Surgical treatment of epilepsy in children with gloneuronal brain tumors: morphology, MRI semiology and factors affecting the outcome

Melikian A.G.; Shishkina L.V.; Vlasov P.A.; Kozlova A.B.; Shul'ts E.I.; Kushel' Iu.V.; Korsakova M.B.; Buklina S.B.; Varyukhina M.D.

doi:https://doi.org/10.17116/neiro2020840116

Abbreviations

AE — antiepileptic

DNT — dysembryoplastic neuroepithelial tumor

GNT — glioneuronal tumor

ECo-G – electrocorticography

Video-EEG – video-electroencephalography

SSEP — somatosensory evoked potentials

BEA — bioelectric activity

Syn — synapotophysin

GFAP — glial fibrillary acidic protein

AHoA — anterior choroidal artery

Introduction

Epileptic seizures often accompany brain tumors. Moreover, this is common symptom in patients with benign intracerebral tumors. The term “Long-term Epilepsy Associated Tumors (LEAT)” is widely accepted in modern literature [1—4]. Tumors is the second cause of epilepsy in all major series of patients undergoing surgical treatment. Glioneuronal tumors (GNT) occupy a special place among above-mentioned neoplasms. These tumors consist of ganglion cells (in gangliogliomas) or altered forms of neurons and specific glioneuronal component (in dysembryoplastic neuroepithelial tumors) in addition to glial tumor cells. GNTs account less than 1.5% of all brain tumors in children [5]. However, unlike other benign gliomas, these tumors are characterized by predominant cortical lesion, almost no progression. GNTs do not cause focal neurological disorders and cerebral symptoms although they are alternated or surrounded by cortical dysplasia in some cases. These tumors are usually diagnosed in children, usually affect temporal lobe and cause only stereotypical intractable epileptic seizures. Encephalopathy and developmental delay are observed in some patients, especially in children up to 2—3 years old. Therefore, surgery is reasonably used for epilepsy in patients with GNT. However, surgical effectiveness and postoperative freedom from seizures vary from 60% to 100% [3, 6—15].

This research is based on a retrospective analysis of our own experience in surgical treatment of epilepsy in children with gangliogliomas and DNTs. The purpose was to analyze their morphology, typical MR signs and predictors of favorable outcome and persistent remission.

Material and methods

Patients were selected from the epilepsy surgery database. This database has been prospectively collected since 2006 in our center. Inclusion criteria: patients younger 18 years old with morphologically verified GNT who underwent surgery in 2006–2017.

There were 152 children (median age 8 years). Redo surgery for persistent or recurrent seizures was required in 7 children (two patients underwent 3 procedures in each case) for the period from 2 weeks to 6 years (mean 11 months). Thus, there were 161 cases of GNT resection in children. The median age of the first clinical manifestations and onset of epileptic seizures was 4 years 7 months. Duration of preoperative epilepsy was 23.5 months. (Fig. 1, 2).

Fig. 1. Number of sick and operated patients aged from 9 months to 17 years.

Fig. 2. Duration of preoperative epilepsy in patients aged from 9 months to 17 years.

MRI and long-term video-EEG monitoring was performed in all children in addition to standard neurological examination. Seizures were recorded in more than 1/3 of cases. All children were examined by epileptologist. Neuropsychologist surveyed 71 patients (with various tests in some cases).

Focal motor and non-motor seizures were observed in the majority of patients (n=119), bilateral tonic-clonic seizures or their combination with focal ones – in 22 cases, epileptic spasms – in 4 patients, seizures with unknown onset – in 4 patients. Ineffective AE management with numerous and unsuccessful attempts to select anticonvulsants were noted in more than ³/₄ of all cases. Moderate (n=38) and severe (n=12) delay of psycho-speech development was observed in 50 patients in addition to epileptic seizures.

Interictal EEG was usually represented by periodic and sometimes continued regional deceleration corresponding to localization of tumor. Structure of deceleration included regional epileptiform activity (spikes, peak-slow and acute-slow wave complexes) with periodic propagation to nearby parts and homologous areas of contralateral hemisphere. Bilateral-synchronous propagation of epileptiform activity with formation of diffuse discharges was noted in some cases. This usually coincided with phenomenology of seizures.

The pattern of ictal epileptiform activity also varied depending on the semiotics of seizures. Localized or lateralized epileptiform patterns were usually observed in patients with focal seizures (acute waves, acute-slow wave complexes, rhythmic activity of alpha-beta diapason with transformation of frequency and amplitude of activity). Seizure pattern rarely remained local. Propagation of epileptic activity through the hemisphere and secondary generalization were more common. Pattern was diffuse in children up to 1.5—2 years old with spasms. Seizure onset zone was unclear in some patients.

Histological type of GNTs and their localization are shown in Tables 1, 2

and Fig. 3.

Fig. 3. Localization of tumors depending on the age of patients.

Temporal tumors were observed in almost 2/3 of cases. Lesion of temporal lobe pole and medio-basal complex was common (n=86), insular cortex was involved in 4 cases. Other segments of limbic lobe (cingulate gyrus in the depth of interhemispheric fissure) were affected in another 6 patients (Table 1). Patients with temporal tumor significantly prevailed in children under 2 years of age (Fig. 3). Incidences of various localizations of tumors were similar (Table 2).

Each patient was discussed at extended conference after examination. Multidisciplinary team determined electro-clinical syndrome and topography of epileptogenic zone, indications for surgery, type of surgical procedure and perioperative risks. The findings were presented to the patient's parents and guardians to obtain informed consent for surgery.

Surgical approach was determined by localization and dimensions of tumor.

