Epilepsy surgery in children operated on before 3 years of age – experience of the Motol epilepsy center

Summary

Aim: The aim of this study was to comprehensively evaluate clinical characteristics, surgical strategies, and outcomes of epilepsy surgery in children with focal structural epilepsy operated on before 3 years of age at the Motol Epilepsy Center, and to identify factors associated with postoperative seizure control and neurodevelopmental outcomes. Methods: A retrospective analysis was performed on a cohort of children who underwent resective or disconnective epilepsy surgery before three years of age between 2000 and 2025. Clinical and electroencephalographic characteristics, neuroimaging findings, type and extent of surgery, histopathological diagnosis, postoperative complications, and long-term seizure outcomes were assessed. Postoperative seizure control was evaluated using the Engel and International League Against Epilepsy (ILAE) classifications. Neurodevelopment was assessed using developmental or intelligence quotient measures before and after surgery. Statistical analysis included non-parametric tests and correlation analyses. Results: A total of 77 surgical procedures were performed in 66 children. The median age at surgery was 22 months. The most common etiologies were malformations of cortical development and tuberous sclerosis complex. Earlier epilepsy onset was observed in patients with a more extensive epileptogenic zone. Long-term seizure freedom (≥ 2 years) was achieved in 77% of patients, of whom 70% were able to discontinue antiseizure medication completely. A longer interval between epilepsy onset and surgery and older age at surgery were associated with poorer postoperative outcomes. Permanent neurological deficits were rare. Lower preoperative developmental quotient was associated with earlier epilepsy onset and a larger epileptogenic zone. Cognitive performance and the prevalence of associated neurodevelopmental disorders remained stable after surgery. Conclusion: Epilepsy surgery in children operated on before three years of age is an effective treatment with a high likelihood of achieving seizure freedom and the potential for complete withdrawal of antiseizure medication. Early surgical indication is a key determinant of favorable outcomes and may contribute to stabilization of the neurodevelopmental profile in this high-risk patient population.

Keywords:

epilepsy surgery – Quality of life – complications – drug-resistant epilepsy – presurgical evaluation – seizure control – cognitive outcomes

This is an unauthorised machine translation into English made using the DeepL Translate Pro translator. The editors do not guarantee that the content of the article corresponds fully to the original language version.

Introduction

Epilepsy surgery in the youngest age group—from birth through the toddler years—clearly illustrates the current trend toward  early      surgical intervention in children with focal structural epilepsy, as well as the breadth of procedures used in modern epilepsy surgery. Children who undergo surgery in the first years of life represent a very specific subgroup of candidates for resective           or disconnective      epilepsy surgery. The structural basis of focal epilepsy at this age consists almost exclusively of congenital lesions, most commonly malformations of cortical development (MCD). The earlier the onset of epilepsy     , the more likely it is that the epileptogenic zone will be extensive or multifocal and located outside the temporal lobes. A significant subgroup of these children consists of patients with hemispheric epileptic syndromes, which may be caused by developmental lesions (hemimegalencephaly, cortical dysplasia, tuberous sclerosis complex [TSC]) as well as destructive lesions (e.g., prenatal ischemia). Hypothalamic hamartomas constitute a specific entity.

The course of the disease in these children is often catastrophic: seizures are very frequent, the epileptogenic zone is extensive, and psychomotor development is dramatically affected by the disease, often leading to stagnation or regression. At the same time, these children have the highest cumulative risk of developing epileptic encephalopathy, lifelong exposure to the adverse effects of antiseizure medications (ASM) [1], and sudden unexpected death in epilepsy (SUDEP) [2]. Therefore, surgical treatment may be indicated even in a small proportion of patients who do not yet meet the criteria for drug resistance; however, a much more common scenario involves semi-acute indications such as an extremely high seizure frequency, EEG status epilepticus, or regression in psychomotor development.

Surgical strategies encompass a full spectrum of procedures—ranging from hemispherotomies, multilobar resections, and disconnections to lesionectomies and      endoscopic procedures indicated for hypothalamic hamartomas. Given the extent and nature of the lesions, the scope of surgery tends to be greater in the youngest patients than in older children or adults. Diagnostic options are limited at this age—some methods cannot be used at all (e.g., functional MRI, stereo-EEG before the age of 2 [3]), while others are feasible only with difficulty (intraoperative mapping of the eloquent cortex).

