Modified 3+3 Design for MTD Re-estimation

Abstract

The 3+3 clinical trial design is one of the most popular dose-finding designs used in phase I oncology trials to identify the maximum tolerated dose (MTD) for new treatment regimens. While this design is widely used due to its simplicity , it has some notable limitations, including a maximum of six patients per dose level and fixed target toxicity rates. To address these issues, we propose a modified 3+3 design that extends the traditional 3+3 design by treating the remaining patients at the MTD level for additional dose-limiting toxicity (DLT) assessment. This modification allows for a more flexible and accurate way to identify the MTD, enhanced by the use of isotonic regression to calculate DLT rates. To compare the modified 3+3 designs and the traditional 3+3 design, computer simulation studies have been carried out under various dose-toxicity scenarios. The results show that the modified 3+3 design yields higher accuracy in MTD identification.