Most clinical drug regimens for irinotecan (CPT-11 [Camptosar]) have been empirically based on classic in vivo pharmacokinetic and pharmacodynamic considerations. We propose an alternative approach that attempts to provide a rationally designed schedule of irinotecan administration based on preclinical data. HL60 cells grown in suspension or as subcutaneously implanted solid xenografts in nude mice served as in vitro and in vivo models to rest the activity of irinotecan or its active metabolite, SN-38. For SN-38, within an effective drug concentration range, scheduling drug administration based on duration of DNA synthesis inhibition significantly potentiated cell kill in vitro, and increasing drug concentrations at suboptimal scheduling did not result in additive cell kill. These data suggested that even though high drug doses may be attainable in vivo, they may not be required to achieve maximum antitumor activity. To test this hypothesis, a sensitive in vivo model to test the toxicity and antitumor activity of CPT-11 is required, which is provided in the human myeloid HL60 xenograft model grown in nude mice. In this model, CPT-11 at a dose 50 mg/kg, daily x5 (MTD) achieved 100% complete tumor regression. This model should be useful to test the hypothesis that for irinotecan, administration of a minimum effective dose (MED) at an optimal schedule can achieve maximum antitumor activity and should therefore prevail over the classic approach of administering the MTD.