Patients younger than one year of age are also at the highest risk for perioperative complications, particularly in connection with blood loss due to a smaller total blood volume [4] and a higher proportion of water in brain tissue. However, these risks are offset to some extent by the extraordinary plasticity and regenerative capacity of the developing brain, which allows for compensation even after more extensive or complicated procedures. Nevertheless, it appears that early surgical treatment without significant delays yields the best long-term outcomes. Large multicenter studies have unequivocally demonstrated that any delay in surgical intervention reduces patients’ chances of achieving seizure freedom [5,6], discontinuing ASM     , and stabilizing or improving cognitive development.

Despite knowledge of these facts, however, cases involving the youngest patients with focal structural epilepsy are often perceived as highly complex, leading multidisciplinary teams to hesitate when deciding on surgical intervention—with all the aforementioned adverse consequences.

Our study aimed      to conduct a comprehensive analysis of a group of children with focal structural epilepsy who underwent resective      or disconnective  epilepsy surgery at the Epilepsy Center of Motol University Hospital before the age of three. We studied the clinical presentation, neurological and neuropsychological findings, EEG results, neuroimaging findings, types of surgical procedures performed, the histological basis of the disease, and the overall outcomes of epilepsy surgery.

 

Materials and Methods

We retrospectively analyzed comprehensive data from children who underwent resective or disconnective      epilepsy surgery at the Epilepsy Center of Motol University Hospital between 2000 and 2025. Basic clinical data were available for all patients, including seizure type and frequency, results of video/EEG monitoring, and brain MRI      performed using a dedicated epilepsy      protocol. In indicated cases, neuroimaging was supplemented with FDG-PET (fluoro-2-deoxy-D-glucose) or SISCOM (subtraction ictal SPECT co-registered to MRI). The first seizure was retrospectively classified according to the current classification of the International League Against Epilepsy (ILAE) [7]; for neonates     , the 2021 ILAE classification [8] was used. To assess cognitive development, either the Developmental Quotient (DQ) from the Gesell Developmental Scale or the Mental Development Index from the Bayley Scales of Infant and Child Development was used; both are referred to here as the Developmental Quotient (DQ). The specific method selected was based on the patient’s clinical condition and ability to cooperate. Some patients did not undergo psychological assessment due to organizational reasons or because of their very young age. A follow-up psychological examination was conducted at least one year after surgery. If the child had reached an age and developmental level permitting the administration of a standardized intelligence scale, intellectual performance (intelligence quotient; IQ) was assessed using a Wechsler scale. Otherwise, developmental scales were used again. Any neurodevelopmental comorbidities were also clinically identified, particularly atypical social interaction (suspected autism spectrum disorder [ASD]) or confirmed ASD     , symptoms of attention-deficit/hyperactivity disorder (ADHD), and  developmental language disorders.

Furthermore, the type and extent of the neurosurgical procedure performed were recorded; in this context, expected postoperative neurological deficits and unexpected complications [9]—including their probable causes—were assessed separately. The extent of the procedures was also evaluated quantitatively. In patients who underwent hemispheric procedures     , brain tissue segmentation was performed on preoperative      MRI, and the proportion      of the resected or disconnected volume to the total cortical volume                was calculated. For the other children, post-resection MRI was used; the extent of resection was delineated in the 3D Slicer software (version 5.2.1) and subsequently compared with the total cortical volume. All MRI data were further processed in MATLAB R2024b (The MathWorks Inc., Natick, MA, USA) using the CAT12.9 tool (Structural Brain Mapping Group, Jena University Hospital, Jena, Germany) [10].

One year after surgery, standard follow-up clinical examinations, EEG, and MRI scans were performed, as well as a follow-up neuropsychological evaluation. Postoperative seizure control was assessed using the ILAE outcome classification      and the Engel classification [11,12]; reductions in or discontinuation of ASM      were also recorded. Reoperations performed before the patient reached three years of age were also included in the evaluation.

For statistical analysis, the Mann-Whitney U test, Spearman’s correlation coefficient, and Fisher’s exact test were used, depending on the type of variables analyzed. The level of significance was set at p < 0.05.

