Correlation of FDG-PET hypometabolism and SEEG epileptogenicity mapping in patients with drug-resistant focal epilepsy

Abstract

Objective: Interictal [18F]fluorodeoxyglucose-positron emission tomography (FDG-PET) is used in the presurgical evaluation of patients with drug-resistant focal epilepsy. We aimed at clarifying its relationships with ictal high-frequency oscillations (iHFOs) shown to be a relevant marker of the seizure-onset zone.

Methods: We studied the correlation between FDG-PET and epileptogenicity maps in an unselected series of 37 successive patients having been explored with stereo-electroencephalography (SEEG).

Results: At the group level, we found a significant correlation between iHFOs and FDG-PET interictal hypometabolism only in cases of temporal lobe epilepsy. This correlation was found with HFOs, and the same comparison between FDG-PET and ictal SEEG power of lower frequencies during the same epochs did not show the same significance.

Significance: This finding suggests that interictal FDG-PET and ictal HFOs may share common underlying pathophysiologic mechanisms of ictogenesis in temporal lobe epilepsy, and combining both features may help to identify the seizure-onset zone.

Keywords: SEEG; Epileptogenicity map; FDG-PET; Focal epilepsy; Ictal HFO.

Figure 1

Examples of single bipolar SEEG recordings in 10 patients (left: TLE; right: extra‐TLE) showing the diversity of seizure‐onset dynamics (time = 0 s). Visual definition of seizure onset was based on the careful inspection of all SEEG channels simultaneously (not shown here) and the presence of fast oscillations and/or repetitive spikes followed or concomitant to fast activity.

Figure 2

Epileptogenicity mapping of patient 14 (see Table 1). (A) Bipolar SEEG recordings of a seizure in (1) amygdala, (2) anterior parahippocampal gyrus, (3) posterolateral temporal gyrus (T3), (4) anterior hippocampus, and (5) posterior hippocampus. (B) Time‐frequency transform of SEEG power of the signals shown in (A). (C) Positions of electrodes in patient's MRI mesh of cortical surface. Numbers indicate locations of electrodes with signals shown in (A–B). (D) Epileptogenicity map in 80–160 Hz frequency band. L, left; R, right.

Figure 3

ROI‐based correlation of FDG‐PET and ictal SEEG power in patient 14 (see Table 1). (A) FDG‐PET map normalized in the MNI space, showing hypometabolism in the left anterior temporal lobe. L, left; R, right. (B) ROI of the Destrieux atlas in the MNI space. (C) ROI‐based correlation of FDG‐PET and ictal SEEG power in different frequency bands. Each dot represents for each ROI the values of ictal SEEG power (t‐value of the comparison of ictal vs. baseline period) and cerebellum‐normalized FDG‐PET. The lines show the linear regression on the data distribution for each frequency band (significance of correlation: p_2–7 = 0.3738; p_8–20 = 0.7324; and p_80–160 = 0.0263).

Figure 4

ROI‐based correlation of FDG‐PET and ictal SEEG power for each patient (see Table 1) in the different frequency bands.