關鍵字: 高原期劑量藥效模型、早期臨床試驗、兩階段試驗設計、標準藥物、 混合藥物 ;Abstract In this paper, we develop a two-stage design to identify the optimal dose of the additive component (OAD) in a combined drug, where the combined drug has toxicity probability less than the targeted toxicity probability (TTP) but most outperforms the standard drug in efficacy. In the first stage, different dose assignments for identifying the MTD are studied based on the CRM design working under the one-parameter logistic regression model for the dose-toxicity relationship. In the second stage, the group sequential design is used for dose escalation where the MTD is adjusted and the OAD is re-estimated under a plateaued dose-efficacy model. A simulation study is conducted to investigate the OAD estimation, over-toxic dose assignment and trial time of the proposed design and the competitive single-step designs. The proposed design does not suggest assign patients to receive the standard drug in the first stage. Moreover, the two-stage design is competitive to the single-stage designs in the OAD estimation, but the single-stage designs are time consuming and need more patients. Key words : Plateau efficacy model, early clinical trial, two-stage design, standard drug, combined drug.
