Merging Traditional Scientific Computing with Data Science to Develop a New Prediction Engine for Brain Cancer

Gbocho Masato Terasaki

University of California, Merced

"Merging Traditional Scientific Computing with Data Science to Develop a New Prediction Engine for Brain Cancer"

Glioblastoma multiforme (GBM) is one of the fastest-growing brain tumors and it has very low survival rates. Mathematical modeling can be used to predict the growth and treatment of brain cancer. However, one of the difficulties lies in the ability to estimate patient-specific parameters in the mathematical model from magnetic resonance imaging (MRI) data. We constructed a numerical solver to simulate tumor growth over a realistic 3D brain geometry derived from segmented-MRI. Then, using information about the size of the different glioma sub-regions, we are developing a method that estimates the patient-specific model parameters to inform the forward simulation. Ultimately, we hope to predict the overall survival of a patient from a single pre- operative scan.

---