No link between piriform cortex subregion resection and seizure freedom in two cohorts with temporal lobe epilepsy

Postoperative seizure freedom in temporal lobe epilepsy (TLE) has been linked to the extent of piriform cortex resection. Earlier work used manual piriform cortex segmentations truncated at the limen insulae and overlapping the amygdala. We re-evaluated this association using an investigator-independent, connectivity-based segmentation.

Patients with TLE who underwent mesiotemporal epilepsy surgery at a single center with > 1-year postoperative follow-up were included retrospectively (discovery cohort). Postoperative images and connectivity-defined piriform cortex subregions were registered to preoperative T1-weighted scans. Additional manual piriform cortex segmentation was performed for comparison. The extent of resection within temporal lobe subregions was compared between seizure-free (ILAE I) and non-seizure-free patients. Results were validated using a large independent cohort. For additional validation, voxel-based lesion-outcome mapping was performed using logistic regression.

Twenty-eight patients (15 female) with TLE of heterogeneous etiologies were included in the discovery cohort. No significant association between ILAE outcome and the extent of resection within manually or automatically defined piriform cortex was observed. Extensive analyses in the validation cohort (n = 305, 205 with hippocampal sclerosis) confirmed this result. Additional voxel-based lesion-outcome mapping (n = 305) showed no significant associations with ILAE I outcome. Resected proportions within other mesiotemporal regions were not associated with ILAE I outcome.

No association between ILAE outcome and the resected proportion of specific mesiotemporal structures was detected. This result was robust across cohorts, piriform cortex definitions and analytical approaches. These results differ from previous reports and suggest that piriform cortex resection should be considered on an individualized basis rather than routinely incorporated into temporal lobe resections.