Long-term monitoring of brain tumors: When is it necessary?


  • Jeffrey Kennedy,

    1. Comprehensive Epilepsy Center, Department of Neurology, Northwestern University Feinberg School of Medicine, Chicago, Illinois, U.S.A
    Search for more papers by this author
  • Stephan U. Schuele

    Corresponding author
    1. Comprehensive Epilepsy Center, Department of Neurology, Northwestern University Feinberg School of Medicine, Chicago, Illinois, U.S.A
    • Address correspondence to Stephan U. Schuele, Department of Neurology, Northwestern University Feinberg School of Medicine, Abbott Hall # 1425, 710 North Lake Shore Drive, Chicago, IL 60611, U.S.A. E-mail: s-schuele@northwestern.edu

    Search for more papers by this author


Tumors, particularly low grade glioma and glioneuronal tumors, account for 25–35% of patients who are undergoing epilepsy surgery for intractable seizures. A comprehensive epilepsy evaluation including video–electroencephalography (EEG) monitoring is useful for most of these patients, to determine the optimal extent of resection for the achievement of seizure-free outcome without causing postoperative deficits. Video-EEG monitoring for patients with brain tumor should also be considered in specific situations, such as patients with new postoperative seizures or advanced tumors with unexplained mental status change.

The dogma that patients undergoing epilepsy surgery need a preceding video–electroencephalography (EEG) evaluation may not hold true for all patients with tumor-related epilepsy (TRE). Seizure-free outcome in patients who are undergoing brain tumor resection is higher than in most epilepsy surgery series—before as well as after video-EEG monitoring became available. In someone found to have a brain tumor after presenting with a seizure, a causal correlation is high, although cases with nonepileptic events and incidental magnetic resonance imaging (MRI) findings have been reported (Vernooij et al., 2007; Scarella et al., 2012). A detailed seizure history and interictal epileptiform EEG findings may suffice for establishing an anatomic connection between the tumor and the origin of the seizure. Not all patients with brain tumors have sufficiently frequent seizures to make recording of ictal events a realistic goal. On the other hand, patients with frequent seizures related to low grade brain tumors benefit from a comprehensive presurgical epilepsy evaluation, including video-EEG monitoring to optimize their chances of a long-term seizure-free outcome.

Aside from a presurgical evaluation, long-term EEG monitoring in patients with brain tumor can be useful for a variety of other reasons. Prolonged ambulatory EEG increases the yield of interictal abnormalities significantly in patients who have a normal routine EEG. Diagnostic video-EEG monitoring can be indicated to distinguish epileptic from nonepileptic seizures. It is also useful in patients at risk for undetected clinical seizures or electrographic seizures as the cause of an unexplained change in cognition and responsiveness.

New-Onset Seizure Evaluation in Patients with Brain Tumors

The incidence of brain tumors in people with epilepsy is around 4% (Olafsson et al., 2005). An epileptic seizure is the presenting clinical symptom in 30–50% of patients with brain tumors. Another 10–30% will develop seizures later in the course of the disease (van Breemen et al., 2007). More than 95% of glioneuronal tumors and >75% of low grade gliomas will develop TRE. Patients with high grade glioma develop seizures in 30–50% and with meningioma in around 25%.

In a patient presenting with new-onset seizures, a brain tumor is typically found by MRI imaging, which is highly recommended for anyone who presents with seizures suggesting a focal onset or who have a certain age of onset of seizures. In any patient presenting with seizure-like events, a description based on patient recollection and a detailed witness report is the crucial part of the initial evaluations. In patients with brain tumors, the clinical history will help to determine if the events were epileptic, nonprovoked, and likely related to the structural lesion. Patients with temporal lobe brain tumors present with a specific aura in >70% of cases (Zaatreh et al., 2003). The aura is often described as an epigastric rising sensation, a depersonalization experience, or other psychic phenomena, all consistent with a temporal lobe seizure origin. An aura in extratemporal lobe epilepsy is less common. Forty percent to 50% of patients with frontal lobe TRE report auras that support a focal onset, but are often nonspecific in terms of localization (Zaatreh et al., 2002).

