Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2017 Jun 26;15:819-831.
doi: 10.1016/j.nicl.2017.06.027. eCollection 2017.

Performance of Unscented Kalman Filter Tractography in Edema: Analysis of the Two-Tensor Model

Affiliations
Free PMC article

Performance of Unscented Kalman Filter Tractography in Edema: Analysis of the Two-Tensor Model

Ruizhi Liao et al. Neuroimage Clin. .
Free PMC article

Abstract

Diffusion MRI tractography is increasingly used in pre-operative neurosurgical planning to visualize critical fiber tracts. However, a major challenge for conventional tractography, especially in patients with brain tumors, is tracing fiber tracts that are affected by vasogenic edema, which increases water content in the tissue and lowers diffusion anisotropy. One strategy for improving fiber tracking is to use a tractography method that is more sensitive than the traditional single-tensor streamline tractography. We performed experiments to assess the performance of two-tensor unscented Kalman filter (UKF) tractography in edema. UKF tractography fits a diffusion model to the data during fiber tracking, taking advantage of prior information from the previous step along the fiber. We studied UKF performance in a synthetic diffusion MRI digital phantom with simulated edema and in retrospective data from two neurosurgical patients with edema affecting the arcuate fasciculus and corticospinal tracts. We compared the performance of several tractography methods including traditional streamline, UKF single-tensor, and UKF two-tensor. To provide practical guidance on how the UKF method could be employed, we evaluated the impact of using various seed regions both inside and outside the edematous regions, as well as the impact of parameter settings on the tractography sensitivity. We quantified the sensitivity of different methods by measuring the percentage of the patient-specific fMRI activation that was reached by the tractography. We expected that diffusion anisotropy threshold parameters, as well as the inclusion of a free water model, would significantly influence the reconstruction of edematous WM fiber tracts, because edema increases water content in the tissue and lowers anisotropy. Contrary to our initial expectations, varying the fractional anisotropy threshold and including a free water model did not affect the UKF two-tensor tractography output appreciably in these two patient datasets. The most effective parameter for increasing tracking sensitivity was the generalized anisotropy (GA) threshold, which increased the length of tracked fibers when reduced to 0.075. In addition, the most effective seeding strategy was seeding in the whole brain or in a large region outside of the edema. Overall, the main contribution of this study is to provide insight into how UKF tractography can work, using a two-tensor model, to begin to address the challenge of fiber tract reconstruction in edematous regions near brain tumors.

Keywords: DTI; Diffusion MRI; Edema; Tractography; White matter.

