Abstracts

Rationale: Voxel-based relaxometry (VBR) is a whole-brain statistical analysis of T2 values from magnetic resonance imaging, allowing for lo-calization of abnormalities associated with epilepsy. A patient may be diagnosed with epilepsy, but there may not be a firm under-standing of the seizure focus. In other cases the basic diagnosis of epilepsy may be uncertain. In any of these cases, additional infor-mation, such as that offered by VBR may provide important information to corroborate, or refute indeterminate information from the other sources.2 Our objective was to assess the performance of single-subject VBR at 3 T as a diagnostic tool for patients whose diag-nosis of epilepsy or seizure focus is uncertain. Methods: We assessed 51 patients and 25 healthy controls. Scanning was performed at 3T with a Carr-Purcell-Meiboom-Gill sequence. Initial diagnoses were based on history, EEG, structural MR, and neuropsychological testing. Patients were classified as having suspected but unconfirmed epilepsy (SE), known epilepsy with unknown focus (KE) or known epilepsy with a suspected, unconfirmed focus (SF). The SF group were determined to have a suspected lobe (SF-L), or a suspected lobe and side of origin (SF-LS). VBR was per-formed with SPM2 (α = 0.05, uncorrected). A VBR severity score was determined based on the presence of VBR significance in any of the following 13 regions for each hemisphere: anterior temporal lobe, posterior temporal lobe, amygdala, hippocampus, frontal lobe, parietal lobe, occipital lobe, caudate, putamen, pallidum, thalamus, internal capsule and insular ribbon. Patients were classified as having high (> 6 areas), medium (3-6 areas), low (1-2 areas) or no VBR severity (0 areas). Results: Fifty-one patients were assessed for VBR abnormalities and twenty-five control subjects were used as a statistical control (mean age 26, range of 18-42). Seventeen of the 27 patients in the SF group (63%) exhibited a VBR abnormality in the suspected focus. In these patients, the VBR findings helped corroborate other findings. Fig 1 shows a case where VBR helped to identify the side of seizure origin in a subject with suspected temporal lobe epilepsy. Compared to control data, all patient groups showed higher VBR severity, with the SF-L group showing the highest proportion of patients with high or moderate VBR severity (Fig 2). Patient groupings exhib-ited more VBR abnormalities than controls, where for example, the median number of VBR abnormalities was 1 for controls versus 3 for the SF group. Conclusions: VBR abnormalities occurred more in patients with better-characterized epilepsy (i.e., with a suspected focus). Single-subject VBR can help identify or confirm a seizure foci in patients in whom conventional investigations have failed to yield a diagnosis. In these pa-tients, VBR can strengthen and hone the diagnosis to direct subsequent treatment and investigations.