Abstracts

Rationale: Gamma Knife radiosurgery (RS) may be an alternative to open surgery for mesial temporal lobe epilepsy (MTLE), but the mechanisms of the anticonvulsant lesion of RS are unclear. Some data suggest a necrotic lesion at the target; others suggest neuromodulatory effects. The symptoms of auras may reflect the extent and localization of the developing antiepileptic lesion.Methods: This prospective trial of RS enrolled patients with unilateral hippocampal sclerosis and concordant video-EEG findings. Number and symptoms of auras and seizures were obtained for 3 months baseline and then every 3 months after RS until month 36. Auras were classified according to previously validated categories by localization: mesial (rising epigastric, olfactory/gustatory, experiential, and fear), nonmesial (somatosensory (tingling/vibration), visual, auditory), and other (indescribable, dizziness, vertigo). T2 volumes of the RS lesion were calculated from MRI obtained at 12 months.Results: 26 of 30 enrolled subjects completed the trial. At baseline, 22 (76%) of patients had auras. The majority (69%) were mesial. The mean rate of auras increased nearly 3-fold from preoperative baseline to a peak between 9-12 months. At peak, the proportion of patients with auras did not change significantly (McNemar s test p value = 0.62); 4 patients lost auras and 2 acquired new auras. At peak, mesial auras comprised the greatest proportion (79%) of type. The rate of aura occurrence correlated with the volume of T2 hyperintensity during the time of peak aura rates at 9-12 months postoperatively (Spearman s rho = .495, p=0.007). Between months 15-36, the mean rate of auras declined coinciding with an inversion in the proportion of mesial (43%) versus nonmesial (57%) auras, along with a precipitious drop in the rate of auras and complex partial seizures. At 36m, the proportion of patients with auras declined to 54% (not statistically significant - McNemar s test p value = 0.27). 4 of the 7 patients who were aura-free before surgery experienced new auras by 36 months; in three of the four these new auras were nonmesial. The proportion of individuals with mesial auras decreased from baseline (70%) to 36 months (30%, McNemar s p = 0.02%). Those patients who experienced seizure remission at 36 months had no significant differences in the presence (14/23. 61%) or absence (6/7, 86%) of auras during the preoperative baseline (Fisher s test p=0.37). Whether baseline auras were mesial or not did not affect subsequent outcome (p=0.99). Conclusions: A transient increase and a subsequent decrease in aura rate, a simultaneous decline in complex partial seizures, and a shift in aura category from mesial to nonmesial signaled development and resolution of the subacute RS lesion. We infer that that the radiosurgical damage to mesial structures caused transient excitation of the pathways responsible for generation of mesial seizures while impairing propagation.