A model-free phase I/II dose optimization design for immunotherapy trials

Area of Science:

• Clinical Trials
• Immunotherapy
• Biostatistics

Background:

• Optimizing immunotherapy dosage is crucial for maximizing efficacy while minimizing toxicity.
• Existing clinical trial designs often rely on complex parametric models, limiting transparency and efficiency.
• Jointly modeling multiple outcomes like toxicity, efficacy, and immune response is challenging but essential for immunotherapy dose optimization.

Purpose of the Study:

• To introduce a novel, model-free phase I/II clinical trial design (UFO design) for optimizing immunotherapy dosage.
• To enable efficient information sharing across dose levels by leveraging correlations between toxicity, efficacy, and immune response.
• To extend the UFO design to accommodate delayed outcome data and demonstrate its favorable operating characteristics.

Main Methods:

• Developed a model-free approach for clinical trial design, termed the UFO design.
• Utilized inherent correlations among toxicity, efficacy, and immune response outcomes.
• Incorporated constrained dose-outcome ordering to facilitate information sharing.
• Extended the design to handle delayed outcomes.

Main Results:

• The UFO design demonstrated efficiency and transparency for clinical implementation.
• Simulation studies showed favorable operating characteristics for the proposed design.
• The model-free approach effectively optimizes immunotherapy dosage by jointly modeling multiple outcomes.

Conclusions:

• The UFO design offers a practical and effective model-free strategy for optimizing immunotherapy dosage in clinical trials.
• This approach enhances efficiency and transparency by leveraging outcome correlations and dose-outcome constraints.
• The availability of an R Shiny app facilitates the simulation and implementation of the UFO design in practice.