Evidence for potentials and limitations of brain plasticity using an atlas of functional resectability of WHO grade II gliomas: Towards a “minimal common brain”

Abstract

Despite recent advances in non-invasive brain mapping imaging, the resectability of a given area in a patient harboring a WHO grade II glioma cannot be predicted preoperatively with high reliability, due to mechanisms of functional reorganization. Therefore, intraoperative mapping by direct electrical stimulation remains the gold standard for detection and preservation of eloquent areas during glioma surgery, because it enables to perform on-line anatomo-functional correlations. To study potentials and limitations of brain plasticity, we gathered 58 postoperative MRI of patients operated on for a WHO grade II glioma under direct electrical cortico-subcortical stimulation. Postoperative images were registered on the MNI template to construct an atlas of functional resectability for which each voxel represents the probability to observe residual non-resectable tumor, that is, non-compensable area. The resulting atlas offers a rigorous framework to identify areas with high plastic potential (i.e. with probabilities of residual tumor close to 0), with low compensatory capabilities (i.e. probabilities of residual tumor close to 1) and with intermediate level of resectability (probability around 0.5). The resulting atlas highlights the utmost importance of preserving a core of connectivity through the main associative pathways, namely, it supports the existence of a “minimal common brain” among patients.

Introduction

Current developments in functional mapping and neuroimaging techniques have radically changed the classical static view on the functional organization of cortical areas, for a new dynamic perspective of the brain (Duffau, 2005). Indeed, many recent investigations have highlighted the dynamic capability of the brain to reorganize itself, both during everyday life (i.e. learning) and after a pathological event (e.g. stroke or glioma). This reorganization would be based on the existence of multiple and overlapping redundancies hierarchically organized (Bavelier and Neville, 2002, Duffau, 2001, Duffau et al., 2000, Rossini et al., 2003, Sanes et al., 1995, Schieber and Hibbard, 1993). These findings have testified that neuronal aggregates, beside or outlying a lesion, can increasingly adopt the function of the damaged area and switch their own activation pattern to substitute the lesioned area while facilitating functional recovery following brain damage (Duffau, 2006a). Advances in neuroimaging have enabled a better comprehension of the dynamic interaction between a tumor and functional cortical sites, usually preoperatively assessed by non-invasive functional examinations such as positron emission tomography (PET), functional magnetic resonance imaging (fMRI), MRI-based diffusion tensor imaging (DTI) and magnetoencephalography (MEG). Nevertheless, despite efforts to improve these techniques, their sensitivity and specificity are still limited due to perturbations induced by tumor on local neurovascular and metabolic coupling (sensitivity for the identification of sensorimotor sites ranges from 82% to 100%, whereas it ranges from 66% to 100% for language sites) (Aubert et al., 2002, Bartos et al., 2009, Roux et al., 2003) and neuroimaging is not able to differentiate essential cortical areas (which should be surgically preserved) from the “modulatory” areas that can be functionally compensated and resected without inducing permanent deficits (Duffau et al., 2003).

Axonal pathways also play a crucial role in glioma surgery, considering their infiltrative growth patterns along white matter fiber tracks (Chen et al., 2010, Mandonnet et al., 2006, Pallud et al., 2005). Recent developments in DTI have allowed to track non-invasively in vivo subcortical fibers (Catani et al., 2002, Catani and Thiebaut de Schotten, 2008) providing information on displacements, infiltrations or disruptions of fibers induced by the tumor (Witwer et al., 2002). Nevertheless, tracking algorithms may strongly influence the anatomical data of DTI (Kinoshita et al., 2005), even if some reports have provided some validation on postmortem studies (Thiebaut de Schotten et al., 2011; Lawes et al., 2008). Finally, DTI is not yet able to highlight the functional role of the tracts.

Considering (1) the large variability in structural and functional networks among healthy volunteers (Brett et al., 2002, Tzourio-Mazoyer et al., 2004), (2) functional limitations in neuroimaging, and (3) functional modifications induced by tumoral growing patterns both at cortical and axonal levels (Duffau, 2006a), the study of the brain functional cortical organization and connectivity is needed for individual patients to both select the best indications for surgery and to perform a resection with the optimal benefit/risk ratio. As a consequence, the use of intraoperative direct electrical stimulation (DES) is considered as the “gold standard” to detect both the eloquent cortical areas and subcortical pathways at the individual level (Duffau et al., 2008b, Mandonnet et al., 2010b). Indeed, DES provides accurate and real-time data on the distribution not only of the cortical eloquent areas (Ojemann et al., 1989, Sanai et al., 2008), but also of the functional white matter bundles (Bello et al., 2007, Duffau et al., 2008b, Sanai and Berger, 2010). Thus, DES allows to tailor the tumoral resection according to individual functional boundaries, maximizing the extent of resection while minimizing the risk of permanent neurological deficits.

Combining intraoperative anatomofunctional data with pre and post-operative fMRI and DTI imaging is currently the best approach to assess the functional role of the cortical areas and the white matter fiber tracts (Kamada et al., 2007). For this reason, we propose in this paper the elaboration of a probabilistic postsurgical residue atlas computed on a series of patients who underwent incomplete tumor resection on the basis of intraoperative DES brain mapping. The anatomo-functional correlations we obtained by combining the DES data with postoperative anatomical MRI findings will provide a greater understanding of the functional limits of surgical removal, and will provide new insights into the potentials and limitations of brain plasticity. Especially, this probabilistic atlas highlights the crucial role of the axonal pathways in the reshaping and reorganization of the brain after a lesion. Finally, beyond its fundamental interest, we hope this atlas will be an essential tool for surgery planning, by allowing an objective pre-operative estimation of the expected extent of the resection.