Epilepsy Research (2015) 112, 92—99
journal homepage: www.elsevier.com/locate/epilepsyres
Surgical treatment for patients with symptomatic generalised seizures due to brain lesions Tao Yu a,∗, Guojun Zhang a, Yuping Wang b, Duanyu Ni a, Liang Qiao a, Wei Du a, Yuanyuan Piao a, Yongjie Li a a b
Beijing Institute of Functional Neurosurgery, Xuanwu Hospital, Capital Medical University, China Comprehensive Epilepsy Center of Beijing, Xuanwu Hospital, Capital Medical University, China
Received 14 November 2014; received in revised form 17 January 2015; accepted 6 February 2015 Available online 6 March 2015
KEYWORDS Generalised seizure; Lesion; Semiology; Epileptogenic zone; Surgery
Summary Object: To study the role of epilepsy surgery for patients with focal lesions who exhibited the semiology of clinically generalised seizures. Methods: From our epilepsy surgery series, we identified 29 patients who underwent surgery for seizures, including certain types of generalised seizures, according to their ictal semiology. We systematically reviewed the brain imaging, video-EEG, surgical operation, and pathological findings data of these patients. Results: All patients had at least one type of generalised seizure according to the semiology; these seizures included epileptic spasms, myoclonic seizures, tonic seizures, atonic seizures and atypical absence seizures. Eight patients had a single type of seizure, 11 patients had two types of seizures, and 10 patients had more than two types of seizures. In addition to symptomatic generalised seizures, complex partial seizures were also recorded in eight patients. In 24 patients, the ictal semiology showed slight asymmetric movements in certain types of seizures. Generalised interictal epileptic discharges were recorded in 24 patients; and generalised ictal epileptic discharges as the only EEG pattern were recorded in 13 patients. Intracranial recording was performed in 20 patients; 10 of whom showed a rhythm of fast activities at the initiation of the seizures. Functional hemispherectomy was performed for three patients with hemispheric lesions. Focal resection of the epileptogenic zone was performed in 26 patients. The resected epileptogenic zones involved a single lobe in 10 patients, two lobes in 11 patients, and three lobes in 5 patients; the parietal lobe was the most commonly involved lobe (in 19 cases).
∗ Corresponding author at: Beijing Institute of Functional Neurosurgery, Xuanwu Hospital, Capital Medical University, No. 45 Changchun Street, Beijing 100053, China. Tel.: +86 10 83163174; fax: +86 10 83163174. E-mail address: [email protected] (T. Yu).
http://dx.doi.org/10.1016/j.eplepsyres.2015.02.001 0920-1211/© 2015 Elsevier B.V. All rights reserved.
Surgical treatment for patients with symptomatic generalised seizures
93
Scar lesions (in 17 patients) were most commonly observed on pathological examination. At the last follow-up (mean 18 ± 8.3 months, range 12—48), 17 (58.6%) patients were seizure-free. Conclusions: Certain patients with local brain lesions can have seizures with specific types of generalised semiology. An appropriate operation may be helpful for a portion of these patients. © 2015 Elsevier B.V. All rights reserved.
The terms ‘‘focal’’ and ‘‘generalised’’ have been used to express a dichotomous classification for both seizures and epilepsy. However, it has been previously noted that focal cortical abnormalities can cause seizures with generalised clinical semiology and, occasionally, EEG findings. For example, absences, epileptic spasms, myoclonic seizures, and even atonic seizures have been reported secondary to focal brain lesions in certain patients (Farwell and Stuntz, 1984; Millan et al., 2001; Chugani et al., 1992; Akiyama et al., 2005; Ramachandrannair et al., 2008; Zhao et al., 2010; Kovac and Diehl, 2012). In this article, we use the term ‘‘symptomatic generalised seizure’’ to describe a seizure that has generalised ictal semiology due to a local brain lesion. From January 2008 to June 2013, we used MRI to image patients with focal lesions; the patients showed ictal semiology of certain types of clinical generalised seizures. We localised the epileptogenic zones and performed respective surgeries in these patients. The goal of this study was to present the outcomes of surgical treatment for these patients and to analyse the relationship between the focal lesions and seizures with generalised semiology.
Patients and methods Patients Patients were selected from a pool of children and adults who underwent presurgical evaluation and surgical treatment for the relief of intractable epilepsy; the study was performed at the Comprehensive Epilepsy Centre of Beijing between January 2008 and June 2013. Among the total 1298 patients who received surgical treatment, 32 patients had local lesions on MRI-associated generalised seizures. The seizure type was classified according to the reports of the ILAE on the classification of seizures and epilepsies (Wieser et al., 2001; Engel, 2006; Berg et al., 2010). Despite the controversy, in this article, spasms were classified as generalised seizures. Patients were selected according to the following criteria: (1) patients with refractory epilepsy; (2) patients with severe seizures, including daily seizures with loss of consciousness, falls, or frequent status epilepticus; (3) patients with one or more types of generalised seizures were recorded by the video-EEG, with or without a partial seizure; (4) brain lesions on the MRI, including multiple lesions or hemispheric lesions; and (5) no contraindication to performing the surgical resection. The final sample consisted of 32 (2.4%) patients who underwent surgical resection. Three patients failed the postoperative follow-up. The seizure histories, imaging and video-EEG monitoring data, operative records and pathological findings of the remaining29 patients were analysed.
Sixteen patients (55.2%) were male, and 13 (44.8%) were female. Patient age at the time of surgery ranged from 3 to 34 years (mean age: 10.8 ± 8.0 years). The average seizure onset age was 3.1 ± 2.6 years (range 0.5—10 years).
