Predicting Post-Stroke Cognitive Impairment Using Lesion Location

Abstract

<h3>Research Objectives</h3> To develop a model to predict specific cognitive impairments following stroke based on the lesion's white matter edge density and location with respect to the left posterior arcuate fasciculus. <h3>Design</h3> Retrospective observational cohort study. <h3>Setting</h3> University medical center, inpatient and outpatient. <h3>Participants</h3> 149 adults (> 18 years old at time of stroke) from the Iowa Lesion Patient Registry. Participants must have had a stroke, an MRI or CT with visible focal brain lesion, and a neuropsychological assessment completed at the medical center. The cohort was 50% female, predominantly White (83%), upper middle-aged (mean age of 59 years old), and had an average of 13 years of education. <h3>Interventions</h3> Not applicable. <h3>Main Outcome Measures</h3> Cognitive impairment was quantified using z-scores for the following neuropsychological assessments, selected based on completion by >50% of participants Trail Making Test B, five subtests of the Wechsler Adult Intelligence Scale-IV, Benton Visual Retention Test, Rey Auditory Verbal Learning Test, Complex Figure Test, and two subtests of the Boston Diagnostic Aphasia Examination. The proportion of impaired tests per participant was also calculated. Lesion overlap with the left posterior arcuate fasciculus (LPAF) and lesion edge density were determined by summation of voxels. <h3>Results</h3> There was a significant though small positive correlation between the proportion of impaired z-scores and lesion overlap with the LPAF (r = 0.33, p <.001), indicating that greater lesion overlap corresponds to impairment on a greater number of neuropsychological test scores. The correlation between edge density and proportion of impaired scores was small and borderline significant (r=0.17, p=0.05). For individual neuropsychological tests, participants with lesion overlap to the LPAF had significantly lower scores than non-overlap participants on the majority (10 of 16) of assessed scores, and edge density was significantly negatively correlated to 7 of the 16 test scores. <h3>Conclusions</h3> Lesion overlap with the left posterior arcuate fasciculus has the potential to be used as a predictive marker of cognitive impairment in a post-stroke cohort and may be a stronger measure of cognitive impairment than edge density. <h3>Author(s) Disclosures</h3> No disclosures.