Clinical oncology generates patient data spanning from the molecular scale to the whole-body scale, which tend to be used in isolation when planning patient care. There is no current technique to quantitatively combine these with novel in vitro experimental data into comprehensive models that can illuminate complex, systems-level emergent phenomena and improve therapeutic and surgical planning. In this talk, we will discuss efforts by my lab, the USC Physical Sciences Oncology Center, and the Consortium for Integrative Computational Oncology to solve these issues. With a focus on patient pathology-calibrated breast cancer modeling and multidisciplinary modeling of liver metastases, we will explore agent-based and continuum model calibration to individual patient data, integration with novel experimental measurements, and emergent predictions of macroscopic and systems-level behavior. We will discuss the implications for making and quantitatively testing biological hypotheses, and the role of computational modeling in facilitating a deeper understanding of biology, pathology, and radiology. More information can be found at MathCancer.org.