Predicting outcome of patients with high-grade gliomas after radiotherapy using quantitative analysis of T1-weighted magnetic resonance imaging

Int J Radiat Oncol Biol Phys. 2007 Apr 1;67(5):1476-83. doi: 10.1016/j.ijrobp.2006.11.020. Epub 2007 Feb 2.

Abstract

Purpose: The aim of this study was to test the hypothesis that measuring quantitative changes in signal intensity early after radiotherapy (RT) in the contrast-enhancing tumor rim and nonenhancing core may be a noninvasive marker of early treatment response in patients with high-grade gliomas.

Methods and materials: Twenty patients with high-grade gliomas had magnetic resonance imaging (MRI) performed 1 week before RT, during Weeks 1 and 3 of RT, and every 1 to 3 months after RT as part of a clinical prospective study. Regions of interest (ROI) including contrast-enhancing rim, and the nonenhancing core were defined automatically based on a calculated image of post- to precontrast T1-weighted MRI. Pretreatment T1-weighted MRI signal intensity changes were compared with Weeks 1 and 3 RT and 1 and 3 months post-RT MRI. Clinical and MRI parameters were then tested for prediction of overall survival.

Results: Regional T1-weighted signal intensity changes in both the contrast-enhancing rim and the nonenhancing core were observed in all patients during Week 1 and Week 3 of RT. Imaging parameters including signal intensity change within the nonenhancing core after Weeks 1 to 2 RT (p = 0.004), Weeks 3 to 4 RT (p = 0.002) and 1 month after completion of RT (p = 0.002) were predictive of overall survival. Using multivariate analysis including RTOG recursive partitioning analysis (RPA) and signal intensity change, only the signal intensity change in the nonenhancing core at 1 month after RT (p = 0.01) retained significance.

Conclusion: Quantitative measurements of T1-weighted MRI signal intensity changes in the nonenhancing tumor core (using ratios of pre-post values) may provide valuable information regarding early response during treatment and improve our ability to predict posttreatment outcome.

Publication types

  • Research Support, N.I.H., Extramural

MeSH terms

  • Adult
  • Aged
  • Brain Neoplasms / pathology
  • Brain Neoplasms / radiotherapy*
  • Female
  • Glioma / pathology
  • Glioma / radiotherapy*
  • Humans
  • Magnetic Resonance Imaging*
  • Male
  • Middle Aged
  • Proportional Hazards Models
  • Prospective Studies
  • Radiotherapy, Conformal
  • Treatment Outcome