PM-06IMPROVED ANATOMICAL MODEL PREDICTION OF GLIOMA GROWTH UTILIZING TISSUE-SPECIFIC BOUNDARY EFFECTS

PURPOSE: Kinetic parameters for glioblastoma multiforme (GBM), derived from clinical scans, can be used to predict the occurrence of hypoxia, necrosis, response to radiation therapy, and overall survival. The virulence of the disease is related to both the motility (invasiveness) and the rate of cellular division (proliferation) of glioma cells. Improving current volumetric estimates for these kinetic parameters can allow for more effective predictions of the course and extent of the individual patient's disease. METHODS: Utilizing the assumption that glioma cells preferentially migrate through white matter as compared to grey matter, we integrate the constraints of anatomical boundaries into current measurement-derived, volumetric estimates of growth and invasion kinetics for eight GBM patients. We then implement these revised kinetic parameters into a mathematical model of GBM growth, in an anatomical atlas, and compare the results to the volumetric measurements of the clinical scans. The effectiveness of this method in recapitulating growth in-silico is compared with estimates of disease kinetics from purely volumetric estimates derived from clinical scans. RESULTS: We found that taking anatomical boundaries to glioma growth into account improved estimation of invasiveness by up to 50%, whereas estimates of proliferation rate remained stable. This, in turn, allowed mathematical models to volumetrically match a patient's clinical scans for pretreatment time points with greater accuracy. CONCLUSIONS: Improving estimation of patient-specific parameters used in mathematical models of GBM, we more effectively predict the course and extent of the disease. More accurate prediction of GBM extent lays a better foundation for planning treatment and evaluating the success of treatment against what would have occurred without treatment.