Abstract:
|
The mTPI design (Ji 2010) has increased in use due to an awareness that it is more reliable than 3+3 at selecting the true MTD, and that it has been shown to be safer (Ji and Wang 2013). However, many practical issues remain to be addressed. These include how to determine sample size and additional escalation conditions that may be desired. Fixing the sample size is a challenge and given that mTPI prefers to enroll at the current MTD (like CRM), it may not push the dose when additional gains in efficacy may be made in the presence of a flatter dose-toxicity curve. In some cases 'capping' dose levels at a maximum sample size as well as changing specific cells of the mTPI design may be desirable. In addition, operating characteristics of the design may depend on whether empirical estimates of toxicity are used vs. posterior estimates for both dose escalation decisions as well as MTD determination. This presentation will address some of these issues. Importantly the operating characteristics of mTPI under some of the additional constraints will be assessed. In addition, change management aspects of moving from 3+3 to mTPI within an organization will be discussed.
|