Resection surgery was initially reported by Horsley in 1886, in which three adult patients with focal onset seizures had localized epileptogenic lesions surgically removed at London’s National Hospital ^[36][74]. To be a suitable candidate for resection surgery, patients with focal epilepsy, that is, with semiological, neurophysiological, neuroimaging, and nuclear medicine imaging findings indicating focal or lateralized epileptogenic regions, are required ^{[37][38][39][40][41][42]}[25,34,35,75,76,77]. This concept of suitable candidates for resection surgery has also been applied to patients with ES or West syndrome ^{[16][17][18][19][20][25][43]}[16,17,18,19,20,64,78]. In previous studies involving resection surgery for patients with ES, most patients had epileptogenic lesions with a positive ratio on MRI ranging from 83% to 100% and comprised 94% of the total patients reported (Table 2). CC studies with countable data have shown that the ratios of patients with brain lesions ranged from 0% to 65%, comprising 17% of the reported patients (Table 1). These ratio differences indicate different surgical adaptations for each surgical strategy. Resection surgeries are usually performed in unilateral hemispheres in each patient, and lateralization signs/findings are crucial for adaptation. Lesions observed on MRI that are concordant with seizure laterality, the distribution of discharges on scalp EEG, or abnormal areas on nuclear scanning, are appropriate lesions for resection regions, as are epileptogenic lesions ^{[37][38][40][41]}[25,34,75,76]. As a representative study, Erdemir et al. reported 70 patients for whom resection strategies were decided based on lesions observed on MRI without invasive EEG monitoring [19]. The seizure-free ratio was 60% in their study, which was similar to that in other resection studies using invasive monitoring [19]. EEG and nuclear medicine imaging are also used to predict outcomes. Concordance between areas with interictal discharges on scalp EEG and MRI lesions or hypometabolism areas on positron emission topography (PET) in presurgical evaluations led to favorable findings in relation to seizure outcomes [20]. PET has also been reported to help with surgical decision-making and with guiding the placement of subdural electrodes in non-lesional MRI cases [16]. From these studies, focal or lateralized signs can be considered good indicators of resection surgery for ES, with the consistency of epileptogenic findings among each modality indicating a favorable predictive factor.

Table 2. Previous studies of resection surgery for epileptic spasms.

Authors		N	Age at Surgery	Seizure Types or Symmetry of ES	MRI Lesion	Procedures of Resection
1	Chugani et al. 2015 [16]	65	5.1 ± 4.4 years (0.2–19 years)	ES only: 23% ES + focal: 58% ES + generalized: 11%	Lesion (+): 92%	Hemispherectomy: 31% Subtotal hemispherotomy: 26% Multilobar: 20% Lobar ± tuberectomy: 14% Tuberectomy: 9%
2	Barba et al. 2016 [17]	80	5.8 ± 4.0 years	Symmetric ES: 49% Asymmetric ES: 24% Both types: 28%	Lesion (+): 96%	Lobectomy: 72.5% Multilobar/hemispherotomy: 27.5%
3	Chipaux et al. 2017 [18]	59	4.6 ± 3.5 years (0.3–16 years)	ES only: 15% ES + others: 85%	Lesion (+): 96%	Hemispherotomy: 34% Lesionectomy: 8% Lesionectomy + cortical resection *: 46% Frontal disconnection: 8% Posterior disconnection: 3% Endoscopic HH removal: 3%
4	Erdemir et al. 2021 [19]	70	1.9 ± 1.6 years	ES only: 46% ES + others: 54% Symmetric: 54% Asymmetric: 40% Unknown symmetry: 6%	Lesion (+): 100% Bilateral: 24% Unilateral: 76%	Hemispherectomy: 44% Lobectomy/lesionectomy: 33% Multilobar resection: 23%
5	Liu et al. 2021 [20]	64	3.3 ± 2.7 years	ES only: 64% ES + others: 36%	Lesion (+): 83%	Hemispherotomy/subtotal hemispherotomy: 27% Multilobar: 11% Lobar: 34% Tuberectomy: 23% Unknown: 3%
Outcomes			Good Prognostic Factor		Others
1	Seizure free: 71%		Short epilepsy duration MRI lesion (+)		22 patients discontinued medication after surgery
2	Seizure free: 61%		Complete resection of SOZ MRI lesion (+) Low age at surgery/short epilepsy duration		ECoG monitoring: 30%
3	Seizure free: 75%		Low age at surgery		Visible epileptogenic lesions on MRI in most cases. Eight patients were declined surgery following presurgical evaluation because of multiple foci.
4	Seizure free: 60% Improvement: 24%		Lobar/sublobar epileptogenic lesion		MRI-oriented surgery No invasive EEG (ECoG)
5	Seizure free: 83% **		Concordances: between MRI and scalp EEG between PET and scalp EEG Frequent interictal gamma activities (scalp EEG) Intraoperative continuous discharges (ECoG)		ECoG monitoring: 95% Stereo-EEG: 20%

Abbreviations: ES, epileptic spasms; HH, hypothalamic hamartoma; SOZ, seizure onset zone; MRI, magnetic resonance imaging; PET, positron emission tomography; ECoG, electrocorticography; EEG, electroencephalogram. * Resection of cortex with seizure onset zone in ECoG array; ** residual aura was accepted.

