Setting a gold standard for quantification of leukoaraiosis burden in patients with ischemic stroke: the Atherosclerosis Risk in Communities Study

J Neurosci Methods. 2014 Jan 15;221:196-201. doi: 10.1016/j.jneumeth.2013.10.009.


Background: Accurate and reliable measurement of leukoaraiosis, or MR-detected white, matter hyper-intensity (WMH) burden in subjects with acute ischemic stroke (AIS) is important for, ongoing research studies and future models of risk and outcome prediction, but the presence of a, cerebral infarct may complicate measurement. We sought to assess accuracy of a volumetric method, designed to measure WMH in AIS subjects as compared to the previously validated protocol.

New method: We randomly selected and equally sampled 120 brain scans from the Atherosclerosis, Risk in Communities (ARIC) MRI Study individuals within designated mild, moderate, and severe, tertiles of WMH volume (WMHV). T2 FLAIR axial images were analyzed using the AIS WMH volumetric, protocol and compared with the ARIC (gold standard) method. Pearson correlation coefficients, linear, concordance correlation coefficient, and Blant–Altman procedures were used to assess measurement, agreements between the two procedures.

Results: Median WMHV determined by using the ARIC method was 7.8 cm3 (IQR 5.7–13.55) vs. 3.54 cm3, (IQR 2.1–7.2) using the AIS WMH method. There was good correlation between the two measurements, (r = 0.52, 0.67, and 0.9 for tertiles 1, 2, and 3 respectively) (p < 0.001).

Comparison with existing method: The AIS WMH protocol was specific for leukoaraiosis in ischemic, stroke, but it appeared to underestimate WMHV compared to the gold standard method.

Conclusions: Estimates of MR-detectable WMH burden using a volumetric protocol designed for, analysis of clinical scans correlate strongly with gold standard measurements. These findings will, facilitate future studies of WMH in normal aging and in patients with stroke and other cerebrovascular, disease.

Publication types

  • Research Support, N.I.H., Extramural
  • Research Support, Non-U.S. Gov't

MeSH terms

  • Aged
  • Brain / pathology*
  • Female
  • Humans
  • Image Interpretation, Computer-Assisted / methods
  • Image Interpretation, Computer-Assisted / standards*
  • Leukoaraiosis / complications*
  • Leukoaraiosis / epidemiology*
  • Leukoaraiosis / pathology
  • Magnetic Resonance Imaging
  • Male
  • Middle Aged
  • Risk Factors
  • Stroke / complications*
  • Stroke / pathology