The presence of interictal epileptiform discharges (IEDs) provides corroborating evidence that the events are epileptic. The maximum of the activity may correlate with the location of the tumor and confirm that the lesion on MRI is epileptogenic, but the absence of IEDs does not rule out an underlying epileptic process (Fig. 1). In a surgical series of patients with temporal lobe TRE, 82.4% of patients had an EEG; 57% showed temporal IED on a routine 20–30 min study, and the yield increased to 75% during long-term monitoring (Zaatreh et al., 2003). The appearance of spikes over the contralateral temporal region can be seen in around one third of patients and does not confer an unfavorable outcome (Morris et al., 1998). According to a surgical series of 37 patients with frontal lobe TRE, interictal EEG was recorded in 78.4% prior to surgery, 52% showed IEDs, of which only one fourth correlated with the tumor site, whereas the majority of the activity was bilateral, generalized, or falsely localizing (Zaatreh et al., 2002).

Figure 1.

Correlation of lesion and symptoms. Thirty-four year-old right-handed woman with onset of nocturnal events 3 years ago and a total of six events. Patient is unaware of the events. Per witness report she suddenly starts stiffening all four extremities and develops low amplitude shaking. The events last 30 s, followed by short confusion and return to sleep. Routine EEG and 24 h ambulatory EEG were both normal. This T1 gadolinium-enhanced MRI shows an enhancing extraaxial lesion, 7 × 4 mm, along the right side of the anterior portion of the falx cerebri consistent with a meningioma. Note: seizure semiology is consistent with the lesion location. Interictal EEG in frontal lobe epilepsies is often normal. Meningiomas are an incidental finding in approximately 1% of MRI scans. Repeat MRI scans have been stable so far. A definite correlation between lesion and seizures is not possible at this point without video-EEG monitoring.

Presurgical Evaluation

Epilepsy surgery

Video-EEG monitoring is at the core of a presurgical evaluation, which has the goal of delineating the epileptogenic zone, the area that is to be removed to render the patient seizure-free. In most patients, a complete resection of the contributing structural lesion and the ictal-onset zone is required and provides the best preoperative estimate for resection. A presurgical evaluation will also define surrounding eloquent cortex and provide information on the maximum extent of resection that can be safely performed to optimize seizure-free outcome without causing a postoperative deficit (Fig. 2).

Figure 2.

Concept of epilepsy surgery. In patients with TRE, the lesion and the ictal-onset zone are often closely related, which explains the good results of lesionectomies.

In patients who are undergoing surgery for chronic epilepsy, brain tumors constitute up to 25–35% of cases, with 70–80% involving the temporal lobe. There is mounting evidence that patients with low grade glioma and glioneuronal tumors have an extended long-term survival with gross total resection (GTR), and seizure control has a major impact on their quality of life. A recent meta-analysis of 1,181 patients across 41 studies with low grade glioma and glioneuronal tumors who underwent resection and had follow-up for at least 6 months reported that >85% of patients had intractable epilepsy, and 78% of the refractory patients became seizure-free after surgery (Englot et al., 2012). According to another systematic review of 910 patients of 39 studies, patients with glioneuronal tumors have a better outcome with GTR over subtotal lesionectomy (odds ratio [OR] 5.34; 95% confidence interval [CI] 3.61–7.89), shorter duration of epilepsy (OR 9.48; 95% CI 2.26–39.66), and absence of secondarily generalized seizures (OR 0.40; 95% CI 0.24–0.66). The authors were also able to show that in the subgroup of patients with temporal lobe epilepsy, GTR plus hippocampectomy, corticectomy, or a combination of both did better than GTR alone. However, the data did not permit the determination of how patients were selected for a more extensive resection and how the resection of mesial structures impacted functional outcome, particularly with dominant temporal lobe epilepsy.

A number of compelling arguments speak for preoperative video-EEG monitoring and a comprehensive epilepsy evaluation in patients with low grade glioma and glioneuronal tumors and intractable epilepsy. The rate of a second pathology coexisting with the MRI-visible brain tumor (cortical dysplasia, hamartia, hippocampal sclerosis, nodular heterotopias) is around 20%, and may be even higher in glioneuronal tumors and temporal lobe TRE (Prayson et al., 2010). The decision to extend a resection across the borders of the imaging lesion bears the risk of postoperative functional decline. Memory in particular cannot be adequately assessed in the operating room. Patients with temporal lobe tumors constitute the majority of patients with TRE undergoing epilepsy surgery. Language laterality affects not only language but also memory dominance and should be known prior to surgery.

