Epidemiological studies have long hinted at the possibility that what we eat greatly influences our state of health, in particular our relative risk of developing cancer. In recent years there has been an exponential increase in the number of studies investigating how individual components of the diet interact at the molecular level to determine the fate of a cell. It is now apparent that many such molecules can preferentially inhibit the growth of tumour cells, by inducing cell cycle arrest or apoptosis. The number of signalling pathways and molecular targets involved is continually expanding. Consequently, the picture is becoming ever more complicated, not least because results often appear to be cell-type specific, dose-response relationships are critical, and any one agent appears to have multiple mechanisms of action. In addition most studies have been conducted in cell culture, often with physiologically unachievable concentrations of single agents, making extrapolation to the clinical situation difficult. In this review the mechanisms of action of a few well-studied dietary polyphenols (curcumin, epigallocatechin gallate and resveratrol) and indole-3 carbinol are considered in the light of these issues.