Value of ultralong‐term subcutaneous EEG monitoring for treatment decisions in temporal lobe epilepsy: A case report

Abstract Treatment decisions in epilepsy critically depend on information on the course of the disease, its severity and options for specific local interventions. We here report a patient with pharmaco‐resistant non‐lesional temporal lobe epilepsy with evidence for predominant right temporal epileptogenesis. While seizure frequency had been grossly underestimated for many years, ultralong‐term monitoring with a subcutaneous EEG device revealed actual seizure frequency (66 over 11 months vs four patient‐documented seizures), providing objective data on treatment efficacy and additional supportive lateralizing information that played a decisive role for the choice of surgical treatment, which had been rejected by the patient prior to this information.


| INTRODUCTION
Treatment decisions in epileptology depend crucially on the assessment of patient-individual burden of disease, considering in particular seizure frequency and severity.So far, the disease evaluation has mostly been based on patient reports on seizure occurrence during a reference period, for example, during the months preceding an outpatient visit at the treating neurologist.][3] For objective seizure documentation, in-hospital monitoring using video-EEG surveillance is considered the gold standard when a discrepancy between patient-based documentation and actual seizure load is suspected.The duration of such in-patient video-EEG monitoring is mostly limited to a period of 1-2 weeks, which again may lead to an undersampling of seizures 4,5 and may not represent the situation under real life conditions.That has justified the need to use ultralong-term monitoring periods of months in order to establish a realistic view on seizure occurrence. 6e here report a clinical case in which ultralong-term monitoring has been performed with a subcutaneously implanted EEG device 7 to ascertain the real seizure frequency and the occurrence of seizure patterns in the right hemisphere in a patient with temporal lobe epilepsy who had an unknown seizure frequency and limited information on the predominant lateralization of seizure patterns.The information was relevant for the assessment of pharmacoresistance and to decide if the seizure load justifies the decision in favor of epilepsy

Abstract
Treatment decisions in epilepsy critically depend on information on the course of the disease, its severity and options for specific local interventions.We here report a patient with pharmaco-resistant non-lesional temporal lobe epilepsy with evidence for predominant right temporal epileptogenesis.While seizure frequency had been grossly underestimated for many years, ultralong-term monitoring with a subcutaneous EEG device revealed actual seizure frequency (66 over 11 months vs four patient-documented seizures), providing objective data on treatment efficacy and additional supportive lateralizing information that played a decisive role for the choice of surgical treatment, which had been rejected by the patient prior to this information.

K E Y W O R D S
epilepsy treatment, seizure lateralization, subcutaneous EEG, ultralong-term EEG monitoring surgery as an alternative treatment.Since the predominant seizure lateralization could only be partly established during in-patient monitoring, seizures arising from the right temporal lobe were analyzed as main contributors to seizure load that proposed the patient as a suitable candidate for possible surgical intervention.Furthermore, the patient became aware of her seizure load only when using the objective EEG-based seizure documentation.

| CASE PRESENTATION
A 52-year-old-female right-handed geriatric nurse presented with seizures occurring from the age of 32 years on.From the disease onset, focal seizures with impaired awareness were reported for which she frequently had a complete amnesia.Others reported staring, oral and manual automatisms, and unresponsiveness lasting for 2-3 minutes.In the first years of disease, focal aware symptoms in the form of epigastric sensations at times had been initial seizure symptoms.Despite various therapeutic regimens with anti-seizure medications (carbamazepine, lamotrigine, oxcarbazepine, and levetiracetam), persisting seizures were reported at monthly intervals.Due to lack of awareness during the seizures, an underreporting was suspected early on after diagnosis.
Given the failure of multiple ASMs to control seizures, a first presurgical evaluation was performed in 2015.During 7 days of video-EEG-monitoring, eight seizures were recorded, six from sleep and two from wakefulness, none of them reported by the patient by pressing an alarm button.The first clinical sign on video-EEG was an arousal or a slight change in behavior.The patient showed a gaze deviation to the right side, right manual automatisms and tonic posturing of the right limbs as well as unresponsiveness to stimuli.The EEG showed rhythmic activity 7-8/s right temporal (Sp2, T2, T4) in all seizures about 20 seconds after clinical seizure onset.Epileptiform discharges appeared frequently over the right or left temporal region in approximately equal distribution.
A high-resolution cerebral 3T MRI did not reveal an epileptogenic lesion.An FDG-PET CT scan showed a small decrease in FDG uptake in the anterior right temporal lobe.Based on these findings, an invasive stereo EEG (SEEG)-evaluation was recommended by the interdisciplinary case conference, yet rejected by the patient.
Following further medication changes to oxcarbazepine and levetiracetam and a withdrawal of her driving license after a traffic accident considered to be seizure-related, she again presented in our outpatient clinic to obtain a driving allowance.A second video-EEG monitoring was performed to verify the reported seizure freedom, yet it revealed two habitual seizures within 2 days on unchanged ongoing medication.The seizures again showed right temporal onset with rapid propagation to the left temporal region.
Despite up-titration of her medication up to adverse effects, no complete seizure control was achieved, the patient continued to report up to three seizures from wakefulness per month.In another non-invasive presurgical evaluation, six seizures with right temporal seizure onset were recorded, none of which with a patient alarm.Neuropsychological assessment revealed frontal, temporal, and parietal dysfunction with emphasis on the non-dominant (right) hemisphere.In a subsequent invasive video-EEG monitoring with nine depth EEG electrodes (six electrodes right temporal, three electrodes left temporal) over 9 days, a total of nine focal aware and unaware epileptic seizures occurred, one of them with transition to a bilateral tonic-clonic seizure; only three of nine seizures were reported by a patient alarm.Intracranially, six seizures orginated in the right amygdala and three seizures in the left amygdala seizure onset.Semiology did not provide lateralizing information in eight of nine seizures, the only lateralizing feature was a sign of four with leftsided arm extension during transition to a bilateral tonic-clonic seizure.
Given a bilateral epileptogenesis, yet a ratio of 22:3 seizures suggesting a predominant right temporal onset, somewhat reduced chances for seizure freedom following epilepsy surgery were communicated to the patient.She did not agree to a surgical intervention.With continued antiseizure medication, she initially reported no seizures for 11 months until another unexplained fall down the stairs occurred and resulted in a coccygeal fracture.At this point, she was offered implantation of a system for ultralong EEG registration (UNEEG®) for a further evaluation of seizure frequency and right temporal involvement.The lead of the subcutaneous EEG device with three EEG channels was implanted over the right temporal lobe of the patient.The data were recorded at Simultaneous subcutaneous and scalp EEG recordings showed a high signal quality of subcutaneous ictal patterns (Figure 1).
During an outpatient monitoring of the initial 7 months, subcutaneous EEG showed 52 right temporal seizure patterns in contrast to two seizures documented in the patient diary which had occurred at daytime and was noted by her partner; 33 (63%) seizures occurred during nighttime (10 pm-6 am).
Following a new fall with a severe seizure-related injury, cenobamate was added to her ASM at a maximal dosage of 150 mg/d, resulting in a decrease in seizure frequency as shown by subcutaneous EEG monitoring (Figure 2), yet with re-occurrence of higher seizure rates after a "honeymoon period" and another fall with injury.
During a total recording period of 11 months, 66 right temporal seizure patterns were documented in subcutaneous EEG recordings, which was in striking contrast to only four seizures reported by the patient herself.Two of these seizures showed a timely coincidence with those detected in the subcutaneous EEG, whereas for the two other seizures, no electrographic ictal pattern was detected on the days reported in the patient's seizure diary.
Given the information provided by the ultralong-term ambulatory subcutaneous EEG recordings, the patient now opted for a two-thirds anterior resection of the right temporal lobe, given the high numbers of seizures with a right temporal ictal EEG patterns for which she was amnestic, and the severity of seizures.