 

Results

Between 2000 and 2025, a total of 77 epilepsy surgeries (resective      and disconnective      procedures) were performed at the Epilepsy Center of Motol University Hospital on 66 children under the age of three. This represents approximately 15% of all procedures; however, in recent years there has been a significant increase in the number of surgeries in this age group, and in 2024, for example, it already accounted for 27% of all procedures.

The median age at the onset of epilepsy (Fig. 1) was lowest      in patients with TSC (1 month; interquartile range [IQR] 0.5–2.8), followed by patients with MCD (2 months; IQR 0.3–6). In children with post-ischemic lesions (with a clinical diagnosis of cerebral palsy: CP), epilepsy developed predominantly around half a year of age (median 5 months; IQR 3.8–9.8). A larger extent of the epileptogenic zone was observed in patients with earlier-onset epilepsy (p = 0.008, Spearman’s r = –0.40).

The median age at the first surgical procedure was 22 months, and the duration of epilepsy prior to surgery was 12.5 months. The most common etiologies were MCD (n = 33) and TSC (n = 15), followed by epilepsy-associated tumors (n = 9). In five cases, these were associated with focal cortical dysplasia (FCD IIIb).

The first seizure was almost uniformly focal with motor manifestations (65/66). Only one seizure was focal without motor manifestations. The specific motor manifestations varied; in 16 cases, they consisted of spasms, while in the remaining cases, tonic or clonic seizures predominated. Spasms were documented as one type of seizure in 34 patients, with a median age of onset at 5.5 months (IQR 4–7). Given the children’s age and limited testing options, impaired consciousness during a seizure could be definitively ruled out in only a handful of patients. Individual patients with known structural pathology (TSC or ischemic lesion) were treated presymptomatically with ASM when significant epileptiform findings were observed on EEG; however, this did not prevent the development of drug resistance in this group. Hormone therapy (adrenocorticotropic hormone, i.m.) was indicated in 16 patients; in 7 patients, it was ineffective or only reduced seizure frequency, whereas 9 patients responded, with the effect lasting a median of 3 months (IQR 2.5–5.0).    

A significant proportion of the initial procedures were indicated on an emergency basis—9/66 due to a life-threatening indication, and 6/66 due to extremely frequent seizures. Seven of these patients and an additional 14/66 patients had EEG findings consistent with status epilepticus.    

Brain magnetic resonance imaging (MRI) revealed a causative structural lesion in all cases. FDG-PET was most commonly indicated in patients with MCD (21/33) and tumors (5/9). SISCOM was typically indicated in patients with TSC (6/15).

Long-term invasive monitoring using intracranial electrodes was performed only rarely (twice using subdural electrodes and once using stereo-EEG electrodes). In contrast, intraoperative electrocorticography (ECoG) was      routinely performed during      all procedures, with the exception of hemispherotomies and hypothalamic hamartoma surgeries.

The extent of surgery varied depending on the etiology. The most extensive procedures were performed in patients with ischemic lesions, all of whom underwent hemispherotomy of an atrophic hemisphere (5/5), and in patients with hemispheric MCD (11/33 patients with MCD). In cases of hemimegalencephaly, the resected or disconnected volume reached up to 0.67 of the total cortical volume.      In patients with TSC, lobar resections or extended lesionectomies were the most common procedures (10/15), followed by multilobar resections (5/15). For tumors, the preferred procedure was extended lesionectomy (5/9), indicated to remove any associated FCD IIIb. Patients with Sturge-Weber syndrome underwent multilobar resection (1/2) or hemispherotomy (1/2).

Unexpected complications occurred in 16 of 77 surgeries (20.8%); in 8 cases, they were classified as “major” (with consequences lasting more than 3 months, life-threatening, or requiring neurosurgical intervention), and in another 8 cases as “minor.” Their nature was highly varied—e.g., 3 cases of hydrocephalus, 2 cases of infection     , 4 cases of hemorrhage, 3 cases of ischemia, 1 case of edema, etc. Two patients had a permanent neurological deficit as a result of an unexpected complication. An expected neurological deficit resulting from the nature of the procedure (i.e., the location and extent of the resection or disconnection) was recorded after 29 of 77 procedures (37.7%). In 19 cases, this involved worsening of preoperative hemiparesis (predominantly in children following hemispheric disconnection); in the remaining 10 cases, it involved an isolated visual field defect (most commonly hemianopsia resulting from a more extensive procedure in the posterior quadrant).