Conventional pterional approach was used in most patients with temporal tumors localized in the pole and medial complex. Supracerebellar transtentorial approach was applied in 9 children with small tumors of posterior parts of parahippocampus and fusiform gyrus at the junction with lingual gyrus of occipital lobe [16, 18]. In recent years, we preferred subtemporal minimally invasive approach for small temporal tumors. This access implies vertical skin incision and circular craniotomy up to 3—3.5 cm (Fig. 4).

Fig. 4. Microsurgical resection of temporal lobe glioneuronal tumors (approaches). a — supracerebellar transtentorial approach for a small tumor in posterior parahippocampus on the right (arrowhead on preoperative T2-weighted MR-image on the left); surgical field (center) – corticotomy in the projection of tumor. Cortex of the right cerebellar hemisphere covered with a cotton wool is under the aspirator tip (below); above — dissected cerebellar tentorium sutured with a ligature. MRI scan in 6 months after surgery (on the right). Total resection of tumor is performed. b — anterior temporal lobectomy via pterional approach in a child with a left medial TL-ganglioglioma. Preoperative T2-weighted image (left); in the center — transcortical approach into the temporal horn through the middle temporal gyrus. Сystic part of the tumor prolapsing through the thinned alveus of the hippocampus (arrowhead); right – subpial dissection of the hippocampus (neocortex is already excised). b — subtemporal minimally invasive approach for corticoamygdalohippocampectomy with resection of left medial temporal lobe tumor. Left — surgical wound before its closure; in the center — relaxed brain and wound towards temporal horn of the lateral ventricle and the bed of excised tumor are visible through a cruciform incision of dura mater; in the center — intradermal suture of the wound, right — postoperative T2-weighted image.

Preoperative MRI and frameless stereotactic navigation were used for craniotomy in patients with frontal and parietal tumors especially if neoplasms were close to functionally important cerebral areas. Cortical mapping with SSEP and electrical stimulation preceded resection of tumor. Deep spread of tumors up to corticospinal tract was observed in several rare cases. In these cases, intraoperative monitoring was carried out using electrical stimulation by the tip of aspirator.

Intraoperative monitoring of EEG and ECoG was used in 81 patients. Subgaleal scalp needle electrodes were arranged in accordance with truncated scheme 10—20. EEG was registered immediately after induction of anesthesia, throughout surgery and was completed before dura mater suturing. ECoG was recorded using multi-contact plate electrodes (with 4, 6, 8, sometimes 20 contacts). The electrodes were placed within the areas of interest (above the tumor and over adjacent brain regions). Enlargement of registration zone was required in some cases considering BEA changes and functional significance of certain brain zones (Figs. 5, 6).

Fig. 5. ECoG before and after resection of superficial parietal lobe tumor (simple DNT). a — contrast-enhanced preoperative T1-weighted image showing a well demarcated left-sided cortical tumor of inferior parietal gyrus with scalloping of inner bone plate above the tumor. b — placement of an 8-contact electrode strip above the tumor and adjacent cortex. Tumor with clear boundaries, characteristic whitish color and impaired vascular pattern rises above the adjacent cortex. c and d — placement of the electrodes after resection of tumor for control ECoG (c — single-contact plate in the bed of excised tumor; d — 8-contact cortical plate in posterior parts of the middle temporal gyrus).

Fig. 6. ECoG before resection of the left medial temporal lobe tumor (pterional approach). a and b — surgical wound (a) and electrode arrangement (b) on the lateral neocortex (8-contact strip) and basal-medial cortex of the left temporal lobe (6-contact strip). c — anterior dissection of the temporal horn; a 4-contact strip is laid on the hippocampus. d — ECoG of the hippocampus. Continued acute-wave activity is visible under the contacts 1 and 2 (bipolar — blue, reference — red, ECG — green).

Resection of tumor within visual boundaries was performed in 78 patients. Extended resection involving adjacent brain regions was applied in other patients.

Specimens of resected tumors were stained with hematoxylin and eosin. We applied immunohistochemical analysis with antibodies to glial fibrillary acidic protein (GFAP), synaptophysin (Syn), neurofilaments (Nf and NeuN), Ki-67 proliferative activity marker and M27 neuronal protein. Immunohistochemical data were evaluated considering presence or absence of cytoplasmic and nuclear expression in tumor cells, as well as percentage of nuclear expression of Ki-67 marker.

CT on the first postoperative day was performed in all patients to exclude complications. Discharge was followed by outpatient examination in face-to-face fashion with repeated video EEG and MRI, as well as telephone interviewing. These data are available in 132 cases. Median of follow-up period was 2 years (range 1 month — 3.5 years). Data of comprehensive X-ray examination were available in 110 cases. X-ray survey was carried out within 2—3 months after surgery (MRI in 101 cases).

Various predictors of treatment outcomes including histological type and localization of tumor, age of patients and duration of disease, type of seizures and type of resection were analyzed using parametric and non-parametric statistic methods (Tables 3 and 4).

Results

Morphology. Gangliogliomas were characterized by heterogeneous structure due to combination of neuronal and glial cells. Dysplastic neurons often lost their usual cytoarchitectonics, were distinguished by significant pleomorphism and often represented by giant dual-core or multi-core forms (Fig. 7a).

