There are two key areas in tumour biology that must be considered when developing models to study the disease – tumour heterogeneity and the tumour microenvironment (TME). Current cancer models including cell lines, organoids or, more recently, in vivo patient-derived xenograft (PDX) models often fail to recapitulate these elements sufficiently, and so numerous new treatments that have been developed based on these models do not perform well when moved forward into the clinic.