Edinburgh Imaging

MSc projects 1617 004

Are MRI features of glioblastoma multiforme indicative of molecular subtype? Radiogenomics of GBM.

  • Background: Glioblastoma Multiforme (GBM) is a highly aggressive and near uniformly fatal brain tumour due to its heterogeneous cellular composition and invasiveness. The United Nations agency for health is called the World Health Organisation (WHO). The 2016 WHO update has introduced molecular markers within the classification of GBM. The molecular markers distinguish types of GBM with different treatment sensitivities and prognosis. The classification has introduced isocitrate dehydrogenase (IDH) mutation as a classifier of GBM. There is potential for expansion of future classifications with the presence of further molecular markers such as MGMT. As knowledge increases of these histopathological tests, imaging biomarkers are being sought to provide correlation.  This literature review assesses quality and clinical relevance of available evidence in context of imaging assessment of molecular markers. Future research needs will be identified from gaps in the literature.
  • Methods: PubMed and MEDLINE were utilised to identify all literature relevant to this question. The population selected were patients with GBM, the intervention selected was MRI and the outcome whether the intervention was indicative for IDH mutation (IDHm), methylated MGMT (mMGMT), 1p19q codeletion, ATRX mutation, and Histone3 K27M mutation. The literature that met all these requirements was included and assessed.
  • Results: Significant imaging features for IDH wildtype (compared to IDHm) include >33% of oedema, increased contrast enhancement, invasiveness, necrosis, tumour blood flow, and presence of multi-enhancing foci. Studies analysing gliomas (including GBM) have found very promising results using 2-hydroxyglutarate (2HG) spectroscopy. This modality may provide a clinical diagnostic tool for IDH status within GBMs. mMGMT shows a predisposition for the left hemisphere and less peripheral enhancement. All other visually assessable features analysed on MRI did not show significant results. Using perfusion metrics has yielded conflicting results and currently cannot be reliably diagnostic for mMGMT.
  • Conclusion: Presently research directed towards radiogenomics using non-invasive MRI is a new and promising field. Research aimed at the highly aggressive GBM is lagging behind studies investigating lower grade gliomas. Qualitative features in keeping with a more aggressive phenotype suggest a lack of IDHm status. More research is needed into GBM, IDH and 2HG spectroscopy as this may provide clinicians with an immediately deployable and highly relevant diagnostic tool. Further investigation into texture analysis and mMGMT may prove significant.  Studies are beginning to look at 1p19q codeletion, Histone3 K27M and ATRX in relation to imaging and low-grade gliomas.
Project type:
  • Analysis of acquired data
  • Literature review
Imaging keywords:
Application / disease keywords:
  • Glioblastoma multiforme (GBM)
  • Gliomas
  • 16-17