Functional deficits after the resection of eloquent cortices should be avoided as much as possible in resection surgeries ^[37][25]. Although the eloquent cortices include anterior or posterior language areas, visual cortices, and the primary motor area among others, hemiparesis after the resection of primary motor cortex is the critical for the patients with ES, who usually have severe cognitive impairment. The patients presenting with ES often have large epileptogenic zones and require multilobar resection or hemispherotomy/hemispherectomy ^{[16][17][18][19][20]}[16,17,18,19,20]. If the patients present early pathologic handedness or hemiparesis, the resection surgery is considered relatively acceptable for the patients. If the epileptogenic zone is spread widely in the unilateral hemisphere, but avoids primary motor cortex on intracranial EEG, subtotal hemispherectomy, which spares the primary sensorimotor cortex, is considered a good procedure for the treatment of ES ^[16][44][16,51].

5.2. Characteristics of Patients Who Undergo Corpus Callosotomy

8.2. Characteristics of Patients Who Undergo Corpus Callosotomy

In previous studies regarding CC, patients tended not to have localized or lateralized signs or findings. The largest study concerning CC surgery for patients with ES by Baba et al. included 56 patients without MRI lesions. They reported that 43% of patients were seizure-free and that 23% of patients had a ≥80% seizure reduction postoperatively [6]. CC has also been performed for patients with bilateral diffuse or multiple lesions in patients with polymicrogyria, cortical tubers, or general atrophy on brain MRI ^{[3][7][9][31]}[3,7,9,70]. Criteria for CC surgery to treat ES have included patients with no lesion or equally bilateral epileptogenic lesions selected for CC surgery. Factors such as presurgical severe developmental delay [6] and symmetrical ictal slow waves [10] on presurgical EEG findings have been associated with favorable CC outcomes. Based on advanced analyses for presurgical EEG, less power and less connectivity of high-gamma activities on presurgical scalp EEG [8] recordings and lower phase lags in ictal slow and gamma waves among bilateral hemispheres [11] have been associated with favorable seizure outcomes.

5.3. Types of Epileptic Spasms in Terms of Cortical Excitations and the Role of Corpus Callosum

8.3. Types of Epileptic Spasms in Terms of Cortical Excitations and the Role of Corpus Callosum

Figure 1 shows four pattern types in relation to cortical excitation and the role of the corpus callosum, and we describe the characteristics of the role of the corpus callosum, clinical features, neuroimaging, EEG, and surgical strategy based on previous knowledge and experience. Focal onset ES (type 1) is caused predominantly by unilateral cortices and tends to show a uniformly lateralized seizure type. These emerge with little or no contribution from the corpus callosum. Unilateral epileptogenic lesions on MRI or hypometabolic areas of nuclear medicine imaging are often detectable. The region or side involved in interictal EEG discharges tends to be concordant with lesions on brain imaging. Resection surgery can be performed based on presurgical evaluations. Potential focal onset ES (type 2) originates from the unilateral hemisphere; however, it is clinically not lateralized due to the rapid propagation of epileptic activity via the corpus callosum. The epileptogenic hemisphere is relatively difficult to identify using presurgical evaluations, which do not clearly show localized/lateralized or discordant localization/lateralization among the studies, and which only became apparent after CC. Additional resection surgery is sometimes performed in the newly evident epileptogenic region. There are two types of generalized onset ES in which epileptogenic cortices are equally distributed in bilateral hemispheres. However, the role of the corpus callosum in the emergence of ES differs between these types. In generalized onset ES or bilaterally independent focal onset (asymmetrical) ES with low callosal modulation (type 3), intrahemispheric modulations in each hemisphere predominates in the emergence of the ES. CC cannot prevent the emergence of ES. However, in generalized onset ES with high callosal modulation (type 4), interhemispheric modulation among the bilateral hemispheres via the corpus callosum predominates in the emergence of the ES, and CC can achieve seizure freedom without additional resection. In these cases, ictal EEG is visually bilateral and synchronous on slow waves [10]. Computed frequency analyses of presurgical scalp EEG recordings have shown small phase lags of ictal slow and beta waves among the bilateral hemispheres and less power and connectivity on interictal records ^[8][11][8,11].

