By Pegah Afra, MD
Associate Professor of Neurology, Weill Cornell Medical College; Associate Attending Neurologist, NewYork-Presbyterian Hospital; Adjunct Associate Professor, University of Utah
Dr. Afra reports she is on the speakers bureau and is a consultant for UCB and Sunovion, and is a consultant for GW Pharmaceuticals and SK-Pharma.
Epilepsy surgery is one of the treatment modalities for intractable pharmacoresistent epilepsy. For seizure freedom to be achieved, the epileptogenic zone (EZ) must be resected successfully. Stereoelectroencephalography is one of the invasive intracranial EEG recording techniques used for EZ localization. The technological advancements in the past two decades have made this procedure safer, more accurate, and easier to perform and therefore more readily available.
Cardinale F, Rizzi M, Vignati E, et al. Stereoelectroencephalography: Retrospective analysis of 742 procedures in a single centre. Brain 2019;142;2688-2704.
This article is the largest retrospective study of stereoelectroencephalography (SEEG) series ever reported. It describes a 20-year history of SEEG activity of a single epilepsy center: the “Claudio Munari” center for epilepsy surgery in Milan, Italy. Cardinale et al provided a recent and current review of indications, efficacy, results, outcome predictors, and complications of SEEG.
From a prospectively maintained database, the authors selected three different subsets of patients (SEEG 713, SEEG 630, and SURG 1851) and retrospectively analyzed them in multiple different subsets as below:
1. SEEG 713: A group of 713 patients (400 male/313 female; mean age 26.2 years and 25.9% younger than 18 years) who had 740 SEEGs between May 1996 and July 2018.1 The dataset was analyzed retrospectively to determine efficacy and morbidity of SEEG. Overall, 93.6% of patients underwent scalp video-EEG (prior to SEEG) and the remainder had seizures captured during routine EEG or had an MRI-visible lesion. The total number of implanted electrodes in these 742 SEEG procedures were 10,084. The mean number of implanted electrodes per subject was 13.59 (range 3-22) with less than six electrodes in only six procedures. The implantation was unilateral in 84% (48% on the right, 36% on the left) and bilateral in 16% (asymmetric in 12%, symmetric in 4%). Following SEEG, 80% of patients were selected for surgery; 15% were not candidates (because the epileptogenic zone [EZ] was not well-defined, was multifocal, or included eloquent cortex), and 5% refused the final treatment option or were not operated on because of SEEG-related complications. The complication rate was 1.8% including 0.5% major events (four patients: one death from massive brain edema/hyponatremia, two intracranial bleedings resulting in permanent contralateral hemiplegia, and one compartment leg syndrome resulting in permanent deficit of foot dorsiflexion).
2. SEEG 630: A group of 630 patients who underwent SEEG up to July 2016 and were analyzed retrospectively to look for predictors of seizure outcome using logistic regression models.2 For the purpose of analysis, a “favorable” outcome was defined as the patient being free of disabling seizures (corresponding to ILAE class 1-2 and Engel classes Ia-Ic). An “unfavorable” outcome was defined as not free of disabling seizures (corresponding to ILAE classes 3-6 and Engel classes II-IV). The data from 630 patients were analyzed in three subsets below:
A) SURG 470: A subset of 470 patients who underwent SEEG-guided resective/disconnection surgery: 59.4% had favorable outcome with 48.9% ILAE class I outcome (4/5 Engel Ia and 1/5 Engel Ic). The remaining 40.6% had disabling seizures. Unfavorable outcome predictors were negative MRI and remnant lesion increasing the risk of seizure recurrence by 1/3. Favorable outcome predictors included type II focal cortical dysplasia (FCD), balloon cells, glioneuronal tumors (GNTs), mesial temporal sclerosis (MTS), and periventricular nodular heterotopia (PNH), decreasing the risk of seizure recurrence by half. Additionally, older age at epilepsy onset reduced the risk of unfavorable outcome.
B) SURG 278: A subset of 278 patients who underwent SEEG-guided resective/disconnection surgery with at least 10 years of follow-up. Unfavorable outcome predictors were negative MRI and remnant lesion, while type II FCD predicted favorable outcome.