Presurgical evaluation MR imaging Each patient underwent a standard MRI protocol that was performed using a 1.5-T or 3.0-T MR scanner (Siemens, Munich/Erlangen/Verio, Germany) and consisted of conventional spin-echo T1-weighted axial, sagittal, coronal and T2-weighted axial sequences (section thickness: 5 mm, image gaps: 1 mm). Additionally, 5-mm-thick Fluid Attenuated Inversion Recovery images were obtained; these images included axial and coronal sections that were perpendicular to the long axis of the hippocampus. Video-EEG monitoring and seizure semiology Interictal/ictal scalp EEGs were recorded using a video-EEG monitoring system (Micromed; Treviso, Italy); the electrodes were arranged according to the international 10—20 system. The duration of video-EEG monitoring ranged from 2 to 7 days, and at least three habitual seizures were recorded for each patient. We attempted to record every type of seizure if the patient had more than one type of seizure. The seizure type was classified by two epileptologists, according to the video-EEG findings (without the MRI information of the patient). If the patient was identified as having a certain type of generalised seizures, we further analysed the clinical semiology and interictal/ictal EEG, and we attempted to identify trends of asymmetry as a clue to the lateralisation of epileptogenic zones. Intracranial EEG monitoring and functional mapping Intracranial EEG monitoring was performed to further localise the epileptogenic zone; exceptions were for patients with a hemispheric lesion that was suitable for hemispherectomy or patients with a focal lesion distant to an eloquent area of the brain. The placement of the grid and strip electrodes was guided by non-invasive exams and intraoperative electrocorticography. The EEG sampling rate was set at 1024 Hz to record more details of seizure propagation. Usually, more than two habitual seizures were required to be recorded; we tried to record every type of seizure for each patient during the long-term intracranial EEG monitoring. Preoperative direct cortical stimulation for functional mapping was performed in patients with epileptogenic zones near the eloquent cortex, such as the motor, sensory or language areas.
94
Surgery and follow-up Hemispherectomy or hemispheric disconnection was performed in patients with hemispheric lesions. Tailored cortical resection was performed in patients with focal brain lesions that were thought to be related to the seizures. Intraoperative electrocorticography monitoring, both before and after cortex resection, was performed in these patients to further tailor the resection. Some patients were awoken from general anaesthesia to perform intraoperative functional cortical mapping or movement monitoring during the process of cortical resection. Patients were followed for at least 12 months to observe surgical outcomes. Using a binary outcome rating, the outcomes were divided into seizurefree (no seizures following surgery, or seizure-free for at least 12 months prior to the latest follow-up) or not seizurefree. All tissue sections from the surgical resection were routinely processed to obtain a detailed pathological diagnoses (Yue-Shan et al., 2010). Histopathological diagnoses were made by at least two neuropathologists.
Results The demographic and clinical characteristics of these patients are listed in Table 1.
Semiology According to the ictal videos of the 29 patients, 8 patients had a single type of seizure, 11 had two types of seizures, and 10 had more than two types of seizures. Epileptic spasms were the most common type of seizure and were recorded in 20 cases; both myoclonic seizures and tonic seizures were the second most common type of seizure and were recorded in 14 cases. Atypical absence seizures were found in four cases, and atonic seizures were found in two cases. In addition to the symptomatic generalised seizures, complex partial seizures (CPS) were also recorded in eight of the cases. CPSs were recorded secondarily to myoclonic seizures in three cases. The ictal semiology of the CPSs helped in the localisation of the epileptogenic zones for these patients. After careful analysis of the ictal semiology, certain asymmetric features of the behaviour in one or more types of seizures (including generalised seizures and CPSs) were observed in 24 of the patients. The asymmetric features of the symptomatic generalised seizures, such as unilateral dominant myoclonic seizures and asymmetric spasm behaviours, were thought to be helpful in lateralising the epileptogenic zones.
Neuroimaging All patients had localised abnormalities on MRI, which included hemispheric lesions in three patients, lesions in the frontal and/or central region in six patients, lesions in the frontal—central—parietal region in three patients, parietal—posterior temporal region lesions in 10 patients, parietal—occipital region lesions in four patients, occipital lobe lesions in one patient, and parietal lobe lesions in two patients (including one patient with a tuberous sclerosis
T. Yu et al. complex; however, the corresponding tubers were localised in the parietal lobe). The presence of a unilateral, congenital, or early-acquired lesion that was present on the brain MRI was helpful in localising the epileptogenic zones.
Scalp EEG Three patients showed dominant hemispheric interictal spikes, and two patients showed local interictal spikes. The remaining 24 patients showed generalised interictal epileptic discharges, such as bilateral sharp waves, spikes, and spike-wave complexes. In these 24 patients, slight asymmetries of the interictal discharges were discovered by a careful comparison of 7 patients. Of all of the ictal records, generalised ictal epileptic discharges (as the only pattern) were recorded in 13 patients, while distinguishable lateralised or localised ictal epileptic discharges were found in the remaining 16 patients. Four patients conformed to the standard of symptomatic Lennox—Gastaut syndrome according to the electro-clinical findings.
Intracranial EEG and cortical mapping Intracranial EEG recording on grid or strip electrodes was performed in 20 patients. Nine patients did not undergo intracranial EEG recordings, including three patients with hemispheric lesions, and six patients with focal lesions distant from eloquent areas of the brain. We analysed the ictal intracranial EEGs of every patient and found that most of the seizures with generalised clinical semiology had localised EEG origins. Low-amplitude fast activities were demonstrated in the beginning of the seizures in 10 patients; spike activities were observed in five patients. In the other five patients, either fast activities or spike activities were recorded in the beginning of the different seizures. The rhythms of fast activities in these patients could involve a few cortical electrodes in the beginning of the EEG recording. However, in some patients, time differences of several hundred microseconds between different cortical electrodes were only found through careful measurement. Usually, clinical behaviours occurred when the original fast activities spread to larger numbers of cortical electrodes. Both generalised seizures and CPSs were recorded in eight of 29 patients. In three of the eight patients, the behaviour and EEG of CPSs were recorded after myoclonic seizures. Preoperative functional mapping was performed in 17 cases, and intra-operative functional mapping was performed in eight cases. In this group, we observed an interesting phenomenon that in the same patient, different clinical ictal behaviours could originate from the same cortical region. For example, one patient (no. 28) showed the clinical behaviour of myoclonic seizures, epileptic spasms or tonic seizures when the ictal high frequency activities originating from almost the same region lasted from 200 ms to more than 3 s.
