Minimally invasive ictal EEG source localization including sphenoidal electrodes - a retrospective study and a meta-analysis.

BACKGROUND

In patients with pharmacoresistant temporal lobe epilepsy, minimally invasive sphenoidal electrodes can enhance presurgical video-EEG monitoring. Placed into the infratemporal fossa, these electrodes record EEG activity from the mesial temporal lobe. We hypothesized that inclusion of the signal from sphenoidal electrodes would improve the diagnostic sensitivity of ictal EEG source localization (ESL).

METHODS

We retrospectively included patients with pharmacoresistant focal epilepsy, presurgical low-density EEG including bilateral sphenoidal electrodes, subsequent resective surgery and favorable 1-year seizure outcome (ILAE class 1-2). Per seizure, ESL was conducted separately with vs. without signals from sphenoidal electrodes. Source maxima were compared to the resected zone. For comparison with the literature, we added a univariate random-effects meta-analysis of previous ictal ESL studies.

RESULTS

ESL was applied to 184 seizures of 42 patients (39 with anterior temporal lobe epilepsy). At seizure level, ESL with vs. without sphenoidal electrodes achieved diagnostic sensitivities of 68 % (95 %-CI: 61-75 %) vs. 46 % (39-53 %; p < 0.001). Patient-level diagnostic sensitivity was 50 % (35-65 %) vs. 19 % (7-31 %; p = 0.002). The meta-analysis of ten studies on 233 patients resulted in a patient-level summary sensitivity of 68 % (44-84 %) albeit with substantial heterogeneity (I2=0.689).

CONCLUSIONS

As a proof of principle, sphenoidal electrodes significantly improved the diagnostic sensitivity of ictal ESL in absence of subtemporal scalp electrodes. Previous studies on ictal ESL yielded substantially heterogenous results, depending on their specific methodology. In presurgical epilepsy cases evaluated with both sphenoidal electrodes and ictal ESL, both techniques should be combined as presented here.