Voxel-based relaxometry for cases of an unresolved epilepsy diagnosis.

PURPOSE Voxel-based relaxometry (VBR) is a technique in which a voxel-level statistical comparison of quantitative MR T2 maps is performed to identify regions with significantly elevated T2 relaxation time. Our objective was to assess the performance of single-subject VBR at 3T as a diagnostic tool for patients whose diagnosis of epilepsy or seizure focus location is uncertain. METHODS Fifty-nine patients with possible epilepsy or known epilepsy, but an unknown focus and forty-five healthy controls were studied. All subjects were scanned at 3T using a Carr-Purcell-Meiboom-Gill MR sequence. Single-subject VBR was performed at a significance level of α=0.001. Patients were classified based on whether the diagnosis of epilepsy was in question and whether there was a suspected focus. A VBR score was determined based on the presence of VBR abnormalities in any of 13 predefined regions per hemisphere. RESULTS All patients exhibited significantly more median VBR abnormalities than controls (p<0.05). VBR abnormalities were seen in 69% and 89% of patients with a normal or questionably abnormal MR scan, respectively. Nineteen of the 27 patients with a suspected focus (70%) had VBR abnormalities in the suspected focus, with additional regions of involvement being elucidated. VBR also correctly predicted the seizure focus in 50% of patients whose seizure foci were confirmed based on follow-up history or clinical investigations. CONCLUSIONS Single subject VBR can help identify potential seizure foci in patients whose seizure foci are uncertain.