Surgical resection and complications Anatomical or functional hemispherectomy was performed for three patients with hemispheric lesions. Focal resection
Number
Demographic/clinical characteristics of the patients.
Age
Onset age
Gender
Seizure type
Unilateral predominate symptom
MRI
EEG
Interictal 1 2
12 8
5 1
M F
3 4
4 15
2 7
5 6 7
7 4 10
8
Surgery
Pathology
Interictal asymmetry − −
Ictal local origin + −
Intracranial ictal EEG Rhythmic spikes n
Resective location follow-up (month) R-P R-P
Follow-up (month) 36 24
F 6 F, 18 N
Tuberous sclerosis Scar lesions
+ −
+ −
n Fast activities
L hemispherectomy RF
24 24
F F
Scar lesions FCDI
+ + −
− + −
Fast activities Rhythmic spikes n
LC L C-posterior F R hemispherectomy
12 18 24
F F 6N, 18 F
FCDII Tuberous sclerosis Scar lesions
−
+
n
L P-posterior F
18
F
Scar lesions
+ +
+ +
Fast activities Rhythmic spikes
R P—O—posterior T RP
12 12
N N
B spike slow waves B spike slow waves L hemisphere spikes B spike waves B spikes or spike waves
−
+
Fast activities
R P—O—posterior T
18
N
Scar lesions Polymicrogyria with heterotopias Scar lesions
−
+
n
L hemispherectomy
18
F
Scar lesions
+
+
Polyspikes
12
F
− −
− −
12 18
F F
Polymicrogyria with heterotopias Scar lesions Scar lesions
Burst of B spike waves Burst of B spike waves B spike waves B slow spike waves
+
+
n Rhythmic spikes/fast activities Fast activities
R F-P—O—posterior T R P—O R P—posterior F
LP
12
N
FCDI
−
+
Fast activities
R P—posterior T
24
F
Scar lesions
+ −
+ −
L T—O R P—O
12 12
F N
Scar lesions Scar lesions
B spike slow waves B spike waves R F-C spikes R hemisphere spikes
−
−
Fast activities Rhythmic spikes/fast activities n
R P—posterior T
12
F
− + +
− + +
n Fast activities Rhythmic spikes/fast activities Polyspikes/fast activities Polyspikes
L O—posterior T RC RP
12 12 24
N N N
Meningeal angiomatosis Scar lesions FCDI FCDI
L P—posterior T
48
F
Scar lesions
L-F
12
N
FCDI
R P—O
12
N
Scar lesions
R peresylvian region
18
F
FCDI
Outcome
+ −
Multiple tubers Lesion in R T—P
F F
Spasm, CPS Spasm, myoclonic, tonic Spasm Spasm
+ +
L hemisphere atrophy FCD in L F
1 2.5 1
M F M
Myoclonic, tonic Tonic, atonic Myoclonic, tonic
+ + −
FCD in L C Lesion in L C R hemisphere atrophy
9
4
M
+
Atrophy in L F—P—O
9 10
8 14
4 3
M F
+ +
Atrophy in R T—P—O Polymicrogyria in R P
11
15
7
F
+
FCD in R F-P
12
29
2
M
Spasm, tonic, atypical absence, CPS Myoclonic, CPS, GTCS Spasm, myoclonic, CPS Myoclonic, atypical absence, CPS Spasm
+
L hemisphere atrophy
13
5
2.5
M
Spasm, myoclonic
+
Lesion in L T—P
14 15
3 13
0.5 5
M M
Spasm, tonic Spasm, tonic
− +
Lesion in R O Atrophy in R F-P
16
8
0.5
M
Tonic
−
Lesion in L P
17
11
3
F
+
Heterotopia in R T—P
18 19
13 34
8 8
F F
GTCS, myoclonic, atypical absence Spasm Spasm, tonic
+ +
Lesion in L T—P Atrophy in R P—O
20
4
1
F
Spasm
−
Lesion in L T—P
21 22 23
5 18 4
2 10 2.5
M M M
Spasm Myoclonic Myoclonic, spasm, tonic
+ + +
Lesion in L T—O FCD in R F Lesion in L post-F-C
24
3
1.5
M
Spasm, tonic
+
Lesion in L T—P
B spike waves
−
−
25
6
1
F
Myoclonic, CPS
+
FCD in L F
−
−
26
10
3
M
Spasm, CPS
+
Lesion in R P—O
+
+
27
6
2
M
+
Lesion in R peresylvian region
−
−
Polyspikes/fast activities Fast activities
28
7
0.5
F
+
−
−
Fast activities
L P—O
12
6 F, 6 N
Scar lesions
29
0.5
F
Lesion in B P—O, L predominate Lesion in R P—O
B spike waves
29
Spasm, myoclonic, tonic, atypical absence Spasm, myoclonic, tonic, atonic Myoclonic, CPS
B spike slow waves R F-P spike or spike rhythm Burst of B spike waves
R hemisphere spikes
+
+
n
R P—O—posterior T
18
F
Scar lesions
+
B frontal spikes B spike slow waves spike rhythm B spike slow waves B spikes B spikes B spike slow waves B posterior spike waves B spike waves B spike waves
95
CPS: complex partial seizure, GTCS: general tonic—clonic seizure, L: left, R: right, B: bilateral, F: frontal lobe, T: temporal lobe, P: parietal lobe, O: occipital lobe, C: central area, F: free, N: not free, FCD: focal cortical dysplasia.