Surface video-EEG monitoring can help to differentiate between a neocortical and mesial temporal lobe epilepsy and can advocate for or obviate the need to extend the resection and risk functional decline (Hamer et al., 1999; Kennedy & Schuele, 2012). An inpatient video-EEG evaluation allows the placement of additional anterior temporal electrodes including sphenoidal electrodes, which can be helpful in differentiating mesial versus neocortical spike populations. Ictal recording of scalp EEG is more reliable in localizing the epileptogenic zone. More than one seizure type has been reported in up to nearly half of patients who present with seizures, and the corresponding electrographic onset of each type should be captured as part of the presurgical evaluation to determine the extent of potentially epileptogenic cortex. False lateralization of seizures on surface recordings can occur and invasive monitoring can be performed to clarify the actual localization. Surface video-EEG monitoring will also help to determine if electrocorticography and mapping is indicated and if the patient should undergo intraoperative or extraoperative electrocorticography. For many patients with extratemporal lobe TRE, mapping of motor, sensory, or speech function can be done intraoperatively. A more comprehensive evaluation is warranted if the epileptogenic zone is not clearly delineated by the lesion and close to essential functional areas (Fig. 3).

Figure 3.

Invasive monitoring in a patient with dominant temporal lobe epilepsy. Fifty-seven-year-old right-handed woman who presented with a witnessed new-onset seizure at work consisting of staring and automatisms. MRI of the brain showed a left tentorial extraaxial enhancing 3.4 × 3.3 cm meningioma (left panel), with moderate mass effect on the left hippocampus. Overnight video-EEG monitoring showed left temporal sharp waves. The patient underwent partial resection of the meningioma 2 days later and continued to have two to three seizures per week and failed four antiepileptic medications in the following 2 years. Noninvasive EMU evaluation showed left temporal sharp waves and captured several typical events with onset from the left temporal area. Language was left dominant. Neuropsychological assessment showed full-scale IQ (FSIQ) in 68th percentile with mild to moderate deficits in verbal encoding and anterograde verbal memory. During an invasive evaluation, interictal and ictal discharges were seen from the left anterior temporal tip and adjacent lateral temporal neocortex (right panel). An extensive anterior temporal resection was performed. The patient developed new-onset episodic speech impairment (stuttering) 1 month postoperatively, which was diagnosed as nonepileptic based on repeat video-EEG monitoring. The events responded to counseling. The patient has otherwise remained seizure-free for now 2.5 years postoperatively.

More than 20% of patients do not become seizure-free after surgery, and a comparison with the preoperative video-EEG evaluation is extremely helpful to determine the reason for seizure recurrence. The most common causes for surgical failures are an insufficient resection of the epileptogenic cortex, a wrong localization, or the wrong diagnosis. De novo postoperative psychogenic seizures are seen in 3–4% of patients who are undergoing brain surgery for epilepsy or other indications (Reuber et al., 2002). Preoperative video-EEG monitoring not only excludes patients with nonepileptic seizures from undergoing unnecessary surgery, it also helps to document if postoperative seizures differ from preoperative events.

Tumor surgery

A logical obstacle in performing video-EEG monitoring is sporadic seizures in some patients, and the yield of capturing events during a typical 5–10 day epilepsy monitoring unit (EMU) stay may be low. In some studies, the average seizure frequency of patients undergoing temporal lobe surgery for TRE was 4.7 per week (range 1 in 2 months to 8 per day) and in patients with frontal lobe tumors, 7.6 per week (range 1 in 3 months to 15 per day; Zaatreh et al., 2002, 2003). These studies suggest that most patients with low grade glioma and glioneuronal tumors who undergo surgery have sufficiently frequent seizures to allow seizure characterization preoperatively with video-EEG monitoring.

In patients with higher grade tumors, complete resection of the epileptogenic zone is often not feasible and surgery may be limited to a biopsy, partial lesionectomy, or debulking. Seizures may have a limited impact on quality of life in that situation. Despite the fact that more than half of patients with higher grade tumors have a seizure frequency of more than one per month on medication and could be monitored, most patients probably choose to avoid a time-consuming video-EEG evaluation (Moots et al., 1995). However, interictal EEG findings on routine or ambulatory EEG can be extremely helpful to tailor even a partial resection and optimize the chance to make an impact on seizure frequency in patients who are significantly affected by their seizures (Fig. 4).

