Clinical manifestations

Presentation and course

	• Seizures are a common presentation for patients with brain tumors. The prevalence of seizures is intimately linked to the tumor type.
	• In general, seizures are more frequently seen in patients with low-grade tumors (ie, dysembryoplastic neuroepithelial tumors) compared to high-grade tumors (ie, glioblastoma).
	• The presence of an isocitrate dehydrogenase (IDH) mutation is associated with a higher risk of seizures in gliomas.
	• Long-term epilepsy-associated tumors (LEATs) have the highest likelihood of seizure freedom after gross total resection among all brain tumors.
	• The extent of resection is significantly related to the likelihood of seizure freedom after resection of a low-grade glioma.

The prevalence of seizures associated with brain tumors is variable and is most significantly associated with the tumor grade; however, advances in molecular genetic classifications suggest that tumor genetics play a significant role in the risk for seizures.

Of all tumor types, epilepsy is seen in 30% or more of patients with brain tumors (40). Low-grade tumors (eg, neuroepithelial tumors) are more epileptogenic, with seizures seen in up to 100% of patients with dysembryoplastic neuroepithelial tumors and in 80% to 90% of patients with gangliogliomas (40). IDH-mutated subtypes of WHO grade 2 or 3 gliomas are associated with seizures more often than IDH wild-type (wt) gliomas, with seizures seen in 50% to 80%. However, the increased frequency of seizures has not been seen in patients with IDH-mutated grade 4 astrocytomas compared to glioblastoma IDHwt (50; 12). In addition, patients with IDH-mutated low-grade tumors are more likely to develop pharmacoresistant seizures compared to those with IDHwt tumors (11). Seizures are seen in 29% to 49% of patients with grade 4 gliomas (independent of IDH mutation status) (40).

Preoperative seizures are reported in 29% to 60% of patients with meningiomas with risk factors, including male sex, absence of headache, and peritumoral edema (40; 18). After meningioma resection, approximately 70% of those with preoperative epilepsy become seizure free. New postoperative seizures have been reported in 0% to 42.9% (18; 24).

With respect to brain metastases, seizures are seen in approximately 20% to 35% of patients (40). Of all patients with systemic cancer, approximately 20% develop brain metastases, with lung, breast, and melanoma being the most common (18; 19; 33). Risk factors associated with brain metastases and seizures may include younger age, parietal location, and recurrent tumor, although these require further validation (33).

Prognosis and complications

Long-term epilepsy-associated tumors (LEATs). LEATs are associated with the best seizure outcome after surgical resection. The largest study of 1493 pathologically confirmed LEATs from the European Epilepsy Brain Bank reported that 75.9% of the patients were free from disabling seizures 5 years after surgery (25).

Low-grade gliomas. Diffuse astrocytomas and oligodendrogliomas will inevitably undergo malignant transformation to a high-grade glioma (WHO grade 3 or 4). As such, when the tumor is still within the low-grade classification (WHO grade 2), treatment goals include delaying time to malignant transformation as well as seizure control. Interestingly, patients with seizures in the context of low-grade gliomas have better prognosis compared to those without a history of seizures. The largest observational study of 1509 patients with low-grade gliomas found that approximately 90% had a history of seizures, and those with seizures at diagnosis had increased malignant, progression-free, and overall survival compared to those without a history of seizures (31). These findings may be influenced by the IDH mutational status of the tumors. IDH-mutated tumors harbor a higher risk of seizures but also a better prognosis. IDH mutational status was not routinely evaluated in all patients at the time of the aforementioned study.

In terms of the prognosis of tumor-related epilepsy, seizure control is significantly related to the extent of resection for patients with low-grade gliomas (17; 49). Approximately 70% of patients who undergo gross total resection (of FLAIR abnormality on MRI) are seizure free (Engel class I), whereas approximately 30% continue to have seizures (Engel class II to IV) (07; 17). One group has proposed that the threshold for the extent of resection is 80% as seizure freedom (for a limited timeframe) is often attained in this population when at least 80% of total tumor volume resection is achieved (49). However, it is estimated that no more than 45% of patients are eligible for gross total resection as low-grade gliomas tend to grow close to or within eloquent regions of the brain (34). Seizure control (Engel class I) has been reported in 43% to 47% of patients with subtotal resection and 8% of patients with biopsy alone (07; 17). Therefore, the goal of resection is to include the greatest volume possible for seizure outcome. Studies of seizure frequency in gliomas are limited by the utilization of patient-reported outcomes of seizure frequency and by the typical lack of lifelong follow-up, which can be of decades’ duration in patients with low-grade gliomas.