Surgical treatment for patients with symptomatic generalised seizures
Table 1
96 of the epileptogenic zone was performed in 26 patients. Because of the location and size of the epileptogenic zone, the focal resection was not limited to a single lobe in some of the patients. In total, the resected epileptogenic zones involved a single lobe in 10 patients, two lobes in 11 patients, and three lobes in 5 patients (involving the parietal lobe in 19 cases, the occipital lobe in 14 cases, the posterior temporal cortex in 10 cases, and the posterior frontal or central cortex in nine cases). One child had a tracheospasm after the operation; the child gradually recovered after trachea cannula and hormonal treatment. One child had status epilepticus after multiple lobe disconnection; however, he demonstrated a ‘‘running down phenomenon’’ in the following months (Salanova et al., 1996); he subsequently became seizurefree 5 months after his operation. Unfortunately, one patient had epilepticus status 1 day after the operation; this was followed by mental disability. The patient continued to have frequent seizures after sedative treatment.
Pathological findings Scar lesions were found in 17 patients, including three patients with a history of encephalitis. Histopathologically, the destruction of the cortices with severe neuron loss, fibrous gliosis and cyst formations were observed. In certain cases, an island distribution of residual neurons, which had an oligodendroglia-like appearance, was observed; this was particularly common in cases secondary to ischaemia/hypoxia. Focal cortical dysplasia (FCD) was found in seven patients, polymicrogyria with heterotopias was found in two patients, tuberous sclerosis was observed in two patients, and meningeal angiomatosis was observed in one patient.
Postsurgical follow-up At the last follow-up (mean 18 ± 8.3 months, range 12—48 months), 17 (58.6%) patients were seizure free. The outcomes changed slightly in the months after the operation. When the patients were followed up for 3 months, 19 (65.5%) were seizure-free. When the patients were followed up for 6 months, 18 (62.1%) were seizure-free. One patient had an obvious increase in tonic seizures shortly after the operation, but the seizures decreased significantly 1 month later; the patient eventually became seizure-free 5 months after the operation. One patient experienced a recurrence of seizures 3 months after the operation. Two patients had seizure-free outcomes in the first 6 months after their operations followed by a recurrence of seizures in the following months; the seizures in these cases were reduced. The outcomes were stable after 12 months in the patients who were followed for more than 1 year. The variable features listed in the table were not significantly associated with seizure-free rates at the last follow-up (Table 2).
Discussion We report on a distinct group of patients with symptomatic generalised seizures and visible local lesions on
T. Yu et al. MRI. The type of generalised seizures included epileptogenic spasms, myoclonic seizures, tonic seizures, atonic seizures, and atypical absence seizures. Most of the local lesions were congenital, perinatal, or early-acquired brain lesions, which were considered to be malformations of the cortical development or scar lesions on pathologic examination. Most of the lesions were localised in the posterior temporal—parietal—occipital region or posterior frontal and central regions. The ages of the patients at seizure onset were relatively young. The interictal and ictal discharges of the EEGs were often generalised without notable traits of localisation. According to the ictal EEG recordings, most of our patients had more than one type of seizure; epileptic spasms were the most common type recorded. On careful analysis of the ictal semiology in these patients, a slight asymmetry of the behaviours could be found in some of these seizures. For example, the spastic behaviour was more obvious on the unilateral limb or shoulder in some patients with epileptic spasms. Unilaterally dominant myoclonic behaviour was also being found in some patients with myoclonic seizures. Despite the subtlety of the lateralising signs, they were important in lateralising the epileptogenic zone. The epileptogenic zones were thought to be in the hemisphere contralateral to the limb demonstrating more obvious behaviour. Furthermore, certain patients also have certain types of partial seizures; this was also helpful for lateralising or localising the epileptogenic zone. In our opinion, the symptomatic generalised seizures might have a similar origination with the partial seizures in the same patient. In our study, most of the patients underwent intracranial recordings; we recorded ictal EEGs to localise the epileptogenic zone. We found that most of the seizures with generalised clinical semiology have a localised EEG origin. Additionally, according to the ictal video-EEG recording, we observed an interesting phenomenon in that some patients with a focal lesion had different types of generalised seizures. Of our patients, 58.6% were seizure-free 1 year after the operation. The electro-clinical features and outcomes of surgical resection support the opinion that some seizures with generalised semiology might have focal origins. This is a topic that has been argued over for many years and has been gradually verified by clinical studies. A typical example is the successful surgery for patients with Lennox—Gastaut syndrome (LGS). It has been verified that a special group of LGS patients may have associated focal brain lesions (Bladin, 1985; Satow et al., 2002; Lee et al., 2014). Similarly, the aetiology of West syndrome has gradually changed from generalised seizures to seizures with multiple origins (Jonas et al., 2005; Komaki et al., 2001; Kang et al., 2006; Wyllie, 1996). However, an interesting and confusing question remains as to how local lesions lead to generalised seizures. It is the opinion of some that a cortical mechanism of secondary bilateral synchrony (SBS) might play an important role in patients with focal lesions who have generalised seizures (Kobayashi et al., 2000; Wyllie et al., 2007; Genton and Dravat, 1997). It has been demonstrated that epileptic spasms and myoclonic seizures could be caused by discharges originating from the focal cortex with extremely rapid secondary generalisation; the interhemispheric time