Figure 1. Four pattern types from the viewpoint of cortical excitation and the role of the corpus callosum and the characteristics, clinical features, neuroimaging, EEG, and surgical strategy. Type 1: Focal onset epileptic spasms (ES) are caused predominantly by unilateral cortices and tend to show a uniformly lateralized seizure type. These ES emerge with little or no contribution from the corpus callosum. Unilateral epileptogenic lesions on MRI or hypometabolic areas of nuclear medicine imaging are often detectable. The region or side of interictal EEG discharges tends to be concordant with lesions on brain imaging. Resection surgery can be performed based on presurgical evaluations. Type 2: Potential focal onset ES originates from the unilateral hemisphere; however, rapid bilateral propagation of epileptic excitations via the corpus callosum often emerge as generalized onset (symmetrical) ES. The epileptogenic hemisphere is relatively difficult to identify with presurgical evaluations, which were revealed to be not clearly localized/lateralized or had discordant localization/lateralization among the studies, and became apparent after corpus callosotomy (CC). Additional resection surgery is sometimes performed in the newly evident epileptogenic region. Type 3: Generalized onset ES or bilaterally independent focal onset (asymmetrical) ES with low callosal modulation can emerge predominantly through intrahemispheric modulations in each hemisphere. CC cannot prevent the emergence of ES. Type 4: Generalized onset ES with high callosal modulation occurs through predominant interhemispheric modulation via the corpus callosum, and CC can achieve seizure freedom without additional resection. Abbreviations: CC, corpus callosotomy; ES, epileptic spasms; MRI, magnetic resonance imaging; NMI, nuclear medicine imaging; NA, not applicable.

8.4. Surgical Selection Flow Proposals

5.4. Surgical Selection Flow Proposals

Proposals for flow of surgical selection (resection surgery or CC) in patients with intractable ES are shown in Figure 2. In the schema, those patients with ES whose treatment had failed or who were not suitable for treatment with ACTH, vigabatrin, or other antiepileptic drugs, were evaluated using scalp video-EEG monitoring, brain MRI, and brain nuclear medicine imaging. If the epileptogenic region is localized in the unilateral hemisphere—which has been confirmed comprehensively using the following lateralization signs: asymmetrical ES (consistently in the same side), an epileptogenic lesion within the unilateral hemisphere, concordance of localization of interictal discharges on EEG and MRI lesions, and concordance of localization of interictal discharges on EEG and hypometabolism in nuclear medicine imaging—then resection surgery should be selected. If patients have no or fewer lateralization signs, that is, no or bilateral multiple epileptogenic lesions, non-lateralized epileptogenic findings on EEG or nuclear medicine imaging, symmetrical ictal slow waves and/or discordance of these signs, then CC is selected as first-line treatment. Patients who continue to have ES can receive additional resection surgery if re-evaluations (mainly neurophysiological examinations) show lateralization signs. If additional resection surgery is not appropriate, treatments using vagus nerve stimulation, dietary therapy, and other antiepileptic drugs can be attempted.

Figure 2. Selection flow of surgery for medically-intractable epileptic spasms (ES). If the epileptogenic region is identified in the unilateral hemisphere on the basis of lateralization signs—asymmetrical ES, epileptogenic lesion localized or lateralized in the unilateral hemisphere, concordance between localization of interictal discharges on EEG and MRI lesions, and concordance between localization of interictal discharges on EEG and hypometabolism in nuclear medicine imaging—the patient is presumed as lateralized case (ES type 1 in Figure 1) and resection surgeries (including frontal/posterior disconnection, subtotal or total hemispherotomy/hemispherectomy) are selected. Patients with early pathologic handedness or hemiparesis before surgery are considered relatively good candidates for resection surgery. If the patients have no, less or discordant lateralization signs—no or bilaterally multiple epileptogenic lesions, non-lateralized epileptogenic findings on EEG or nuclear medicine imaging, symmetrical ictal slow waves and/or discordance among findings of EEG/MRI and nuclear medicine imaging—the patient is presumed as non-lateralized case (ES type 2–4 in Figure 1) and corpus callosotomy (CC) is selected first. The ratio of seizure freedom in CC is estimated at 25–79% from previous studies (Table 1). If the patient is under 10 years of age, presenting severe cognitive impairment or unilateral widespread cerebral lesion, total CC is selected. In other patients, anterior, or posterior is initially recommended (the patients with unilateral widespread cerebral lesion can be considered for hemispherotomy/hemispherectomy). Additional total CC (two-staged CC) is considered in the patients with residual ES. Type 4 ES mostly ceases only by CC. If re-evaluations (mainly neurophysiological examinations) show lateralization signs, the ES is thought of as type 2 and additional resection surgery is recommended (seizure freedom: 43–71%). If the ES remains after CC (no change, bilaterally independent or milder ES than before), other treatments including antiepileptic drugs, dietary therapy and/or vagus nerve stimulations are selected.