Epilepsy affects over 150,000 people in the United States. Thirty percent of epilepsy is refractory to pharmacotherapy [1]. Surgical intervention in refractory epilepsy is frequently curative and visualization of epileptogenic lesions using MRI is correlated to successful surgical outcomes [2]. MRI-negative (MRI^-) epilepsy patients, those who do not have an identifiable lesion on MRI, are less likely to be considered candidates for surgery and have inferior outcomes when compared to MRI-positive (MRI⁺) patients [3]. However, post-surgical analysis of resected tissue has shown that a fully resected focal lesion, identified by histology, is a good predictor of an eventual determination of surgical success, even if the subject was referred to surgery as MRI^- [4]. This suggests that there are focal epileptogenic lesions, amenable to surgery, which are not visualized by current imaging protocols.

Ultra-high field MRI scanners, such as those operating at 7 Tesla (7T), can offer higher signal to noise ratio (SNR) and enhanced contrast which may be used to improve the resolution of images in order to increase the conspicuity of epileptogenic lesions and provide more accurate delineation of lesion boundaries. We previously reported the results of preliminary case studies, evaluating the feasibility of using 7T imaging to improve lesion detection in epilepsy patients who were non-lesional in their diagnostic scans [5]. Here, we report results for a larger patient study, with comparison to healthy controls, to assess the value of 7T imaging to reveal subtle abnormalities acting as epileptogenic foci in patients with focal epilepsy who have non-lesional diagnostic MRI scans.

17 healthy controls (12 male, ages 20-56 years) and 20 epilepsy patients (14 male, ages 20-57 years) with previous normal MRI scans were recruited for the 7T imaging study. Each subject had indications of focal epilepsy, identified through semiology and EEG, of unknown etiology with no record of family history, traumatic brain injury, brain infection, or febrile incident. The images were acquired on a 7T MAGNETOM scanner (Siemens, Erlangen) using a 32-channel receive Nova Medical head coil. The MRI protocol consisted of: A) 4 sequences acquired at a coronal oblique angle perpendicular to the hippocampus: MP-RAGE (TR 3000 ms, TI 1050 ms , TE 2 ms, voxel 0.7x0.7x0.7 mm³), MP2RAGE [6] (TR 6000 ms, TI1 1050 ms , TI2 3000 ms, TE 5.1 ms, voxel 0.8x0.8x0.8 mm³), T₂ TSE (TR 6000 ms, TE 69 ms, voxel 0.4x0.4x2.0 mm³), and FLAIR (TR 9000 ms, TI 2600 ms, TE 123 ms, voxel 0.7x0.7x3 mm³); and B) 2 sequences acquired axially: Susceptibility weighted imaging (SWI) (TR 23 ms, TE 14 ms, voxel 0.21x0.21x1.5 mm³), and T₂ TSE (TR 6000 ms, TE 69 ms, voxel 0.4x0.4x2.0 mm³). The 7T images were inspected by an experienced neuroradiologist, and abnormalities were reported to the patient's referring epileptologist. The relationship of the subject's epilepsy to the results of the 7T scan were classified into 4 categories: Definite, possible, uncertain, and none.

The 7T imaging protocol enabled the detection of 13 potentially epileptogenic abnormalities that were undetectable or overlooked at lower field strength. Seven of the lesions did not co-localize to the suspected seizure onset zone and were categorized with an uncertain relationship to the patient's epilepsy. Six lesions did co- localize to the known seizure onset zone and two of these abnormal findings were definitely related to the subject's epilepsy and contributed directly to subsequent surgical intervention and improved patient prognosis. The remaining four are possibly related to the subject's epilepsy and are under further investigation. A range of structural and vascular abnormalities were found in the 7T images of epilepsy patients including hippocampal asymmetry (n=5), hippocampal architecture disruption (n = 2) (Figure 1), bilaterally small hippocampi (n=1), cavernoma (n=3) (Figure 2), developmental venous anomalies (n=4), and a polymicrogyria (n=1) (Figure 3). Abnormalities were also detected in healthy controls, but found to be less numerous extensive and not clinically significant. The frequency with which abnormal findings occurred in both control subjects and patients with epilepsy is provided in Figure 4).

The improved resolution and contrast offered by 7T MRI revealed several subtle structural features that were undetectable at lower field strengths and aided in the localization of epileptogenic foci in previously non-lesional patients. Studying the full set of abnormalities associated with epilepsy when compared to healthy controls may also impart valuable new insights into the etiology of the disease.