Surgical treatment for patients with symptomatic generalised seizures Table 2
97
Surgical outcome at last follow-up in relation to variables. N
Seizure freeN (%)
Not seizure freeN (%)
P value (Chi-square)
Age at surgery
journal homepage: www.elsevier.com/locate/epilepsyres
Surgical treatment for patients with symptomatic generalised seizures due to brain lesions Tao Yu a,∗, Guojun Zhang a, Yuping Wang b, Duanyu Ni a, Liang Qiao a, Wei Du a, Yuanyuan Piao a, Yongjie Li a a b
Beijing Institute of Functional Neurosurgery, Xuanwu Hospital, Capital Medical University, China Comprehensive Epilepsy Center of Beijing, Xuanwu Hospital, Capital Medical University, China
Received 14 November 2014; received in revised form 17 January 2015; accepted 6 February 2015 Available online 6 March 2015
KEYWORDS Generalised seizure; Lesion; Semiology; Epileptogenic zone; Surgery
Summary Object: To study the role of epilepsy surgery for patients with focal lesions who exhibited the semiology of clinically generalised seizures. Methods: From our epilepsy surgery series, we identified 29 patients who underwent surgery for seizures, including certain types of generalised seizures, according to their ictal semiology. We systematically reviewed the brain imaging, video-EEG, surgical operation, and pathological findings data of these patients. Results: All patients had at least one type of generalised seizure according to the semiology; these seizures included epileptic spasms, myoclonic seizures, tonic seizures, atonic seizures and atypical absence seizures. Eight patients had a single type of seizure, 11 patients had two types of seizures, and 10 patients had more than two types of seizures. In addition to symptomatic generalised seizures, complex partial seizures were also recorded in eight patients. In 24 patients, the ictal semiology showed slight asymmetric movements in certain types of seizures. Generalised interictal epileptic discharges were recorded in 24 patients; and generalised ictal epileptic discharges as the only EEG pattern were recorded in 13 patients. Intracranial recording was performed in 20 patients; 10 of whom showed a rhythm of fast activities at the initiation of the seizures. Functional hemispherectomy was performed for three patients with hemispheric lesions. Focal resection of the epileptogenic zone was performed in 26 patients. The resected epileptogenic zones involved a single lobe in 10 patients, two lobes in 11 patients, and three lobes in 5 patients; the parietal lobe was the most commonly involved lobe (in 19 cases).
∗ Corresponding author at: Beijing Institute of Functional Neurosurgery, Xuanwu Hospital, Capital Medical University, No. 45 Changchun Street, Beijing 100053, China. Tel.: +86 10 83163174; fax: +86 10 83163174. E-mail address: [email protected] (T. Yu).
http://dx.doi.org/10.1016/j.eplepsyres.2015.02.001 0920-1211/© 2015 Elsevier B.V. All rights reserved.
Surgical treatment for patients with symptomatic generalised seizures
93
Scar lesions (in 17 patients) were most commonly observed on pathological examination. At the last follow-up (mean 18 ± 8.3 months, range 12—48), 17 (58.6%) patients were seizure-free. Conclusions: Certain patients with local brain lesions can have seizures with specific types of generalised semiology. An appropriate operation may be helpful for a portion of these patients. © 2015 Elsevier B.V. All rights reserved.
The terms ‘‘focal’’ and ‘‘generalised’’ have been used to express a dichotomous classification for both seizures and epilepsy. However, it has been previously noted that focal cortical abnormalities can cause seizures with generalised clinical semiology and, occasionally, EEG findings. For example, absences, epileptic spasms, myoclonic seizures, and even atonic seizures have been reported secondary to focal brain lesions in certain patients (Farwell and Stuntz, 1984; Millan et al., 2001; Chugani et al., 1992; Akiyama et al., 2005; Ramachandrannair et al., 2008; Zhao et al., 2010; Kovac and Diehl, 2012). In this article, we use the term ‘‘symptomatic generalised seizure’’ to describe a seizure that has generalised ictal semiology due to a local brain lesion. From January 2008 to June 2013, we used MRI to image patients with focal lesions; the patients showed ictal semiology of certain types of clinical generalised seizures. We localised the epileptogenic zones and performed respective surgeries in these patients. The goal of this study was to present the outcomes of surgical treatment for these patients and to analyse the relationship between the focal lesions and seizures with generalised semiology.
Patients and methods Patients Patients were selected from a pool of children and adults who underwent presurgical evaluation and surgical treatment for the relief of intractable epilepsy; the study was performed at the Comprehensive Epilepsy Centre of Beijing between January 2008 and June 2013. Among the total 1298 patients who received surgical treatment, 32 patients had local lesions on MRI-associated generalised seizures. The seizure type was classified according to the reports of the ILAE on the classification of seizures and epilepsies (Wieser et al., 2001; Engel, 2006; Berg et al., 2010). Despite the controversy, in this article, spasms were classified as generalised seizures. Patients were selected according to the following criteria: (1) patients with refractory epilepsy; (2) patients with severe seizures, including daily seizures with loss of consciousness, falls, or frequent status epilepticus; (3) patients with one or more types of generalised seizures were recorded by the video-EEG, with or without a partial seizure; (4) brain lesions on the MRI, including multiple lesions or hemispheric lesions; and (5) no contraindication to performing the surgical resection. The final sample consisted of 32 (2.4%) patients who underwent surgical resection. Three patients failed the postoperative follow-up. The seizure histories, imaging and video-EEG monitoring data, operative records and pathological findings of the remaining29 patients were analysed.
Sixteen patients (55.2%) were male, and 13 (44.8%) were female. Patient age at the time of surgery ranged from 3 to 34 years (mean age: 10.8 ± 8.0 years). The average seizure onset age was 3.1 ± 2.6 years (range 0.5—10 years).
