CFO: Calibration-free odds design for phase I/II clinical trials

Abstract

Recent revolution in oncology treatment has witnessed emergence and fast development of the targeted therapy and immunotherapy. In contrast to traditional cytotoxic agents, these types of treatment tend to be more tolerable and thus efficacy is of more concern. As a result, seamless phase I/II trials have gained enormous popularity, which aim to identify the optimal biological dose (OBD) rather than the maximum tolerated dose (MTD). To enhance the accuracy and robustness for identification of OBD, we develop a calibration-free odds (CFO) design. For toxicity monitoring, the CFO design casts the current dose in competition with its two neighboring doses to obtain an admissible set. For efficacy monitoring, CFO selects the dose that has the largest posterior probability to achieve the highest efficacy under the Bayesian paradigm. In contrast to most of the existing designs, the prominent merit of CFO is that its main dose-finding component is model-free and calibration-free, which can greatly ease the burden on artificial input of design parameters and thus enhance the robustness and objectivity of the design. Extensive simulation studies demonstrate that the CFO design strikes a good balance between efficiency and safety for MTD identification under phase I trials, and yields comparable or sometimes slightly better performance for OBD identification than the competing methods under phase I/II trials.

Time‐to‐event calibration‐free odds design: A robust efficient design for phase I trials with late‐onset outcomes

“…Because they are model-free and calibration-free, CFO and TITE-CFO are robust and easy-to-use. 28 Through our comprehensive numerical studies, TITE-CFO demonstrates high robustness in comparison to both the rulebased TITE-BOIN and model-based TITE-CRM designs. Due to the model-based nature, the TITE-CRM design is sensitive to the selection of the model skeleton, which barricades its usage in practice.…”

“…where x ko , m ko ð Þare the number of DLT outcomes and the number of patients who have completed the DLT assessment at dose level k. By default, we choose α ¼ ϕ and β ¼ 1 À ϕ, 28 where ϕ is the target DLT rate.…”

“…Repeat steps 2-3 until the prespecified maximum sample size is reached and select the MTD using the isotonic regression. 29 As a common practice in phase I designs, 17,19,28 we impose an overdose control rule to guard patient safety. When the data suggest the current dose level is over-toxic with high probability, we eliminate the current dose level as well as those higher dose levels.…”

“…To address the calibration problem and ease the implementation of a phase I design, a novel calibration‐free odds (CFO) design has been proposed 28 . As indicated by the name, the CFO design does not require to calibrate any essential design parameter.…”

“…To address the calibration problem and ease the implementation of a phase I design, a novel calibration-free odds (CFO) design has been proposed. 28 As indicated by the name, the CFO design does not require to calibrate any essential design parameter. To make it calibration-free, the CFO design adopts the game competition idea, which compares the evidence supporting dose de-escalation and escalation at the current dose level and its two neighboring doses.…”