T2 MRI texture analysis is a sensitive measure of tissue injury and recovery resulting from acute inflammatory lesions in multiple sclerosis

Neuroimage. 2009 Aug 1;47(1):107-11. doi: 10.1016/j.neuroimage.2009.03.075. Epub 2009 Apr 8.

Abstract

T2 hyperintensity is pathologically non-specific in multiple sclerosis (MS) yet lesion analysis remains an important disease outcome. Texture analysis based on the polar Stockwell Transform (PST) was performed on twelve acute T2 lesions before, during, and after the development of gadolinium-enhancement. When regular myelin structure is disrupted coarse texture increases because tissue becomes disorganized. Coarse texture was quantified as the sum of low frequency energy (sumLFE). Matching regions of contralateral and general normal appearing white matter (NAWM) and chronic T2 lesions were analyzed in parallel as controls. Coarse texture increased in acute lesions during enhancement (p<0.05) then variably recovered. It remained stable in NAWM and tended to increase in chronic T2 lesions. Seven of twelve acute lesions persisted visually at 8 months and the sumLFE was higher in these visually persistent lesions (p<0.05) than in resolved lesions. The sumLFE at month 8 correlated with that in pre-lesional NAWM and in acute lesions (p<0.05) and was independent of lesion volume, signal intensity (SI), and location. This study suggests that PST texture analysis extracts more information about tissue integrity than conventional MRI analysis that relies on lesion size and SI. Texture analysis also appears to identify abnormalities in pre-lesional NAWM, to measure tissue injury in acute lesions, predict poor recovery, and detect mild ongoing tissue injury in chronic T2 lesions. PST texture analysis using conventional MRI may therefore provide valuable new insights into lesion pathology by measuring tissue integrity. This small longitudinal study supports further validation of the PST technique.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Adult
  • Analysis of Variance
  • Brain / pathology*
  • Female
  • Humans
  • Longitudinal Studies
  • Magnetic Resonance Imaging / methods*
  • Male
  • Middle Aged
  • Multiple Sclerosis, Relapsing-Remitting / diagnosis
  • Multiple Sclerosis, Relapsing-Remitting / pathology*
  • Prognosis
  • Signal Processing, Computer-Assisted*
  • Young Adult