Figure 4.

Patient with malignant transformation of an oligodendroglioma. Thirty-eight year-old right-handed man who initially presented 6 years ago with recurrent episodes of a sudden bad taste in his mouth and a feeling of lightheadedness, lasting 2–3 min. Initial MRI showed a right temporal low grade oligodendroglioma, which was partially resected. One year later he developed mostly nocturnal generalized convulsions that were eventually controlled with two antiepileptic medications. He returned with increasing auras over the past 4 months associated with fear and problems with coordination and concentration, occurring several times a week and affecting his ability to work as a consultant. Ambulatory EEG (bipolar montage, left over right) showed right mid-temporal spikes (T8, A), occasionally with a more extensive right hemispheric (B) or generalized field (C). Intraoperative electrocorticography showed repetitive spikes from a right lateral temporal plate (D, upper part), corresponding with the transition area between abnormal and normal signal change on MRI (T2 sequence, D, lower part). Patient has been seizure-free for 2 months.

Diagnostic Long-Term Monitoring

Epileptic versus nonepileptic events

Patients with nonepileptic events represent 20–25% of referrals to tertiary epilepsy centers for intractable seizures. Incidental tumors or cystic lesions on MRI are not infrequent, and about 1% of MRIs show a meningioma or an arachnoid cyst that may not necessarily be the cause of the patient's seizures (Fig. 1; Vernooij et al., 2007). The rate of nonepileptic seizures in patients with highly epileptogenic brain lesions is unknown, but cases of patients with hypothalamic hamartoma and psychogenic gelastic seizure or coexisting nonepileptic and epileptic seizures have been reported (Zaatreh et al., 2002; Scarella et al., 2012).

On the other hand, frontal lobe seizures in particular can have an unusual semiology with a normal interictal EEG and may be misdiagnosed as nonepileptic seizures without the help of video-EEG monitoring. Not infrequently, MRI is either not performed in such patients or abnormalities including glioneuronal tumors are not detected until the patient undergoes video-EEG evaluation and is found to have a focal epilepsy from a specific region (Winston et al., 2013).

Unrecognized clinical seizures

Clinical seizures go unrecognized >50% of the time. According to one study, 73% of focal seizures with impaired and 26% of focal seizures without impaired consciousness, and 42% of secondarily generalized seizures go unrecognized with significant impact on treatment and clinical decision making (Hoppe et al., 2007). This is particularly important when patients with epilepsy are evaluated for driving privileges or report unexplained cognitive and memory problems.

Acute postoperative seizures

Acute postoperative seizures in patients who are undergoing epilepsy surgery are seen in around 25% of cases and carry a poor long-term prognosis, unless the seizures can be considered acute symptomatic from the surgery, occurred within the first 24–48 h, and differed in semiology from the preoperative events. In patients undergoing surgery for supratentorial tumors, new-onset postoperative clinical seizures within the first week are seen in 8–15% of patients (Hardesty et al., 2011). Prophylactic antiepileptic drug (AED) treatment provides no benefit and has been linked to an increase in postoperative adverse events and generalized seizures. Postoperative seizures can be recurrent and difficult to detect or diagnose (Ericson et al., 2011). Video-EEG monitoring is highly recommended for any postoperative neurosurgical patient with unexplained transient or persistent neurologic symptoms.

Status epilepticus in patients with advanced brain tumors

Brain tumors account for 4–12% of cases of convulsive status epilepticus (Neligan & Shorvon, 2010). A change in seizure frequency or type after a postsurgical period of improvement is often related to tumor recurrence or bleeding, particularly if not explained by poor medication compliance. Nonconvulsive status has been described in patients with primary brain tumors, metastasis, and patients with systemic cancer (Cavaliere et al., 2006). The rate of nonconvulsive status is unknown but long-term video-EEG monitoring should be considered in patients with brain tumor who present with altered mental status that is unexplained by MRI imaging or systemic metabolic disturbance.


Dr. Schuele is on the speaker's bureau for GlaxoSmithKline and Lundbeck. Dr. Kennedy reports no potential conflict of interest in relation to this work. We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.