High-grade gliomas. The largest series to date examining features of tumor-related epilepsy in patients with malignant gliomas (WHO grade 3 and 4) evaluated 648 patients (505 with glioblastoma, 143 with astrocytoma WHO grade 3) (06); 24% of the patients presented with seizures, and a total of 77% were seizure-free 12 months after surgery. Most importantly, seizure recurrence in patients who had postoperative seizure control was significantly associated with tumor recurrence. This study did not incorporate the contemporary molecular markers (eg, IDH mutational status).

Meningiomas. Prior observational studies have reported a linear growth rate of 2 to 4 mm per year for asymptomatic meningiomas; however, some meningiomas have a nonlinear growth rate, and some do not grow over time (04). Surveillance imaging is key to understanding the behavior of a patient’s meningioma. The true growth rate of symptomatic meningiomas is unknown as the majority of symptomatic lesions are treated with resection. Grade 2 and 3 meningiomas have a high risk of recurrence after gross total resection: 50% for grade 2 and 90% for grade 3 tumors at 5 years (04).

With respect to meningioma-related epilepsy, the best treatment strategy is often resection as seizure freedom is estimated to occur in approximately 70% of patients with preoperative epilepsy (16). However, an estimated 12% have new-onset epilepsy postoperatively. After resection, patients with preoperative seizures, in-hospital seizures, and medical or surgical complications are at a higher risk for postoperative epilepsy (09).

Brain metastases. Risk factors for postoperative seizures after resection of metastatic lesions include younger age, recurrent tumor, supratentorial location, and incomplete resection (33; 48). Interestingly, peritumoral edema and extent of resection are not associated with postoperative seizures (33). In a study of patients who were seizure-free prior to resection with a single supratentorial metastatic lesion, 30% of the patients who had a biopsy or subtotal resection developed seizures after surgery; 5% of the patients who underwent a gross total resection of the metastatic lesion developed seizures (48).

Seizures may be secondary to causes beyond the tumor itself, including radionecrosis after radiotherapy, cavernomas secondary to whole-brain radiotherapy, opportunistic CNS infections, or drugs that lower the seizure threshold (eg, bupropion, cisplatin) (40).

Biological basis

	• Seizures do not often arise from the tumor itself. They arise from the peritumoral tissue and, rarely, from distant locations outside of the peritumoral region.
	• The epileptogenicity of the peritumoral tissue is likely due to multiple factors, including but not limited to mechanical disruption, cortical damage, or deposition of hemosiderin.
	• The peritumoral environment is hyperexcitable, and this is due, in part, to an imbalance of excessive glutamate and reduced inhibitory GABA transmission. The relationship between this imbalance and the tumor genetic mutations identified in brain tumors is still being explored.
	• Additional hypotheses for causes of tumor-related epilepsy include free radical formation, alterations in neuronal function and connectivity, and alterations in expression of specific genes and proteins.
	• Tumor-related epilepsy is often refractory to the currently available antiseizure medications. This is due, in part, to the presence of multidrug resistance proteins.

Etiology and pathogenesis

Seizures do not arise from the tumor as gliomas have no intrinsic neural component. The epileptogenic cortex often lies in the peritumoral region and, in some cases, may be located beyond the peritumoral region (29; 32). Furthermore, it is clear that the cause of tumor-related epilepsy is multifactorial as no single tumor entity or features are unequivocally associated with seizures.

There are multiple theories explaining tumor-related epilepsy, although epileptogenicity likely arises at the intersection of these theories. These hypotheses are centered around structural causes and the microcellular environment.

The mechanical disruption of cortical tissue, whether it be from the tumor itself, peritumoral edema, cerebral ischemia, intracerebral hemorrhage, or vascular compression, may promote epileptogenicity (41; 32). Rapidly growing tumors (eg, glioblastoma, brain metastases) may cause seizures by more invasive methods of cortical damage, such as necrosis, infiltrative neuronal growth, mass effect, or deposition of hemosiderin (41).

