Abstracts

Rationale: Surgical management of epilepsy of frontal origin carries unique difficulties, associated with the ill-defined frontal seizure semiology, the intrinsic frontal lobe epileptogenic synchronicity, and the large frontal lobe volume. Surgical outcome of frontal lobe epilepsy is worse than the one of temporal epilepsy. Identification of any factors affecting outcome of frontal epilepsy surgery is of paramount importance for improving outcome. Methods: Between 1981 and 2006, 79 patients undergoing 81 resections as treatment for medically refractory frontal epilepsy were included in our retrospective study. Variables studied in 31cases with normal imaging studies were (1) spatial concordance of non-invasive and invasive electrographic data, (2) invasive electrographic data alone, (3) age range at surgery, (4) seizure frequency, and (5) seizure duration. Variables studied in 50 cases with structurally abnormal imaging studies were (1) spatial concordance of non-invasive and invasive electrographic data, (2) invasive electrographic data alone, (3) age range at surgery, (4) seizure frequency, and (5) seizure duration. Extent of resection was also studied. Results: Of 12 non-lesional cases with concordance (C) or partial concordance (PC) of electrographic data, 8 (67%) were Class I outcomes. Of those 15 cases with non-concordance (NC) or discordance (DC) of electrographic data, 1 (6.7%) was a Class I outcome. Of 10 cases with subdural /- depth electrode localization, 5 (50%) were Class I. Of 17 with depth electrode localization, 4 (23.5%) were Class I. Of 9 cases ? 18 y.o., 3 (33%) were Class I. Of 21 cases > 18 y.o., 6(28.5%) were Class I. Median seizure duration was 12 yrs in seizure free cases and 18 yrs in non-seizure free cases (Table 1). Median seizure frequency was 30/mo in both seizure free and non-seizure free cases (Table 2). There were 34 cases with non-tumor structural lesions and 16 tumor cases. SD /- depth monitoring was done in 11 cases and depth monitoring alone was done in 2. There was insufficient concordance data for analysis. There was also insufficient invasive monitoring data for analysis (ictal localization in only 8 of 13 cases). Of the non-tumor lesional cases, 5 (71.4%) of 7 ? 18 y.o. were Class I and 10 (41.7%) of 24 > 18 y.o. were Class I. Of 12 tumor cases 9 (75%) were Class I (only 3 were ? 18 yo). Of 22 complete resections in non-tumoral cases 13 (59%) were Class I (2 incomplete resections with follow-up data). Three (60%) of 5 complete tumor resections were Class I and 6 (85.7%) of 7 incomplete resections were Class I. Conclusions: Concordance of non-invasive and invasive electrographic data was the most important determinant of outcome in non-lesional cases. Completeness of lesion resection in this series did not yield significantly better outcome. Younger patients were more likely to have Class I outcomes in non-tumor lesional cases. Tumor cases were more likely to have Class I outcomes compared to non-tumor lesional cases. Shorter seizure duration favored Class I outcomes in the non-lesional and non-tumor lesional cases.