OP04. PREDICTING SITES OF TREATMENT FAILURE AND TUMOUR RECURRENCE IN THE INVASIVE MARGIN OF GLIOBLASTOMAS (PRAM-GBM) - TRIAL

AIM: The invasive nature of glioblastomas ensures that residual disease is inevitable following surgery. As this cannot be identified with conventional imaging, radiotherapy volumes must include normal brain to treat this occult invasion. We need an imaging biomarker to identify this residual tumour and predict the site of tumour progression. Such a biomarker would redefine surgical targets and allow personalisation of radiotherapy treatment volumes. Diffusion tensor imaging (DTI) is one potential biomarker as it can identify a larger abnormality than conventional MRI and biopsies of these abnormalities confirm invasive tumour. This study aims to qualify an imaging biomarker that can accurately predict sites of glioblastoma progression, allowing personalisation of radiotherapy and surgical targets. METHOD: We have developed a multicentre, pragmatic, longitudinal observational cohort study of 120 patients from 6 centres. Patients with a glioblastoma suitable for radical resection and chemoradiotherapy, aged 16–75 years with good performance status and no contraindication to contrast-enhanced MRI will be recruited. Patients will be imaged pre-operatively and pre-radiotherapy with DTI and perfusion MR. Primary endpoint is the diagnostic accuracy of DTI to predict site of progressive tumour. Secondary endpoints will explore time-to-progression, location of DTI/perfusion abnormalities in relation to radiotherapy treatment volumes, potential for extended surgical resection and comparison of central vs. local analysis of data. RESULT: Our previous work suggests DTI can identify a larger abnormality than conventional MRI and biopsies of these abnormalities confirm invasive tumour. Follow up studies show it predicts the site of tumour progression with 89% sensitivity. Interestingly the extent of resection appears to be a key contributor to the pattern of recurrence. In a small study 6 patients (40%) had residual contrast enhancement on immediate post-operative MRI. This group appeared to progress sooner than the complete resection group (PFS 283 days vs. 507 days; P2cm from the edge of the resection cavity. Technical development of DTI as a biomarker has shown that it can be acquired on any MR system at both 1.5 and