Visual Rating Scales of White Matter Hyperintensities and Atrophy: Comparison of Computed Tomography and Magnetic Resonance Imaging

J Stroke Cerebrovasc Dis. 2018 Jul;27(7):1815-1821. doi: 10.1016/j.jstrokecerebrovasdis.2018.02.028. Epub 2018 Mar 23.


Goal: Magnetic resonance imaging (MRI) is the preferred modality for research on structural age-related brain changes. However, computed tomography (CT) is widely available and has practical and cost advantages over MRI for large-scale brain imaging research studies in acutely unwell patients. However, the relationships between MRI and CT measures of white matter hyperintensities (WMH) and atrophy are unclear. We examined the relationships between visual ratings of WMH, atrophy, and old infarcts in patients who had both CT and MRI scans.

Materials and methods: Patients who had both CT and MRI scans in the International Stroke Trial-3 were studied. In both modalities, 2 raters independently completed standardized visual rating scales for WMH, and for central and superficial atrophy using a 5-point scale. In addition, 1 rater recorded old infarcts according to size and location.

Findings: Seventy patients with a mean age of 69 years were studied. There were moderate to substantial intrarater CT-MRI agreements for periventricular components of WMH scales (weighted Κappa = .55-.75). Agreements for basal ganglia ratings were lower (weighted Κappa = .18-.44), partly because of the misclassification of prominent perivascular spaces. Atrophy scales showed moderate to substantial CT-MRI agreements (weighted Κappa = .44-.70). MRI was more sensitive in the detection of smaller infarcts and cavitated lesions.

Conclusions: Standardized visual rating scales of white matter lesions and atrophy mostly show substantial agreement between CT and MRI. Clinical CT scans have a strong potential for wider exploitation in research studies, particularly in acutely unwell populations.

Keywords: Computed tomography; cerebral atrophy; magnetic resonance imaging; validity; white matter lesions.

Publication types

  • Comparative Study
  • Multicenter Study
  • Randomized Controlled Trial

MeSH terms

  • Aged
  • Brain / diagnostic imaging*
  • Brain / pathology
  • Humans
  • Magnetic Resonance Imaging*
  • Observer Variation
  • Tomography, X-Ray Computed*
  • White Matter / diagnostic imaging*
  • White Matter / pathology