The Randomized CRM: An Approach to Overcoming the Long-Memory Property of the CRM

TL;DR

This paper introduces the randomized CRM (rCRM), a modification of the standard CRM designed to address its long-memory issue by incorporating random dose adjustments, thereby improving variability without sacrificing accuracy.

Contribution

The paper proposes the rCRM, a novel variation of the CRM that adds random escalation and de-escalation to mitigate long-memory effects in dose-finding trials.

Findings

rCRM maintains similar MTD identification rates as CRM

Reduces variability in the number of cohorts at the true MTD

Shows comparable safety profiles in simulations

Abstract

The primary object of a phase I clinical trial is to determine the maximum tolerated dose (MTD). Typically, the MTD is identified using a dose-escalation study, where initial subjects are treated at the lowest dose level and subsequent subjects are treated at progressively higher dose levels until the MTD is identified. The continual reassessment method (CRM) is a popular model-based dose-escalation design, which utilizes a formal model for the relationship between dose and toxicity to guide dose-finding. Recently, it was shown that the CRM has a tendency to get "stuck" on a dose-level, with little escalation or de-escalation in the late stages of the trial, due to the long-memory property of the CRM. We propose the randomized CRM (rCRM), which introduces random escalation and de-escalation into the standard CRM dose-finding algorithm, as an approach to overcoming the long-memory property of the CRM. We discuss two approaches to random escalation and de-escalation and compare the operating characteristics of the rCRM to the standard CRM by simulation. Our simulation results show that the rCRM identifies the true MTD at a similar rate and results in a similar number of DLTs compared to the standard CRM, while reducing the trial-to-trial variability in the number of cohorts treated at the true MTD.