Presurgical evaluation MR imaging Each patient underwent a standard MRI protocol that was performed using a 1.5-T or 3.0-T MR scanner (Siemens, Munich/Erlangen/Verio, Germany) and consisted of conventional spin-echo T1-weighted axial, sagittal, coronal and T2-weighted axial sequences (section thickness: 5 mm, image gaps: 1 mm). Additionally, 5-mm-thick Fluid Attenuated Inversion Recovery images were obtained; these images included axial and coronal sections that were perpendicular to the long axis of the hippocampus. Video-EEG monitoring and seizure semiology Interictal/ictal scalp EEGs were recorded using a video-EEG monitoring system (Micromed; Treviso, Italy); the electrodes were arranged according to the international 10—20 system. The duration of video-EEG monitoring ranged from 2 to 7 days, and at least three habitual seizures were recorded for each patient. We attempted to record every type of seizure if the patient had more than one type of seizure. The seizure type was classified by two epileptologists, according to the video-EEG findings (without the MRI information of the patient). If the patient was identified as having a certain type of generalised seizures, we further analysed the clinical semiology and interictal/ictal EEG, and we attempted to identify trends of asymmetry as a clue to the lateralisation of epileptogenic zones. Intracranial EEG monitoring and functional mapping Intracranial EEG monitoring was performed to further localise the epileptogenic zone; exceptions were for patients with a hemispheric lesion that was suitable for hemispherectomy or patients with a focal lesion distant to an eloquent area of the brain. The placement of the grid and strip electrodes was guided by non-invasive exams and intraoperative electrocorticography. The EEG sampling rate was set at 1024 Hz to record more details of seizure propagation. Usually, more than two habitual seizures were required to be recorded; we tried to record every type of seizure for each patient during the long-term intracranial EEG monitoring. Preoperative direct cortical stimulation for functional mapping was performed in patients with epileptogenic zones near the eloquent cortex, such as the motor, sensory or language areas.
94
Surgery and follow-up Hemispherectomy or hemispheric disconnection was performed in patients with hemispheric lesions. Tailored cortical resection was performed in patients with focal brain lesions that were thought to be related to the seizures. Intraoperative electrocorticography monitoring, both before and after cortex resection, was performed in these patients to further tailor the resection. Some patients were awoken from general anaesthesia to perform intraoperative functional cortical mapping or movement monitoring during the process of cortical resection. Patients were followed for at least 12 months to observe surgical outcomes. Using a binary outcome rating, the outcomes were divided into seizurefree (no seizures following surgery, or seizure-free for at least 12 months prior to the latest follow-up) or not seizurefree. All tissue sections from the surgical resection were routinely processed to obtain a detailed pathological diagnoses (Yue-Shan et al., 2010). Histopathological diagnoses were made by at least two neuropathologists.
Results The demographic and clinical characteristics of these patients are listed in Table 1.
Semiology According to the ictal videos of the 29 patients, 8 patients had a single type of seizure, 11 had two types of seizures, and 10 had more than two types of seizures. Epileptic spasms were the most common type of seizure and were recorded in 20 cases; both myoclonic seizures and tonic seizures were the second most common type of seizure and were recorded in 14 cases. Atypical absence seizures were found in four cases, and atonic seizures were found in two cases. In addition to the symptomatic generalised seizures, complex partial seizures (CPS) were also recorded in eight of the cases. CPSs were recorded secondarily to myoclonic seizures in three cases. The ictal semiology of the CPSs helped in the localisation of the epileptogenic zones for these patients. After careful analysis of the ictal semiology, certain asymmetric features of the behaviour in one or more types of seizures (including generalised seizures and CPSs) were observed in 24 of the patients. The asymmetric features of the symptomatic generalised seizures, such as unilateral dominant myoclonic seizures and asymmetric spasm behaviours, were thought to be helpful in lateralising the epileptogenic zones.
Neuroimaging All patients had localised abnormalities on MRI, which included hemispheric lesions in three patients, lesions in the frontal and/or central region in six patients, lesions in the frontal—central—parietal region in three patients, parietal—posterior temporal region lesions in 10 patients, parietal—occipital region lesions in four patients, occipital lobe lesions in one patient, and parietal lobe lesions in two patients (including one patient with a tuberous sclerosis
T. Yu et al. complex; however, the corresponding tubers were localised in the parietal lobe). The presence of a unilateral, congenital, or early-acquired lesion that was present on the brain MRI was helpful in localising the epileptogenic zones.
Scalp EEG Three patients showed dominant hemispheric interictal spikes, and two patients showed local interictal spikes. The remaining 24 patients showed generalised interictal epileptic discharges, such as bilateral sharp waves, spikes, and spike-wave complexes. In these 24 patients, slight asymmetries of the interictal discharges were discovered by a careful comparison of 7 patients. Of all of the ictal records, generalised ictal epileptic discharges (as the only pattern) were recorded in 13 patients, while distinguishable lateralised or localised ictal epileptic discharges were found in the remaining 16 patients. Four patients conformed to the standard of symptomatic Lennox—Gastaut syndrome according to the electro-clinical findings.
Intracranial EEG and cortical mapping Intracranial EEG recording on grid or strip electrodes was performed in 20 patients. Nine patients did not undergo intracranial EEG recordings, including three patients with hemispheric lesions, and six patients with focal lesions distant from eloquent areas of the brain. We analysed the ictal intracranial EEGs of every patient and found that most of the seizures with generalised clinical semiology had localised EEG origins. Low-amplitude fast activities were demonstrated in the beginning of the seizures in 10 patients; spike activities were observed in five patients. In the other five patients, either fast activities or spike activities were recorded in the beginning of the different seizures. The rhythms of fast activities in these patients could involve a few cortical electrodes in the beginning of the EEG recording. However, in some patients, time differences of several hundred microseconds between different cortical electrodes were only found through careful measurement. Usually, clinical behaviours occurred when the original fast activities spread to larger numbers of cortical electrodes. Both generalised seizures and CPSs were recorded in eight of 29 patients. In three of the eight patients, the behaviour and EEG of CPSs were recorded after myoclonic seizures. Preoperative functional mapping was performed in 17 cases, and intra-operative functional mapping was performed in eight cases. In this group, we observed an interesting phenomenon that in the same patient, different clinical ictal behaviours could originate from the same cortical region. For example, one patient (no. 28) showed the clinical behaviour of myoclonic seizures, epileptic spasms or tonic seizures when the ictal high frequency activities originating from almost the same region lasted from 200 ms to more than 3 s.
