Abstracts

Rationale: MSI is being developed into a standard clinical tool for presurgical functional mapping in epilepsy and brain tumor patients. The relationship between receptive language localization using MSI and cortical mapping with ECS is not well understood. This study describes our recent experience comparing the two language mapping methods for anatomical concordance. Methods: The patients were 7 males and, 12 females ranging in age from 4 to 52 years (average 21 yrs). All were candidates for epilepsy or brain tumor surgery. MSI receptive language mapping was performed using an auditory word recognition task[1, 2]. MEG data were acquired with a 148 MEG channel whole head system and analyzed using single equivalent current dipole model. Receptive language areas were defined by dipole clusters, each made up of at least 5 dipoles within 1 centimeter range. Electrical stimulation mapping was performed with surgically implanted subdural electrode arrays using a formal language protocol that briefly tested naming, auditory comprehension, repetition and reading. 3D MSI images (language dipoles on a 3D MR series in DICOM format) and a 3D CT or MR scan after grid placement were fused into a set of hybrid images with a neuronavigation workstation. The distance of the active electrodes and MSI language dipoles were measured based on these combined images. Language sites from the two methods were considered concordant if they were within 1 cm on the lateral cortex or if MSI dipoles were deep to a cortical stimulation site on the same gyrus or an adjacent sulcus. Sites that did not overlap using these criteria were considered to be disconcordant. The degree of correlation between MSI and ECS was divided into 3 categories: 1. Complete concordance - all MSI language dipoles overlapped with ECS; 2. Partial concordance; 3. Discordant-no language dipoles overlapped with ECS. Results: Temporal language regions were identified with ECS in 15/19 patients. Receptive language mapping could not be generated via ECS in four patients and these patients proceeded to surgery with the guidance of MSI language mapping. In the 15 patients with both MSI and ECS , 3/15 were complete concordant, 12 of 15 patients were partially concordant. Majority of MSI activity lies deep in the sulcus, while ECS mapped the surface cortex. In these partially concordant cases, resection spared the deeper language sites identified by MSI and the surface cortical regions defined by ECS. No significant language deficits developed post-operatively in any patient. Conclusions: Based on our data, we believe: 1. MSI is a valuable tool for receptive language mapping. 2. MSI and ECS are complimentary. 3. Utilizing both methods may generate a more complete language map to minimize potential deficit. 4. MSI should be regarded as equally valuable in identifying language areas as compared to ECS. Reference: 1.Papanicolaou, A.C., et al., J Neurosurg, 2004. 100(5): p. 867-76. 2.Simos, P.G., et al., J Neurosurg, 1999. 91(5): p. 787-96.