Abstracts

Rationale: A few recent studies reported the usefulness of in vivo structural 7-Tesla (7T) MRI to noninvasively identify focal cortical dysplasia (FCD). Initial success has been reported in some cases with negative 3T MRI. However, no previous studies investigated MRI post-processing on 7T for lesion detection. Here we present a large array of epileptogenic pathologies imaged at 7T on 70 patients with drug-resistant epilepsy undergoing presurgical evaluation. We report initial results of the impact of 7T visual and post-processing analyses on clinical care, as well as confirmation of the epileptogenicity and pathology of 7T findings by ICEEG and surgery. Methods: This study was approved by the Cleveland Clinic institutional review board. Patients were included if they had a clear or equivocal lesion on 3T MRI, or had negative 3T MRI but clinical and other imaging data strongly suggested focal epilepsy. Post-processing of the T1w MP2RAGE sequence was carried out using the morphometric analysis program (MAP) in SPM (Figure 1). Visual analysis of 7T MRI and review of MAP results were done by a dedicated neuroradiologist (SEJ); additional information from 7T (in comparison with the 3T MRI) was communicated to the patient management team, and interpreted in the context of other presurgical evaluation data. Concordance of 7T findings, ICEEG and surgical resection was assessed using previously published scheme (Murakami et al., Brain, 2016). All available microscopic slides from surgical resections were reviewed by a dedicated neuropathologist (IB). Seizure outcome was classified according to Engel’s class. Results: A total of 70 patients were included, consisting of three groups: patients with completely normal lower-field MRI (nonlesional group, N=26), patients with equivocal lesions on lower-field MRI (subtly lesional group, N=13), and patients with definite lower-field MRI lesions (lesional group, N=31). In the nonlesional group, visual analysis revealed additional findings in 7 patients, and MAP identified subtle FCD in an additional 9 patients. In the 10 patients who had positive 7T findings (visual and MAP) and ICEEG, good concordant was observed in 9 (7 concordant, 2 concordant-minus; a concordant example is shown in Figure 2). In all the patients who had positive 7T findings (visual and MAP) and underwent surgery, 8 underwent complete resection of the abnormality, and had favourable seizure outcomes (7 class I, 1 class II); 2 had partial or no resection of the abnormality, and had class III outcome. The subtly lesional group mostly included patients with suspected MTS and FCD. In only two patients, the suspected lesion (FCD) was confirmed. MAP yielded additional FCD in 2 patients. In the lesional group, the majority of cases had no additional information generated by 7T; however, in 4 patients with extensive cortical abnormalities, visual analysis of the 7T better delineated the extent of lesions. In two patients, additional FCD was identified by visual analyses of 7T. MAP did not generate additional findings in this group. Histopathology work is ongoing. The main weakness of our current setup included imaging artefacts associated with field inhomogeneity, motion and pulsation. Conclusions: Our initial experience showed that 7T MRI is useful to: (1) detect subtle FCD lesions in 3T-negative patients; (2) confirm or refute suspicious equivocal findings on 3T scans; (3) delineate the extent of pathology in the presence of 3T-visible lesions. The use of 7T does not replace MRI post-processing; instead, guidance from MRI post-processing is even more appreciated for FCD detection, due to larger number of imaging slices. Funding: JoshProvides Epilepsy Assistance Foundation Research Grant