Surgical resection and complications Anatomical or functional hemispherectomy was performed for three patients with hemispheric lesions. Focal resection
Number
Demographic/clinical characteristics of the patients.
Age
Onset age
Gender
Seizure type
Unilateral predominate symptom
MRI
EEG
Interictal 1 2
12 8
5 1
M F
3 4
4 15
2 7
5 6 7
7 4 10
8
Surgery
Pathology
Interictal asymmetry − −
Ictal local origin + −
Intracranial ictal EEG Rhythmic spikes n
Resective location follow-up (month) R-P R-P
Follow-up (month) 36 24
F 6 F, 18 N
Tuberous sclerosis Scar lesions
+ −
+ −
n Fast activities
L hemispherectomy RF
24 24
F F
Scar lesions FCDI
+ + −
− + −
Fast activities Rhythmic spikes n
LC L C-posterior F R hemispherectomy
12 18 24
F F 6N, 18 F
FCDII Tuberous sclerosis Scar lesions
−
+
n
L P-posterior F
18
F
Scar lesions
+ +
+ +
Fast activities Rhythmic spikes
R P—O—posterior T RP
12 12
N N
B spike slow waves B spike slow waves L hemisphere spikes B spike waves B spikes or spike waves
−
+
Fast activities
R P—O—posterior T
18
N
Scar lesions Polymicrogyria with heterotopias Scar lesions
−
+
n
L hemispherectomy
18
F
Scar lesions
+
+
Polyspikes
12
F
− −
− −
12 18
F F
Polymicrogyria with heterotopias Scar lesions Scar lesions
Burst of B spike waves Burst of B spike waves B spike waves B slow spike waves
+
+
n Rhythmic spikes/fast activities Fast activities
R F-P—O—posterior T R P—O R P—posterior F
LP
12
N
FCDI
−
+
Fast activities
R P—posterior T
24
F
Scar lesions
+ −
+ −
L T—O R P—O
12 12
F N
Scar lesions Scar lesions
B spike slow waves B spike waves R F-C spikes R hemisphere spikes
−
−
Fast activities Rhythmic spikes/fast activities n
R P—posterior T
12
F
− + +
− + +
n Fast activities Rhythmic spikes/fast activities Polyspikes/fast activities Polyspikes
L O—posterior T RC RP
12 12 24
N N N
Meningeal angiomatosis Scar lesions FCDI FCDI
L P—posterior T
48
F
Scar lesions
L-F
12
N
FCDI
R P—O
12
N
Scar lesions
R peresylvian region
18
F
FCDI
Outcome
+ −
Multiple tubers Lesion in R T—P
F F
Spasm, CPS Spasm, myoclonic, tonic Spasm Spasm
+ +
L hemisphere atrophy FCD in L F
1 2.5 1
M F M
Myoclonic, tonic Tonic, atonic Myoclonic, tonic
+ + −
FCD in L C Lesion in L C R hemisphere atrophy
9
4
M
+
Atrophy in L F—P—O
9 10
8 14
4 3
M F
+ +
Atrophy in R T—P—O Polymicrogyria in R P
11
15
7
F
+
FCD in R F-P
12
29
2
M
Spasm, tonic, atypical absence, CPS Myoclonic, CPS, GTCS Spasm, myoclonic, CPS Myoclonic, atypical absence, CPS Spasm
+
L hemisphere atrophy
13
5
2.5
M
Spasm, myoclonic
+
Lesion in L T—P
14 15
3 13
0.5 5
M M
Spasm, tonic Spasm, tonic
− +
Lesion in R O Atrophy in R F-P
16
8
0.5
M
Tonic
−
Lesion in L P
17
11
3
F
+
Heterotopia in R T—P
18 19
13 34
8 8
F F
GTCS, myoclonic, atypical absence Spasm Spasm, tonic
+ +
Lesion in L T—P Atrophy in R P—O
20
4
1
F
Spasm
−
Lesion in L T—P
21 22 23
5 18 4
2 10 2.5
M M M
Spasm Myoclonic Myoclonic, spasm, tonic
+ + +
Lesion in L T—O FCD in R F Lesion in L post-F-C
24
3
1.5
M
Spasm, tonic
+
Lesion in L T—P
B spike waves
−
−
25
6
1
F
Myoclonic, CPS
+
FCD in L F
−
−
26
10
3
M
Spasm, CPS
+
Lesion in R P—O
+
+
27
6
2
M
+
Lesion in R peresylvian region
−
−
Polyspikes/fast activities Fast activities
28
7
0.5
F
+
−
−
Fast activities
L P—O
12
6 F, 6 N
Scar lesions
29
0.5
F
Lesion in B P—O, L predominate Lesion in R P—O
B spike waves
29
Spasm, myoclonic, tonic, atypical absence Spasm, myoclonic, tonic, atonic Myoclonic, CPS
B spike slow waves R F-P spike or spike rhythm Burst of B spike waves
R hemisphere spikes
+
+
n
R P—O—posterior T
18
F
Scar lesions
+
B frontal spikes B spike slow waves spike rhythm B spike slow waves B spikes B spikes B spike slow waves B posterior spike waves B spike waves B spike waves
95
CPS: complex partial seizure, GTCS: general tonic—clonic seizure, L: left, R: right, B: bilateral, F: frontal lobe, T: temporal lobe, P: parietal lobe, O: occipital lobe, C: central area, F: free, N: not free, FCD: focal cortical dysplasia.
