Connectome analysis for pre-operative brain mapping in neurosurgery

Abstract

Object: Brain mapping has entered a new era focusing on complex network connectivity. Central to this is the search for the connectome or the brains ‘wiring diagram’. Graph theory analysis of the connectome allows understanding of the importance of regions to network function, and the consequences of their impairment or excision. Our goal was to apply connectome analysis in patients with brain tumours to characterise overall network topology and individual patterns of connectivity alterations. Methods: Resting-state functional MRI data were acquired using multi-echo, echo planar imaging pre-operatively from five participants each with a right temporal–parietal–occipital glioblastoma. Complex networks analysis was initiated by parcellating the brain into anatomically regions amongst which connections were identified by retaining the most significant correlations between the respective wavelet decomposed time-series. Results: Key characteristics of complex networks described in healthy controls were preserved in these patients, including ubiquitous small world organization. An exponentially truncated power law fit to the degree distribution predicted findings of general network robustness to injury but with a core of hubs exhibiting disproportionate vulnerability. Tumours produced a consistent reduction in local and long-range connectivity with distinct patterns of connection loss depending on lesion location. Conclusions: Connectome analysis is a feasible and novel approach to brain mapping in individual patients with brain tumours. Applications to pre-surgical planning include identifying regions critical to network function that should be preserved and visualising connections at risk from tumour resection. In the future one could use such data to model functional plasticity and recovery of cognitive deficits.

Overall survival time prediction for high-grade glioma patients based on large-scale brain functional networks

“…While a tumor is growing, penetrating, and taking places with more surrounding brain tissues, its influence on the brain connectomics could widely spread. The brain may evolve and adapt to the neoplasm, leading to other indirect, plastic changes at "normal appearing" regions or even on the unaffected side (Hart et al 2016a;Maesawa et al 2015;Ghumman et al 2016;Wang et al 2010). Hart et al (2016b) suggested that different tumor locations might affect different functional networks.…”

“…By calculating FCs for all pairs of the brain regions, a largescale whole-brain functional network (functional connectomics) can be constructed, and the network properties can be quantified using graph-theoretic analysis (Van Den Heuvel and Pol 2010;Rubinov and Sporns 2010;Bullmore and Sporns 2009;Wee et al, 2012). Brain functional connectomics features have been extensively used to reveal abnormalities caused by various neurological and psychiatric diseases (Rosazza and Minati 2011;Fox and Greicius 2010), including brain tumors (Hart et al 2016a(Hart et al , 2016bAerts et al 2016;van Dellen et al 2013;Xu et al 2013;Huang et al 2014). These studies have demonstrated that local lesion(s) could potentially have a much wider influence on the large-scale brain networks, causing network topographical changes and other system-level consequences (Wang et al 2009b;Wang et al 2010;Zhang et al 2009;Tijms et al 2013).…”

“…Similarly, gliomas may also lead to topographical alterations to the whole-brain FC connectomics (Hart et al 2016b(Hart et al , 2016aAerts et al 2016;Ghumman et al 2016;Bartolomei et al 2006;Bosma et al 2009;Cochereau et al 2016). It is probable that different gliomas could cause distinct influences to the brain functional connectomics, resulting in different aftermaths from multi-dimensional behavioral abnormalities to varied survival time.…”

Multi-Atlas Segmentation of MR Tumor Brain Images Using Low-Rank Based Image Recovery

“…This is because pathologies can cause difficulties in registration of normal brain atlases to the target image. Despite this, atlas based image segmentation like MAS is still widely used to segment pathological brain images [8–10]. …”

Pre-operarative Multivariate Connectome Analysis for Glioma Patients

“…Despite advances in the new science of connectomics, techniques to directly study the function of white matter tracts in vivo have proved ambiguous. Cortical stimulation [7] and functional magnetic resonance imaging (fMRI) [12] are useful but constrained to identifying a focus of maximal activation. A purely anatomy-based approach such as tractography is challenging due to the definition of region of interest (ROI) and distortion of anatomical landmarks caused by lesions and brain plasticity [20].…”