Small Vessel Disease Score Can Improve Dementia Risk Prediction


An analysis of pooled data from 3 longitudinal clinical trials suggests that SVD score, which can be compiled via a rapid visual assessment of clinical MRI scans, may be able to improve the prediction of dementia risk.

Dr Hugh Markus

Hugh S. Markus, FMedSci, professor of stroke medicine, department of clinical neurosciences, University of Cambridge

Hugh S. Markus, FMedSci

New study data suggest that a simple small vessel disease (SVD) score can aid the prediction of future dementia risk. The score can be compiled via a rapid visual assessment of clinical magnetic resonance imaging (MRI) scans.1

All told, the score improved the prediction of dementia (area under the curve [AUC], 0.85; 95% CI, 0.81—0.89) compared to clinical risk factors alone (AUC, 0.76; 95% CI, 0.71–0.81), and more severe SVD was associated with higher predictive performance. Power calculations showed selecting patients with a higher score reduced sample sizes required for hypothetical clinical trials by 40% to 66%, depending on the outcome measure used.

The study, conducted by Hugh S. Markus, FMedSci, professor of stroke medicine, department of clinical neurosciences, University of Cambridge, and colleagues, included a pooled analysis of 3 longitudinal cohort studies—SCANS (St George’s Cognition and Neuroimaging in Stroke; n = 121), RUN DMC (Radboud University Nijmegen Diffusion Imaging and Magnetic Resonance Imaging Cohort; n = 503), and ASPS (Austrian Stroke Prevention Study; n = 1218)—consisting of more than 1800 total patients.2-4

READ MORE: Small Vessels, Big Impact: The Cerebrovascular Condition You Should Watch For

“Whereas the results were broadly consistent across populations, the score appeared more predictive in studies with patients with more severe SVD, defined in SCANS as the presence of confluent WMH,” Markus and colleagues wrote, noting that several possible explanations exist. First, that the range MRI scores in ASPS were low, and that in cohorts with symptomatic SVD, dementia is likely to be related to the SVD, and be vascular in nature.

“In contrast, in population-based cohorts such as ASPS, many cases of dementia will be due to nonvascular causes such as Alzheimer disease, to which SVD will make a lesser contribution,” they detailed. “A clinical implication is that the MRI score will be most useful in patients who already have significant SVD, although in view of the association between WMH and Alzheimer disease it would be interesting to formally test its predictive value in this population.”

SVD scores ranged from 0—4, based on the MRI presence of lacunes, white matter hyperintensities (WMH), cerebral microbleeds (CMB), and perivascular spaces, independently, with each feature summed in an ordinal score. Simple SVD score, which accounted for the lack of perivascular space data in RUN DMC, ranged from 0–3. This simple SVD score was also further modified to include more information on the severity of MRI features. In this amended score, WMH were graded from 0–3 using the Fazekas scale, and the number of lacunar infarcts was graded from 0 to 3 (0 = none; 1 = 1 to 2; 2 = 3 to 5; 3 = >5), giving the amended SVD score a range from 0 to 7.

The simple SVD score, when added to a model to include variables such as age, sex, and education years improved the prediction, with AUC improving from 0.76 (95% CI, 0.72—0.80) to 0.81 (95% CI, 0.77–0.85) in a model with the simple SVD score (P = .098) and to 0.83 (95% CI, 0.80—0.87) with the amended score (P = .011).

When vascular dementia cases were used as the outcome, the prediction became slightly stronger (AUC, 0.85 95% CI, 0.81—0.89) with the simple score (P = .005), and also stronger (AUC, 0.86; 95% CI, 0.82—0.90) with the amended score (P = .002). The amended score appeared to increase the AUC in SCANS (0.71 vs 0.74) more so than it did for RUN DMC (0.81 vs 0.82) and ASPS (0.83 vs 0.84).

“The strengths of this study are that we determined whether SVD scores could predict progression to dementia in longitudinal studies,” Markus and colleagues concluded. “Furthermore, we replicated results across multiple cohorts and included cohorts with differing severities of SVD. This allowed us to determine whether the predictive value differed based on the severity of SVD.”


1. Al Olama AA, Wason JMS, Tuladhar AM, et al. Simple MRI score aids prediction of dementia in cerebral small vessel disease. Neurology. 2020;94:e1-e9. doi:10.1212/WNL.0000000000009141

2. Lawrence AJ, Patel B, Morris RG, MacKinnon AD, Rich PM, Barrick TR, Markus HS. Mechanisms of cognitive impairment in cerebral small vessel disease: multimodal MRI results from the St George's cognition and neuroimaging in stroke (SCANS) study. PLoS One. 2013;8(4):e61014. doi: 10.1371/journal.pone.0061014

3. van Uden IW, van der Holst HM, Tuladhar AM, et al. White Matter and Hippocampal Volume Predict the Risk of Dementia in Patients with Cerebral Small Vessel Disease: The RUN DMC Study. J Alzheimers Dis. 2016;49(3):863-73. doi: 10.3233/JAD-150573

4. Freudenberger P, Petrovic K, Sen A, et al. Fitness and cognition in the elderly: The Austrian Stroke Prevention Study. Neurology. 2016;86(5):418-24. doi: 10.1212/WNL.0000000000002329

Related Videos
Shahid Nimjee, MD, PhD
Peter J. McAllister, MD, FAAN
Video 6 - "Utilization of Neuroimaging in Alzheimer’s Disease"
Video 5 - "Contribution of Multiple Pathways to the Development of Alzheimer’s Disease"
Michael Levy, MD, PhD
Michael Levy, MD, PhD
Michael Kaplitt, MD, PhD
© 2024 MJH Life Sciences

All rights reserved.