Article
Probabilistic fibertracking in deep brain stimulation: Is it worth the effort?
Search Medline for
Authors
-
Juergen Schlaier
- Klinik und Poliklinik für Neurochirurgie, Universitätsklinikum Regensburg, Germany
-
Judith Anthofer
- Klinik und Poliklinik für Neurochirurgie, Universitätsklinikum Regensburg, Germany
-
Kathrin Steib
- Klinik und Poliklinik für Neurochirurgie, Universitätsklinikum Regensburg, Germany
-
Max Lange
- Klinik und Poliklinik für Neurochirurgie, Universitätsklinikum Regensburg, Germany
-
Alexander Brawanski
- Klinik und Poliklinik für Neurochirurgie, Universitätsklinikum Regensburg, Germany
| Published: | June 8, 2016 |
|---|
Outline
Text
Objective: Several Groups suggested to integrate diffusion-weighted-imaging-based fibertracking (DTI) into the target planning for deep brain stimulation. Probabilistic fibertracking is thought to be superior to deterministic DTI. A higher number of read-out directions is believed to provide better results than only few read-out directions. However, probabilistic DTI is much more complex to perform and vastly more time consuming than deterministic DTI. The aim of our study was to compare the quality and overlap of fiber tracts as well as the time duration of the tracking procedures between probabilistic and deterministic DTI.
Method: The dentate-rubro-thalamic tract (DRTT) was investigated in 12 cerebral hemispheres of 6 patients using probabilistic and deterministic fibertracking (FT). Both tracking algorithms were performed on diffusion weighted images with 12 and 64 read-out directions. For each hemisphere, the resulting 4 DRTTs (deterministic12, deterministic64, probabilistic12, probabilistic64) were compared concerning their quality and overlap of fiber tracts as well as the time duration of the tracking procedures.
Results: Probabilistic FT with 64 read-out directions provided DRTTs of significantly higher quality than deterministic FT with 64 read-out directions (p=0.002). However, deterministic FT with 12 read-out directions almost reached the quality of probabilistic FT-64. The tracts of probabilistic FT with 64 read-out directions overlapped consistently with the deterministic FT (12 read-out directions). The spatial course of the other tracts showed inconsistently minor and major deviations. Probabilistic FT was much more time consuming than deterministic FT: 1971 minutes vs. 16 minutes.
Conclusions: Concerning quality and spatial courses, there are significant differences between probabilistically and deterministically determined fiber tracts, depending on the amount of read-out directions of the diffusion-weighted images. Which algorithms provide the most reliable and clinically relevant results has to be evaluated still.
