The brain connectome as a personalized biomarker of seizure outcomes after temporal lobectomy

Abstract

Individual variations in connectome topography, combined with presurgical clinical data, may be used as biomarkers to better estimate surgical outcomes in patients with TLE.

Deep learning applied to whole‐brain connectome to determine seizure control after epilepsy surgery

“…In fact, a discriminant function using several relevant clinical variables in the presurgical assessment of TLE patients showed approximately 70% accuracy, which was further reduced when considering cross‐validation. The predictive accuracy of our deep‐learning model appears to be driven by a number of networks involving both temporal and extratemporal structures, as shown in Figure , and in line with prior findings using other statistical approaches . These widespread connections were both ipsilateral and contralateral to seizure focus, extending beyond the mesial temporal region thought to be pathologic.…”

“…We note that patients included in this study correspond to an ongoing cohort of individuals treated surgically for epilepsy at Medical University of South Carolina (MUSC). Individuals from this cohort have been reported in other group studies …”

Preoperative automated fibre quantification predicts postoperative seizure outcome in temporal lobe epilepsy

“…Although our sample is one of the largest to date that has investigated the relationship between preoperative DTI and postoperative seizure outcome (Bonilha et al , 2013, 2015; Ji et al , 2015; Keller et al , 2015 b ; Munsell et al , 2015), it is small in context of epidemiological studies of outcome, and therefore caution should be exercised when interpreting the relationship between clinical data and outcomes. We do report a significant effect of sex on outcome, with males being more likely to attain complete seizure freedom compared to females, which is consistent with other larger epidemiological studies (Burneo et al , 2006; Aull-Watschinger et al , 2008).…”

“…Although patients with TLE and neuroradiological evidence of hippocampal sclerosis have improved postsurgical outcomes relative to patients with TLE and no MRI lesion (Berkovic et al , 1995; McIntosh et al , 2004), between two-thirds and one-half of patients with hippocampal sclerosis will experience postoperative seizures (Berkovic et al , 1995; Janszky et al , 2005). Current suggestions for why these persistent postoperative seizures occur include a combination of insufficient resection of mesial temporal lobe tissue (Bonilha et al , 2004; Bonilha and Keller, 2015), mesial temporal lobe pathology existing outside the margins of resection (Babb et al , 1984; Holmes et al , 2000; Prasad et al , 2003; Keller et al , 2007), contralateral temporal lobe seizure involvement (Hennessy et al , 2000; Lin et al , 2005; Keller et al , 2007), occult extra-temporal lobe involvement, including temporal-plus epilepsy (Sisodiya et al , 1997; Ryvlin and Kahane, 2005; Kahane et al , 2015; Barba et al , 2016), structural network alterations (Bonilha et al , 2015; Keller et al , 2015 b ), and atypical subtypes of TLE that may be particularly resistant to conventional temporal lobe surgery (Blumcke et al , 2007; Thom et al , 2010; Bonilha et al , 2012). The development of predictive biomarkers for the future success of surgical intervention in epilepsy represents an important research endeavour, particularly as a reliable prognostic marker could inform patient clinical management and surgical decision-making.…”

“…This may be partly due to the fact that sophisticated DTI acquisitions are not incorporated into routine preoperative evaluation in a clinical setting. However, the application of graph theoretical methods to determine alterations in structural network topology is growing in TLE (Bernhardt et al , 2015), and there have been recent attempts to correlate preoperative structural connectomes with postoperative seizure outcome in small groups of patients with TLE (Bonilha et al , 2013, 2015; Munsell et al , 2015). Despite the interest in developing potential prognostic markers of outcome using preoperative connectomes, the underlying biological significance and anatomical specificity of such data are difficult to interpret.…”