Abstract #0756
Predicting Recurrence Patterns of Gliomas Using Diffusion Tensor Imaging
Carpenter T, Jena R, Burnet N, Price S, Pickard J, Gillard J
University of Cambridge
Gliomas are one of the most heterogeneous tumours yet we treat them all the same. If we could identify different phenotypes it might allow us to individualise treatment. By splitting the diffusion tensor into its isotropic and anisotropic components we have identified three groups of abnormality that can predict recurrence patterns in 92% of cases. Diffuse abnormalities usually predict a general increase in tumour size. Localised abnormalities usually predict recurrence in the direction of this abnormality. For the few patients with minimal abnormality this appears to predict slow progression and may be a marker of long-term survival.