Interictal blood-brain barrier dysfunction in piriform cortex of people with epilepsy

Abstract

Objective: The piriform cortex is considered to be highly epileptogenic. Its resection during epilepsy surgery is a predictor for postoperative seizure freedom in temporal lobe epilepsy. Epilepsy is associated with a dysfunction of the blood-brain barrier. We investigated blood-brain barrier dysfunction in the piriform cortex of people with temporal lobe epilepsy using quantitative T1-relaxometry.

Methods: Gadolinium-based contrast agent was administered ictally and interictally in 37 individuals before undergoing quantitative T1-relaxometry. Postictal and interictal images were co-registered, and subtraction maps were created as biomarkers for peri-ictal (∆qT1_{interictal-postictal}) and interictal (∆qT1_{noncontrast-interictal}) blood-brain barrier dysfunction. Values were extracted for the piriform cortex, hippocampus, amygdala, and the whole cortex.

Results: In temporal lobe epilepsy (n = 14), ∆qT1_{noncontrast-interictal} was significantly higher in the piriform cortex than in the whole cortex (p = 0.02). In extratemporal lobe epilepsy (n = 23), ∆qT1_{noncontrast-interictal} was higher in the hippocampus than in the whole cortex (p = 0.05). Across all individuals (n = 37), duration of epilepsy was correlated with ∆qT1_{noncontrast-interictal} (ß = 0.001, p < 0.001) in all regions, while the association was strongest in the piriform cortex. Impaired verbal memory was associated with ∆qT1_{noncontrast-interictal} only in the piriform cortex (p = 0.04). ∆qT1_{interictal-postictal} did not show differences in any region.

Interpretation: Interictal blood-brain barrier dysfunction occurs in the piriform cortex in temporal lobe epilepsy. This dysfunction is linked to longer disease duration and worse cognitive deficits, emphasizing the central role of the piriform cortex in the epileptogenic network of temporal lobe epilepsy.

Figure 1

(A) Rotation of the nonenhanced T1‐weighted MRI relative to the anterior commissure by +20°. (B) Sagittal, coronal, and axial view of the piriform masks of a representative subject with epilepsy on a structural T1‐weighted MRI. Color scheme: Blue = frontal PIC; pink = temporal PIC.

Figure 2

(A) & (B) ∆qT1_ni in cortex, hippocampus, amygdala, and PIC of people with TLE and ETLE, respectively. (C) & (D) ∆qT1_ip in cortex, hippocampus, amygdala, and PIC of people with TLE and ETLE, respectively. This figure indicates that interictal blood–brain barrier dysfunction occurs in the piriform cortex of people with temporal lobe epilepsy and in the hippocampus of people with temporal and extratemporal lobe epilepsy (please note the borderline‐significant p‐values for the hippocampus). Peri‐ictal blood–brain barrier dysfunction is not found. ETLE, extratemporal lobe epilepsy; PIC, piriform cortex; TLE, temporal lobe epilepsy. p‐Values <0.1 are shown.

Figure 3

(A) ∆qT1_ni in several regions of interest (cortex, hippocampus, amygdala, and PIC) plotted against the duration of epilepsy (in years). (B) ∆qT1_ni of people with impaired verbal memory scores in several left hemisphere regions of interest (cortex, hippocampus, amygdala, and PIC). This figure illustrates how duration of epilepsy is correlated with interictal blood–brain barrier in all regions across all individuals, while the association is strongest in the piriform cortex. Also, it demonstrates that people with epilepsy and impaired verbal memory only show interictal blood–brain barrier dysfunction in the piriform cortex. For the purpose of visualization, a smaller scale of the y‐axis in (A) is displayed. Memory impairment was assumed for scores less than one standard deviation below the mean of the normative sample (<90). PIC, piriform cortex. p Values <0.1 are shown. The brains above the diagrams indicate the region under investigation.