Perampanel in brain tumor‐related epilepsy: Observational pilot study

Abstract Objective Possible loss of efficacy and potential interactions between antiepileptic drugs (AEDs) and chemotherapy could complicate the management of patients with brain tumor‐related epilepsy (BTRE) that may expose patients to an increased risk of adverse events. Perampanel (PER) is a highly selective, noncompetitive, alpha‐amino‐3‐hydroxy‐5‐methyl‐4‐isoxazole propionic acid (AMPA)‐type glutamate receptor antagonist. This study evaluates the effectiveness, QoL, cognition, and mood of PER in add‐on therapy in BTRE patients. Material and Methods Observational pilot study on the effectiveness of PER as add‐on therapy in BTRE patients with uncontrolled seizures with a 6‐month follow‐up. Results We recruited 26 BTRE patients. During the follow‐up, 16 underwent chemotherapy and 11 radiotherapy; 11 had disease progression. Five patients dropped out. Mean daily PER dosage was 6.6 mg in the 21 patients who completed the follow‐up and 6.4 mg in the ITT population. The mean number of seizures/month decreased from 10.8 ± 15.03 at baseline to 1.7 ± 4.34 in the 21 patients who reached the final follow‐up. Responder rate was 88.4%: Eight patients were seizure‐free, 15 had ≥50% seizure reduction, and 3 remained stable. Four patients (15.4%) reported AEs: 2 required PER dose reduction, and 2 dropped out. Neuropsychological, mood, and QoL questionnaires were not statistically different compared to baseline. There were no significant differences in seizure control in patients with/without IDH1 mutation and with/without MGMT methylation. Conclusions Perampanel proved to be effective on seizure control in BRTE patients and to be well tolerated without negative effects on cognition and QoL. Perampanel could be a valid therapeutic option in BTRE.


| INTRODUC TI ON
In patients with brain tumor, epilepsy represents one of the most important symptoms. It has been estimated that seizures occur in rates varying from 20% to 40% of patients with brain tumors; seizures are the onset symptoms in a significant proportion of cases (Maschio et al., 2014;Perucca, 2013;Van Breemen, Wilms, & Vecht, 2007).
Approximately 10% of patients with brain tumor do not have seizures as first symptom, but develop seizures at a later stage (Maschio, 2012).
Therefore, the selection of the appropriate AED therapy in BTRE patients should be driven by multiple factors, which include not only efficacy in the specific type of seizure to be treated, but also tolerability and drug-interaction potential.
Perampanel is approved as add-on therapy for partial-onset seizures with or without secondary generalization in adult patients with epilepsy (Raedler, 2016).
Perampanel pharmacokinetic profile is characterized by good bioavailability, rapid absorption, and a long elimination half-life.
Three randomized, double-blind, placebo-controlled trials (French et al., , 2013Krauss et al., 2012) and their extension study (Krauss et al., 2014), proved that, in patients with focal seizures with/without secondary generalization, PER was efficacious and well tolerated, and the proportion of patients that remained on PER treatment was comparable to that of placebo.
For these reasons, we decided to study PER as add-on therapy in BTRE patients to evaluate the efficacy on seizure control, as well as tolerability and impact on QoL, cognition and mood for 6 months (European Medicine Agency, 2018).

| MATERIAL S AND ME THODS
Observational pilot cohort study was approved by the Ethical Committee (Prot. n° 0004691.12_04_2017). The study conforms to recognized standards of European Medicines Agency Guidelines for Good Clinical Practices.

| Primary aim
To evaluate efficacy of PER as add-on therapy on seizure control in BTRE patients, after 6 months.
2. To detect the presence of PER-related side effect during followup period compared with baseline.
3. To monitor PER impact on cognition, mood, and Quality of life after 6 months of therapy, compared with baseline.

| Population
BTRE patients with uncontrolled seizure activity despite AED therapy with adequate dosages. Perampanel was added as first or second add-on. Patients were monitored for 6 months and underwent functional and neurological evaluation, count of seizure frequency, monitoring of adverse events, administration of QoL, mood, and cognitive questionnaires. All patients were treated with current standard care for brain tumor, including neuroradiological follow-up. Any patient that experienced uncontrolled seizures with PER at maximum tolerated doses and needed a new AED was considered treatment failure and analyzed as such (Last Observation Carried Forward "LOCF" will give seizure frequency at AED change).

| Investigational product
Perampanel as first or second add-on therapy at dosage ranging from 4 to 12 mg/die was taken as a single oral dose at bedtime. The starting dosage is 2 mg/die with a slow increasing schedule as per label (with a weekly increase of 2 mg/day). Minimal effective dose is 4 mg/day. Depending on seizure control and on occurrence of adverse events, in order to achieve freedom from seizures, PER was titrated up to the maximum approved dosage of 12 mg/day.

