Abstracts

Rationale: Resective surgery is an important consideration for patients with treatment resistant epilepsy as it may offer the best chance for seizure freedom. Initial patient counseling and the decision to refer a patient to an epilepsy surgery center is usually based on the history and physical exam, routine electroencephalogram (EEG), and brain MRI. The aim of this review is to summarize the literature and identify factors typically available in the general office setting that are reliable predictors of epilepsy surgery outcome. Methods: A literature search was performed using Pubmed and Embase. Inclusion criteria included: English language, sample size ?20 patients, MRI performed on ?90% of patients, a median age ?16 years, average of ?1 year follow-up, and predictive value assessment of clinical factors, routine EEG and/or MRI brain for outcome in epilepsy surgical resections. Articles were independently reviewed by at least 2 authors and data on study design and predictive factors were abstracted. Results: Of the 2,248 articles related to predictors of epilepsy surgical outcome identified, 123 met all inclusion criteria. Only 12 articles had a prospective study design. Study populations varied in size and epilepsy characteristics with the majority focusing on mesial temporal lobe epilepsy or lesional epilepsies. The studies focused almost exclusively on patients that had undergone resection rather than all patients considered for surgery or undergoing invasive procedures (intention-to-treat analysis). Predictive factors were not uniformly assessed in most studies, nor were they uniformly defined across studies. For example, in various studies, unilateral routine EEG epileptiform activity was variably categorized as "only ipsilateral spikes", ">70% ipsilateral", ">80% ipsilateral", or ">90% ipsilateral". Although many studies used the Engel or modified Engel classification systems, many utilized non-standardized outcome determinations (e.g. "good"). Only 6 studies had a masked assessment of seizure outcome following surgery. Conclusions: The heterogeneous patient populations, methodologies and outcome determinations significantly limit the existing literature's ability to predict the likelihood of a patient achieving seizure freedom from resective surgery based upon pre-operative data, especially in patients with extra-temporal non-mesial epilepsy. Prospective multicenter studies based upon intention to treat (i.e. enrolling patients prior to invasive procedures to determine the number of patients that are considered for epilepsy surgery that do not progress to a resective procedure) are necessary to better facilitate and encourage early referral to comprehensive epilepsy centers, counsel patients and families and identify new strategies for improving outcomes. By establishing and utilizing accepted, standardized definitions (e.g. treatment resistant epilepsy), one can improve the generalizability of findings across patients and centers.