Real-time multi-peak tractography for instantaneous connectivity display

Abstract

The computerized process of reconstructing white matter tracts from diffusion MRI (dMRI) data is often referred to as tractography. Tractography is nowadays central in structural connectivity since it is the only non-invasive technique to obtain information about brain wiring. Most publicly available tractography techniques and most studies are based on a fixed set of tractography parameters. However, the scale and curvature of fiber bundles can vary from region to region in the brain. Therefore, depending on the area of interest or subject (e.g., healthy control vs. tumor patient), optimal tracking parameters can be dramatically different. As a result, a slight change in tracking parameters may return different connectivity profiles and complicate the interpretation of the results. Having access to tractography parameters can thus be advantageous, as it will help in better isolating those which are sensitive to certain streamline features and potentially converge on optimal settings which are area-specific. In this work, we propose a real-time fiber tracking (RTT) tool which can instantaneously compute and display streamlines. To achieve such real-time performance, we propose a novel evolution equation based on the upsampled principal directions, also called peaks, extracted at each voxel of the dMRI dataset. The technique runs on a single Computer Processing Unit (CPU) without the need for Graphical Unit Processing (GPU) programming. We qualitatively illustrate and quantitatively evaluate our novel multi-peak RTT technique on phantom and human datasets in comparison with the state of the art offline tractography from MRtrix, which is robust to fiber crossings. Finally, we show how our RTT tool facilitates neurosurgical planning and allows one to find fibers that infiltrate tumor areas, otherwise missing when using the standard default tracking parameters.

Keywords: HARDI; diffusion MRI; free open source software; medical visualization; neurosurgical planning; tractography.

Figure 1

Maxima map of upsampled fODFs showing multiple directions at each voxels. The color code is a mapping of their local coordinates (x, y, z) to the red-green-blue channels [R, G, B]. The length of the maxima can also be adjusted interactively by the user. (HC dataset).

Figure 2

Overview of the MultiPeak-RTT panel. The main fiber tracking parameters can be tuned interactively by the user, allowing a more in-depth comprehension of their effects on the resulting streamlines.

Figure 3

Whole brain fiber tractography obtained via the shell-seeding option. In (A) an isosurface is generated from an anatomical T1 map to fit the boundaries of GM/WM matter interface. Seed points are then launched at every vertex of the surface (B). (C) Shows approximately 200,000 fibers generated with this technique (<10 s). These fibers can then be selected with a VOI for precise exploration as seen in (D). (HC dataset).

Figure 4

Streamlines obtained with HARDI (left column) and DTI (right column) tractography. HARDI reconstructions reveals lateral fibers which are not present in the DTI reconstruction. Bottom row reveals the bending part of the Cg, a part that is often missing when performing DTI-based tractography. (HC dataset).

Figure 5

(A) The seven FiberCup reference bundles (Poupon et al., ; Côté et al., 2013). (B) ROIs used for score computation. Bottom row shows valid connections for best MRtrix parameters (C) and associate MultiPeak bundles (D).

Figure 6

Qualitative analysis of 1000 streamlines generated from VOIs placed in specific regions of the brain (left column). iFOF: VOI was placed in the inferior frontal lobe. CC: ROI located at the middle of the CC. CST: Elongated VOI located at the pontine nuclei level. FX: Seeds where initiated in the middle body of the FX. Center column shows the result of MultiPeak-RTT, while on the right column, the results were obtained using MRtrix (SD_STREAM) with default parameters as in Tournier et al. (2012). (HC dataset).

Figure 7

Reconstructions of four different fiber bundles (FX, CST, iFOF, CC) using a mixture of parameters. Some of these parameters are best suited for long bundles (CST, iFOF), while others parameters favors the reconstruction of curved fibers. (HC dataset).

Figure 8

Neurosurgical application of the MultiPeak-RTT method. (A) T1-weighted image revealing an anaplastic astrocytoma tumor extending across the left hemisphere of the brain. (B) Coherent structure within the tumor region in accordance to local orientations (zoomed rectangle). In (C) one can observe a premature termination of the fiber tracts inside the tumor (red lines). By lowering the FA threshold to 0.06, we allow the tracts to momentarily step over small local changes in the diffusivity and continue their course toward the cortex (D). (TP dataset).

Figure 9

Graphical representation of the frame per second (FPS) performance regarding the number of seed per axis present within the VOI. For the default 1000 seeds proposed (10 × 10 × 10), the mean FPS value is over 20.