Personalized Glioblastoma Multiforme Growth Modeling by Integrating Clinical Genomic and Phenotypic Data

Abstract Glioblastoma multiforme (GBM) is a highly aggressive brain tumor characterized by significant morbidity and mortality, alongside a complex and heterogeneous nature that complicates outcome prediction and treatment optimization. Here, we present a patient-specific simulation framework for GBM tumor growth, integrating genomic data from pathological biomarkers in biopsy specimens with phenotypic data derived from standard MRI sequences (T1Gd and T2). Model precision was enhanced by adding a patient-specific Pathological Coefficient, derived via a parameter sweeping analysis, to the model's proliferation term. Our findings demonstrate a substantial improvement in simulation accuracy with the incorporation of this genomic information, particularly the Ki-67 proliferation index. Specifically, tumors exhibiting higher Ki-67 expression correlated with increased Pathological Coefficients and accelerated growth dynamics. Mutant IDH-1 type GBM models further demonstrated greater sensitivity to these coefficients compared to wild-type tumors. The personalized nature of these models facilitates more accurate predictions of overall survival and supports informed clinical decision-making, enabling oncologists to simulate individual tumor behavior and tailor therapies for optimized patient outcomes.