Three-dimensional (3-D) images of 14 patients, in whom subdural electrodes were implanted for epilepsy surgery evaluation, were produced by fusing computerized tomography (CT) scans and magnetic resonance images (MRI) to determine the exact location of the electrodes. In 25% (198 of 806) of the subdural electrodes an exact location for resection strategy was not evident from the X-rays. The location of these electrodes ('doubtful location', DL) was compared to the 3-D images. Intraoperative inspection served as the gold standard. Concordance score was defined as good for 75-100% concordance, moderate for 50-75%, and poor for less than 50% concordance of the DL electrodes to the intraoperatively determined location. A comparative analysis of the 3-D images and X-ray films showed a highly significant difference (P < 0.0002) in favor of 3-D images. The concordance score for the DL electrodes in the 3-D images was good in 93% of the patients and moderate in 7%. In contrast, the concordance score was good in only 7% of the patients for the two-dimensional (2-D) skull films, moderate in 57%, and poor in 36%. Interobserver reliability was better for 3-D images (93%) than for the 2-D X-rays (43%). These findings suggest that 3-D images aid preoperative planning for resective epilepsy surgery.