Lennox-Gastaut syndrome (LGS) is considered an epileptic encephalopathy and is characterized by a triad of intractable seizures (in particular tonic seizures during sleep, but atonic and atypical absence seizures are also common), cognitive and behavioral impairments and diffuse slow spike-and-wave (SSW) and paroxysms of fast activity (PFA) on electroencephalography (EEG) (1). LGS is estimated to affect between 1 and 2% of all patients with epilepsy (2). Generally, LGS often occurs in young children (3).

The etiology of LGS can be classified as cryptogenic or symptomatic. LGS of symptomatic etiology may be secondary to hypoxic ischemic encephalopathy, congenital brain malformation, vascular malformation, genetic conditions such as tuberous sclerosis, trauma, brain tumor, or perinatal meningoencephalitis (4). Among them, viral encephalitis (VE) is a known cause of LGS. VE is a severe disease of the central nervous system that usually presents with seizure attacks and progressive neurological deficits in the acute phase (5, 6). Intractable epilepsies that occur several months to years after acute encephalitis are called postencephalitis epilepsies (PEs), some of which present as LGS (7, 8).

Although the majority of patients with LGS have diffuse EEG patterns, some focal changes causing secondary generalized epileptic encephalopathy can be identified on neuroimaging (1). Patients with such focal lesions can be treated surgically through comprehensive preoperative evaluation. However, studies on the surgical treatment of LGS after VE are rare. Therefore, the objective of this study was to discuss surgical strategies for secondary LGS of VE origin in children.

Lennox-Gastaut syndrome is one of the most serious forms of intractable epilepsy in children. Generally, during the course of LGS, more than 70% of patients show no response to anti-seizure medicines, and less than 10% of patients achieve a seizure-free status (10). It has been reported in the literature that for LGS children with focal lesions on MRI, despite a lack of localized epileptic patterns on EEG, epilepsy surgery may be a treatment strategy (11, 12).

Malformation during cortical development is the most common cause of surgeries for children with LGS; other reasons include tuberous sclerosis complex, perinatal complications, craniocerebral trauma, and infection (13). The reported risk of late unprovoked seizures in population-based cohorts of cerebral infection survivors from developed countries is estimated at 7–8% (5). Studies have reported that the main risk factors for epilepsy after VE include epileptic seizures in the acute phase of encephalitis, status epilepticus, and abnormally hyperintense T2/FLAIR signals on MRI, often involving the temporal and frontal lobes (14, 15).

Children with LGS secondary to VE often present with a number of unique characteristics. The diagnosis of LGS is much more complicated in the clinic. The inclusion criteria included seizures, particularly generalized tonic, atonic and myoclonic seizures as well as atypical absence seizures and spasms, generalized SSW and PFA on EEG and progressive developmental regression (16). However, it can be difficult to find all the conditions that fulfill the diagnostic criteria. Moreover, patients with LGS secondary to VE often have obvious structural brain lesions on MRI, despite diffuse or generalized epileptiform discharge patterns on EEG. Discordance in the investigations may lead to more difficult localization of the epileptogenic zone.

Despite the high prevalence of drug-resistant epilepsies after cerebral infection, especially in patients with MRI-identifiable lesions, only a small number of patients undergo epilepsy surgery (14). The effect of surgical treatment of epilepsy secondary to VE is often poor because of the large extent of brain damage during acute VE, which often involves bilateral brain regions. In the acute phase, MRI is often inconsistent with that in the residual phase. However, the surgical outcomes in certain subgroups of patients with epilepsy after VE are encouraging (17). At present, there is no literature reporting the surgical outcomes of patients with LGS secondary to VE.

Among our cases, all of our patients had obvious lesions on MRI, despite the lack of localized epileptic patterns on interictal and ictal EEG. These patients all underwent different surgical approaches, and the prognoses were different. The surgical strategies were mainly based on neuroimaging data.