Fig. 7. Morphology of glioneuronal tumors. a — ganglioglioma. Left – large dysmorphic neurons, multinuclear ganglion cells and eosinophilic granular body are in fibrillary glial tumor matrix with perivascular lymphoid infiltration. H&E x200. In the center – dysplastic neurons with MAP 2 expression, x400. Right — Syn expression (synaptophysin) with characteristic annular marker accumulation, x 200 (top) and GFAP expression in glial tumor matrix (bottom), x 200. b — simple form of DNT. Left — isomorphic glioneuronal cell component of the tumor with mucinous contents in single cysts. H&E x 200; In the center — diffuse expression of Syn, x200. Right — neurofilament protein expression with single floating neurons, x200. c — complex form of DNT. Left — subcortical nodules in DNTs consist of oligodendroglial-like cells, H&E, x200. In the center — astrocytic differentiation of tumor, H&E, x 400. Right — GFAP expression in single multipolar astrocytes; x 200. d — non-specific diffuse form of DNT. Left — numerous neuronal cells in glial matrix; H&E x200. In the center — MAP2 expression in neuronal component of tumor, x400. Right — focal expression of CD34 x200 (top); irregular diffuse expression of neurofilament protein with few dysplastic neurons, x200 (bottom).

Glial component was represented by fibrillary or piloid astrocytoma. Tumor tissue was characterized by significant dystrophic glial changes with multiple eosinophilic bodies and Rosenthal fibers, as well as various degree of perivascular lymphoid infiltration. Microcalcifications were detected in some cases. Immunohistochemical study of gangliogliomas showed clear expression of Syn as characteristic annular membrane accumulation in ganglion cells (Fig. 7a on the right). GFAP expression was found in glial component. CD34 marker expression was observed in about 50% of cases (both in tumor cells and adjacent cortex). Proliferative activity did not exceed 5%.

In accordance with the current WHO classification of CNS tumors (2016), 3 forms of dysembryoplastic neuroepithelial tumors were distinguished [19].

Histological structure of simple DNT was represented by intracortical nodules consisting of small, round, oligodendrocyte-like cells (the so-called specific glioneuronal component). Sometimes, these cells were oriented into microcolumns immersed in fibrillary microcystic or myxoid matrix with separate floating neurons (Fig. 7b).

In case of complex DNT, morphological picture was determined by combination of specific glioneuronal and glial components. Therefore, tumor had heterogeneous and often multinodular structure. Glial component was represented by astrocytic differentiation and looked like piloid astrocytoma in most cases with various cellular and nuclear polymorphism and vascular endothelial proliferation in some cases (Fig. 7c). We found no signs of atypia and necrotic changes in any case.

Non-specific and diffuse DNTs included neoplasms with diffuse growth, morphological and immunohistochemical signs of glioneuronal tumors. However, clear foci of specific glioneuronal component and other obvious diagnostic criteria were absent (Fig. 7d). These tumors looked like combined neoplasms with areas similar to either ganglioglioma or other gliomas.

We used the term “NOS glioneuronal tumor” in small number of patients (n=15) with insufficient amount of specimen or its deformation after coagulation.

MRI. In most cases, gangliogliomas were characterized by focal zones of enhanced MR signal in T2 and T2-FLAIR images. Propagation into the white matter was observed in some cases. These tumors were isointensive in T1-weighted images and heterogeneous contrast enhancement was noted in 1/4 of cases. Stromal petrificates were visualized in some cases. These foci were usually placed closer to pia-arachnoid membrane involved into pathological process. Solid part of the tumor was combined with one or more cysts of different dimensions in about 50% of cases (Fig. 8a).

Fig. 8. MR features of glioneuronal tumors. (a—c) a — left medial temporal lobe ganglioglioma. Heterogeneous hyperintense signal in T2WI (left) and homogeneous hyperintense signal in FLAIR (in the center) images without clear demarcation from lateral neocortex. Hypointense tumor with a small nodule of contrast enhancement in anterior uncus (right T1WI). b — simple DNT. A small cyst-like lesion with clear boundaries affecting posterior part of the right fusiform gyrus. No any mass-effect upon surrounding brain. Tumor tissue is hyperintense in T2 (left) and FLAIR (in the center) images, but homogeneous hypointense in T1WI (right). c — complex DNT with lesion of the right temporal lobe neocortex, infiltration of entorhinal area and subcortical white matter. Tumor tissue is mostly hyperintense on T2 and FLAIR sequences (left, center), but intermingled with iso- (T2WI) and hypointense (FLAIR) areas giving an impression of polycystic mass with hypointense signal in T1WI (right). Cystic walls rarely accumulate contrast agent (right).

Fig. 8. MR features of glioneuronal tumors. (d—e) d — nonspecific diffuse DNT followed by cortical lesion of the left anterior and medial temporal lobe structures. Tumor tissue with unclear boundaries is mostly hyperintense on T2WI (left) and FLAIR images (in the center). Signal is predominantly isointense on T1WI with focal hypointense zones. Contrast-enhanced examination revealed subpial focal contrast accumulation (on the right). e — NOS glioneuronal tumor. A tiny tumor in the right cingulate gyrus with DNT signaling characteristics, first of all there is complex form (bright signal on T2WI (left) and irregularly enhanced signal — on FLAIR image (center)). The tumor actually “disappeared” in sucker during dissection. Therefore, it was not possible to clarify the morphological diagnosis.

Dysembryoplastic neuroepithelial tumors were characterized by hypointensive foci in T1 and T2-FLAIR images and hyperintensive ones in T2-weighted images. These zones were commonly located in the cortex and rarely in subcortical spaces. Multiple cystic chambers with enhanced signal along their contour in T2 and FLAIR images reflected the structure of these tumors (cysts and pseudocysts) and usually corresponded to simple and complex DNT (Fig. 8b, c). Diffuse nonspecific DNTs were characterized by alternation of cystic structures with solid parts or dysplastic cortex (Fig. 8d). Only peripheral contrast enhancement of pseudocyst was observed (ring-shaped accumulation). Thinning and prolapse of internal bone plate over the affected cortex was diagnosed in some cases. These were patients with simple and clearly focal tumors as a rule (Fig. 5a).

