gms | German Medical Science

Objective: Clinicians who want to apply DTI fiber tracking information for planning of neurosurgical procedures or intraoperative neuronavigation need precise anatomical information concerning the location and course of fiber tracts in 3-D space. Here it is essential to visualize the entire extension, the “3-D envelope” of a specific fiber tract. Depiction of the corticospinal tract is usually performed using 2-D regions of interest, e.g., the posterior limb of the internal capsule. The advantage of this time-saving choice is impaired by the fact that only parts of the tract are visualized no matter which tracking parameters were used. The line propagation algorithm detects fibers mainly running in dorso-caudal direction. Compared to known anatomy, the fibers which originate in the lateral motor area might not be visualized and therefore, tract volume might be underestimated. Using a 3-D volume of interest as the seed region as well as optimized tracking parameters is a simple solution; at least results can be significantly enhanced, and the motor pathway can be depicted to a larger extent.

Methods: We performed single-shot diffusion weighted spin-echo EPI sequences on 10 healthy subjects using a 3.0T clinical MR-scanner. 32 gradient axes were measured. Additional T1-weighted magnetization prepared 3-D turbo fast-field gradient echo (TFE) sequences were acquired for anatomical reference. Prior to tracking with the newly developed "StealthDTI" tracking program affine registrations were applied compensating for geometrical distortions. We used 3-D volumes of interest covering the subcortical white matter of the precentral gyrus to define the start regions for the tracking algorithm. Results were qualitatively and quantitatively compared.

Results: Fiber tract portrayal of the corticospinal tract could be significantly enhanced with high data quality and an optimized tracking strategy. For the first time we could show a “complete” tract picture even displaying its lateral parts, resembling a result that comes close to its anatomical template.

Conclusions: With the knowledge of how to bypass sources of errors in DTI and by applying optimized tracking parameters the automatic depiction of the corticospinal tract can be significantly improved revealing an anatomically accurate fiber tract portrayal. This enables a neurosurgeon to gain additional information when performing operations in the eloquent motor region e.g. by implementation of this information into a neuronavigation system.