Among the 57 patients with at least 2 years of postoperative follow-up, 44 (77%) were seizure-free (Table 1). ASM was completely discontinued in 31 of these 44 patients. Seizures persisted postoperatively in 11 patients, and two additional patients underwent reoperation after 3 years of age. Nine patients had been followed for less than 2 years, and long-term outcomes were therefore not yet available.      

A longer interval between the onset of epilepsy and epilepsy surgery was associated with poorer postoperative seizure control (Mann-Whitney test; p = 0.01); this was also observed in patients who underwent surgery at a later age (p = 0.02). By contrast, age at epilepsy onset was not associated with postoperative seizure outcome.      Neither the type of pathology nor the preoperative EEG findings had a significant effect on postoperative seizure control. Seizure control was also not dependent on the extent of resection (p = 0.14); however, there was a noticeable trend toward smaller resections in patients with postoperative seizure recurrence.

Preoperative DQ was assessed in 56 of 66 patients, with a median of 75.0 (54.5–90.0). Of this group, postoperative DQ was available for 22 patients, with a median of 70.0 (42.0–90.0). Postoperative IQ was available for 28 of 56 patients, with a median of 76.0 (68.0–91.5). The relationship between these quotients with respect to individual structural causes of epilepsy is shown in Fig. 2. For the remaining 6 patients in this group, neither DQ nor IQ was assessed postoperatively. A lower preoperative DQ was associated with a larger extent of the epileptogenic zone (p = 0.0002, Spearman’s rho = –0.47) and earlier onset of seizures (p = 0.0005, Spearman’s rho = 0.45). A trend toward higher DQ scores prior to surgery was also observed in patients who remained seizure-free postoperatively (Mann-Whitney; p = 0.11). We observed a high prevalence of associated neurodevelopmental disorders: ASD (23%), ADHD (33%), and developmental language      disorders (33%). There was no significant change in their prevalence after surgery.

Figures 3–5 show the findings for the patients depicted in Figure 1.

Discussion

Our cohort demonstrates a high success rate of surgical treatment for epilepsy in the youngest age group of children—those operated on before the age of 3—which, in terms of postoperative seizure control, is entirely comparable to the success rate of surgeries in older children,      despite the limitations of presurgical evaluation      (     surgeries performed urgently, a limited range of preoperative diagnostic tests, including invasive monitoring) and the necessity of performing the most technically complex procedures (hemispherotomy, multilobar resection, or disconnection).

The number of functional tests performed (SISCOM, FDG-PET) was lower in our cohort than in older children [13]. This is likely related to the more frequent detection of structural MRI lesions in the youngest patients. Furthermore, for certain diagnoses—such as gliotic lesions      following prenatal stroke—metabolic or perfusion methods typically do not provide significant new information.

The age at epilepsy onset fell within a relatively narrow range  for MCD, TSC, and prenatal ischemia, whereas for tumors associated with epilepsy, it was distributed almost evenly across the entire age range studied     .

An association has been demonstrated between a larger extent of the epileptogenic zone and earlier onset of seizures. Surprisingly, this association is less pronounced in patients diagnosed with TSC, for whom relatively limited procedures are often indicated at our center; however, even these procedures have the potential to disrupt complex epileptogenic networks in patients with this diagnosis. Furthermore, a larger extent of the epileptogenic zone was observed in patients with a lower preoperative DQ. These are often patients with extensive, frequently hemispheric lesions, where the developmental level also reflects the presence of a motor disorder in the form of hemiparesis, which was often part of the preoperative neurological findings. The association between earlier onset of epilepsy  and more pronounced developmental delay at the time of the preoperative evaluation further illustrates this relationship. The relatively low preoperative DQ indicates that a large portion of the study cohort is at risk of developing an intellectual disability; moreover, a relatively high percentage of patients also exhibit symptoms of other neurodevelopmental disorders. Stable postoperative cognitive performance and a similarly unchanged prevalence of associated neurodevelopmental comorbidities demonstrate the potential of surgical intervention to prevent  further deterioration      in cases of ongoing epileptic encephalopathy. These results are consistent with the findings of studies of similar cohorts abroad [14].

