Combining Neurovascular and Neurodegenerative Magnetic Resonance Imaging Measures in Stroke.

Background and Purpose- Individual markers of cerebral small vessel disease and cerebral atrophy explain a small proportion of variance in vascular risk factors and cognitive function. Combining these markers into a single measure of neurovascular and neurodegenerative disease may be more powerful. We assessed this using data contained in the Virtual International Stroke Trials Archive - Prevention sub-archive. Methods- We extracted white matter hyperintensities (WMH) and cerebrospinal fluid (CSF) volumes from 317 people with ischemic stroke or transient ischemic attack who had baseline magnetic resonance imaging. We assessed progression of volumes in 208 people who had 2-year follow-up magnetic resonance imaging. WMH and CSF volumes were segmented from fluid attenuated inversion recovery and T1 images. The combined neurovascular and neurodegenerative measure was the sum of WMH and CSF volume normalized by intracranial volume. We assessed (1) the relationship between baseline vascular risk factors and imaging markers; and (2) the relationship between baseline imaging markers and Mini-Mental State Examination score at follow-up using multiple linear regression. We also assessed implications for sample size calculations using n=208 participants with follow-up magnetic resonance imaging. Results- Vascular risk factors accounted for 7%, 11%, and 12% of the variance in WMH, CSF, and combined volume, respectively (all P<0.001). The association between baseline combined volume and 6-month follow-up Mini-Mental State Examination (ß=-0.442; SE, 0.07; P<0.0001) was 32% greater than WMH (ß=-0.302; SE, 0.06; P<0.0001) and 12% greater than CSF (ß=-0.391; SE, 0.07; P<0.0001) alone. The combined volume required between 207 and 3305 (20%) fewer patients per arm than WMH alone to detect reductions of 10% to 40% in volume progression over 2 years. Conclusions- A combined neurovascular and neurodegenerative magnetic resonance imaging measure including WMH and CSF volume was more closely related to vascular risk factors and cognitive function than either WMH or CSF volume alone. The combined volume may be a more sensitive measurement for clinical trials.

Cortical thickness, white matter hyperintensities, and cognition after stroke.

BACKGROUND: A thinner cerebral cortex is associated with higher white matter hyperintensity burden and cognitive impairment in community-dwelling and dementia cohorts. It is important to assess these associations in people with ischemic stroke because their cerebrovascular disease profiles are different to these cohorts. AIMS: We aimed to determine whether cortical thickness was related to white matter hyperintensity burden and cognition after ischemic stroke. METHODS: We measured cortical thickness using advanced normalization tools' "KellyKapowski" function in 244 patients with ischemic stroke or transient ischemic attack from the Virtual International Stroke Trials Archive. We measured white matter hyperintensity burden via quantitative volumes and Fazekas score. We extracted data on vascular risk factors at baseline and Mini Mental State Examination scores at one year. We assessed associations between imaging and clinical data using correlation and multiple linear regression. RESULTS: Pairwise correlation showed that higher white matter hyperintensity Fazekas score was associated with a thinner cortex (rho = -0.284, P < 0.0001). White matter hyperintensities were generally distributed adjacent to and above the lateral ventricles. Voxel-wise analyses showed statistically significant negative associations between cortical thickness and white matter hyperintensities across fronto-temporal and inferior parietal cortical regions. Mean cortical thickness was positively related to Mini Mental State Examination in pair-wise correlation (r = 0.167, P = 0.0088) but there was no independent association after adjustment for age and white matter hyperintensities (beta = 0.016, P = 0.7874). CONCLUSIONS: Cortical thickness was not an independent predictor of cognition after ischemic stroke. Further work is required to understand how white matter hyperintensities are associated with a thinner cortex in temporal regions but less so in more superior regions where white matter hyperintensities are generally found in people with stroke.