Abstracts

Rationale:
Coagulating much of the epileptogenic lesion (EL)’s cortex is a favorable prognostic factor for radiofrequency thermocoagulation (RFTC) for focal epilepsy. However, trajectory limitations exist due to the risk of brain tissue and vascular injury around coagulated lesions. Therefore, it is difficult to coagulate the entire focus when many lesions cannot be created or the focus extends over the eloquent area. We propose an RFTC focal disconnection technique that minimizes the number of lesions and trajectories through stereoelectroencephalography (SEEG) monitoring and cortical stimulation mapping (CSM) to estimate the abnormal network consisting of epileptogenic zones and propagation patterns, and intensively coagulating its anomaly network.

Methods:
A 17-year-old right-handed woman with drug-resistant epilepsy. Her habitual seizure was a focal tonic-clonic seizure in the right upper limb, occurring several times weekly, following an aura of discomfort in the right side of the abdomen, without impaired awareness. MRI showed EL in the left cingulate gyrus (CgG), paracingulate gyrus (PaCgG), and posterior part of the supplementary motor area (SMAp), suggesting focal cortical dysplasia (FCD) (Figure 1A). SEEG targeting EL and additional areas (detailed in Figure 1B, C) recorded the habitual seizures. All seizure onsets were observed in CgG and PaCgG, followed by SMAp and then PrCG hand, with no propagation to the anterior and posterior area of the lesion (SMAa, ACC, PCC) (Figure 1C). CSM using SEEG electrodes induced right upper limb discomfort and tonic seizures in the right hand with SMAp stimulation. High-frequency stimulation of PaCgG induced a weakness attack in the right upper limb and low-frequency stimulation induced discomfort in the right side of the abdomen (Figure 1B). Based on these results, the habitual seizure’s main pathway was considered as a network starting from the MRI lesions of CgG and PaCgG to PrCG hand. Therefore, coagulated lesions were created to disconnect the network between CgG (PaCgG) and PrCG hand. The cortical area and network of CgG (PaCgG)-ACC and PCC and SMAa were intentionally spared because they were not involved in seizure propagation (Figure 2A).

Results:
RFTC was performed under local anesthesia with neurological symptom monitoring. Test stimulation of the EL’s cortex induced habitual seizures and SMA symptoms, confirmed by CSM. Thirteen coagulation lesions were created and disconnected in six trajectories per the preoperative plan (Figure 2A). The EL’s cortex in this case was 2.13 cm³, and the coagulated lesion 0.79 cm³, much less than the coagulated volume required for conventional RFTC. Postoperative MRI showed no complications (Figure 2B). Postoperative assessment showed no higher brain function deficits. The patient remains seizure-free 18 months postoperatively.

Conclusions:

A good seizure outcome was obtained by SEEG identification and disconnection of the EL network in the eloquent area. This method suggests that improvements in the identification of epileptogenic networks may lead to a new strategy for epilepsy surgery: network disconnection.

Funding: JSPS KAKENHI [Grant Number 23K15663] (principal investigator; Tomotaka Ishizaki)