|
Abstract:
|
To co-develop a CDx and the corresponding therapeutic product, ideally the CDx should be used for trial enrollment. But this is not realistic due to various reasons such as CDx readiness and other operational constraints. Instead of the CDx, a CTA, which can be the research version of the CDx or a local lab test, is typically used to determine patients' biomarker status for enrollment eligibility. This approach is subject to sampling bias in an enrichment trial design because patients with desired CDx biomarker status are excluded from the trial if the CTA gives the opposite biomarker status. Li (2015) has proposed a bridging study framework to correct for the sampling bias when there is only one CTA. In practice, more than one CTAs may be used to expedite trial enrollment or to meet site specific requirement. Furthermore, the CTAs may have different concordance rates to the CDx. We propose two methods to adjust the sampling bias due to multiple CTAs when estimating the therapeutic efficacy in the population determined by the CDx. A Monte-Carlo simulation was conducted to evaluate the performance of the proposed methods. The simulation result will be discussed in the presentation.
|
