Pipeline for Measuring Residual Tumour Volume in Patients With Glioblastoma Based on a Diffusion Tensor Tissue Signature

AIMS Glioblastoma is the most common malignant primary brain tumour in adults. It has a poor prognosis despite surgery and chemoradiotherapy. Extent of resection is an important modifiable prognostic factor. The latest Response Assessment in Neuro-Oncology (RANO) categories found that post-operative residual tumour volume was negatively associated with patient outcomes. Our study aimed to set up a pipeline for measuring the pre-operative and post-operative brain volumes and tumour volumes based on anatomical magnetic resonance imaging (MRI) brain and diffuse tensor imaging (DTI) images. This would allow subsequent analysis of the association between residual tumour volume and patient outcomes. METHODS Pre-operative and post-operative MRI brain images, including T1 contrast and DTI, were obtained from patients recruited in the ‘Predicting Sites of Tumour Progression in the Invasive Margin of Glioblastomas’ (PRaM-GBM) trial. The diffusion tensor tissue signature involved the splitting of tensor information into isotropic diffusion (p) and anisotropic diffusion (q) components. RESULTS Automated skull-stripping of the MRI brain was done by HD-BET and FSL brain extraction tools and the resulting images were reviewed manually on ITK-SNAP. DTI images were processed by FSL. DTI metrics and maps were generated. After coregistration, segmentation of the q map was performed to identify areas with q abnormality. The following volumes were obtained by FSL and relevant calculations: 1.Pre-operative brain volume, excluding CSF spaces and contrast-enhancing tumour; 2.Post-operative brain volume, excluding CSF spaces and contrast-enhancing tumour; 3.Pre-operative T1 contrast-enhancing tumour volume; 4.Pre-operative q-abnormality-defined tumour volume; 5.Volume of resected non-enhancing tissue, by subtracting (2) from (1); 6.Overall resection volume, by adding (3) and (5); and 7.Post-operative residual tumour volume defined by the q abnormality, by subtracting (6) from (4). CONCLUSION Overall, we established a pipeline for measuring the residual tumour volume defined by the q abnormality in DTI. Future work will focus on correlating the residual tumour volume with patient outcomes.