Abstract 129: Predicting Patients Who Will Have Cognitive Decline After Carotid Endarterectomy or Stenting Using Structural Connectivity Graph Analysis
Objectives: Many CEA and CAS patients experience postoperative neurocognitive decline. We sought to apply structural connectivity metrics to identify patients at increased risk for postoperative decline based solely on preoperative imaging.
Methods: Under an IRB approved protocol 28 patients underwent presurgical T1 structural and 30 direction diffusion tensor imaging (DTI) MRI and neuropsychological tests before and 1 month after surgery. Patients with decline showed decreased performance on the Rey-AVLT on 1 month follow up. The T1 images were processed using FreeSurfer 5.3, with resulting segmentations reviewed and edited as needed under neuroradiologist supervision. Whole brain tractography was performed using Diffusion Toolkit and visually inspected. Connectivity matrices were then generated, and graph metrics were computed using the Brain Connectivity Toolbox.
Results: Controlling for age, classifiers using the graph analysis metrics “weighted optimal community structure” & “binary component sizes” were able to identify patients that would experience cognitive decline with 81% sensitivity 83% and specificity (p<.05, false discovery rate .05). These two measures were computed at 10 proportion edge thresholds from .1 to 1 at intervals of .1 in weighted and binary networks respectively.
Conclusions: Applying preoperative structural connectivity analysis in CEA and CAS patients may identify patients at increased risk for postoperative cognitive decline, and in so doing may help risk stratify patients and guide them to preventive interventions.
Author Disclosures: S. Soman: None. G. Prasad: None. E. Hitchner: None. A. Rosen: None. W. Zhou: None.
- © 2014 by American Heart Association, Inc.