Intraoperative ECoG. ECoG pattern was usually characterized by reduced amplitude and impaired cortical activity over the tumor zone. Separate and grouped acute epileptiform potentials, spike-wave complexes and rarely acute-wave bursts were recorded towards periphery from the central zone. In rare cases, persistent epileptiform activity was observed beyond the resection zone.

Quality of resection. Control scanning (MRI in 101 cases) revealed missed small superficial tumor in 1 case. Subtotal or partial resection with clear residual neoplasms was confirmed in 39 patients, total resection – in 82 cases. Only early postoperative CT data are available for other 39 patients.

Complications. Surgical, neurological and somatic complications were considered (Table 3).

Bone flap osteomyelitis was diagnosed in 4 and 9 months after resection of temporal lobe tumor (after the third surgery through the same pterional approach in one case). Hematomas (epidural in one case and within the tumor bed in another one) were recognized and removed in the evening on the day of surgery without complications.

Anterior temporal lobectomy for advanced temporal lobe tumors was performed in all cases complicated by complete hemianopsia. All cases of cranial nerve neuropathy were represented by transient paresis of the third nerve after resection of temporal lobe tumor and coincided with injury of pia-arachnoid membrane of the crural cistern. Transient motor disorders in extremities were associated with resection of tumors of the central gyrus or parietal region. Motor disorders were associated with PHoA spasm and regional ischemia in 3 patients with temporal tumors. Somatic complications (water-electrolyte disturbances in 1 child, pneumonia in another one) in early postoperative period delayed discharge for 2 and 3 weeks, respectively.

Epileptic seizures. Freedom from seizures was achieved in 107 out of 132 patients (81%) with available follow-up data (Engel class IA). No need for anticonvulsant therapy was confirmed in 60 children (45%) at the last examination.

Favorable outcome (Engel class IA) was observed in 86 (84%) out of 102 patients with available follow-up data for 1 year and over (median – 2 years). AE treatment was not necessary in 55 of them (54%) at the last examination. Moreover, improved cognitive status was noted in 18 (39%) out of 46 children with preoperative developmental delay. It was confirmed by examination and parents’ opinion.

Postoperative outcomes in these 102 patients and predictors of favorable treatment outcome are shown in Tables 4 and 5.

It is important to emphasize that complete cessation of seizures was considered a favorable outcome (Engel class IA). Class IB with auras and especially IC and D (auras with possible rare disabling seizures), as well as all other outcomes (class II—IV) were recognized as failures. Epileptic spasms were classified as generalized and 4 children with seizures without clear onset were excluded from the analysis.

We considered exclusively postoperative MRI to determine quality of resection (83 out of 102 patients). There were several repeated examinations over 2—3 years in most cases. Ratio of radical (gross-total) and partial resections was 59/43.

There were no recurrent tumors. Partial resection of temporal lobe ganglioglioma and nonspecific diffuse DNT of the occipital lobe was followed by progression of residual neoplasms in 2 cases, respectively. Redo surgery was indicated despite the absence of seizures.

Discussion

Epileptogenicity of glioneuronal tumors has been confirmed by various authors [5, 11, 20, 21]. The same data were obtained in those trials devoted to registration of epileptiform activity by deep immersion electrodes [10]. The authors consider attractive but not yet confirmed hypothesis about the role of neuronal component of tumor [21] and its alternation with foci of cortical dysplasia [5, 22, 23]. However, the key role in epileptogenesis is usually assigned to metabolic features of GNT and increased expression of glutamine receptors [20, 24, 25]. Apparently, these statements explain more significant damage to abnormal epileptogenic neuronal network (zone) after advanced resection of tumor. Our data confirm this assumption (Table 4).

Another predictor of treatment outcome is age of patient by the moment of intervention. Younger age is associated with better results. Multiple attempts to select AE therapy for a long time are accompanied by worse outcomes. The last is true even if these attempts were partially successful (Table 4).

Finally, patients with focal seizures have a slightly better prognosis than children with generalized seizures ceteris paribus (Table 4). In general, our results are similar to those reported in various trials [3, 4, 13, 15, 26—32].

There is a question regarding the role of preoperative epileptological examination. It should be remembered that epileptogenic zone topography is not always identical to the boundaries and volume of tumor. In some cases, this zone is formed outside the tumor, for example, in medial structures of the temporal lobe in a patient with ipsilateral neocortical process (the so-called “double pathology”) [33, 34]. In other cases, contralateral temporal lobe, unilateral insula or orbitofrontal cortex and sometimes posterior temporal-occipital neocortex may be involved in epileptogenesis despite the MR signs of unilateral damage to only medial-temporal structures (the so-called “temporal plus” forms) [35]. Parietal-occipital tumors affecting precuneus can simulate electro-clinical picture of frontal epilepsy, etc. [36]. Therefore, epileptologist is required in the management of these patients even in case of well-experienced surgeon. A thorough professional analysis of electro-clinical patterns of stereotypic seizures and their focal onset is the main objective of preoperative epileptological examination.

Intraoperative monitoring of ECoG and scalp EEG is a logical intraoperative extension of interaction between surgeon and epileptologist/physiologist. Intraoperative monitoring is a fairly common technique. However, appropriateness of this approach has been discussed for the last years. Some researchers almost abandoned these methods [3], while others consider these techniques very useful [37—39], especially if resection beyond the tumor borders is required.

