy: surgical outcome in children versus adults. Can J Neurol Sci 2009;36:187–91. [36] Lee YJ, Kang HC, Bae SJ, et al. Comparison of temporal lobectomies of children and adults with intractable temporal lobe epilepsy. Childs Nerv Syst 2010;26:177–83. [37] Lee YJ, Lee JS. Temporal lobe epilepsy surgery in children versus adults: from etiologies to outcomes. Korean J Pediatr 2013;56:275–81. [38] Hauptman JS, Pedram K, Sison CA, et al. Pediatric epilepsy surgery: long-term 5-year seizure remission and medication use. Neurosurgery 2012;71:985–93.
