An ever-increasing amount of data is available in cancer research, much of it not yet fully used, provided by the different ‘omics’ technologies (genomics, radiomics, proteomics, etc.). cancer types for diagnosis, therapy planning, response assessment and follow-up. We study the properties of current quantifiers of tumor properties (such as heterogeneity or morphological features) to determine which can be surrogates of the disease status. We also define new properties based on mathematical models that could be used as prognostic and response biomarkers. We further study how to combine the biomarkers into more complex prognostic and predictive models.