In our sample, extended resection was often discussed in children with medial-temporal epilepsy (61 out of 83 patients). This was mainly due to above-mentioned patterns of progressive course and the tendency to involve distant intact structures into the epileptogenic neural network (hippocampus first of all) [40, 41]. Abnormal hippocampal changes (from deformation and disorganization of the layers to typical sclerosis) were histologically confirmed in 26 of our 36 cases despite any MR and macroscopic disorders were not observed in these specimens. It is worth to note that postoperative freedom from seizures was achieved in all 26 patients.

Is invasive EEG necessary and appropriate in all patients with GNT? Obvious evidence of its usefulness is not reported in our research (Table 4). Moreover, certain drawbacks including general anesthesia and short survey time are clear. Nevertheless, we believe that ECoG is still necessary and valuable in some patients. To date, our opinion largely coincides with the recommendations of Rosenow et al. [42]. Intraoperative ECoG is advisable in patients with extratemporal tumor if visual boundaries of tumor are unclear or adjacent to functionally important areas. In such cases, monitoring may be valuable to enlarge resection or vice versa restrict surgery for prevention of neurological deficit. Advanced resection (anterior temporal lobectomy) is always desirable in children with temporal tumors within the pole and medial complex. Therefore, simultaneous resection of hippocampus is advisable in children with frequent and severe seizures and tumor in non-dominant hemisphere regardless MRI and macroscopic intraoperative data. ECoG is not required in this situation. On the contrary, hippocampectomy is usually unjustifiable in patients with tumor in dominant hemisphere and intraoperative ECoG may be extremely valuable in these cases.

Invasive stereo-EEG with multiple submersible multi-contact electrodes outside of anesthesia and surgery seems more reliable. Chassoux et al. [10, 43] used this method in patients with various histological types of temporal DNT. The authors revealed different epileptogenic activity of tumor tissue and proved probability of this activity in adjacent cortex. Comparison of electrographic, histological and MRI data confirmed that epileptogenic zone is usually limited by tumor boundaries in patients with complex and simple forms of DNT. Thus, intraoperative ECoG is not required in this situation. However, this survey is desirable in patients with other types of tumors because extended resection is usually required.

Our findings partially confirm these data (Table 5). The trend towards more favorable prognosis in children with simple forms of DNT is obvious. Can we recognize these forms considering only MRI data? Apparently, yes, we can because simple forms of DNT looked like hyperintensive cystic lesions in T2 and FLAIR images with rare exceptions. Borders of these foci were clearly recognized in MRI scans and during surgery (Fig. 8).

Comparison of MR-features of GNT and morphological characteristics of these tumors is one of the goals of the study. We tried to reflect this information in the figures (Fig. 7, 8). It is important to note that GNT nomenclature is constantly being updated. The last edition of WHO classification includes diffuse leptomeningial glioneuronal tumor [19] in addition to gangliogliomas described by Perkins in 1926 [6], dysembryoplastic neuroepithelial tumors (DNT) described by Daumas-Duport in 1988 [44] and other rarer forms (papillary glioneuronal tumors, rosette-forming glioneuronal tumors, desmoplastic infantile gangliogliomas). Certain hybrid forms of tumors with symptoms of gangliogliomas and DNT are still discussed [5, 45—47]. Moreover, generally accepted histological and immunohistochemical criteria were developed for simple and complex forms of DNT, as well as gangliogliomas. At the same time, nonspecific diffuse DNTs have turned out to be the most difficult for diagnosis. Additional molecular genetic examination is required for these tumors (BRAF mutations, IDH 1 (R132H) and IDH2 (R172K) mutations, deletion 1 p/19q). The term "NOS glioneuronal tumors" was introduced to define tumors which do not fit into existing nosological forms of GNTs.

Preoperative mental-speech developmental delay was observed in about a third of children (50 out of 161 cases). Temporal tumors prevailed in these children (39 out of 50 patients). Mild and severe mental retardation was diagnosed in 38 and 12 patients, respectively. Severe disorders prevailed in children with temporal tumors (9 out of 12 cases). Median age of disease onset was 1 year. Annual postoperative follow-up data on mental development are available in 32 children. Improved cognitive status was observed in 17 patients. Postoperative freedom from seizures was 100%. Anticonvulsant therapy was unnecessary in 11 children.

Only 7 out of 12 children with severe preoperative mental retardation were followed-up. Two patients had the same mental status and need for further AE therapy for more than 1 year after surgery (including 1 child with persistent seizures). Mild progress in skill acquisition and formation of speech was observed in 5 cases despite the absence of seizures.

Developmental delay in children with epilepsy is traditionally and reasonably associated with epileptic seizures and progressive epileptic encephalopathy. Predominant early manifestation of epilepsy and developmental delay may be also associated with etiopathogenesis of this disease. So, glioneuronal tumors are often determined as developmental tumors in the literature considering their “favorite” localization (medial temporal complex with the presence of a pool of pluripotent neuroblasts in hippocampus and dentate gyrus). In addition, their frequent combination with cortical dysplasia is also considered. Another argument in favor of the fact that GNT is viciously formed brain area (hamartoma) rather a tumor is positive tissue reaction with CD34 protein (up to 80% of gangliogliomas and majority of diffuse non-specific DNTs) [5]. Above-mentioned statements seem logical, but do not explain all the facts. Firstly, development of these tumors in dorsal cortex of the frontal and parietal lobes is unclear. We did not find any obvious differences between children with improved mental development and no seizures and those with persistent mental retardation despite total resection of tumor followed by complete postoperative freedom from epilepsy. Perhaps, cognitive deficit is caused not only by neoplasm and adverse effect of epileptiform activity on the brain. There is more common anatomical and physiological defect in some children with GNT despite the same morphological diagnosis and similar topography of lesion. Apparently, we cannot recognize these abnormalities. These issues require further study.

