Tumor progression depends on the intricate interplay between biological processes that span the molecular and macroscopic scales. A mathematical agent-based model is presented to describe the 3-D (three-dimensional) progression of a brain tumor type (i.e., glioblastoma multiforme) as the collective behavior of individual tumor cells whose fate is determined by intracellular signaling pathways (i.e., MAPK pathway) that are governed by the temporal-spatial distribution of key biochemical cues (i.e., growth factors, nutrients). The model is used to investigate how tumor growth and invasiveness depend on the response of migrating tumor cells to chemoattractants. Simulation results suggest that individual cell sensitivity to chemical gradients is necessary to generate in silico tumors with the irregular shape and diffusive tumor-stroma interface characteristic of glioblastomas. In addition, vascular network damage influences tumor growth and invasiveness. The results quantitatively recapitulate the central role that nutrient availability and signaling proteins have on tumor invasive properties. (C) 2010 American Institute of Chemical Engineers AIChE J, 57: 778-792, 2011