The use of mathematical modelling in cancer research has just passed its 50-year anniversary. Investigators used this approach to explain the age-dependent incidence of cancer in the 1950s — this has recently been extended to address the age-specific acceleration of cancer (see highlight on page 174). It wasn't until Al Knudson formulated the two-hit hypothesis of tumour-suppressor genes 20 years later, however, that this type of quantitative approach was put on the map. But what has happened in mathematical modelling since then? Does research in dry labs explain results from wet labs, or can it be used instructively to indicate new avenues of research? We explore these questions, and more, in this issue.

Martin Nowak and colleagues use the principles of mutation and selection to describe the dynamics of cancer progression on page 197. They initially focus on the activation of oncogenes and the inhibition of tumour-suppressor genes, but later examine the controversial issue of genetic instability. Mathematical modelling sheds new light on the role of genetic instability in tumorigenesis, and how features such as tissue organization might affect it.

Genetic instability not only increases the rate at which tumour suppressors are inactivated, it also helps a tumour develop resistance against anticancer drugs. This form of 'robustness' is discussed by Hiroaki Kitano on page 227. He suggests that the robust circuits that are used by a tumour can only be manipulated with a good understanding of system dynamics. Computational studies provide theoretical results about which feedback loops should be removed, but they require verification.

So, it seems that mathematical modelling can provide a framework from which to explore new ideas and hypotheses. The next 50 years might allow cancer researchers to understand other complex problems, such as angiogenesis and metastasis.