Optimizing Schedule of Administration in Phase I Clinical Trials

In conventional phase I studies, patients are assigned to a dose of an experimental agent under study, receiving either a single administration or a course consisting of several administrations of the agent. Patients are followed for a short period of time during which a dose-limiting toxicity (DLT) may occur. At the end of the follow-up period, a binary outcome indicating the presence or absence of DLT is recorded. Patients who experience a DLT after one or more treatment modifications will not fit the simplified framework of phase I trial designs unless the DLT is associated with the originally planned treatment schedule. This relies on the implicit assumption that altering a patient's treatment schedule has no effect on the probability of a DLT. The maximum tolerable schedule (MTS) method is adaptive in its selection of the schedule assigned to each patient, using Bayesian methods to continually update the MTS.