Abstracts

Rationale: Laser interstitial thermal ablation (LITT) is becoming widely used as a minimally invasive alternative to classical antero-mesial temporal resections in patients with mesial temporal lobe epilepsy (MTLE). A major theoretical advantage of this approach is that it could result in fewer cognitive deficits, especially when treating MTLE in the language dominant hemisphere. We evaluated ablation volumes in subregions of the mesial temporal lobe and their relationship with changes in cognitive outcomes, in 43 consecutive patients undergoing selective laser ablations of amygdala and hippocampus. Methods: All patients underwent formal pre and 6 month post-operative neuropsychological assessment using the Boston Naming Test (BNT) and 4 tests from Wechsler Memory Scale-III (Logical Memory (LM) I & II Recall, Visual Reproduction (VR) I & II Recall). 15 out of 26 patients for whom we measured ablation volumes, underwent SEEG localization prior to LITT. The pre-ablation MRI was processed in FreeSurfer to obtain cortical and subcortical segmentation of the brain, including hippocampal subfields. Ablation volumes were outlined using a post-ablation T1 MRI with contrast registered with the pre-ablation MRI. The percentage of amygdalar, hippocampal and entorhinal subfields that were included in the ablation volume, were computed. These volumetric measures were regressed against changes in performance on memory tests post-ablation. Surgical outcome was ascertained in all cases following LITT at most recent f/u. Results: At a mean follow-up interval of 25.4 months (± 15.2), an Engel IA outcome was obtained in 72% (n= 43), while 16% of patients had Engel II-IV. The distribution of surgical outcome was similar for patients who underwent LITT with or without prior SEEG. No surgery-related complications were noted. In the 26 patients for whom we quantified ablation volumes, a median of 73.4% ± 19.6 of amygdala, 71.4± 16.8 of hippocampus and 38.1± 20.6 of entorhinal cortex was ablated. No significant changes were seen in the BNT at 6 months, only the LM I memory test showed a significant deficit (p 2=0.5,p A=-0.19,βHc=-0.14, p Hc=0.12,p CA2/3=0.04), while the ablation of CA1 correlated with VR I  (βCA1=0.22) and VR II scores (βCA1=0.23), p Conclusions: Laser ablation is a good alternative to traditional amygdalo-hippocampectomy, in carefully selected cases, as it provides good seizure outcomes while being significantly less invasive. Our results confirm that the CA1 plays a role in visual delayed recall. The deficit analysis on our complete LITT dataset (43 patients) will yield important insights into the relative contributions of mesial temporal structures to naming and to verbal memory. Funding: N/A