Nonlesional epilepsy—a case report in adult epileptology

We report the case of a right-handed, 33-year-old woman suffering from magnetic resonance imaging (MRI)-negative pharmacoresistant epilepsy since the age of 6 years. The semiology of her epileptic seizures has remained stable since their first occurrence. They start with an expression of disgust without facial motor asymmetry, followed by oral automatisms and mild grunting. Subsequently, the patient experiences transient deficits of comprehension and language production. The patient generally presents with an alteration of consciousness during the event, although she continues repetitive activities. Postictal confusion is short lasting, from 15 to 180 s. She has never had a bilateral tonic–clonic evolution. Frequency of occurrence, at the time of our evaluation, was quite variable, from 2–3 times per month to 10 times per day. The epilepsy was pharmacoresistant despite polytherapy.

The patient’s personal history was otherwise unremarkable and family history was negative for brain disorders. There was no complaint of cognitive impairment or mood disorder.

The patient underwent a presurgical evaluation. At the time of the non-invasive presurgical evaluation, the patient was receiving carbamazepine (800 mg/day), levetiracetam (1000 mg/day), lacosamide (250 mg/day), and (on-demand) clobazam. Long-term video-electroencephalogram (EEG) revealed focal interictal epileptiform discharges over the left frontotemporal lobe (F3-C3 and T7-P7; see Fig. 1). We also observed diffuse and bilateral synchronous epileptiform discharges, with slight asymmetry of amplitude and/or onset in favor of the left side (Fig. 2). Electric Source Imaging (ESI) was localized in the left temporal pole for the focal discharges. We also recorded eight seizures. At seizure onset, the patient showed a face of disgust (Fig. 3) followed occasionally by a right-hand wiping and then bilateral nasolabial myoclonus. The patient had two seizures with bilateral tonic–clonic evolution in the context of medication reduction. One of them was preceded by right hemi-facial myoclonus and forced rotation of the head with the eyes toward the right. The EEG lateralization was less clear than during interictal activity (see the diffuse, bilateral spike-and-wave discharges at the onset of the second seizure, Fig. 4). However, in three seizures, a left-sided rhythmic activity was identified later during the seizure (Fig. 5). Brain 3T Magnetic Resonance Imaging (MRI) with a dedicated epilepsy protocol showed a discrete hippocampal asymmetry, with a smaller size on the left side, but without a clear sign of hippocampal sclerosis. Voxel-based morphometry and 7T MRI did not find a significant abnormality. FluoroDeoxyGlucose-Positron Emission Tomography ([¹⁸F]FDG-PET) showed a left cortical temporopolar hypometabolism and the maximal hyperperfusion on ictal Single Photon Emission Computed Tomography (SPECT) and subtraction analysis (Subtraction Ictal SPECT COregistered to MRI: SISCOM) was localized in the left cortical temporopolar region. Neuropsychological testing (performed 1 year before our evaluation) revealed a slight deficit for elaborate language and short-term verbal memory with no evolution. Psychiatric evaluation was unremarkable.

At our multidisciplinary case review, we considered a left hemispheric, nonlesional epilepsy involving the medial left temporal lobe, cingulate cortex, and/or insula, with a global picture of a “temporal plus” epilepsy and the expression of disgust with some aspects of ictal pouting (“chapeau de gendarme”) described in similar cortical structures [1]. The occasional bilateral synchronous discharges suggested the involvement of a deep or midline structure. The clear localization toward the left of focal discharges, as well as the ictal discharges concentrated over the left hemisphere after seizure onset (see above and Fig. 5), the semiology before bilateral tonic–clonic evolution, and the results of PET and SPECT imaging and neuropsychological testing all indicated a left hemispheric onset. We proposed an invasive recording with stereotactic depth electrodes to investigate the suspected regions discussed above (temporal pole, amygdala, hippocampus [including lateral temporal contacts], superior posterior temporal gyrus, anterior, middle and posterior insula, anterior, middle and posterior cingulate, orbitofrontal, inferior frontal gyrus and supplementary motor area) including one contralateral region (anterior hippocampus). See Fig. 6 for an overview of the implantation scheme.

During SEEG, the patient had more than 20 focal seizures. Seizures started in the lateral (i.e., cortical) left temporopolar region and, for some of them, with a simultaneous onset in the superior posterior temporal gyrus. Cortical mapping revealed that high-frequency stimulation of the superior temporal area was sufficient to elicit speech deficit already at low intensity (50 Hz, 4 mA, 2 s). Interictal epileptiform discharges occurred predominantly on left temporolateral contacts and less often left insular contacts; bursts of diffuse discharges were also recorded and showed a slightly earlier onset in the left temporal pole contacts (Fig. 7).

Based on the ictal pattern, we concluded that the seizure-onset zone covered a large network involving predominantly the left lateral temporal pole. Resection of the temporal pole, as well as the left amygdala, but not the left hippocampus, was proposed. The left superior temporal gyrus was not considered for resection, due to the risk of language impairment. The rapid involvement of the posterior superior temporal lobe as well as the bilateral synchronous interictal discharges were considered to somewhat reduce the odds of good outcome but were not a contraindication for surgery.

The patient accepted the resection of the left temporal pole and left amygdala. Anatomopathological studies showed a focal cortical thickening with dysmorphic neurons and no balloon cell, compatible with a focal cortical dysplasia type IIa (ILAE classification; Fig. 8). The patient has not presented a recurrence of seizure since surgery (last follow-up at 12 months) and her EEG showed no postoperative epileptiform abnormality. Postoperative medication comprised carbamazepine (800 mg/day) and levetiracetam (1000 mg/day), the latter being currently progressively withdrawn given the 12-month seizure-free follow-up.

The lack of identification of a potential epileptogenic lesion on imaging should not prevent a presurgical evaluation of epilepsy. On the contrary, although conflicting data exist regarding the link between disease duration and surgical outcome [2, 3], some evidence shows that shorter epilepsy duration is associated with a better outcome in focal cortical dysplasia [4], which underlines the importance of promptly referring patients to presurgical evaluation of epilepsy when pharmacoresistance is confirmed or suspected.

Diffuse, bilateral epileptic activity can be challenging when discussing the perspective of epilepsy surgery but can represent fast bilateral propagation from a focal epileptic activity. Bilateral implantations are associated with higher risk for patients and should be carefully planned [5]. However, even when EEG cannot clearly inform the localization of epilepsy, other diagnostic tools can be decisive. In the past few decades, more widespread use of several imaging tools for multimodal evaluations has improved the surgical outcome [6]. In the present case, the localization value of ESI, PET, SPECT and neuropsychological testing was indeed particularly helpful in the elaboration of the intracranial implantation scheme. In the present case, bilateral implantation was thus largely avoided (except for one right temporal electrode) thanks to the multimodal concordance, decreasing the perioperative risk. Altogether, neither a lack of a lesion on MRI nor the presence of bilateral synchronous EEG activities represents a definite contraindication to an invasive presurgical monitoring.

This clinical case illustrates the difficulties raised by the lack of an epileptogenic lesion on imaging studies and partly bilateral synchronous interictal discharges on EEG. Patients with nonlesional epilepsy remain potential candidates for epilepsy surgery and should thus be evaluated accordingly. In this case, some localizing EEG features and lateralized results in functional imaging (PET, SPECT) were critical to support the presurgical evaluation. Other diagnostic tools and a thorough evaluation of the patient during long-term monitoring can indeed improve the management of epilepsy and should thus be offered to all patients with suspected or confirmed pharmacoresistant epilepsy.