Surgical treatment for patients with symptomatic generalised seizures
Table 1
96 of the epileptogenic zone was performed in 26 patients. Because of the location and size of the epileptogenic zone, the focal resection was not limited to a single lobe in some of the patients. In total, the resected epileptogenic zones involved a single lobe in 10 patients, two lobes in 11 patients, and three lobes in 5 patients (involving the parietal lobe in 19 cases, the occipital lobe in 14 cases, the posterior temporal cortex in 10 cases, and the posterior frontal or central cortex in nine cases). One child had a tracheospasm after the operation; the child gradually recovered after trachea cannula and hormonal treatment. One child had status epilepticus after multiple lobe disconnection; however, he demonstrated a ‘‘running down phenomenon’’ in the following months (Salanova et al., 1996); he subsequently became seizurefree 5 months after his operation. Unfortunately, one patient had epilepticus status 1 day after the operation; this was followed by mental disability. The patient continued to have frequent seizures after sedative treatment.
Pathological findings Scar lesions were found in 17 patients, including three patients with a history of encephalitis. Histopathologically, the destruction of the cortices with severe neuron loss, fibrous gliosis and cyst formations were observed. In certain cases, an island distribution of residual neurons, which had an oligodendroglia-like appearance, was observed; this was particularly common in cases secondary to ischaemia/hypoxia. Focal cortical dysplasia (FCD) was found in seven patients, polymicrogyria with heterotopias was found in two patients, tuberous sclerosis was observed in two patients, and meningeal angiomatosis was observed in one patient.
Postsurgical follow-up At the last follow-up (mean 18 ± 8.3 months, range 12—48 months), 17 (58.6%) patients were seizure free. The outcomes changed slightly in the months after the operation. When the patients were followed up for 3 months, 19 (65.5%) were seizure-free. When the patients were followed up for 6 months, 18 (62.1%) were seizure-free. One patient had an obvious increase in tonic seizures shortly after the operation, but the seizures decreased significantly 1 month later; the patient eventually became seizure-free 5 months after the operation. One patient experienced a recurrence of seizures 3 months after the operation. Two patients had seizure-free outcomes in the first 6 months after their operations followed by a recurrence of seizures in the following months; the seizures in these cases were reduced. The outcomes were stable after 12 months in the patients who were followed for more than 1 year. The variable features listed in the table were not significantly associated with seizure-free rates at the last follow-up (Table 2).
Discussion We report on a distinct group of patients with symptomatic generalised seizures and visible local lesions on
T. Yu et al. MRI. The type of generalised seizures included epileptogenic spasms, myoclonic seizures, tonic seizures, atonic seizures, and atypical absence seizures. Most of the local lesions were congenital, perinatal, or early-acquired brain lesions, which were considered to be malformations of the cortical development or scar lesions on pathologic examination. Most of the lesions were localised in the posterior temporal—parietal—occipital region or posterior frontal and central regions. The ages of the patients at seizure onset were relatively young. The interictal and ictal discharges of the EEGs were often generalised without notable traits of localisation. According to the ictal EEG recordings, most of our patients had more than one type of seizure; epileptic spasms were the most common type recorded. On careful analysis of the ictal semiology in these patients, a slight asymmetry of the behaviours could be found in some of these seizures. For example, the spastic behaviour was more obvious on the unilateral limb or shoulder in some patients with epileptic spasms. Unilaterally dominant myoclonic behaviour was also being found in some patients with myoclonic seizures. Despite the subtlety of the lateralising signs, they were important in lateralising the epileptogenic zone. The epileptogenic zones were thought to be in the hemisphere contralateral to the limb demonstrating more obvious behaviour. Furthermore, certain patients also have certain types of partial seizures; this was also helpful for lateralising or localising the epileptogenic zone. In our opinion, the symptomatic generalised seizures might have a similar origination with the partial seizures in the same patient. In our study, most of the patients underwent intracranial recordings; we recorded ictal EEGs to localise the epileptogenic zone. We found that most of the seizures with generalised clinical semiology have a localised EEG origin. Additionally, according to the ictal video-EEG recording, we observed an interesting phenomenon in that some patients with a focal lesion had different types of generalised seizures. Of our patients, 58.6% were seizure-free 1 year after the operation. The electro-clinical features and outcomes of surgical resection support the opinion that some seizures with generalised semiology might have focal origins. This is a topic that has been argued over for many years and has been gradually verified by clinical studies. A typical example is the successful surgery for patients with Lennox—Gastaut syndrome (LGS). It has been verified that a special group of LGS patients may have associated focal brain lesions (Bladin, 1985; Satow et al., 2002; Lee et al., 2014). Similarly, the aetiology of West syndrome has gradually changed from generalised seizures to seizures with multiple origins (Jonas et al., 2005; Komaki et al., 2001; Kang et al., 2006; Wyllie, 1996). However, an interesting and confusing question remains as to how local lesions lead to generalised seizures. It is the opinion of some that a cortical mechanism of secondary bilateral synchrony (SBS) might play an important role in patients with focal lesions who have generalised seizures (Kobayashi et al., 2000; Wyllie et al., 2007; Genton and Dravat, 1997). It has been demonstrated that epileptic spasms and myoclonic seizures could be caused by discharges originating from the focal cortex with extremely rapid secondary generalisation; the interhemispheric time
Surgical treatment for patients with symptomatic generalised seizures Table 2
97
Surgical outcome at last follow-up in relation to variables. N
Seizure freeN (%)
Not seizure freeN (%)
P value (Chi-square)
Age at surgery