Not only are mechanical factors at play, but there are likely a combination of factors promoting epileptogenicity at the cellular level. Future studies will help create a better understanding of tumor genetics and the relationship to a hyperexcitable tumor microenvironment. It has become clear that a mutation in isocitrate dehydrogenase (IDH) correlates with tumor-related epilepsy. The product of a mutant IDH is D-2-hydroxyglutarate (D2HG), which is structurally similar to glutamate. It has been shown that D2HG acts as a mimicker to glutamate and binds to the NMDA receptor, thereby increasing neuronal activity and leading to seizures (08).

Glutamate, in its true state, also likely plays a role in the increased risk of seizures with brain tumors. Glutamate, the primary excitatory neurotransmitter in the brain, is increased due to upregulation of glutamine, glutamine synthetase protein, and glutamine synthetase enzyme activity in glioblastoma (32). Furthermore, gliomas release neurotoxic levels of glutamate, which has been associated with the high expression of system xc^-, an antiporter that is responsible for extracellular glutamate release in exchange for cysteine uptake (03; 38). Although the tumor itself is releasing glutamate, peritumoral astrocytes are less able to remove glutamate; at the same time, microglia may release glutamate in response to glioma cell signaling (03). The presence of glutamate not only causes excitability but also contributes to cell death, allowing tumor growth (38).

In addition, there is a direct relationship formed by the synaptic connections between glioma cells and the surrounding neurons, which promotes hyperexcitability and seizures. Venkatesh reported on the presence of AMPA receptor-dependent synapses between neurons and gliomas (45). Intratumoral gap junction–mediated connections also amplify the potassium currents derived from neuronal activity. Through multiple means of connection, glioma cells become integrated into the neural circuitry and likely contribute not only to seizures, but also to glioma proliferation.

Furthermore, brain tumors frequently disrupt the blood-brain barrier, and this can be linked to its epileptogenic effect. In terms of gliomas, the blood-brain barrier becomes more permeable, allowing glutamate into the peritumoral region (03).

Although glutamate plays a role in NMDA receptor activation, reduced inhibitory GABAergic transmission is also a factor (03), likely by downregulation of KCC2, a potassium chloride co-transporter. KCC2 functions to maintain a chloride gradient across neurons, thereby regulating GABA (gamma-aminobutyric-acid) function. In mouse models of gliomas, KCC2 expression is downregulated in the peritumoral tissue (as well as nontumoral epileptic cortex), which leads to increased intracellular chloride and thereby inhibits GABA activity, resulting in a lower seizure threshold (38).

The morphological changes that are present in peritumoral tissue also likely play a role in epileptogenicity. These changes can include aberrant neuronal migration, enhanced intracellular communication through gap junction channels, imbalance between local inhibitory and excitatory mechanisms (including glutamate concentrations as described), and morphological alterations of synaptic vessels (41).

Tumoral epilepsy is often refractory to the currently available antiseizure medications. Potential mechanisms include multidrug-resistance proteins and potential interactions between antiseizure medications and chemotherapeutic agents, with the latter being less common with newer-generation antiseizure medications. Multidrug resistance proteins include those that belong to the ATP-binding cassette transporter family. These proteins function to facilitate or block transport over the endothelial cell membrane and are often overexpressed with gliomas, gangliogliomas, and focal cortical dysplasia (40).

Epidemiology

	• Seizures are seen in approximately 30% to 100% of patients with brain tumors, depending on the tumor type.
	• Seizures are the presenting symptom for 20% to 40% of patients with brain tumors.
	• Low-grade tumors are more likely to be associated with seizures.

Of all tumors, brain tumors make up approximately 1% to 2% of tumors in adults (27). Among all patients with brain tumors, seizures are seen in approximately 30% to 100%, depending on the tumor type (19). Combining all tumor types, seizures are the onset symptom in 20% to 40% of patients with brain tumors. Of all patients with epilepsy, brain tumors are the cause of seizures in 6% to 10% of patients (27).

Low-grade lesions are more epileptogenic than high grade, although genetic features play a role. Patients who present with seizures are at a higher risk for continued seizures throughout their disease course, regardless of tumor type, antiseizure medication, and treatment (40). Brain tumors that are supratentorial, cortically based, and located in the frontotemporal regions are more commonly associated with seizures (40).