| DISCUSSION
This case shows that subcutaneous ultralong-term EEG monitoring can provide relevant additional information beyond what patients can report on their seizures.Subcutaneous EEG provided high-quality EEG recordings over a period of 11 months, confirming earlier reports on long-term stability of recordings, 8 on its good tolerability and safety, 9 and allowing to clearly identify ictal electrographic patterns and to base seizure counting thereon.Added value of the diagnostic procedure led to the identification of a more than 10-fold higher seizure frequency than reported by the patient herself in her seizure diary, thus confirming a gross underreporting of seizures as had been suggested by only 2 of 25 seizures reported also during in-hospital monitoring.This led to a more valid judgment on the severity of epilepsy in this patient, and added evidence for insufficient efficacy of pharmacotherapy.Overall, this supported the decision in favor of alternative, non-pharmacological treatment approaches.This can be epilepsy surgery 10 or neuromodulatory treatment approaches (eg, VNS, 11 DBS, 12 RNS, 13 or FCS 14 ).
Moreover, device implantation over the temporal lobe in a patient considered a possible candidate for a surgical intervention provided additional evidence on the involvement of the right temporal lobe in focal seizures.The total number of electrographic seizure patterns recorded over the right temporal lobe was far higher than during three in-patient video-EEG monitoring (66 vs 25 Seizures), thus strongly suggesting an at least predominant role of this temporal lobe in seizure generation.Short durations of in-patient monitoring have been shown to considerably undersample seizures and may provide insufficient information on the lateralization of seizure origin. 15,16A correct lateralization of at least the predominant side of seizure generation is, however, a prerequisite for resective epilepsy surgery, laser-based thermocoagulation 17 or epicranial Focal Cortex Stimulation. 14Whereas the unilateral implantation used here could be based on prior imaging and EEG findings, other patients in whom lateralization of seizures is more unclear may particularly profit from future devices for ultralong-term monitoring presently under investigation. 15,18The case here shows that in patients in whom additional converging evidence is needed, a unilateral implantation may already provide relevant information which can contribute to decisions in favor of a focal intervention, as offered here in the form of a two-thirds anterior temporal lobe resection.
In two seizures reported only by the patient, inaccuracy in seizure reporting cannot be ruled out as a reason for the discrepancy with subcutaneous recordings.There is thus a limitation in the interpretation of these two seizures that it is not possible to exclude that EEG-based seizure detection was not sensitive enough or that the right temporal lobe was not affected in these cases.As muscle activity occurs not infrequently in temporal lobe seizures, particularly during oral automatisms, ictal EEG patterns may be obscured by this, suggesting that two-channel EEG recordings may not always allow to detect more subtle ictal EEG patterns.

F I G U R E 1
Simultaneous registration of a seizure in surface EEG (top) and subcutaneous EEG (bottom) at seizure onset (A, red arrow), during seizure evolution (B) and at seizure end (C) on the right temporal lobe.Both the initial rhythmic temporal theta pattern and the late rhythmic delta pattern are clearly visible, intermediate parts are obscured by muscle artifacts (B) in subcutaneous recordings, and the propagation to the left temporal lobe is visible only with the bilateral scalp recordings.207 Hz and bandpass filtered 0.5-48 Hz equiripple FIR filter with a sidelobe attenuation of 40 dB and passband ripple of <0.1 dB.

F I G U R E 2
Comparison of the number of seizures documented by the patient and right temporal seizure patterns detected by subcutaneous EEG (subQ-EEG) monitoring.