Article
Fiber density mapping in patients with gliomas: histopathologic evaluation of a novel approach for post-processing of DTI data
Search Medline for
Authors
-
Oliver Ganslandt
- Neurochirurgische Klinik, Universität Erlangen-Nürnberg, Deutschland
-
Michael Buchfelder
- Neurochirurgische Klinik, Universität Erlangen-Nürnberg, Deutschland
-
Andreas Stadlbauer
- Neurochirurgische Klinik, Universität Erlangen-Nürnberg, Deutschland; MR Physics Group, Institut für Radiologie, Landesklinikum St. Pölten, Österreich
| Published: | September 16, 2010 |
|---|
Outline
Text
Objective: Fiber density mapping (FDM) is a new approach for post-processing of DTI data and is an indicator of the density of white matter fibers within the bundle passing through a voxel or a region of interest. FDM is a 2-step post-processing strategy: first, there is a reconstruction of all fiber paths for the whole brain from the volumetric DTI data using a tracking algorithm and tracking thresholds and second, the depiction of the individual FD values for each voxel as parametric maps of the whole brain is performed. The purpose of this study was the histopathological evaluation of FDM in glioma patients to assess the extent of destruction of white matter structures.
Methods: We correlated FDM data and histopathological findings from MRI-guided stereotactic biopsies of 20 patients with a supratentorial glioma WHO grade II or III. DTI data were obtained using a diffusion-weighted EPI with six diffusion directions and isotropic voxel size of 1.9 mm3. Fiber tracking using DTI-Studio and the Fiber Assignment by Continuous Tracking (FACT) method was performed to calculate FD values for all voxels within the VOI. Coregistration of FDM data with a 3D MPRAGE data set and a T2-weighted TSE data set, which were used for stereotactic brain biopsies, allowed correlation of FD values with histopathological findings expressed as % tumor infiltration and tumor cell number (tumor CN).
Results: The histopathologic findings of 78 MR image-guided stereotactic biopsies from all 20 patients were correlated with the corresponding FD values at the biopsy site. For FD we found a strong negative logarithmic correlation with both the % tumor infiltration and the number of tumor cells. Complete destruction of white matter structures, i.e. FD = 0, was found for a tumor cell infiltration ≥60% and a tumor CN ≥150. A FD value of 18 (i.e. the lower limit for the FD values in cNWM) is related to a tumor infiltration of approx. 16% as calculated from our histopathology-versus-fiber density model.
Conclusions: We correlated FDM data with histopathological findings from MRI-guided stereotactic biopsies and found a logarithmic correlation with both the % tumor cell infiltration and tumor CN. For a tumor infiltration of >60% (150 tumor cells) no fiber structures remain. In regions with a tumor infiltration of <60% and >16%, fiber structures still exists, but they are certainly functionally not important. However, in tumor regions with <16% tumor cells, functionally important fiber structures may still exist.
