Abstract:
|
The 3+3 design has been the standard for dose escalation studies in early phase oncology. But as the clinical research in oncology expanded through the 1990s and continued expanding in the current century, concern grew over whether 3+3 was finding the true maximum tolerate dose (MTD), the primary objective of dose escalation in oncology. The introduction of the CRM by O’Quigley et al was a great advance in identifying the true MTD, but its aggressive dose escalation approach raised concerns of giving highly toxic doses to sick patients. Modifications to the original CRM were soon made to alleviate these concerns, but the reluctance to move beyond 3+3 and to accept these innovative designs remain, due to the perception that more patients in the Bayesian designs will experience dose limiting toxicities remains in the minds of many oncologists who do plan and do these studies. In this presentation, we will compare the most well know Bayesian dose escalation designs-CRM, mTPI-2, BOIN, and N-CRM-to each other and to the 3+3 design. The comparisons will give statisticians and clinical trialists an idea of how different designs compare and help with the choice of a dose escalation design.
|