Long-term epilepsy-associated tumors (LEATs). Seizures are common in LEATs and occur in almost 100% of dysembryoplastic neuroepithelial tumors and in 80% to 90% of gangliogliomas. The mean age of onset for dysembryoplastic neuroepithelial tumors is 15 years of age and for gangliogliomas is 16 to 19 years of age, although they can be first identified at later ages as well. The average age for undergoing tumor resection for a WHO grade 1 tumor is 30 years of age (19).

Low-grade gliomas. Seizures are seen in 65% to 90% of patients with low-grade gliomas. A systematic review and meta-analysis identified predictors of postoperative seizure control in patients undergoing surgical resection of low-grade gliomas (35). Improved seizure outcome was seen in patients 45 years of age or older and with gross total resection compared to subtotal resection. Seizures with an onset 1 or more years prior to surgery as well as focal seizures (compared to generalized convulsions) were associated with poorer seizure control.

High-grade gliomas. Seizures are seen in 25% to 60% of patients with high-grade gliomas (15). Tumor location is a significant risk factor for seizures in patients with high-grade gliomas, specifically, superficial cortical areas or tumors in the temporal lobe, frontal lobe, or insula (13). In a voxel-based study of patients with no history of seizures in glioblastoma multiforme, tumor presence in the superior frontal, inferior occipital, and inferoposterior temporal lobes was associated with an increased risk of seizures after surgery (05). In addition, patients with multifocal disease have a higher risk of seizures compared to those with a singular tumor (15). This may merely represent a greater amount of cortex affected by the tumor.

Prevention

	• Based on the current AAN recommendations, there is no recommendation for the use of prophylaxis medications to prevent seizures in patients with brain tumors who have not already experienced seizures. However, this statement was made in 2000 and has not been re-evaluated using newer antiseizure medications.
	• Perioperative antiseizure medications are commonly used for prophylaxis with surgery for brain tumors, although the data are conflicting.

In 2000, the Quality Standards Subcommittee of the American Academy of Neurology (AAN) performed a meta-analysis of four randomized clinical trials and eight observational studies and reported that there appears to be no benefit for the prophylactic use of antiseizure medications in patients with brain tumors (22). None of the included trials showed a statistically significant difference in seizure incidence versus seizure freedom between patients with and without antiseizure medication prophylaxis. It is important to note that the meta-analysis included studies evaluating the use of phenytoin, carbamazepine, phenobarbital, and valproic acid. This meta-analysis did not review the outcome with newer antiseizure medications (ie, levetiracetam, lacosamide). The AAN guideline has not yet been updated; however, in 2021, the Society for Neuro-Oncology (SNO) and European Association of Neuro-Oncology (EANO) updated their recommended practice parameter for the use of antiseizure medications in patients with brain tumors (46). Based on an updated meta-analysis using studies examining newer, nonenzyme-inducing agents (ie, levetiracetam), the same recommendation was made. Providers should not prescribe antiseizure medication prophylaxis to seizure-naive patients with brain tumors.

The use of prophylactic antiseizure medications around the time of tumor resection is common, although the data regarding the benefit of using antiseizure medications perioperatively in patients without a history of seizures are conflicting. A meta-analysis investigating outcomes in seizure-naive patients given perioperative antiseizure medications was completed in 2019 (47). The most commonly used medications were phenytoin, phenobarbital, levetiracetam, and valproate. The authors reported no significant difference in postoperative seizures with prophylactic antiseizure medications for either early (ie, within the first postoperative week) or late (ie, after the first postoperative week) seizures. The aforementioned SNO/EANO practice parameter also indicated that based on their analysis, there is insufficient evidence (level C) to recommend that antiseizure medications be used in the perioperative period (46).

Diagnostic workup

	• The steps included in a diagnostic evaluation depend on the treatment goals.
	• The first step should always be to verify the concordance between a patient’s epilepsy and the brain tumor.
	• For low-grade lesions, diagnostic workup is often driven by the patient’s epilepsy, whereas oncologic goals drive the decision making for high-grade lesions.

The diagnostic evaluation for patients with tumor-related epilepsy significantly depends on the tumor type and treatment goals. In all cases with lesional epilepsy, the first question that should be addressed is whether a histological diagnosis is necessary. If a tissue diagnosis is needed, the surgery is often optimized with a surgical approach that will render the patient seizure free.