If the MRI shows obvious bilateral lesions, such as in case 2 in our series, palliative surgery should be performed. Both CC and vagus nerve stimulation (VNS) are safe and effective treatments for LGS patients (18, 19). Some meta-analyses have supported that CC has a better seizure outcome than VNS (20, 21). Therefore, in our case, CC was performed, and a more than 80% seizure reduction was reported. Both intellectual and motor development were improved postoperatively, which is consistent with previous literature (22–24).

If instead, MRI reveals obvious unilateral lesions, as in cases 1, 3, and 4 in our series, curative surgery may be a promising treatment. However, some important tips should be noted. First, although the three cases presented with prominent lesions on one side, mild abnormalities can sometimes be observed on the contralateral side. Surgery on the prominent side is intended to reduce the frequency and severity of seizure attacks. In case 3, left TPO disconnection was performed, but a mild abnormal signal was observed in the right posterior insular lobe. Seizures recurred 1 year postoperatively, but tonic and atypical absence seizures were not observed. Therefore, in these situations, the surgeon and parents should have extensive discussions regarding the advantages and disadvantages of surgery. Second, MRI in the residual phase should be compared with that performed in the acute phase. Sometimes, more obvious lesions can be seen in FLAIR/diffusion-weighted imaging in the acute phase, while in the residual phase, there are no obvious lesions in the same region on MRI.

Although MRI provides much more localization or lateralization information, PET images can provide supplementary information for the former. In some studies, inconsistency between PET and MRI often heralded a poor outcome (17, 25). In case 4, MRI revealed encephalomalacia in the left temporal lobe; however, PET showed hypometabolism that was more extensive than the encephalomalacia observed on MRI. Left temporal lobectomy was performed, but the seizures recurred 1 month postoperatively. Therefore, PET should be carefully analyzed to better determine the border of the lesions.

The medical history should be noted as well. Cases 2, 3 and 4 had a history of febrile convulsions in the acute phase. And all of them had poor outcome. However, previous studies found that febrile convulsions were a predictor of good surgical outcome (26, 27). Different etiology may explain this apparent discrepancy. Our study included only patients with LGS secondary to VE. Previous studies included patients with mesial temporal lobe epilepsy secondary to hippocampal sclerosis. Moreover, immune state also should be noted. Anti-N-methyl-D-aspartate receptor (NMDAR) encephalitis after virus infection is a recently identified constellation that should be distinguished from drug-resistant epilepsy secondary to VE. NMDAR antibodies have been frequently detected in patients with virus encephalitis. Anti-NMDAR encephalitis after virus infection often received immunosuppressive therapy instead of surgery, because surgical treatment for anti-NMDAR encephalitis after virus infection often associated with poor outcome.

There are several limitations in our study. This was a retrospective study, and only four cases were included; therefore, statistical analysis could not be performed. More patients will be included in further investigations to analyse the prognostic factors for the outcome of treatments for LGS.

Surgery is a challenging treatment method for patients with LGS secondary to VE. The surgical outcomes can differ due to anatomo-electro-clinical correlations. Generally, compared to malformation of cortical development, VE results in more obvious lesions on MRI. However, it should be emphasized that careful analysis of MRI features of MRI in both the acute and residual phases is very crucial. Concordance between MRI lesions and PET imaging findings corresponded to better surgical outcomes. In addition, generalized seizures and diffuse scalp EEG findings were not contraindications to epilepsy surgery. Therefore, surgical decisions regarding LGS secondary to VE should be determined very cautiously.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

This study was approved by the Institutional Review Board of the Ethics Committee of Peking University First Hospital. Written informed consent to participate in this study was provided by the participants' legal guardian/next of kin. Written informed consent was obtained from the individual(s), and minor(s)' legal guardian/next of kin, for the publication of any potentially identifiable images or data included in this article.

LC, YJ, and XL designed the study and revised the paper. NW, CL, and QL analyzed the data and drafted and revised the paper. CL, HY, and YS collected the data. SW and TJ helped to select the patients. GY helped to interpret the EEG data. YW performed the patient follow-up. All authors contributed to the article and approved the submitted version.