Distinct approaches to language pathway tractography: comparison of anatomy-based, repetitive navigated transcranial magnetic stimulation (rTMS)-based, and rTMS-enhanced diffusion tensor imaging-fiber tracking.

OBJECTIVE Visualization of subcortical language pathways by means of diffusion tensor imaging-fiber tracking (DTI-FT) is evolving as an important tool for surgical planning and decision making in patients with language-suspect brain tumors. Repetitive navigated transcranial magnetic stimulation (rTMS) cortical language mapping noninvasively provides additional functional information. Efforts to incorporate rTMS data into DTI-FT are promising, but the lack of established protocols makes it hard to assess clinical utility. The authors performed DTI-FT of important language pathways by using five distinct approaches in an effort to evaluate the respective clinical usefulness of each approach. METHODS Thirty patients with left-hemispheric perisylvian lesions underwent preoperative rTMS language mapping and DTI. FT of the principal language tracts was conducted according to different strategies: Ia, anatomical landmark based; Ib, lesion-focused landmark based; IIa, rTMS based; IIb, rTMS based with postprocessing; and III, rTMS enhanced (based on a combination of structural and functional data). The authors analyzed the respective success of each method in revealing streamlines and conducted a multinational survey with expert clinicians to evaluate aspects of clinical utility. RESULTS The authors observed high usefulness and accuracy ratings for anatomy-based approaches (Ia and Ib). Postprocessing of rTMS-based tractograms (IIb) led to more balanced perceived information content but did not improve the usefulness for surgical planning and risk assessment. Landmark-based tractography (Ia and Ib) was most successful in delineating major language tracts (98% success), whereas rTMS-based tractography (IIa and IIb) frequently failed to reveal streamlines and provided less complete tractograms than the landmark-based approach (p < 0.001). The lesion-focused landmark-based (Ib) and the rTMS-enhanced (III) approaches were the most preferred methods. CONCLUSIONS The lesion-focused landmark-based approach (Ib) achieved the best ratings and enabled visualization of the principal language tracts in almost all cases. The rTMS-enhanced approach (III) was positively evaluated by the experts because it can reveal cortico-subcortical connections, but the functional relevance of these connections is still unclear. The use of regions of interest derived solely from cortical rTMS mapping (IIa and IIb) leads to cluttered images that are of limited use in clinical practice.