For many oncological agents, myelosuppression is the dose-limiting toxicity and the quantitative characterisation of the relationship between drug dose, plasma concentration and haematological toxicity is of importance in the drug development. Mechanism-based population pharmacokinetic-pharmacodynamic models have been developed for this purpose and the applications of these in candidate selection, first-in-man studies, prodrug and formulation development, dose finding, schedule optimisation, assessing influence of modifying agents, drug combination studies, subgroup identification and feedback individualisation are reviewed.Population pharmacokinetic analysis of (sparse) concentration-time data have for a decade provided useful and sometimes crucial information during oncological drug development whereas there are far fewer reports on the use of population pharmacodynamic models for the same purpose. Although pharmacokinetic/pharmacodynamic models have been identified as having a potentially important role in oncological drug development (Aarons et al. 2001a), the only area where there has been considerable use of the methodology is in characterising haematological toxicity. In these models, the time course of leukocytes or absolute neutrophil counts after chemotherapy is characterised. A clear trend in time has been that the mechanistic basis for haematological toxicity ( fig. 1) has been increasingly considered in the structure of the models developed, as described in a recent review of the evolution of pharmacokinetic/pharmacodynamic models for myelosuppression (Friberg & Karlsson 2003). Although generally developed on data from treatment with drugs already on the market, there is, partially realised, usefulness of such predictive models for a dose-limiting toxicity, such as frequently occurring myelosuppression. We have used one of the recently published models (Friberg et al. 2002) (fig. 2) in clinical drug development. In this presentation, we will first briefly present the model, thereafter discuss the different uses of it and finally make some general remarks. Mechanism-based population pharmacokinetic/ pharmacodynamic model for haematological toxicityThe mechanism-based model describes the system in terms of five sequential compartments ( fig. 2). The first compartment in the chain represents the pool of proliferative cells, which are capable of self-replication. These cells are susceptible to drug-induced cell death but will otherwise generate non-mitotic cells that after maturation will be released to the systemic circulation and subsequently taken up by tissues. The physiological system is quantified in terms of the baseline circulating cell count (Baseline), the mean maturation time and a parameter governing the increase in self-replication that occurs when circulating cells are depleted (g), mimicking the stimulating endogenous granulocyte colony stimulating factor (G-CSF) interaction. An additional parameter, the first-order rate constant for loss from the circulating pool was found n...