Conclusion

Morphological diagnosis of GNT requires detailed immunohistochemical verification and molecular genetic examination if it is necessary. These measures are required for differential diagnosis with diffuse glial tumors that is extremely important for further adequate treatment.

MRI is preferred to diagnose glioneuronal tumors. This survey is usually valuable to localize tumor and clearly determine its histological type.

Total and early resection of tumor is desirable in children with glioneuronal tumors complicated by epileptic seizures. The risk of complications is small in these cases while possible postoperative neurological disorders are usually transient.

Close cooperation with an epileptologist at all stages of treatment is essential to achieve favorable result, since epileptogenic zone topography and necessary resection can go beyond the boundaries of tumor.

Authors’ participation

Concept and design of the study — A.M., L.SH.

Collection and analysis of data — A.M., L.SH.,

E.SH., A.K., P.V., YU.K., M.K., S.B., M.V.

Statistical analysis — A.M.

Writing the text — A.M., L.SH.,

Editing — A.M., L.SH. E.SH., A.K., P.V.

The authors declare no conflicts of interest.

Commentary

The article is devoted to the problem of surgical treatment of glioneuronal tumors (GNT) in pediatric patients. This issue is extremely important considering high epileptogenicity of these tumors and their resistance to antiepileptic therapy. The outcomes of surgical treatment are discussed taking into account morphological features of tumors, their localization and MR characteristics in 152 patients. The authors emphasize high sensitivity of MRI. This survey is usually valuable to differentiate histological type of the tumor and suggest necessary boundaries of resection.

Intraoperative ECoG monitoring is applied for accurate verification of epileptogenic zone. However, it is unclear how this method affected type of resection and what criteria were used for enlargement of resection. The conclusion on the role of ECoG in decision making on hippocampectomy in patients with tumor of dominant hemisphere is also controversial. In our opinion, preoperative neuropsychological testing may be usually sufficient to predict mnestic disorders after resection of dominant hippocampus. ECoG is justified in case of epileptic patterns under the grid. Can we clearly say that hippocampus is not epileptogenic if these patterns are absent?

Freedom from epileptic seizures is the main criterion for evaluation of effectiveness of surgical treatment of glioneuronal tumors. The authors reported high efficiency of surgical treatment (Engel class IA in 84% of cases) in 102 patients who were followed-up for 1 year and over. The applied criteria for abolition of antiepitheptic therapy would be interesting for the reader. A fairly low postoperative morbidity and comprehensive analysis of complications should be emphasized.

The authors discuss the importance of preoperative examination and teamwork of neurosurgeon and epileptologist. An importance of pre- and intraoperative neuromonitoring for determining the exact boundaries of resection is reported considering original own and foreign data.

In our opinion, the article is of undoubted scientific and practical interest and will be interesting for various specialists involved in the treatment of epilepsy

A.A. Zuev, N.P. Utyashev (Moscow, Russia)

References:

