Volumetric analysis of the piriform cortex in temporal lobe epilepsy

Abstract

The piriform cortex, at the confluence of the temporal and frontal lobes, generates seizures in response to chemical convulsants and electrical stimulation. Resection of more than 50% of the piriform cortex in anterior temporal lobe resection for refractory temporal lobe epilepsy (TLE) was associated with a 16-fold higher chance of seizure freedom. The objectives of the current study were to implement a robust protocol to measure piriform cortex volumes and to quantify the correlation of these volumes with clinical characteristics of TLE. Sixty individuals with unilateral TLE (33 left) and 20 healthy controls had volumetric analysis of left and right piriform cortex and hippocampi. A protocol for segmenting and measuring the volumes of the piriform cortices was implemented, with good inter-rater and test-retest reliability. The right piriform cortex volume was consistently larger than the left piriform cortex in both healthy controls and patients with TLE. In controls, the mean volume of the right piriform cortex was 17.7% larger than the left, and the right piriform cortex extended a mean of 6 mm (Range: -4 to 12) more anteriorly than the left. This asymmetry was also seen in left and right TLE. In TLE patients overall, the piriform cortices were not significantly smaller than in controls. Hippocampal sclerosis was associated with decreased ipsilateral and contralateral piriform cortex volumes. The piriform cortex volumes, both ipsilateral and contralateral to the epileptic temporal lobe, were smaller with a longer duration of epilepsy. There was no significant association between piriform cortex volumes and the frequency of focal seizures with impaired awareness or the number of anti-seizure medications taken. Implementation of robust segmentation will enable consistent neurosurgical resection in anterior temporal lobe surgery for refractory TLE..

Keywords: Hippocampus; Piriform cortex; Temporal lobe epilepsy.

Fig. 1

3-Dimensional reconstruction of the Piriform Cortex.

Fig. 2

A: Right (Green) and Left (Red) Segmentations of Piriform Cortex in Healthy Individual on MRI. B: Protocol for delineating the piriform cortex. Step 1. Slice 1: The most anterior slice that includes the piriform cortex is that in which the limen insulae, the white matter tracts connecting insula and temporal lobe is evident). This marks the start of the piriform cortex and the cortical amygdala (PCA). On this slice, the PCA is located at the confluence of the frontal lobe and the temporal stem where it contours the inferior and superior banks of the entorhinal sulcus forming a C-shape. The frontal part projects from the core of the entorhinal sulcus and extends till 50–75% of the distance to the olfactory tubercle (Vaughan et al., 2014; Galovic et al., 2019). In pre-slice 1, the red circle is the olfactory tubercle and in the green rectangle is the optic chiasm for reference. Anatomically, the olfactory tubercle is anterior to the chiasm which can be used as a reference to locate it on sagittal MRI imaging. Following that, it was easily identified on the coronal plane and marked on the first slice. The temporal part of the piriform cortex lines the lower half of the C-shape i.e., inferior bank of the entorhinal sulcus and covers 30% of the distance to the apex of gyrus semilunaris. The apex of the gyrus semilunaris is the most prominent medial point of the gyrus. The voxel size is 1 mm isotropic. The thickness of the PCA was kept to two voxels thick in the absence of grey-white border, otherwise, it followed the grey-white interface if this was evident. The curvature of the junction between the frontal and temporal parts of the PCA always showed a clear gray-white matter differentiation laterally, therefore the thickness of this area was dependent on this differentiation across all slices. In contrast, in the portions where the PCA joins the amygdala, gray-white differentiation disappears, and the 2-voxel thickness rule applies. Step 2. Slice 2: From here, the frontal section of the PCA was maintained as above, whilst the temporal part was extended to 50% of the distance to the apex of the gyrus semilunaris. Slice 3: The temporal part was extended to 75% of the distance to the apex of the gyrus semilunaris on the following slice. Step 3. Slice 4: In the following slice, the temporal part of the PCA was extended to cover the gyrus semilunaris entirely, therefore representing 100% of the distance from the core of the entorhinal sulcus to the apex. This usually corresponded with the appearance of the temporal horn. However, as brain asymmetry is common, this was not the case in every patient. Step 4. Slice 5: The temporal part of the PCA was then contoured around the entire length of the gyrus semilunaris and terminated at the sulcus semiannularis. The sulcus semiannularis (see arrow) is denoted by a subtle dip at the lower boundary of the gyrus semilunaris. The temporal extension of the PCA to the sulcus semiannularis continues until the end slice of the piriform cortex. If the sulcus semiannularis was not clearly visible, the white matter beneath the amygdala was used to indicate the tail end of the PCA. Step 5. Slice 10: As the entorhinal sulcus shortens and the optic tracts start to move more laterally and closer to the PCA, the frontal extension becomes shorter until it disappears. This is because, as explained before, the frontal part projects from the core of the entorhinal sulcus and extends till 50–75% of the distance to the olfactory tubercle, which is kept as a constant landmark. Step 6. Last Slice: The last slice is marked as the one slice before the complete fusion of the cerebral peduncle with the pons. The thickness of the PCA was kept to 1 voxel thick here.

