Introduction: Glioblastomas are the most common primary malignant brain tumour. Concurrent temozolominde and radiation therapy-STUPP protocol, gives a median survival of 14.6 months, however responses to treatment and prognosis are variable. There is limited literature on whether analysis of molecular features and MRI parameters combined would advance prognosis estimation. In this study, six MRI parameters were investigated in glioblastomas divided into five molecular subtypes as identified by our research group. Three of the five groups were positive for O(6)-methylguanine-DNA methyltransferase (MGMT)..

Methods: Pre-surgical MRI of 81 retrospectively selected patients with a histological diagnosis of glioblastoma was studied. Analysed MRI radiological features included: tumour margin regularity, cystic component, necrosis, haemorrhage, peritumoral oedema grade (I–III) and unifocal versus multifocality. Two neuroradiologists, blinded from patients’ treatment regime, length of survival and tumour molecular subtype, read the images independently. The cases were grouped into 5 molecular subtypes according to immunohistochemical status.

Results: Tumour with regular margin (p<0.01), cystic component (p<0.001), no necrosis (P<0.001) and low-grade peritumoral oedema (p<0.001) were features of good survival; tumours with necrosis (P<0.001), haemorrhage (p<0.05), and high-grade peritumoral oedema grade II or III (p<0.0001) were features of poor survival, provided the tumour was MGMT negative. In the MGMT positive subtypes, no differences in survival were found based on radiological features. Interestingly MGMT positivity is a poor prognostic factor irrespective of TMZ treatment. Temozolomide treatment and radiation therapy have been considered as potential confounding factors.

Conclusions: Radiological features predict survival in MGMT negative, but not MGMT positive glioblastomas. Tumour regular margin, cystic component and low-grade peritumoral oedema are associated with longer survival, whereas tumour necrosis, haemorrhage and high grade peritumoral oedema were negative prognostic features in MGMT negative subtypes.

Patient Care: Better understanding of GBMs radiologically to predict tumour behaviour and classification

Learning Objectives: To test the associations between radiological features on MRI and tumour molecular subtype. Further, to identify features that predict patient prognosis.

References: 1. Antonio Omuro, MD1; Lisa M. DeAngelis, MD1 Glioblastoma and Other Malignant GliomasA Clinical Review JAMA. 2013;310(17):1842-1850. doi:10.1001/jama.2013.280319. 2. Stupp R1, Mason WP, van den Bent MJ, Weller M, Fisher B, Taphoorn MJ, Belanger K Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. N Engl J Med. 2005 Mar 10;352(10):987-96. 3. American Brain Tumour Association 2014