Luyken C, Blumcke I, Fimmers R, Urbach H, Elger CE, Wiestler OD, Schramm J. The spectrum of long-term epilepsy-associated tumors: long-term seizure and tumor outcome and neurosurgical aspects. Epilepsia. 2003;44(6):822-830.
Holthausen H. Benign tumors. In: Arzimanoglou A.Ch, Gaillard W., Holthausen H., Jayakar P., Kahan P., Mathern G., ed. Pediatric Epilepsy Surgery. Montrouge, France: John Libbey Eurotext; 2016.
Giulioni M, Marucci G, Pelliccia V, Gozzo F, Barba C, Didato G, Villani F, Di Gennaro G, Quarato PP, Esposito V, Consales A, Martinoni M, Vornetti G, Zenesini C, Efisio Marras C, Specchio N, De Palma L, Rocchi R, Giordano F, Tringali G, Nozza P, Colicchio G, Rubboli G, Lo Russo G, Guerrini R, Tinuper P, Cardinale F, Cossu M; Commission for Epilepsy Surgery of the Italian League Against Epilepsy. Epilepsy surgery of «low grade epilepsy associated neuroepithelial tumors»: A retrospective nationwide Italian study. Epilepsia. 2017;58(11):1832-1841. https://doi.org/10.1111/epi.13866
Baud MO, Perneger T, Racz A, Pensel MC, Elger C, Rydenhag B, Malmgren K, Cross JH, McKenna G, Tisdall M, Lamberink HJ, Rheims S, Ryvlin P, Isnard J, Mauguière F, Arzimanoglou A, Akkol S, Deniz K, Ozkara C, Lossius M, Rektor I, Kälviäinen R, Vanhatalo LM, Dimova P, Minkin K, Staack AM, Steinhoff BJ, Kalina A, Krsek P, Marusic P, Jordan Z, Fabo D, Carrette E, Boon P, Rocka S, Mameniškienė R, Vulliemoz S, Pittau F, Braun KPJ, Seeck M. European trends in epilepsy surgery. Neurology. 2018;91(2):96-106. https://doi.org/10.1212/WNL.0000000000005776
Thom M, Blumcke I, Aronica E. Long-term epilepsy-associated tumors. Brain Pathology. 2012;22(3):350-379.
Zentner J, Wolf HK, Ostertun B, Hufnagel A, Campos MG, Solymosi L, Schramm J. Gangliogliomas: clinical, radiological, and histopathological findings in 51 patients. Journal of Neurology, Neurosurgery, and Psychiatry. 1994;57(12):1497-1502.
Nolan MA, Sakuta R, Chuang N, Otsubo H, Rutka JT, Snead OC, 3rd, Hawkins CE, Weiss SK. Dysembryoplastic neuroepithelial tumors in childhood: long-term outcome and prognostic features. Neurology. 2004;62(12):2270-2276.
Sandberg DI, Ragheb J, Dunoyer C, Bhatia S, Olavarria G, Morrison G. Surgical outcomes and seizure control rates after resection of dysembryoplastic neuroepithelial tumors. Neurosurgical Focus. 2005;18(6a):E5.
Adachi Y, Yagishita A. Gangliogliomas: Characteristic imaging findings and role in the temporal lobe epilepsy. Neuroradiology. 2008;50(10):829-834.
Chassoux F, Daumas-Duport C. Dysembryoplastic neuroepithelial tumors: where are we now? Epilepsia. 2013;54(suppl 9):129-134. https://doi.org/10.1111/epi.12457
Giulioni M, Marucci G, Martinoni M, Marliani AF, Toni F, Bartiromo F, Volpi L, Riguzzi P, Bisulli F, Naldi I, Michelucci R, Baruzzi A, Tinuper P, Rubboli G. Epilepsy associated tumors: Review article. World Journal of Clinical Cases. 2014;2(11):623-641. https://doi.org/10.12998/wjcc.v2.i11.623
Ranger A, Diosy D. Seizures in children with dysembryoplastic neuroepithelial tumors of the brain — A review of surgical outcomes across several studies. Child’s Nervous System: ChNS: Official Journal of the International Society for Pediatric Neurosurgery. 2015;31(6):847-855. https://doi.org/10.1007/s00381-015-2675-9
Bonney PA, Boettcher LB, Burks JD, Baker C, Conner AK, Fujii T, Mehta VA, Briggs RG, Sughrue ME. Rates of Seizure Freedom After Surgical Resection of Diffuse Low-Grade Gliomas. World Neurosurgery. 2017;106:750-756. https://doi.org/10.1016/j.wneu.2017.06.144
Nguyen HS, Doan N, Gelsomino M, Shabani S. Dysembryoplastic Neuroectodermal Tumor: An Analysis from the Surveillance, Epidemiology, and End Results Program, 2004-2013. World Neurosurgery. 2017;103:380-385. https://doi.org/10.1016/j.wneu.2017.04.093
Faramand AM, Barnes N, Harrison S, Gunny R, Jacques T, Tahir MZ, Varadkar SM, Cross HJ, Harkness W, Tisdall MM. Seizure and cognitive outcomes after resection of glioneuronal tumors in children. Epilepsia. 2018;59(1):170-178. https://doi.org/10.1111/epi.13961
Ture U, Harput MV, Kaya AH, Baimedi P, Firat Z, Ture H, Bingöl CA. The paramedian supracerebellar-transtentorial approach to the entire length of the mediobasal temporal region: an anatomical and clinical study. Laboratory investigation. Journal of Neurosurgery. 2012;116(4):773-791. https://doi.org/10.3171/2011.12.JNS11791
Weil AG, Middleton AL, Niazi TN, Ragheb J, Bhatia S. The supracerebellar-transtentorial approach to posteromedial temporal lesions in children with refractory epilepsy. Journal of Neurosurgery Pediatrics. 2015;15(1):45-54. https://doi.org/10.3171/2014.10.PEDS14162
Konovalov AN, Pitskhelauri DI, Melikyan AG, Shishkina LV, Serova NK, Pronin IN, Eliseeva NM, Shkatova AM, Samborskiy DY, Bykanov AE, Golovteev AL, Grinenko OA, Kopachev DN. Supracerebellar transtentorial approach to tumors of the posterior portions of the medial temporal region. Voprosy neirokhirurgii im. N.N. Burdenko. 2015;79(4):38-47. (In Russ.) https://doi.org/10.17116/neiro201579438-47
Louis DN, Ohgaki H, Wiestler OD, Cavernee WK, eds. WHO Classification of Tumours of the Central Nervous System. Lyon: LARC; 2016.
Aronica E, Yankaya B, Jansen GH, Leenstra S, van Veelen CW, Gorter JA, Troost D. Ionotropic and metabotropic glutamate receptor protein expression in glioneuronal tumours from patients with intractable epilepsy. Neuropathology and Applied Neurobiology. 2001;27(3):223-237.
Blumcke I, Wiestler OD. Gangliogliomas: an intriguing tumor entity associated with focal epilepsies. Journal of Neuropathology and Experimental Neurology. 2002;61(7):575-584.
