Abstracts

RATIONALE: Two dimensional (2D) interpretation of MRI and FDG-EPT has limitations such as a difficulty of sulco-gyral analysis and an inaccuracy of PET reading due to frequently rotated or tilted PET image and lack of anatomical information. We evaluated the value of three-dimensional(3D) rendering of PET and MRI images for localizing epileptic focus. METHODS: Twenty normal controls and 17 patients with intractable epilepsy were included. Six had temporal lobe epilepsy (TLE) and 11 had extra-TLE. All subjects had FDG-PET and volumetric MRI (1.6 mm thickness). PET-MRI co-registration was performed by automated image registration (AIR). 3D rendering were done using Analyze 7.5 software (Mayo, Rochester). Two reviewers (SBH, WCS) interpreted 2D and 3D PET and MRI with no clinical information. RESULTS: 3D PET of normal controls showed a higher glucose metabolism on frontal lobe and polar and medial occipital regions, but a lower metabolism on parietal, temporal, and lateral occipital regions. However, right and left hemispheres had symmetric glucose metabolism. In the patients, 3D PET localized an epileptic focus in 14/17 (82.4%) while 2D PET found an epileptic focus in 10/17 (58.8%). 3D PET provided a precise anatomical location of hypometabolic zone. In 7 patients, 3D rendering of both PET and MRI were performed. There were 3 neocortical TLE and 4 extra-TLE. 3D MRI found an epileptic lesion in 4/7 (57.1%). 3D PET localized a hypometabolic zone in 5/7 (71.4%), but one of them had multifocal hypometabolic zones. Either 3D PET or 3D MRI localized epileptic focus in 6/7 (85.7%). CONCLUSIONS: 3D PET rendering improved the sensitivity of FDG-PET in localizing epileptic focus and provided a clear anatomical location of functional deficit zone. 3D PET and 3D MRI were complementary for determining an epileptic focus. Hypometabolic zone of FDG-PET included irritative zone and ictal onset zone.