The youngest children have a high likelihood           of completely discontinuing      ASM      after surgery. The importance of early surgical intervention in this patient group is underscored by the significant proportion of cases requiring acute or even life-saving surgery due to extremely frequent or life-threatening seizures. Our study confirms the relationship between early  indication for surgery and a higher likelihood of achieving a seizure-free state. The extensive nature of the      procedures and their urgent      indications—in a situation where most patients have very frequent seizures and/or continuous epileptiform EEG activity at the time of surgery—are associated with a relatively higher number of complications and expected postoperative deficits than in older children or adults [13]. However, permanent neurological deficits resulting from unexpected complications were rare in our cohort (in 2 of 66 cases).

The growing experience of the multidisciplinary team at the Epilepsy Center of Motol University Hospital has resulted in an increasing number of procedures being indicated at such an early age     . We believe that optimizing both diagnostic work-up      and surgical treatment will lead to better long-term outcomes across all variables describing surgical outcomes (seizure control, ASM  use, cognitive function, and quality of life).

 

Ethical Aspects

This study was conducted in accordance with the 1975 Declaration of Helsinki and its revisions in 2004 and 2008. The study was approved by the Ethics Committee of Motol and Homolka University Hospitals and the Second Faculty of Medicine, Charles University in Prague (February 11, 2026, EK-35/26).

 

Funding

Supported by the Ministry of Health of the Czech Republic in cooperation with the Czech Health Research Agency under projects No. NW25-04-00427 and NU23-08-00528, the conceptual development project of the research organization 00064203-6005, the Ministry of Education, Youth, and Sports of the Czech Republic as part of the ERDF project “Brain Dynamics –⁠ CZ.02.01.01/00/22_008/0004643,” co-financed by the European Union, and the National Recovery Plan project LX22NPO5107 (MŠMT), funded by the European Union. All rights reserved under intellectual property laws.

 

Conflict of Interest

The authors declare that they have no conflict of interest regarding the subject matter of the study.

 

	S eizure-free postoperatively (Engel I) (n, %) / median (IQR)	Incomplete seizure control Engel II+ (n, %) / median (IQR)	p      value
Onset of epilepsy (months)	2.5 (0.6–8.5)	5 (3.2–8.2)	0.31
Duration of epilepsy (months)	11 (5.0–19.0)	21 (13.8–26.8)	0.01*
Age at surgery (months)	22 (8.0–28.0)	30 (22.0–33.2)	0.02*
DQ before surgery	75 (55–90), 39/44	55 (44–76), 9/11	0.11
Neurodevelopmental comorbidities before🡪 after
PAS	7/33 → 9/33	3/7 → 3/7
ADHD	9/30 → 12/30	3/6 → 2/6
speech disorder	9/33 → 8/33	1/7 → 4/7
Histological diagnosis
–⁠ MCD	24 (82%)	5 (17%)
–⁠ TSC	8 (73%)	3 (27%)
–⁠ scar	4 (80%)	1 (20%)
–⁠ tumor	8 (89%)	1 (11%)
–⁠ Sturge-Weber	0	1 (100%)
Extent of surgery
–⁠ hemispheric	14 (88%)	2 (12%)
–⁠ multilobar	7 (70%)	3 (30%)
–⁠ Unilobar resection	4 (57%)	3 (43%)
–⁠ focal resection	19 (86%)	3 (14%)
–⁠ resection volume	0.11 (0.04–0.43)	0.05 (0.03–0.15)	0.14
Interictal EEG abnormality
–⁠ generalized	2 (50%)	2 (50%)
–⁠ hemispheric	10 (90%)	1 (10%)
–⁠ multiregional	16 (70%)	7 (30%)
–⁠ regional	14 (100%)	0 (0%)
–⁠ none	2 (100%)	0

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 1. Comparison of clinical data with postoperative seizure control. Only patients followed for 2 or more years after surgery are included.

ADHD –⁠ attention-deficit/hyperactivity disorder; DQ –⁠ developmental quotient; IQR –⁠ interquartile range; MCD –⁠ cortical developmental malformation; ASD –⁠ autism spectrum disorder; TSC –⁠ tuberous sclerosis complex