Abstracts

Rationale:
The Piriform Cortex (PC), which surrounds the banks of the Entorhinal Sulcus, has been shown by multiple studies to play an important role in the pathophysiology of Temporal Lobe Epilepsy (TLE). A recent study reported that resection of more than half of the PC resulted in 16-fold increase odds of becoming seizure-free.1 Hence, we wanted to to assess the feasibility of manual segmentation of the PC and application of the Geodesic Information Flows (GIF) algorithm for automated segmentation, to guide surgical resection of the PC.
Method:
Two blinded independent examiners manually segmented the PC in 60 patients with TLE (55% Left TLE, 52% Women) with a median age of 35 years (IQR, 29-47 years) and 20 controls (60% Women) with a median age of 39.5 years (IQR,31-49) by using ITK-SNAP (Version 3.8.0-beta; www.itksnap.org).2 Both examiners were blinded to the side of disease and to each other’s segmentations. A total of 160 segmentations were done to calculate inter-rater variability and a second set of 160 segmentations for calculating intra-rater variability. All segmentations followed a rigorous, predefined segmentation protocol based on clear anatomical landmarks provided by previous research.1,3
Results:
Right PC was systematically larger in patients and controls and automatic parcellation was applied using the Geodesic Information Flows (GIF) algorithm for automatic segmentation to guide surgical resection. This automatic algorithm resulted in the creation of two independent masks. The first, delineating the frontal portion of the PC located in the superior bank of the Entorhinal Sulcus, which should not be resected due to the close proximity of the Lenticulostriate Arteries. Second, the temporal portion of the PC extending towards the Gyrus Semilunaris, representing the area that should be resected. This parcellation of the PC along with tractography imaging of the optic radiation and inferior frontal occipital fasciculus was used to guide anterior temporal lobe resection in a patient with right hippocampal sclerosis and right language dominance using interventional MRI (Figure 1 and Figure 2).  Postoperatively there was complete seizure freedom and no visual field or language deficit.
Conclusion:
We developed a pipeline to segment the PC in individual patients using GIF, and to present this data to the operating surgeon through the operative microscope during resection in the intraoperative MRI (iMRI) suite. Therefore, reliable segmentation of PC is feasible and can be applied prospectively and used to guide neurosurgical resection that increases the chances of a good outcome from temporal lobe resection for TLE. References: •Galovic M, Baudracco I, et al. Association of piriform cortex resection with surgical outcomes in patients with temporal lobe epilepsy. JAMA Neurol. 2019;76(6):690-700. doi:10.1001/jamaneurol.2019.0204 •Yushkevich PA, Piven J, Hazlett HC, et al. User-guided 3D active contour segmentation of anatomical structures: Significantly improved efficiency and reliability. Neuroimage. 2006;31(3):1116-1128. doi:10.1016/j.neuroimage.2006.01.015 •Pereira PMG, Insausti R, Artacho-Pérula E, Salmenperä T, Kälviäinen R, Pitkänen A. MR volumetric analysis of the piriform cortex and cortical amygdala in drug-refractory temporal lobe epilepsy. Am J Neuroradiol. 2005;26(2):319-332.
Funding:
:N/A