A software tool for both the maximum tolerated dose and the optimal biological dose finding trials in early phase designs

Abstract

Background:: Phase I and/or I/II oncology trials are conducted to find the maximum tolerated dose (MTD) and/or optimal biological dose (OBD) of a new drug or treatment. In these trials, for cytotoxic agents, the primary aim of the single-agent or drug-combination is to find the MTD with a certain target toxicity rate, while for the cytostatic agents, a more appropriate target is the OBD, which is often defined by considering of toxicity and efficacy simultaneously. Accessible software packages to achieve both these aims are needed. Results:: The objective of this study is to develop a software package that can provide tools for both MTD- and OBD-finding trials, which implements the Keyboard design for single-agent MTD-finding trials as reported by Yan et al. (2017), the Keyboard design for drug-combination MTD-finding trials by Pan et al. (2020), and a phase I/II OBD-finding method by Li et al. (2017) in a single R package, called Keyboard. For each of the designs, the Keyboard package provides corresponding functions such as get.boundary (⋯) for deriving the optimal dose escalation and de-escalation boundaries, select.mtd (⋯) for selecting the MTD when the trial is completed, select.obd (⋯) for selecting the OBD at the end of a trial, and get.oc (⋯) for generating the operating characteristics. Conclusion:: The Keyboard R package developed herein provides convenient tools for designing, conducting and analyzing single-agent, drug-combination, phase I/II OBD-finding trials.

Keywords

• Maximum tolerated dose
• Optimal biological dose
• Dose-finding
• Phase I/II trials
• Bayesian adaptive design