Related Experiment Video
Updated: Dec 27, 2025
Assessing Cortical Cerebral Microinfarcts on High Resolution MR Images
Published on: November 20, 2015
Simple MRI score aids prediction of dementia in cerebral small vessel disease
Ali Amin Al Olama1, James M S Wason1, Anil M Tuladhar1
1From the Stroke Research Group (A.A.A.O., H.S.M.), Clinical Neurosciences, University of Cambridge; MRC Biostatistics Unit (J.M.S.W.), Institute of Public Health, Cambridge; Institute of Health and Society (J.M.S.W.), Newcastle University, UK; Department of Neurology (A.M.T., E.M.C.v.L., F.-E.d.L.), Radboud University Nijmegen Medical Centre, Donders Institute for Brain, Cognition and Behaviour, Centre for Medical Neuroscience, Nijmegen, the Netherlands; Division of Neurogeriatrics (M.K., E.H., R.S.), Department of Neurology, Medical University of Graz; Institute for Medical Informatics (E.H.), Statistics and Documentation, Medical University of Graz, Austria; and Department of Psychology (R.G.M.), King's College, Institute of Psychiatry, Psychology and Neuroscience, London, UK.
Objective:
To determine whether a simple small vessel disease (SVD) score, which uses information available on rapid visual assessment of clinical MRI scans, predicts risk of cognitive decline and dementia, above that provided by simple clinical measures.
Methods:
Three prospective longitudinal cohort studies (SCANS [St George's Cognition and Neuroimaging in Stroke], RUN DMC [Radboud University Nijmegen Diffusion Imaging and Magnetic Resonance Imaging Cohort], and the ASPS [Austrian Stroke Prevention Study]), which covered a range of SVD severity from mild and asymptomatic to severe and symptomatic, were included. In all studies, MRI was performed at baseline, cognitive tests repeated during follow-up, and progression to dementia recorded prospectively. Outcome measures were cognitive decline and onset of dementia during follow-up. We determined whether the SVD score predicted risk of cognitive decline and future dementia. We also determined whether using the score to select a group of patients with more severe disease would reduce sample sizes for clinical intervention trials.
Results:
In a pooled analysis of all 3 cohorts, the score improved prediction of dementia (area under the curve [AUC], 0.85; 95% confidence interval [CI], 0.81-0.89) compared with that from clinical risk factors alone (AUC, 0.76; 95% CI, 0.71-0.81). Predictive performance was higher in patients with more severe SVD. Power calculations showed selecting patients with a higher score reduced sample sizes required for hypothetical clinical trials by 40%-66% depending on the outcome measure used.
Conclusions:
A simple SVD score, easily obtainable from clinical MRI scans and therefore applicable in routine clinical practice, aided prediction of future dementia risk.