Shamji MF, Fric-Shamji EC, Benoit BG. Brain tumors and epilepsy: pathophysiology of peritumoral changes. Neurosurgical Review. 2009;32(3):275-284; discussion 84-86.
Pallud J, Capelle L, Huberfeld G. Tumoral epileptogenicity: how does it happen? Epilepsia. 2013;54(suppl 9):30-34. https://doi.org/10.1111/epi.12440
Aronica E, Crino PB. Epilepsy related to developmental tumors and malformations of cortical development. Neurotherapeutics : the Journal of the American Society for Experimental NeuroTherapeutics. 2014;11(2):251-268. https://doi.org/10.1007/s13311-013-0251-0
de Groot M, Reijneveld JC, Aronica E, Heimans JJ. Epilepsy in patients with a brain tumour: focal epilepsy requires focused treatment. Brain: a Journal of Neurology. 2012;135(Pt 4):1002-1016. https://doi.org/10.1093/brain/awr310
Bonney PA, Boettcher LB, Conner AK, Glenn CA, Briggs RG, Santucci JA, Bellew MR, Battiste JD, Sughrue ME. Review of seizure outcomes after surgical resection of dysembryoplastic neuroepithelial tumors. Journal of Neuro-Oncology. 2016;126(1):1-10. https://doi.org/10.1007/s11060-015-1961-4
Englot DJ, Han SJ, Berger MS, Barbaro NM, Chang EF. Extent of surgical resection predicts seizure freedom in low-grade temporal lobe brain tumors. Neurosurgery. 2012;70(4):921-928; discussion 8.
Behrens E, Schramm J, Zentner J, Konig R. Surgical and neurological complications in a series of 708 epilepsy surgery procedures. Neurosurgery. 1997;41(1):1-9; discussion -10.
Schramm J, Aliashkevich AF. Surgery for temporal mediobasal tumors: experience based on a series of 235 patients. Neurosurgery. 2007;60(2):285-294; discussion 94-95.
Harvey AS, Cross JH, Shinnar S, Mathern GW; ILAE Pediatric Epilepsy Surgery Survey Taskforce. Defining the spectrum of international practice in pediatric epilepsy surgery patients. Epilepsia. 2008;49(1):146-155.
Bjellvi J, Flink R, Rydenhag B, Malmgren K. Complications of epilepsy surgery in Sweden 1996-2010: a prospective, population-based study. Journal of Neurosurgery. 2015;122(3):519-525. https://doi.org/10.3171/2014.9.JNS132679
O’Brien DF, Farrell M, Delanty N, Traunecker H, Perrin R, Smyth MD, Park TS; Children’s Cancer and Leukaemia Group. The Children’s Cancer and Leukaemia Group guidelines for the diagnosis and management of dysembryoplastic neuroepithelial tumours. British Journal of Neurosurgery. 2007;21(6):539-549.
Blumcke I. Neuropathology of focal epilepsies: a critical review. Epilepsy and Behavior. 2009;15(1):34-39.
Prayson RA, Fong J, Najm I. Coexistent pathology in chronic epilepsy patients with neoplasms. Modern Pathology. 2010;23(8):1097-1103. https://doi.org/10.1038/modpathol.2010.94
Kennedy J, Schuele SU. Long-term monitoring of brain tumors: when is it necessary? Epilepsia. 2013;54(suppl 9):50-55. https://doi.org/10.1111/epi.12444
Yu T, Wang Y, Zhang G, Cai L, Du W, Li Y. Posterior cortex epilepsy: diagnostic considerations and surgical outcome. Seizure. 2009;18(4):288-292.
Sugano H, Shimizu H, Sunaga S. Efficacy of intraoperative electrocorticography for assessing seizure outcomes in intractable epilepsy patients with temporal-lobe-mass lesions. Seizure. 2007;16(2):120-127.
Southwell DG, Garcia PA, Berger MS, Barbaro NM, Chang EF. Long-term seizure control outcomes after resection of gangliogliomas. Neurosurgery. 2012;70(6):1406-1413; discussion 13-14. https://doi.org/10.1227/NEU.0b013e3182500a4c
Wray CD, McDaniel SS, Saneto RP, Novotny EJ, Jr, Ojemann JG. Is postresective intraoperative electrocorticography predictive of seizure outcomes in children? Journal of Neurosurgery Pediatrics. 2012;9(5):546-551. https://doi.org/10.3171/2012.1.PEDS11441
Ghareeb F, Duffau H. Intractable epilepsy in paralimbic Word Health Organization Grade II gliomas: should the hippocampus be resected when not invaded by the tumor? Journal of Neurosurgery. 2012;116(6):1226-1234. https://doi.org/10.3171/2012.1.JNS112120
Blumcke I, Aronica E, Urbach H, Alexopoulos A, Gonzalez-Martinez JA. A neuropathology-based approach to epilepsy surgery in brain tumors and proposal for a new terminology use for long-term epilepsy-associated brain tumors. Acta Neuropathologica. 2014;128(1):39-54. https://doi.org/10.1007/s00401-014-1288-9
Rosenow F, Menzler K. Invasive EEG studies in tumor-related epilepsy: when are they indicated and with what kind of electrodes? Epilepsia. 2013;54(suppl 9):61-65. https://doi.org/10.1111/epi.12446
Chassoux F, Landre E, Mellerio C, Laschet J, Devaux B, Daumas-Duport C. Dysembryoplastic neuroepithelial tumors: epileptogenicity related to histologic subtypes. Clinical Neurophysiology. 2013;124(6):1068-1078. https://doi.org/10.1016/j.clinph.2012.11.015
Daumas-Duport C, Scheithauer BW, Chodkiewicz JP, Laws ER, Jr, Vedrenne C. Dysembryoplastic neuroepithelial tumor: a surgically curable tumor of young patients with intractable partial seizures. Report of thirty-nine cases. Neurosurgery. 1988;23(5):545-556.
Anyanwu CT, Robinson TM, Huang JH. Rosette-forming glioneuronal tumor: an update. Clinical and translational oncology: official publication of the Federation of Spanish Oncology Societies and of the National Cancer Institute of Mexico. 2019. [Epub ahead of print]. https://doi.org/10.1007/s12094-019-02179-8
Singh V, Gupta K, Salunke P, Kumar S, Radotra BD, Vasishta RK. Rosette-forming and papillary glioneuronal tumors — a clinicopathological and molecular analysis. Clinical Neuropathology. 2019;38(4):180-188. https://doi.org/10.5414/NP301164
Yang C, Fang J, Li G, Li S, Ha T, Wang J, Yang B, Yang J, Xu Y. Histopathological, molecular, clinical and radiological characterization of rosette-forming glioneuronal tumor in the central nervous system. Oncotarget. 2017;8(65):109175-109190. https://doi.org/10.18632/oncotarget.22646