Rolling continual reassessment method with overdose control: An efficient and safe dose escalation design

Area of Science:

• Clinical pharmacology
• Biostatistics
• Drug development

Background:

• Phase 1 clinical trials aim to determine the maximum tolerated dose (MTD) using dose-limiting toxicity (DLT) data.
• Traditional methods like the continual reassessment method (CRM) often fail to account for late-onset DLTs, posing safety risks.
• Existing CRM extensions do not explicitly control the risk of overdosing.

Purpose of the Study:

• To propose a novel dose escalation design, the rolling-CRM, that incorporates both DLT timing and overdose control.
• To evaluate the performance of the rolling-CRM design against existing dose escalation strategies.

Main Methods:

• The rolling-CRM design utilizes a two-parameter logistic regression model for DLT probability and a piecewise exponential model for DLT time.
• A comprehensive simulation study was conducted to compare rolling-CRM with other dose escalation designs.

Main Results:

• The rolling-CRM design significantly shortens trial duration compared to traditional CRM designs.
• The proposed design effectively controls overdose risk.
• The rolling-CRM may require a larger sample size than traditional CRM designs.

Conclusions:

• The rolling-CRM design offers an improved approach for phase 1 dose escalation by addressing late-onset toxicities and overdose risks.
• This novel design enhances patient safety and trial efficiency in early-phase drug development.