| Inclusion criteria
Patients age ≥ 18 years ≤ 75 (both sexes) with primary (low-and high-grade WHO gliomas) or secondary brain tumor, with biopsy or surgical resection; in a stable phase of disease (evidenced by unchanged neuroradiological examinations), with/without chemotherapy (CT), radiotherapy (RT), and corticosteroids started before PER introduction. Structural epilepsy characterized by focal onset (aware/unaware seizures) (Scheffer et al., 2017); ≥4 seizures in the last month, despite 1-2 AEDs at the maximum tolerated stable dosages. Seizure count in the last month before the enrollment. All persons gave their informed consent prior to their inclusion in the study.

| Study design
Trial duration for patient: 24 weeks + 1 week for titration of PER up to 4 mg/day.
Total trial duration including follow-up: 49 weeks, 24 weeks for recruitment.
We decided a follow-up period of 6 months, according to European Medicine Agency-EMA guidelines (European Medicine Agency, 2018).
The study schedule is the following: Patients were given a seizure diary to compile in order to register any information about seizure occurrence in the last month. In order to attest seizure occurrence, patients were requested to call the center after each episode and calls were recorded and compared with questionnaire responses. Perampanel initial dosage of 2 mg/day.

Patients underwent a neuropsychological battery of tests, including
Quality of life and behavioral assessment.  Seizure diary and Investigational Product check. Patients with improved seizure control continued PER therapy at the appropriate dosage with the marketed drug. If during PER treatment period at maximum tolerated dosage, patients reported seizure activity that requires the introduction of another AED, it would be considered as treatment failure and therefore analyzed as such (seizure frequency would be considered equal to that at study entry).

| Primary efficacy variables
To evaluate efficacy of PER, we used the mean seizure frequency in comparison with baseline during the period of 6 months treatment, after having reached minimal effective dose of 4 mg/day and freedom from seizure rate at 6 months of treatment.

| Secondary efficacy variables
Responders are defined as patients who obtained a ≥50% reduction

| Safety variables
Incidence of adverse events using PER evaluated by AEP described later.
An "adverse event" (AE) is whichever unfavorable and unintended sign, symptom, or disease associated temporally with the use of a medical treatment or procedure that might or might not be considered related to the aforementioned treatment or procedure.
Disease progression is not considered an AE. Patients who are administered at least one dose of drug will be included in the toxicity analysis. AEs (spontaneously reported or observed) will be recorded in association with details of time of onset and resolution, intensity, need for treatment, and possible connection with the ongoing treatment in the investigator's opinion.

| Side effect evaluation
Presence of side effects was assessed using Adverse Events Profile (Gilliam et al., 2004): a self-report multi-item questionnaire specific

| Quality of life and psychological state evaluation
Patients' perceived quality of life was evaluated by Quality of life in Epilepsy Inventory (QOLIE 31P-v2) (Cramer & Van Hammée, 2003) a 31-item self-administered questionnaire for epileptic patients.
Psychological state was assessed by Symptom Checklist-90 (SCL-90) (Derogatis & Savitz, 2000) a self-administered 90-item questionnaire, that assesses the intensity and frequency of psychopathological symptoms in the last week.

| Population size and statistical analysis
All enrolled patients will be considered as intention-to-treat population (ITT). Safety and efficacy of PER will be evaluated in this patients' population.
The primary endpoint will be the mean difference in the number of seizure pretreatment and after 6 months. We will use the t test for paired data.
Patients who do not reach 6 months of therapy will be evaluated in the last month of follow-up available.
Based on an earlier study of a large population of drug-resistant patients (Kanwaljit et al., 2016), we estimated an average seizure rate of 4 per month before the introduction of Perampanel; assuming that the treatment gives a reduction in the mean seizures number equal to 2 and estimating, from data of the preceding series, that this difference has a standard deviation (SD) of 2.8, 17 patients will be needed to obtain a statistical power of 80% to a level of significance of 5%. LGG: low-grade glioma;

TA B L E 1 (Continued)
The mean daily PER dosage was 6.6 mg in the 21 patients who In the ITT population (26 patients), the mean number of seizures reduced from 10.6 ± 14.27 at baseline to 2.3 ± 5.3 (p = .004) (last follow-up available 2.4 months) (see Figure 1). Responder rate at 6 months was 95.2%: seven patients seizure-free, 13 with a reduction ≥50%, and 1 remained stable (Table 1).  (Table 2).
In order to evaluate a potential correlation between seizure frequency during PER therapy and issues related to the oncological disease, we compared the decreasing number of seizures depend- In all these, comparisons no significant differences were observed. We reported the results of an observational pilot study on 26 BTRE patients treated for 6 months with PER as add-on therapy.