Further neurophysiologic studies should be considered if there are elements that are discordant when obtaining seizure history, including an EEG, ambulatory EEG, or possibly an epilepsy monitoring unit (EMU) admission. Similar neurophysiologic studies should be considered in pharmacoresistant epilepsy in the context of a lesion as epilepsy surgery may be warranted.

Meningiomas. In the majority of cases, patients will undergo meningioma resection for reasons other than epilepsy (eg, functional deficits, headaches). No prospective studies have looked at the value of EEG in patients with meningiomas undergoing evaluation or surgery. Although the sample size was small, one study evaluated the role of ECoG in patients with preoperative seizures undergoing meningioma resection and found that epileptiform discharge frequency pre-resection and post-resection did not correlate with postoperative epilepsy (20).

Low-grade lesions. Clinical decision making in cases of low-grade lesions is often guided by the patient’s epilepsy. In cases of pharmacoresistant epilepsy, additional neurophysiologic studies as well as imaging studies can help guide resection and increase the likelihood of seizure freedom. More specifically, patients may undergo an EEG and an EMU admission with continuous scalp video-EEG recording to confirm that seizure semiology, neurophysiologic data, and imaging are concordant. In guiding epilepsy surgery, a PET, SPECT, and MEG may be considered. Furthermore, a comprehensive neuropsychological evaluation may have value. Invasive EEG may be warranted in patients whose data are discordant or for whom suspected focal cortical dysplasia is associated with the lesion. Importantly, this testing should not significantly delay the timing of surgery as prolonged duration of seizures has been associated with poorer seizure outcome after surgery (26; 35).

High-grade lesions. These patients will often immediately go to surgery on discovery of the lesion. Immediate diagnostic work-up for epilepsy purposes is often unwarranted, although intraoperative ECoG could be considered. There are conflicting data regarding the value of ECoG in high-grade gliomas, although ECoG has not been prospectively studied in this patient population.

Management

	• No standard guidelines for the medical management of tumor-related epilepsy exist.
	• The most commonly used antiseizure medication is levetiracetam due to its tolerability, lack of interactions with concurrent medications, and dosing regimen. However, mood side effects are common, and patients should be screened.
	• The role of surgery depends on the tumor type. In general, seizure frequency often improves with resection; however, ECoG may be considered.
	• When appropriate, chemotherapy and radiation may improve seizure frequency, particularly in infiltrating gliomas.

Medical management

There are no standard guidelines for the medical management of tumor-related epilepsy. Several goals drive treatment decisions in patients with tumor-related epilepsy, regardless of the type of tumor, including improvement of seizure control and achieving seizure freedom; improvement of quality of life; and improvement of survival (30).

Randomized controlled trials comparing antiseizure medications in this population are rare. The choice of antiseizure medication is often guided by concurrent medications, potential side effects, and efficacy. The most common side effects with antiseizure medications in this population include cognitive side effects, skin hypersensitivity, and bone marrow toxicity (43).

The most commonly used medication in patients with tumor-related epilepsy is levetiracetam as it does not have significant drug interactions, can be given as a loading dose, is safe in pregnancy overall, and is generally well-tolerated. However, mood side effects can be common and are more commonly seen in patients with frontal lobe tumors (51).

Lamotrigine is similarly well-tolerated, but it takes time to increase the dose to a therapeutic dose. Therefore, it is not appropriate as a starting agent (without a concurrent medication). As the brain tumor patient population is at risk for frequent hospitalizations, this medication may not be an optimal choice.

When levetiracetam is not tolerated or effective, lacosamide or valproic acid are often the next agents of choice. One prospective study comparing lacosamide to levetiracetam in patients with tumor-related epilepsy found improved efficacy without mood side effects or impact on quality of life (28).

Zonisamide, clobazam, or perampanel are additional medications to consider, keeping in mind potential side effects for the individual patient (eg, avoid zonisamide in patients with prior nephrolithiasis or clobazam in patients planning pregnancy).

In this population, it is important to re-evaluate the antiseizure medication(s) being used when a patient is on concurrent chemotherapy, tyrosine-kinase inhibitors, or steroids (43).

Surgical management

The role of surgery is most significantly related to tumor type.

