PhD Markus D. Schirmer | Spatial effects of white matter hyperintensity disease burden from clinical stroke populations

Guest Lecture

  • Date: Dec 13, 2018
  • Time: 01:00 AM - 02:00 AM (Local Time Germany)
  • Speaker: PhD Markus D. Schirmer
  • J. Philip Kistler Stroke Research Centre, Massachusetts General Hospital, USA Harvard Medical School, Boston, USA
  • Location: MPI for Human Cognitive and Brain Sciences
  • Room: Wilhelm Wundt Room (A400)
  • Host: Department of Neurology
The identification of biomarkers. which can help predict disease outcome. remains one of the most promising research areas across a variety of diseases. Particularly. studying the spatial distribution of underlying disease burden may provide important insights into pathological patterns. Stroke is one of the leading causes of death and disability worldwide. however. it remains largely understudied in the acute setting due to time restrictions during data acquisition and the resulting low resolution magnetic resonance images. This has made outcome prediction particularly difficult. as it leads to heterogeneity in the data and/or methodologies between studies. One of the key phenotypes of stroke research is a patient's white matter hyperintensity (WMH) burden. most commonly assessed using FLAIR images. which has been linked to stroke outcome. Studies often rely on measuring the total burden and summarize it as a single number: it’s volume. based on manual outlines on a patient’s scan. The lack of an automated segmentation methodology for clinical data has so far hindered large-scale. reproducible investigations. In the first part of this talk. we present an automated pipeline for monomodal automatic WMH segmentation. which can alleviate some of these challenges. Specifically. we demonstrate it's efficacy for volume estimation in a cohort of 2.533 patients. showing the association between higher WMH burden and poorer outcome after stroke (p<0.001). In the second half of this talk. we demonstrate the use of WMH segmentations for investigating spatial WMH disease burden and how other clinical variables can modify these patterns. In particular. we demonstrate effects for hypertension and smoking status. and show that these clinical variables lead to a shift of disease burden from posterior to anterior vascular regions (p<0.05 and p<0.01. respectively). This illustrates the potential of uncovering spatial variations of disease patterns by using large-scale cohorts.
Go to Editor View