Abstracts

Rationale: Patients with MRI-negative focal epilepsies are challenging since they are considered less favourable candidates for resective epilepsy surgery. Several studies have shown the benefit of using a morphometric analysis program (MAP) on T1-weighted MRI scans to detect subtle lesions, such as focal cortical dysplasia. MAP findings allow a focussed re-evaluation of the original MRI scans, to ultimately identify previously overlooked, structural epileptogenic lesions. Most studies address the diagnostic gain of second review after MAP analysis. There is little data on patients where even the second look after MAP analysis did not reveal any structural lesions. In this study we assess the diagnostic yield of morphometric analysis in a group of “second look” MRI-negative patients.

Methods: We identified 25 patients with MRI-negative focal epilepsies with a clear localization of the epileptogenic zone based on intracranial EEG or postoperative seizure free outcome. Data of patients and 64 healthy controls was aquired on a GE Signa 3T scanner, including a high resolution 3D T1 dataset. MRI negativity was defined by an experienced neuroradiologist, confirmed by an experienced epileptologist and reviewed in a patient management conference. Morphometric analysis was performed with MAP18 software, creating 4 different z-score maps, reflecting a patients deviation from the control population. Ten brain regions were specified (left and right frontal, central, temporal, parietal and occipital) to determine whether MAP18 findings were located in the correct region. ROC analyses were performed to identify the optimal z-score threshold for each map.

Results: The 4 resulting z-score maps and their respective diagnostic performance were: 1) Junction image, based on blurring of the grey-white matter junction (optimal threshold: 5,3; sensitivity: 64%; positive predictive value (PPV): 18%; i.e. 18% of the identified alterations were located in the region containing the epileptogenic zone). 2) Extension image, based on increased depth of sulci (threshold: 4,0; sensitivity: 67%; PPV: 19%). 3) Thickness image, based on abnormal cortical thickness (threshold: 3,2; sensitivity: 64%; PPV: 15%). 4) Combined image, that combines the highest scores of each of the three aforementioned maps for each voxel (threshold: 5,3; sensitivity: 73%; PPV: 18%). Histopathology of the 15 patients who already underwent resections showed mild malformations of cortical development (mMCD II) in 13 and focal cortical dysplasia in 2.

Conclusions: In contrast to previous studies, this study showed that while MAP18 software is useful in detecting previously overlooked subtle structural lesions, the additional diagnostic yield in a "second look“ MRI-negative patient group is limited. Further analyses, including the size of z-score clusters and individual thresholding could improve the diagnostic performance.

Funding: Please list any funding that was received in support of this abstract.: none.