Random walk model based on DTI for predicting the microscopic spread of gliomas

The current methods of determining the treatment margins for Stereotactic Radiotherapy (SRT) are often inadequate as recurrences and/or secondary tumors occur at/near the boundaries of the treatment margin. If our hypothesis that paths of elevated water diffusion provide a preferred route for migration of cancer cells is correct then future SRT treatment volumes would be modified to provide elongated margins along the paths of elevated water diffusion; thereby reducing the incidence of recurrence. We hypothesize that the migration of tumor cells from the surface of the tumor can be predicted using a random walk model constrained by the local diffusion environment. In our implementation the diffusion environment is represented by the Principal Diffusion Direction (PDD) obtained from Diffusion Tensor Imaging (DTI). The results based on the analysis of DTI datasets of patients who had recurrences show a high correlation between areas of high cell concentration determined using the random walk model and the location of recurrences and/or secondary tumors.