High frequency oscillations in epilepsy surgery

Abstract

Epilepsy surgery is the sole treatment to cure patients with focal epilepsy. The chances on seizure freedom after surgery is determined by how well we can pinpoint the epileptic focus and delineate the margins of the disease tissue for resection. Intra-operative invasive EEG, measuring brain signals directly from the cortex, can help to exactly localize the epileptic tissue to guide the surgeon. This is called tailored surgery. Recording the EEG at a high sampling rate (>2000 Hz) enables identification of newly discovered high frequency oscillations (HFOs, >80-500 Hz) that seem more precise biomarkers of epileptic tissue than the currently usedspikes (<80 Hz). The aim of this thesis was to lay a foundation for the clinical use of HFOs in the intra-operative EEG for tailored surgery to increase the chance of surgical success. As a first step we studied retrospectively how to interpret these HFOs, subdivided into ripples (80 -250 Hz) and fast ripples (250-500 Hz), measured in the intra-operative EEG before and after surgery. We compared them to the gold standard, spikes. We found that incomplete resection of spikes, ripples or fast ripples in pre-resection EEG alone does not predict the persistence of seizures. Surgical manipulation did not cause new HFOs on the edge of the resection, which an undesired behavior reported for spikes. Only fast ripples that remained in the post-resection EEG predicted seizure recurrence. The second part of the thesis describes “The HFO trial”; a Dutch single-blinded multi-center RCT to determine prospectively if intra-operative ECoG tailored surgery using HFOs, instead of inter-ictal spikes, is feasible and will lead to an equal or better post-operative seizure outcome (non-inferiority). Approval by the medical ethical board of the University Medical Center Utrecht and start of this interventional study was established in 2014. In total 78 participants will be recruited and results are expected in 2018. Third, a key issue for worldwide clinical acceptance of HFOs for tailored surgery is simple and automated HFO analysis instead of tedious visual analysis. We explored methods to optimize the yield of HFO detection procedures. A promising method is single pulse electrical stimulation (SPES), in which brief electrical pulse are delivered to the brain that evoke pathological delayed spike-like responses. These responses contain evoked HFOs that are specific for the seizure origin. SPES could be a tool to get on-demand HFOs to delineate the epileptic focus during surgery. Second, we found that non-harmonicity in the EEG signal can identify electrodes covering epileptogenic tissue, without the need for detecting each individual HFO. Intra-operative tailoring based on HFOs, and fast ripples in particular, next to spikes could increase seizure freedom after surgery. Our findings highlight the importance of a final post-resection recording, as this provides the best prognosis for surgical success. The use of HFOs proofs to be feasible in clinical practice, although caution is advised with respect to non-pathological HFOs that can occur. The presented innovative signal analytical approaches for HFOs detection in the intra-operative EEG strengthen the basis of clinical use of HFOs for epilepsy surgery.