C) RF-THC 153: A subset of 153 patients who underwent SEEG-guided radiofrequency thermal coagulation. For the purpose of analysis of this subset, optimal responders were defined as patients who had favorable outcomes or were not keen to undergo resective/disconnective surgery because they were satisfied with clinical results (regardless of outcome class). In these 153 patients, 83.7% had unsatisfactory results and 16.3% were optimal responders (4/5 in ILAE class 1-2 [i.e., favorable outcome] and 1/5 in ILAE class 3). A higher seizure rate predicted unsatisfactory outcome, and PNH at MRI was significantly associated with better outcome. Male gender resulted in better outcome, with trends toward significance. There was 2% incidence of major events/complications.
3. SURG 1851: A group of 1,851 patients who underwent resective/disconnective surgery up to July 2018. This analysis was done to report on SEEG indications. Overall, the results showed that throughout the years, the bulk of MRI-negative SEEG-studied subjects progressively increased while the proportion of subjects with common causes (MTS, GNTs, and FCD) decreased. Regardless, the proportion of seizure-free patients did not decrease throughout the years. Additionally, the proportion of SEEG-studied patients who underwent resective surgery in Rolandic-perirolandic regions progressively decreased.
A) SURG 1598: A subset of 1,598 patients who underwent resective/disconnective surgery and had up to 24 months of follow-up. This subset was analyzed to compare seizure outcomes between subjects who were studied by SEEG and subjects who were not studied by SEEG (by means of contingency table). After surgery, disabling seizures recurred in 40.6% of patients who underwent SEEG and in 18.4% of patients who did not have SEEG.
B) SURG 767: A subset of patients who underwent resective/disconnective surgery with at least 10 years of follow-up. The subset was analyzed to determine if the chance of being free of disabling seizures was different between the subjects who were or were not studied with SEEG. The results showed that the recurrence of disabling seizures (unfavorable outcome; classes 3-6) was more than double for patients who underwent SEEG.
COMMENTARY
The SEEG data confirm that indications for SEEG have not changed in more than 50 years. In summary, SEEG (as well as any other invasive intracranial EEG) is indicated when analysis of anatomico-electro-clinical correlations obtained by noninvasive methods are insufficiently concordant or inconclusive and, therefore, do not support a unified hypothesis on the location of the EZ and surgical proposal or when EC is involved early. Of the 742 SEEG procedures carried out on 713 patients (with total implantation of 10,084 electrodes), there were four clinically relevant intracranial bleeds, making an implantation-related intracranial bleeding rate of 0.04% per electrode. The overall complication rate for 742 procedures was 1.8%, including 0.5% major events (death or permanent neurological sequel). All of these data point to the fact that SEEG is a safe and well-tolerated neurosurgical procedure.
After SEEG evaluation, resective surgery was indicated in 79.9% of the patients. Of the patients who underwent surgery, 59.4% were free of disabling seizures at least two years after surgery. This suggests that when noninvasive localization methods (clinical data, scalp video EEG, and neuroimaging) are insufficient for localization of EZ, SEEG is an effective method in further evaluating patients with intractable epilepsy for the purpose of resective surgery. There was a progressive increase of MRI-negative cases throughout the years; however, the proportion of seizure-free patients did not decrease throughout the years. This is a subtle suggestion that technological advancements may have had some positive effect on the surgical outcomes by improving accuracy and therefore outcome.
In the presented data, MRI-negative and post-surgical lesion remnant were significant risk factors for seizure recurrence, while type II FCD, balloon cells, GNTs, MTS, older age at epilepsy onset, and PNH were significantly associated with seizure freedom. This makes absolute sense, since SEEG is only a diagnostic modality, and the ultimate outcome of epilepsy surgery will depend on the population that is being treated. For example, patients with MRI-negative lesions are more likely to have hard to localize EZ and require bilateral electrodes, while patients with lesions in the EC are more likely to undergo partial resection. Therefore, it only makes sense that these were predictors of unfavorable outcome. On the other hand, patients with well-defined lesions, like type II FCD, balloon cells, GNTs, and MTS, are less likely to have bilateral or extensive EZ; therefore, this pathology correlates with favorable surgical outcome.
REFERENCES
- Cardinale F, Casaceli G, Raneri F, et al. Implantation of stereoelectroencephalography electrodes: A systematic review. J Clin Neurophysiol 2016;33:490-502.
- Minotti L, Montavont A, Scholly J, et al. Indications and limits of stereoelectroencephalography (SEEG). Neurophysiol Clin 2018;48:15-24.
