London, United Kingdom 2013 Poster Session Red Cerebrovasc Dis 2013; 35 (suppl 3)1-854 531 459 Small vessel stroke and white matter disease Development and validation of a computational method to quantify perivascular spaces on MRI in cerebral small vessel disease. X. WANG1, M.C. Valdés Hernánde2, F.M. Chappell3, F.N. Doubal4, J.M. Wardlaw5 School of Medicine and Veterinary Medicine, University of Edinburgh, Edinburgh, UNIT-ED KINGDOM1, Brain Research Imaging Centre, University of Edinburgh, Edinburgh, UNIT-ED KINGDOM2, Brain Research Imaging Centre, University of Edinburgh, Edinburgh, UNITED KINGDOM3, Brain Research Imaging Centre, University of Edinburgh, Edinburgh, UNITED KINGDOM4, Brain Research Imaging Centre, University of Edinburgh, Edinburgh, UNITED KINGDOM5 Background: Perivascular spaces (PVS) are associated with ageing, cerebral small vessel disease, cognitive im-pairment, inflammation and possibly increased blood brain barrier permeability. Most studies to date use visual rating scales to assess PVS: these have limitations: including being prone to observer variation. A reliable computational method is desirable to quantify PVS precisely. Methods: We used 16 test cases from an ageing study to develop and test the method, including testing ob-server variability. This method chooses a standard slice at both the basal ganglia (BG) and centrum semiovale (CS) regions. Early test steps showed that CS measurements had low agreement and were hard to quantify, so further optimization used two ovoid BG regions of interest (ROI). This method used T2-weighted (T2W) MR images with a normalised intensity range. After manually selecting the BG ROIs, PVS were automatically extracted from 2 combined T2W normalised images using Analyze. The PVS volume and count were obtained using a maximum of 3 thresholds for all 16 cases. This method was then applied to 96 mild stroke patients’ images. All images also had visual PVS ratings. We used linear regression to test associations between BG PVS count, volume and vi-sual rating scores and coefficient of variation (CV) to evaluate model fit in 96 patients. Results: For the 16 test cases, the BG PVS count intra (mean=18.97, SD=5.67) and inter (mean=0.5, SD=4.87) observer agreement was good, as well as the PVS volume. In the 96 cases, PVS count increased significantly with PVS volume (0.0099, 95%CI 0.0085-0.0113, CV=30%). PVS volume also increased with PVS score (0.023, 95%CI 0.013-0.032, CV=48%). PVS count showed similar association to PVS volume with visual rating scores. Conclusion: This semi-automatic method is efficient in quantifying the PVS by applying one threshold in most cases and avoiding ceiling and floor effects. It shows good agreement with a visual rating scale. 460 Small vessel stroke and white matter disease The total burden of brain MRI markers of small vessel disease is associated with cognitive function in patients with cerebral small vessel disease M. Huijts1, A.A. Duits2, R.J. Van Oostenbrugge3, A.A. Kroon4, P.W. De Leeuw5, J. Staals6 Maastricht University Medical Centre, Maastricht, THE NETHERLANDS1, Maastricht Uni-versity Medical Centre, Maastricht, THE NETHERLANDS2, Maastricht University Medical Cen-tre, Maastricht, THE NETHERLANDS3, Maastricht University Medical Centre, Maastricht, THE NETHERLANDS4, Maastricht University Medical Centre, Maastricht, THE NETHERLANDS5, Maastricht University Medical Centre, Maastricht, THE NETHERLANDS6 Introduction – White matter lesions (WMLs), asymptomatic lacunar infarcts, brain microbleeds (BMBs) and enlarged perivascular spaces (EPVS) have been identified as silent lesions due to ce-rebral small vessel disease (cSVD). All these markers have been individually linked to cognitive functioning, but are also strongly correlated with each other. The combined effect of these markers on cognitive function has never been studied and would possibly provide more useful information on the effect on cognitive function. Methods – We included 189 patients with a high prevalence of cSVD (112 hypertensive patients and 77 first-ever lacunar stroke patients). Patients underwent brain MRI and extensive neuropsychological assessment. We rated the presence of any asymptomatic la-cunar infarct, extensive WMLs, any deep BMB, and moderate to extensive EPVS in the basal gan-glia. One point was awarded for the presence of each of these markers, with a minimum score of 0 and a maximum of 4. Associations with domains of cognitive function were analyzed with correla-tion analyses. Results – Correlation analyses revealed significant associations between cSVD cate-gory and all cognitive domains (all </=.001). Results remained significant for information process-ing speed (r=-.181, p=.013) and overall cognition (r=-.178, p=.017), after correction for age and sex. Testing of trend using linear regression analyses revealed the same results. Discussion – We demonstrated that after adjustments for age and sex, a higher total burden of cSVD is associated with decreased performance on tests of information processing speed and overall cog-nition.
Karger_ESC London_2013
To see the actual publication please follow the link above