| D ISCUSS I ON
We observed a reduction in seizure frequency in the 21 patients Responder rate at 6 months (21 patients) was 95.2, with 33.3% of patients who were seizure-free after 6 months. Seizure frequency remained stable in 1 out of 21 patients, and none had seizures worsening.
Literature data on BTRE patient populations treated with PER as add-on are very few; however, they indicated a good seizure response rate; Vecht and colleagues in a prospective study on 12 patients with low-and high-grade gliomas and drug-resistant epilepsy, assuming PER for 6 months for a median daily dose of 8 mg, reported an high seizure response rate in 9 out of 12 patients (75%), seizure freedom in 6 out of 12 patients (50%), improvement in cognitive functions and acceptable safety profile (Vecht et al., 2017).
Izumoto and colleagues evaluated a case series of 12 patients with uncontrollable epilepsy related to both low-and high-grade gliomas treated with PER as add-on therapy and obtained 10 patients who achieved more than 50% seizure reduction and seizure freedom in 6 patients (60%) (Izumoto et al., 2018). Our results are in line with these evidences, indicating a good seizure response to PER treatment as add-on in this patient population.
Regarding patients' subjective reports, 4 out of 26 patients (15.4%) referred presence of side effects. In two patients that reported vertigo (common AEs reported with PER) (Fycompa, 2017), PER was reduced, and in 2 patients that reported aggressiveness, PER was withdrawn. Concerning the evaluation of AEDs related physical side effects (AEP), we observed no statistically significant difference in AEP profile mean scores between basal and final follow-up which values remain stable indicating the presence of moderate physical side effects (see Table 2). Regarding neuropsychiatric side effects (SCL-90), we observed a significant decrease in psychosomatic symptoms' scale mean scores and stability in all domains explored by questionnaire, which values remain stable within normal values at basal and at final follow-up, indicating low incidence of neuropsychiatric disturbances (see Table 2). As for the possibility that PER efficacy remains stable during oncological disease progression, our results indicate that despite F I G U R E 1 Comparison in mean seizure number/month between basal and final follow-up evaluation in total population (n = 21) and in ITT population (n = 26) the high number of patients with progression disease during the follow-up (11 patients, 42.3%), the efficacy of PER on seizure control remains high, also in the group of patients with progression disease (Table 1).
Furthermore, our results did not evidence that the efficacy of PER on seizure control could be influenced by factors related to brain tumors, such as systemic therapy, oncological progression, different histology and malignancy, surgical procedures. This results should be cautiously considered because of low power due to small sample size and short follow-up (only 6 months).
Regarding the molecular indices analysis, in the different two groups (IDH1-mutated/wild type, MGMT with or without promoter methylation) we did not observe significative differences.
Our results differ from the study results of Dunn-Pirio et al. (2018) in which they found that, between patients with a decrease in seizure activity, the majority had IDH1-mutant tumors. This difference could be caused by our low patient number, because for just 15 out of 26 patients we had the analysis for IDH1 and for 11 patients for MGMT.
Regarding QoL evaluation, we did not observe statistically significant differences in QOLIE 31 P mean scores, which values remain within normal ranges at basal and at final follow-up evaluations (Table 2).
Literature data on BTRE patients indicate a correlation between good seizure control and improved scores in Quality of life questionnaire (Maschio & Dinapoli, 2012). Probably, in our sample we did not observe significant improvements in questionnaire mean scores because 9 out of 14 patients also performed concomitant oncological treatments (6 out of 14 CT; 2 out of 14 RT) and 2 other patients were in disease progression. However, the good seizure control obtained by patients with PER in add-on at TA B L E 2 Comparison between quality of life, neuropsychological, psychological tests and AEP profile before and after 6 months of treatment with PER in add-on therapy Regarding cognitive performances, we observed stability in patients' mean scores and at final follow-up compared to baseline. This result is in line with the only one study in literature which shows a low impact of PER on cognitive functions in glioma patients, tested with a short computerized battery (Vecht et al., 2017).
However, our data on tests and questionnaires cannot be generalized due to the small number of patients who repeated control tests at 6 months. With reference to this specific aspect, it could be useful to administer a brief neuropsychological test battery or brief and specific questionnaires in order to avoid the high number of patients with poor compliance, as suggested by literature data (Newton & Maschio, 2015).

| CON CLUS ION
In our study, we observed good efficacy on seizure control without negative effects on cognition and on QoL of PER in patients with BTRE.
Despite the limitations due to the small number of patients, PER could be a therapeutic option in BTRE patients due to responders' high rate and number of seizure-free patients.
These results need further studies with a longer follow-up to confirm this high responder rate and the possible correlation with molecular indices, which could be a starting point for truly tailored therapies for patients with BTRE.

ACK N OWLED G M ENTS
We thank Dr. Selvaggia Camilla Serini for reviewing this manuscript.

At the Biostatistic Unit of Regina Elena National Cancer Institute all
dataset analyzed for the current study is available.