Cancer survival may be inversely related to the speed at which a primary tumour grows and disseminates. Assessment of prognosis using surgical and/or radiological definition of disease extent, i.e. staging, has thus become a standard intervention in newly diagnosed patients, with the most popular framework being the tumour-node-metastasis (TNM) system. However, increasing use of biomarkers--non-TNM factors that predict therapeutic benefit, rather than adverse disease outcome--has weakened the decision-making dominance of TNM. This shift from risk-led to benefit-led practice is now starting to blur the time-honoured qualitative distinction between curable (M(0), early stage, adjuvant) and incurable (M(1), early metastatic, palliative) disease treatment strategies; the same biologic drug strategy may improve average survival outcomes by similar increments for two patients, one of whom is 'adjuvant' and the other 'metastatic'. Plausibly, then, biomarker-positive patients presenting with high-TNM (M(1)) disease may enjoy the same, if not more, disease-free and/or overall survival benefit as conventional low-TNM (M(0)) patients when treated with standard adjuvant interventions. Conversely, M(0) patients concerned by quality-of-life issues such as alopecia may in future be able to choose better-tolerated personalized drug regimens similar to those now used with survival benefit in palliative settings, even if such adjuvant regimens have not yet been validated by level 1 data. To these ends, a modernised decision-oriented disease staging system called METS (molecular/extra-primary/tumour/symptoms) is presented here.