Meningiomas. The decision to proceed with resection for meningiomas is often related to whether the patient is symptomatic from the lesion (ie, neurologic deficits, seizures). Patients with asymptomatic meningiomas may undergo routine surveillance imaging to monitor growth rate but may not pursue surgery, especially when a gross total resection is not feasible given neuroanatomic limitations. Approximately 30% of patients with supratentorial meningiomas have preoperative seizures, with one in three patients continuing to have seizures after resection. Approximately 10% of patients without preoperative seizures have focal epilepsy after resection (16).

Long-term epilepsy-associated tumors (LEATs). Given that LEATs are extremely unlikely to undergo malignant transformation, the surgical goal is often to treat the patient’s epilepsy. In the largest retrospective cohort study to date, 77.5% (1027 of 1325) of patients with LEATs were free from disabling seizures 2 years after surgery (25).

Low-grade gliomas. Given that low-grade gliomas often present as expansile tumors on imaging, histological diagnosis is commonly warranted. If the distinction between low-grade glioma and focal cortical dysplasia is not clear, surveillance imaging (ie, repeat imaging with tumor volumetrics every 3 months) is required. The patient should undergo surgery if any change in volume is noted. Gross total resection portends the greatest likelihood for seizure freedom, with 80% of patients becoming seizure free for some interval after gross total resection (14).

High-grade lesions. Tumor resection is essential in the treatment of high-grade lesions. In grade 3 and 4 gliomas, approximately 77% of patients are transiently seizure free after the initial tumor resection (06; 44). The utility of intraoperative ECoG remains unclear in this patient population.

Metastatic lesions. Treatment options for brain metastases include medication, surgery, stereotactic radiosurgery, and whole-brain radiation therapy. There have been no trials comparing seizure outcomes for all of the various treatment options; however, one study evaluated seizure outcomes for radiation therapy and found no difference between whole-brain radiation therapy plus stereotactic radiosurgery versus whole-brain radiation therapy alone or stereotactic radiosurgery alone (21).

Concurrent oncologic treatment

Chemotherapy for low-grade gliomas. There are growing data to support the use of temozolomide to reduce the frequency of seizures (36; 10). The use of upfront temozolomide in low-grade gliomas has been associated with more than 50% seizure reduction in 51% to 59% of patients and seizure freedom in 13% to 55% (44).

Radiation therapy for low-grade gliomas. In a randomized European Organisation for Research and Treatment of Cancer trial evaluating the use of radiation therapy in low-grade gliomas, 75% of patients who had radiation therapy were rendered seizure-free during follow-up (44).

Glioblastoma. The data regarding the benefits of standard oncologic therapy, specifically on seizure control in patients with glioblastoma, are challenging. The largest study to date, which evaluated seizure outcome in 648 patients with high-grade gliomas (both astrocytomas WHO grade 3 and glioblastoma combined), found that 87% (79 of 91) of patients with preoperative seizures were seizure-free at 6 months and 77% (51 of 66) were seizure-free at 12 months after surgery (06). Recurrence or worsening of seizures after patients have completed primary treatment is associated with progression of disease in approximately two-thirds of glioblastoma patients (44).

Special considerations

Pregnancy

Christopher Bonfield and Jonathan Engh provide a good overview of brain tumors and pregnancy (02). There are no specific guidelines for pregnant women with tumor-related epilepsy. Therefore, the recommendations are extrapolated from the literature on women with epilepsy and pregnancy. If a woman is of childbearing age and with epilepsy, folic acid is recommended by both the American Academy of Neurology and the Epilepsy Foundation (23). The dosing is not entirely clear. Generally, 1 mg is recommended for women with epilepsy; however, those taking valproic acid should take 4 mg due to the higher risk of congenital malformations with valproic acid.

Most importantly, seizure medications should not be stopped if a patient discovers she is pregnant as a seizure may pose a major health risk to the mother and the fetus. Lamotrigine and levetiracetam are associated with the lowest risk of congenital malformations (39). As such, these medications are often chosen for women of childbearing age with epilepsy. However, lamotrigine requires weeks to titrate to a therapeutic dose. Therefore, levetiracetam is often the first choice. Given that there is a risk of worsening seizures if an antiseizure medication level decreases more than 35% of the baseline (nonpregnant) level, it is important to monitor drug levels during pregnancy (39).