Correlation of MRI sequences to assess progressive glioblastoma multiforme treated with bevacizumab.

Published

Journal Article

In the context of bevacizumab therapy for the treatment of progressive malignant gliomas, it is currently unclear how different magnetic resonance imaging (MRI) sequences correlate with each other over time. The objective of this study was to determine if a reliable and predictable relationship over time exists between post-gadolinium based contrast agent T1-weighted (T1 + GBCA), T2-weighted, and FLAIR MRI in patients with progressive glioblastoma multiforme (GBM) receiving bevacizumab and chemotherapy. The MRI lesion volumes of 10 patients with progressive GBM that received bevacizumab plus chemotherapy were manually calculated by two independent reviewers. T2 and FLAIR volumes were analyzed by analysis of covariance (ANCOVA) as a function of T1 + GBCA lesion enhancement volume, reviewer, and time interval between MRI acquisitions. Pearson product moment correlation (r) was used to compare pre and post treatment volumes for the group of 10 patients and for each individual patient over their treatment course. ANCOVA demonstrated a significant association between T1 + GBCA and T2-weighted volumes (P = 0.0006) and between T1 + GBCA and FLAIR volumes (P < 0.0001). These associations remained constant over time. The correlation between T1 + GBCA and both T2-weighted and FLAIR volumes improved after bevacizumab treatment. Individual correlations between T1 + GBCA and FLAIR were strong (r ≥ 0.63) with one exception, while correlations between T1 + GBCA and T2 were more variable (r = 0.18-0.99). These findings suggest that FLAIR MRI should be evaluated in addition to T1 + GBCA MRI when evaluating GBM responses.

Full Text

Duke Authors

Cited Authors

  • Thompson, EM; Dosa, E; Kraemer, DF; Neuwelt, EA

Published Date

  • June 2011

Published In

Volume / Issue

  • 103 / 2

Start / End Page

  • 353 - 360

PubMed ID

  • 20848300

Pubmed Central ID

  • 20848300

Electronic International Standard Serial Number (EISSN)

  • 1573-7373

Digital Object Identifier (DOI)

  • 10.1007/s11060-010-0397-0

Language

  • eng

Conference Location

  • United States