本文針對合併兩種藥物治療腫瘤的複合藥進行早期臨床試驗設計，其中可能包含二種化療藥或一種化療藥與一種標靶藥。上述複合藥的毒性會因劑量增加而增強，但是標靶藥的遞增藥效可能在較高的劑量時趨於平穩而呈現高原現象，因此本文建議整合毒性與藥效機率設立效用函數，然後尋找該複合藥合乎毒性要求且具有最高效用的最佳生物劑量組合。本文針對二元毒性與藥效反應提出兩種早期臨床試驗設計，其中一種屬於同時考量毒性與藥效的單階升降劑量設計 (cBOIN12)；另一種則為兩階段的升降劑量設計(cUBOIN)。最後本文在多種不同的毒性與藥效機率組合情境之下進行模擬，研究上述複合藥早期臨床試驗設計在選擇最佳生物劑量組合的表現與毒性控管的成效，同時也審視兩種設計的試驗時間。;This thesis considers early clinical trial designs for the combination of two drugs in the treatment of oncology, which may include two chemotherapies or a chemotherapy along with a Molecularly Targeted Agent (MTA). The toxicity of the combined drugs increases with higher doses, but the efficacy of the MTA may plateau at higher doses. Therefore, in this thesis, a utility function that integrates the toxicity and efficacy probabilities is used to find the optimal dose combination (ODC) of the combined drugs, where the ODC meets the toxicity requirements and has the largest utility. Two early clinical trial designs are proposed for binary toxicity and efficacy responses, including a single-stage dose-escalation design (cBOIN12) that considers both toxicity and efficacy simultaneously, and a two-stage dose-escalation design (cUBOIN). Finally, we conduct a simulation study under various combinations of toxicity and efficacy probabilities to investigate the performance of estimating ODC and controlling toxicity as well as the required trial time of the proposed early clinical trial designs for combined drugs.