Figures

Fig. 1
Fig. 1
Synthetic data phantom. Left: grayscale FA images in 3D show the location of the simulated edema (the dark region in the phantom center). The simulated fiber tracts run anterior-posterior, in the orientation of the pink line. Center: FA values are plotted along the central axis of the phantom (this axis is shown as a pink line in the leftmost image). Right: The MD values are plotted along the central axis of the phantom.
Fig. 2
Fig. 2
(a) Axial image from Patient 1 with a left fronto-parietal glioblastoma multiforme tumor and peritumoral edematous zone. (b) Axial image from Patient 2, with two metastatic lesions of unclassified pleomorphic sarcoma in the left frontal lobe and peritumoral edema. Gadolinium was given before the MRI scans.
Fig. 3
Fig. 3
Recovery of simulated edematous fiber tracts. Three tractography methods were seeded in the synthetic edema phantom with minimum FA of 0.2. (a) A phantom containing parallel fibers running anterior-posterior (indicated by green color) with a region of synthetic edema in the phantom center. Simulated edematous tracts were recovered (yellow fibers) using default parameters for tractography. (b) The single-tensor streamline tractography that used independent single-tensor estimation at each voxel (least-squares) followed by Runge-Kutta order two integration for fiber tracking in 3D Slicer. (c) and (d) The two UKF methods (single tensor with and without free water model) that performed model estimation during tracking, using a Kalman filter.
Fig. 4
Fig. 4
Tractography seeding scenarios for tracking through edema. Left column: two-tensor UKF tractography was seeded inside synthetic edema (pink seed region). Right column: two-tensor UKF tractography was seeded within simulated healthy white matter (blue seed region). (c) and (d): Traced fibers shown in yellow. (e) and (f): Tensor one that is the tensor followed during fiber tracking. The calculated tensor model is displayed along the fibers as ellipsoids colored by FA. Higher FA is green and blue, while lower FA (such as that in the edema) is orange.
Fig. 5
Fig. 5
Comparison of single-tensor (a) and two-tensor (b) UKF tractography in the arcuate fasciculus of Patient 1. Two ROIs were applied to select the AF anatomy from the whole brain tractography. However, no fibers were found connecting the two ROIs using the single-tensor UKF tractography. To assess if the AF could be partially traced, we employed each ROI separately (left and center images), in conjunction with expert removal of fibers that appeared not to form part of the AF such as short U fibers.
Fig. 6
Fig. 6
Comparison of single-tensor (a) and two-tensor (b) UKF tractography in the corticospinal tract of Patient 2. Two ROIs were applied to select the CST anatomy from the whole brain tractography. The two-tensor method (right image) can track more fibers for depiction of lateral connections.
Fig. 7
Fig. 7
Two-tensor UKF tractography traverses edema but is affected by initial seeding location in Patient 1. In all views, the background anatomical images are the diffusion baseline images from Patient 1. Tensors are colored by FA, where blue represents a higher FA and orange/yellow is a lower FA. (a)-(d): Tractography (yellow fibers) is seeded from three different regions (cyan) within the arcuate fasciculus (AF). Seed region 1 is outside of the edema, while regions 2 and 3 are inside the edema. The translucent model represents the edematous region delineated by a clinical expert. (e) and (f): Zoomed-in views that show the two-tensor model when seeded in region 1. (g) and (h): Zoomed-in views that show the two-tensor model when seeded in region 3.
Fig. 8
Fig. 8
Two-tensor UKF tractography traverses edema but is affected by initial seeding location in Patient 2. In all views, the background anatomical images are the diffusion baseline images, shown at a location behind the fibers. Tensors are colored by FA, where blue represents a higher FA and orange/yellow is a lower FA. (a)–(d): Tractography (yellow fibers) is seeded from three different regions (cyan) within the corticospinal tract (CST). Regions 1 and 2 are outside of the edema, while region 3 is inside the edema. The translucent model represents the edematous region delineated by a clinical expert. (e) and (f): Zoomed-in views that show the two-tensor model when seeded in region 2. (g) and (h): Zoomed-in views that show the two-tensor model when seeded in region 3.
Fig. 9
Fig. 9
Two-tensor UKF tractography reconstructions of the AF from 22 different parameter settings in Patient 1. The background image is a diffusion b0 image, medial to the AF. Six FA and GA threshold settings are demonstrated in columns. Four qL settings are shown in rows. The top row displays the results without including the free-water model and the three bottom rows include it.
Fig. 10
Fig. 10
Two-tensor UKF tractography reconstructions of the CST from 22 different parameter settings in Patient 2. The background image is a diffusion b0 image, posterior to the CST to avoid occluding fibers. Six FA and GA threshold settings are demonstrated in columns. Four qL settings are shown in rows. The top row displays the results without including the free-water model and the three bottom rows include it.
Fig. 11
Fig. 11
Quantitative measures of tractography sensitivity based on patient-specific fMRI demonstrate the effect of varying UKF tractography parameters. The coverage ratio is measured as the percentage of the target fMRI activation that is reached by the fiber tracts. (a) Reconstructed AF coverage ratio of Broca's area from all parameter settings experiments in the data of Patient 1. (b) Reconstructed CST coverage ratio of the face/foot/hand motor areas from all parameter settings experiments in the data of Patient 2.

Similar articles

See all similar articles

Cited by 6 articles

See all "Cited by" articles

References

    1. Abhinav K., Yeh F.-C., Mansouri A., Zadeh G., Fernandez-Miranda J.C. High-definition fiber tractography for the evaluation of perilesional white matter tracts in high-grade glioma surgery. Neuro-oncology. 2015;17(9):1199–1209. - PMC - PubMed
    1. Aja-Fernández S., Alberola-López C., Westin C.-F. Noise and signal estimation in magnitude MRI and Rician distributed images: a LMMSE approach. Image Processing, IEEE Transactions on. 2008;17(8):1383–1398. - PubMed
    1. Akai H., Mori H., Aoki S., Masutani Y., Kawahara N., Shibahara J., Ohtomo K. Diffusion tensor tractography of gliomatosis cerebri: fiber tracking through the tumor. J. Comput. Assist. Tomogr. 2005;29(1):127–129. - PubMed
    1. Alexander A.L., Hasan K.M., Lazar M., Tsuruda J.S., Parker D.L. Analysis of partial volume effects in diffusion-tensor MRI. Magn. Reson. Med. 2001;45(5):770–780. - PubMed
    1. Basser P., Pajevic S., Pierpaoli C., Duda J., Aldroubi A. In vivo fiber tractography using DT-MRI data. Magn. Reson. Med. 2000;44:625–632. - PubMed

Publication types

Feedback