Recognition of white matter bundles using local and global streamline-based registration and clustering

Abstract

Virtual dissection of diffusion MRI tractograms is cumbersome and needs extensive knowledge of white matter anatomy. This virtual dissection often requires several inclusion and exclusion regions-of-interest that make it a process that is very hard to reproduce across experts. Having automated tools that can extract white matter bundles for tract-based studies of large numbers of people is of great interest for neuroscience and neurosurgical planning. The purpose of our proposed method, named RecoBundles, is to segment white matter bundles and make virtual dissection easier to perform. This can help explore large tractograms from multiple persons directly in their native space. RecoBundles leverages latest state-of-the-art streamline-based registration and clustering to recognize and extract bundles using prior bundle models. RecoBundles uses bundle models as shape priors for detecting similar streamlines and bundles in tractograms. RecoBundles is 100% streamline-based, is efficient to work with millions of streamlines and, most importantly, is robust and adaptive to incomplete data and bundles with missing components. It is also robust to pathological brains with tumors and deformations. We evaluated our results using multiple bundles and showed that RecoBundles is in good agreement with the neuroanatomical experts and generally produced more dense bundles. Across all the different experiments reported in this paper, RecoBundles was able to identify the core parts of the bundles, independently from tractography type (deterministic or probabilistic) or size. Thus, RecoBundles can be a valuable method for exploring tractograms and facilitating tractometry studies.

Tractostorm: The what, why, and how of tractography dissection reproducibility

“…Automatic segmentation methods are becoming more widespread. Methods such as, but not limited to, (Chekir, Descoteaux, Garyfallidis, Côté, & Boumghar, 2014;Garyfallidis et al, 2017;Guevara et al, 2011;O'Donnell et al, 2017;O'donnell, Golby, & Westin, 2013;Wassermann et al, 2016;Wasserthal, Neher, & Maier-Hein, 2018;Yendiki et al, 2011;Zhang et al, 2018) aim to simplify the work of raters. The typical standard of most automatic segmentation method is to reach the accuracy of raters, thus it is crucial to truly quantify human reproducibility in manual tasks.…”

“…Some measures related to streamlines require the datasets to be exactly the same, for example, Dice score, as streamline reconstructions are sets of discrete points with floating point coordinates and not discrete grids like 3D images. For this reason, comparison of streamlines is more challenging and datasets that do not originate from the same source distance in millimeters is often the only available solution (Garyfallidis et al, ; Maier‐Hein et al, ). Automatic segmentation methods are becoming more widespread.…”

“…Binary classification measures are available such as sensitivity or specificity. However, several other measures are available and each is a piece of the puzzle leading to a more accurate interpretation(Chang, Zhuang, Valentino, & Chu, 2009;Garyfallidis et al, 2017;Schilling et al, 2018).…”

Test–retest reproducibility of white matter parcellation using diffusion MRI tractography fiber clustering

“…There are two popular strategies for automated white matter parcellation (O'Donnell, Golby, & Westin, ): (a) fiber clustering that groups white matter fibers according to their geometric trajectories, aiming to reconstruct tracts corresponding to the white matter anatomy (Ding, Gore, & Anderson, ; Garyfallidis et al, ; Garyfallidis, Brett, Correia, Williams, & Nimmo‐Smith, ; Guevara et al, ; Jin et al, ; Kumar, Desrosiers, Siddiqi, Colliot, & Toews, ; O'Donnell & Westin, ; Prasad et al, ; Siless, Chang, Fischl, & Yendiki, ; Visser, Nijhuis, Buitelaar, & Zwiers, ; Wassermann, Bloy, Kanterakis, Verma, & Deriche, ; Zhang, Wu, Norton, ) and (b) cortical‐parcellation‐based that parcellates tractography according to a cortical parcellation, focusing on the structural connectivity among different brain regions of interest (ROIs) (Bassett & Bullmore, ; Bastiani, Shah, Goebel, & Roebroeck, ; Bullmore & Sporns, ; Gong et al, ; Ingalhalikar et al, ; Sporns et al, ; Wakana et al, ; Wassermann et al, ; Yeh, Badre, & Verstynen, ; Zalesky et al, ; Zhang et al, ).…”

Classifyber, a robust streamline-based linear classifier for white matter bundle segmentation

“…Unsupervised streamline-based methods, such as those in Brun et al (2004); Maddah et al (2005); O' Donnell and Westin (2007); Guevara et al (2012); Tunç et al (2014); Siless et al (2016); Zhang et al (2018), perform whole brain segmentation through clustering, without prior knowledge about the anatomy of the bundles and without leveraging examples of expert-made segmented bundles, limiting the quality of segmentation. In contrast, supervised streamline-based methods require one or more examples of the bundle to learn from, in order to segment such bundle in the target subject, such as those in Mayer et al (2011); Olivetti and Avesani (2011); Vercruysse et al (2014); Yoo et al (2015); Labra et al (2016); Garyfallidis et al (2018) and Sharmin et al (2018). It has been shown that streamline-based methods like those presented in Garyfallidis et al (2018), referred to as RecoBundles, and in Sharmin et al (2018), referred to as LAP, outperform connectivity based methods in terms of quality of segmented bundles.…”