Tumor heterogeneity has major implications for tumor development and response to therapy. We relate intra-tumor heterogeneity to morphology as determined by MRI (or CT) images by applying a biologically-inspired model of tumor growth. This model represents the development of a tumor as a set of expanding spheres, each representing a distinct clonal center, with the sprouting of new spheres representing new centers. Each clonal center may possess different characteristics, including genetic composition, growth rate and response to treatment. A major implication of our work is that assessment of a tumor using volume measurements (RECIST, WHO) may not accurately reflect the response to treatment. The application of this new concept to clinical imaging studies has confirmed its ability to track and inform clinicians more accurately about changes in tumor size and morphology due to growth or therapeutic intervention.