Abstract P56: Diffusion Tensor Imaging at the Time of Stroke is Associated With Cognitive Performance 4 Months Later

Abstract

Introduction: Cognitive impairment after stroke is associated with stroke severity and baseline brain health. We hypothesized that acute diffusion tensor imaging (DTI) metrics would identify patients at risk for post-stroke cognitive impairment. Methods: Patients were enrolled prospectively in an observational study that involves serial MRI and cognitive testing in patients with recent stroke and moderate white matter disease on MRI but without dementia. DTI was performed at the time of stroke; cognitive testing with the MOCA and the Telephone Interview for Cognitive Status (TICS) were performed 3 months later. DTI was used to calculate Peak Skeletonized Mean Diffusivity (PSMD), a measure of global white matter microstructural integrity previously validated in cerebral small vessel disease. Fractional anisotropy maps were skeletonized (figure panel A) and a histogram of the corresponding MD values was used to calculate the peak width in the non-stroke hemisphere (panel B). Linear regression was used to test whether acute PSMD in the non-stroke hemisphere, acute stroke volume, or baseline NIHSS predicted cognitive performance 3 months later. Results: Fourteen patients followed-up at a median of 123 days. Patients had a median age of 73 years, mean baseline NIHSS of 1.2 (IQR 0-1.75), mean infarct volume of 4cc (range 0-16cc), mean MOCA of 25 (range 19-30), mean TICS of 33 (range 23-41), and 50% were women. Using multivariable linear regression, only acute PSMD predicted follow-up MOCA (std beta= -0.64, adj R 2 = 0.37, p= 0.013) while compared to baseline NIHSS, PSMD showed a stronger association with follow-up TICS score (std beta= -0.57 vs -0.44, p= 0.017; model adj R 2 = 0.476, p= 0.011)(Panel C). Conclusions: In this cohort of patients with small strokes we found that acute contralateral PSMD provided a measure of brain health that appears to predict cognitive performance at 3 months better than stroke size or severity. These are preliminary findings from an ongoing study.