Drug dosage is of paramount importance in the treatment of cancer, the aim being to optimise drug exposure with a view to maximising antitumour effect and minimising normal tissue toxicity. Pharmacokinetic parameters of anticancer drugs vary considerably from patient to patient. Most clinically useful drug regimens consist of a cocktail of drugs with different mechanisms of action and hence different toxicity profiles. Therefore, it is even more difficult to optimise drug dosage for individual patients. Variability in the pharmacokinetic profile of anticancer agents in individual patients can be further complicated by pharmacokinetically based drug interactions between different anticancer drugs or anticancer drugs and other concomitant medication. Most of the reported studies provide useful information and identify major interactions, but many also demonstrate the difficulty in identifying therapeutically important drug interactions in patients. Even with all the problems associated with acquiring suitable data from cancer patients it is clear that drug interactions do occur and that these can be clinically significant. It is important that potential interactions are identified early in the drug development of new anticancer drugs. This may be made possible by the rapid improvements in analytical techniques and the availability of more appropriate clinically relevant model systems. Therefore, the therapeutic significance of any detected interactions may be assessed, and steps to avoid them may be established, before the drug is under clinical investigation.