Abstracts

Rationale: Patients with intractable focal epilepsy and normal MRIs have generally a less favourable surgical outcome compared with those with lesional epilepsy. This is usually explained by the lack of a reliable marker of the epileptic tissue. Since HFOs may have clinical value as biomarkers for epileptogenesis, we hypothesized that their presence would provide a more accurate localization of the epileptogenic tissue, favoring a better prognosis. Methods: Intracerebral EEGs (iEEGs) were recorded in 17 patients with intractable focal seizures and normal MRIs. The iEEG was low-pass filtered at 500 Hz and sampled at 2000 Hz. Spikes and HFOs were visually marked in 5 or 10-min slow wave sleep samples. High-pass filters at 80 and 250 Hz were used to identify ripples and Fast Ripples (FRs). We analyzed the rates of HFOs and the number of channels with HFOs inside and outside the seizure onset zone (SOZ). We assessed the specificity, sensitivity and accuracy of ripples, FRs and spikes to determine the SOZ and their relation with post-operative outcome. We considered Engel class I and II as good outcome and Engel class III and IV as poor outcome. Results: The SOZ was temporal (71%), occipital (18%) or frontal (6%) in the 16 patients where it was defined. Seizures originated from more than one distinctive area in seven patients (41%). Thirteen patients (76.5%) underwent surgery: 6 (46%) had a good and 7 a poor outcome. The mean rate of spiking was higher in the SOZ channels than in the non-SOZ channels (16.03 10.50/min vs. 4.11 3.33/min; p=0.0002); similarly for ripples (43.4 32.7/min vs. 10.8 11.6/min; p=0.0016) and FRs (10.2 11.0/min vs. 2.0 3.5/min, p=0.0047). The sensitivity, specificity and accuracy to identify the SOZ were for spikes 91%, 29% and 44%; for ripples 91%, 42% and 54%; and for FRs 64%, 80% and 76%. We found no correlation between the number of channels and rates of ripples and FRs inside and outside the SOZ and postoperative outcome. In the seven patients with more than one SOZ, 5 with bitemporal lobe epilepsy showed a clear preponderance of one generator. They underwent a resection on the side of the main SOZ. We analyzed the relationship between the relative rates of ripples and FRs in the secondary SOZ and post-operative outcome. When relative rates of HFOs in the non-resected SOZ were high, the post-operative outcome was poor. Conclusions: Analysis of interictal HFOs during 5-10 min of sleep recording is a good tool to localize the SOZ in patients with epilepsy and normal MRI, given that the rates of HFOs were significantly higher in SOZ than in non-SOZ channels. We did not find any correlation between the post-operative outcome and the rates of HFOs. Interestingly, in patients with more than one identified SOZ when the relative values of HFOs were high in the non-resected SOZ, the post-operative outcome was poor. This suggests a participation of these secondary SOZs to the continuation of seizures after surgery. Supported by CIHR MOP-10189