Fig. 2

A: Right (Green) and Left (Red) Segmentations of Piriform Cortex in Healthy Individual on MRI. B: Protocol for delineating the piriform cortex. Step 1. Slice 1: The most anterior slice that includes the piriform cortex is that in which the limen insulae, the white matter tracts connecting insula and temporal lobe is evident). This marks the start of the piriform cortex and the cortical amygdala (PCA). On this slice, the PCA is located at the confluence of the frontal lobe and the temporal stem where it contours the inferior and superior banks of the entorhinal sulcus forming a C-shape. The frontal part projects from the core of the entorhinal sulcus and extends till 50–75% of the distance to the olfactory tubercle (Vaughan et al., 2014; Galovic et al., 2019). In pre-slice 1, the red circle is the olfactory tubercle and in the green rectangle is the optic chiasm for reference. Anatomically, the olfactory tubercle is anterior to the chiasm which can be used as a reference to locate it on sagittal MRI imaging. Following that, it was easily identified on the coronal plane and marked on the first slice. The temporal part of the piriform cortex lines the lower half of the C-shape i.e., inferior bank of the entorhinal sulcus and covers 30% of the distance to the apex of gyrus semilunaris. The apex of the gyrus semilunaris is the most prominent medial point of the gyrus. The voxel size is 1 mm isotropic. The thickness of the PCA was kept to two voxels thick in the absence of grey-white border, otherwise, it followed the grey-white interface if this was evident. The curvature of the junction between the frontal and temporal parts of the PCA always showed a clear gray-white matter differentiation laterally, therefore the thickness of this area was dependent on this differentiation across all slices. In contrast, in the portions where the PCA joins the amygdala, gray-white differentiation disappears, and the 2-voxel thickness rule applies. Step 2. Slice 2: From here, the frontal section of the PCA was maintained as above, whilst the temporal part was extended to 50% of the distance to the apex of the gyrus semilunaris. Slice 3: The temporal part was extended to 75% of the distance to the apex of the gyrus semilunaris on the following slice. Step 3. Slice 4: In the following slice, the temporal part of the PCA was extended to cover the gyrus semilunaris entirely, therefore representing 100% of the distance from the core of the entorhinal sulcus to the apex. This usually corresponded with the appearance of the temporal horn. However, as brain asymmetry is common, this was not the case in every patient. Step 4. Slice 5: The temporal part of the PCA was then contoured around the entire length of the gyrus semilunaris and terminated at the sulcus semiannularis. The sulcus semiannularis (see arrow) is denoted by a subtle dip at the lower boundary of the gyrus semilunaris. The temporal extension of the PCA to the sulcus semiannularis continues until the end slice of the piriform cortex. If the sulcus semiannularis was not clearly visible, the white matter beneath the amygdala was used to indicate the tail end of the PCA. Step 5. Slice 10: As the entorhinal sulcus shortens and the optic tracts start to move more laterally and closer to the PCA, the frontal extension becomes shorter until it disappears. This is because, as explained before, the frontal part projects from the core of the entorhinal sulcus and extends till 50–75% of the distance to the olfactory tubercle, which is kept as a constant landmark. Step 6. Last Slice: The last slice is marked as the one slice before the complete fusion of the cerebral peduncle with the pons. The thickness of the PCA was kept to 1 voxel thick here.

Fig. 3

Control Data. Fig. 3a: Correlation of control data (mm³) with total intracranial volume. Fig. 3b: Corr. (I) mm³ =Corrected Metric (I) mm³ adjusted for intracranial volume. Fig. 3c: Correlation of control data (mm³) with age. Fig. 3d: Corr. (II) mm³ =Corrected Metric (II) mm³ adjusted for age and intracranial volume. L=Left, R=Right, Vol=Volume.

Fig. 4

Correlation of subject data with duration of epilepsy. Fig. 4a, b: Corr. (II) Piriform Vol = Corrected (II) piriform cortex volume adjusted for age and intracranial volume. Fig. 4c, d: Corr. (I) Hippocampus Vol = Corrected (I) hippocampus volume adjusted for intracranial volume. Fig. 4e, f: Corr.(III) Piriform Vol =Corrected (III) piriform cortex volume adjusted for age, intracranial volume and asymmetry. Fig. 4g, h: Hippocampus Vol= Corrected (II) hippocampus volume adjusted for intracranial volume and asymmetry. L=Left, R =Right.