The five examples given here illustrate new cytotoxic agents at different stages of evaluation. In all cases, considerable effort has gone into detailed pharmacokinetic studies conducted before and during the clinical phase I studies. Has this effort contributed significantly to the development of these agents? At present, it has to be said that the contribution made in the case of these particular agents has been modest. For the anthrapyrazoles, the availability of the pharmacokinetic data did not permit a pharmacokinetically guided dose escalation to be performed because of non-linear kinetics, and a similar comment can be made for rhizoxin, since the human plasma AUC values at the MTD were much lower than in the mouse. For the camptothecin analogues, a detailed knowledge of the kinetics of the closed and open forms of the various agents did not influence the way in which the studies were conducted, nor did pharmacokinetic information appreciably do so for EO9, although some comfort was gained by clinical investigators when the short half-life seen in preclinical species was also observed in humans. For suramin, therapeutic drug monitoring is clearly essential, although toxicity remains a problem. Of course, a proper understanding of the pharmacokinetics and metabolism of these agents greatly improves the interpretation of the clinical observations made and is often critical in planning the next stages of development. This is more clearly seen with agents that have unusual forms of toxicity, such as flavone acetic acid, for which the achievement of notional target concentrations is a key element in clinical trials (Kerr et al, 1987; Maughan et al, 1992). Moreover, as reviewed elsewhere (Graham and Workman, 1992; see also Graham and Kaye, this volume), there are several other instances where pharmacokinetically guided dose escalation has greatly improved the conduct of a phase I study. Good examples of this are iododoxorubicin (Gianni et al, 1990), mitotic inhibitor CI-980 (Brodfuehrer et al, 1992) and DNA intercalator CI-958 (Whitfield et al, 1992). Not surprisingly then, pharmacokinetics can help guide early clinical studies of some compounds but not others and whether they will be of value can only be determined by carrying out the pharmacokinetic measurements. The real value of the pharmacokinetic studies for the five compounds reviewed may not yet have been seen. Interpatient variations in drug handling can play a major part in determining levels of anti-tumour activity as well as toxicity.(ABSTRACT TRUNCATED AT 400 WORDS)