SEARCH

SEARCH BY CITATION

Keywords:

  • Cryptogenic focal epilepsy;
  • Double inversion recovery;
  • Fast FLAIR T2;
  • Magnetization transfer ratio;
  • Voxel-based morphometry

Summary: Purpose: Conventional optimal MRI is unremarkable in 20%–30% of patients with intractable focal epilepsy. These MRI-negative patients are the most challenging in surgical programs. Our aim was to evaluate the yield and utility of quantitative MRI with novel contrasts in MRI-negative patients with refractory focal epilepsy, who were potential surgical candidates.

Methods: Ninety-three consecutive potential surgical candidates with refractory focal epilepsy, 44 with temporal lobe epilepsy, and 49 with frontal lobe epilepsy as determined with ictal scalp video-EEG; and normal optimal conventional MRI, including hippocampal volumes and T2 measures were investigated with quantitative MRI contrasts. The contrasts comprised fast fluid attenuated inversion recovery based T2 measurement (FFT2), double inversion recovery (DIR), magnetization transfer ratio (MTR), and voxel-based morphometry of gray matter (VBM). Voxel-based analyses of whole brain data were used to compare each patient with a control group.

Results: In patients with a putative single focus on scalp video-EEG telemetry, 16% had concordant FFT2 abnormalities, as did 16% with DIR, 5% with MTR and 9% with VBM. The greatest agreement in the localization of abnormalities was between FFT2 and DIR. Altogether, 31% patients had a focal abnormality with at least one contrast in the lobe of seizure onset. Signal changes outside the lobe of the putative focus were found with FFT2 in 36% patients, with DIR in 42%, with MTR in 6% and with VBM in 7%.

Discussion: Quantitative analysis of MRI contrasts had a low yield of identifying focal abnormalities concordant with putative epileptic foci in patients with unremarkable conventional MRI. Specificity was low for FFT2 and DIR. With the low specificity, data must be interpreted with caution, but in some patients may assist in creating a hypothesis for testing with intracranial electrodes.