When designing a dose escalation plan in a phase I trial, we usually follow the guiding principles as follows: avoiding unnecessary exposure of patients to sub-therapeutic dose of an agent (this means treating maximum number of patients within the therapeutic dose range); keeping safety; maintaining rapid accrual. There are two basic types of dose escalation methods used in phase I cancer clinical trials, that is, the rule-based designs and the model-based designs. Rule-based designs, such as the traditional 3+3 design (Figure 1A) and the accelerated titration design (ATD, Figure 1B), are designs that recruit patients to different dose levels according to the prespecified rules acquiring from actual observations of target. Actually, ATD is a combination of variations of the traditional 3+3 design and the model-based design. However, we still use prespecified rules to guide patients in assigning doses, so we consider that ATD is a rule-based design. Simon et al. designated grade 2 toxicity as moderate toxicity. Design 1 (a traditional 3+3 design) was designated as a standard phase I design with 40% escalation steps between dose ranges. Design 2, 3 and 4 were designated as accelerated designs.
Figure 1. Graphical depiction of dose escalation methods for phase I clinical trials. Each box represents a cohort comprising the indicated number of patients treated at a given dose level. A) Traditional 3+3 design. B) Accelerated titration design. DLT = dose-limiting toxicity; SD = starting dose; RD = recommended dose.
To be specific, in design 1, three patients as a cohort experienced a 40% dose escalation until dose-limiting toxicity (DLT) is observed. When there is only one DLT observed, the dose of another three patients will be added to the current dose level and if no additional DLT is observed, the escalation will continue. When there are two DLTs observed at the current dose level, the escalation will be stopped and the prior dose level will be declared as the maximum tolerated dose-MTD (only if six patients have been treated and only one DLT is observed at that level, otherwise the de-escalation will continue until such a dose level is determined). In design 2, the dose escalation is still 40%, but there is only 1 patient at each dose level. If one DLT or two grade 2 toxicity events occur in the first course of treatment (cycle 1), the dose escalation in the accelerated phase will be terminated, and another two patients at current dose level will be recruited before the design 1 is initiated. In design 3, there is a 100% dose escalation between individual patient cohorts. When a DLT or two grade 2 toxicity events are observed during cycle 1, the dose escalation will revert to the more conservative one used in design 1. In design 4, the trial with a 100% dose escalation in the accelerated phase will revert to design 1 as soon as a DLT or two grade 2 toxicity events are observed during any cycle (not just during cycle 1). Besides, intrapatient dose escalation is allowed during the accelerated phase of designs 2, 3, and 4. As we can see, in all three accelerated titration designs, the traditional 3+3 design used after the accelerated phase as a stopping rule, accompanying the recommended usage of the described model are used to estimate the MTD using all toxicity data collected during the trial. However, researchers often determine the MTD based on the traditional 3+3 escalation rule instead of fitting trial data to the model at the end of the trial. Consequently, the original model-based accelerated titration designs are primarily applicable to rule-based designs in clinical practice.
In the acceleration phase of the accelerated titration design, the advantage is that there is only one patient per dose group, and intrapatient dose escalation reduces the number of patients exposed to the ineffective dose, ensuring that patients can receive higher and more effective doses. However, the disadvantage of intrapatient dose escalation is that it masks the cumulative effect of drug treatment, and its long-term toxic effects and delayed toxic effects are also difficult to distinguish. At the same time, since the data for each dose level is only obtained from one patient, there are some difficulties in explaining and interpreting the results of the trial. Besides, it should be noted in ATD that if the toxicity of the drug is delayed toxicity, this design approach is not suitable.
At CD BioSciences, we can not only help you with the accelerated titration design, but also help you make an appropriate choice of statistical strategy. If you have any questions, please feel free to contact us.
References:
1. Simon, R., Rubinstein, L., Arbuck, S. G., Christian, M. C., Freidlin, B., & Collins, J. (1997) ‘Accelerated titration designs for phase i clinical trials in oncology’, J Natl Cancer Inst, 89(15), 1138-1147.
2. Le, T. C., Lee, J. J., & Siu, L. L. (2009) ‘Dose escalation methods in phase i cancer clinical trials’, Journal of the National Cancer Institute,101(10), 708-720.
3. Janet, D., Boris, F., & Larry, R. (2006) ‘Accelerated Titration Designs’, Statistical Methods for Dose-Finding Experiments.