Preoperative language mapping by repetitive navigated transcranial magnetic stimulation and diffusion tensor imaging fiber tracking and their comparison to intraoperative stimulation

Introduction Repetitive navigated transcranial magnetic stimulation (rTMS) can be used for preoperative language mapping, but it still suffers from comparatively high sensitivity and low specificity when compared to direct cortical stimulation (DCS). Therefore, this study evaluates whether the additional consideration of rTMS-based diffusion tensor imaging fiber tracking (DTI FT) for identifying language-positive brain regions improves specificity when compared to DCS. Methods We performed rTMS, rTMS-based DTI FT, and DCS during awake surgery combined with object naming in 20 patients suffering from left-sided perisylvian brain lesions. For rTMS, different error rate thresholds (ERTs) and error types were considered, and DTI FT was conducted with individualized fractional anisotropy thresholds (FATs). Then, receiver operating characteristics (ROC) for rTMS vs. DCS, rTMS-based DTI FT vs. DCS, and rTMS spots confirmed by rTMS-based DTI FT vs. DCS were calculated. Results In general, rTMS vs. DCS was in good accordance with previous literature (sensitivity/specificity: 92.7/13.3 % for all naming errors without ERT). In addition, rTMS-based DTI FT vs. DCS led to balanced results when tracking was based on all errors as well (sensitivity/specificity: 62.8/64.3 % for 100 % FAT). However, rTMS combined with rTMS-based DTI FT vs. DCS did not lead to any improvement in specificity when compared to rTMS vs. DCS alone. Conclusion The additional use of rTMS-based DTI FT to rTMS did not improve the identification of DCS-positive language areas during awake surgery. Yet, concerning rTMS-based DTI FT, this new technique must be validated itself by intraoperative subcortical stimulation.