探討第二期癌症臨床試驗之兩階段設計

Some discussions on the two-stage designs for Phase II cancer clinical trials

葉怡君(Yi-Jyun Ye)

指導教授 ： 黃怡婷

國立臺北大學/商學院/統計學系/碩士(2009年)

摘要

第二期臨床試驗主要目的是在於瞭解藥品的有效性與安全性，並決定用藥的方式、劑量等項目，以做為是否進行第三期臨床試驗的依據。然而，試驗進行中在未確保藥物安全性的情況下，對於讓受試者面臨無法預知的風險，基於道德上的考量，為了避免讓受試者處於無療效或者副作用過高的治療太久，在欲及早知道新藥的療效、縮短臨床試驗時間、有效減少受試者數目的情況下，便延伸出兩階段試驗設計。

兩階段試驗設計最早是由 Gehan 於 1960 年提出一個最粗略的想法，但此時尚未架構出一個完整而明確的流程。至 1982 年Fleming 方才提出了多重階段試驗設計。之後在 1989 年 Simon 提出完整的兩階段設計試驗流程，因選擇樣本數的準則不同又分作 optimal 和 minimax 兩種設計，此設計因施行簡易，所以被廣泛的應用在第二期癌症臨床試驗中。 Fleming 和 Simon 所提出的設計在檢視第一階段試驗結果時，均先終止試驗的進行，直至評斷出藥物的療效與否後再決定是否進行第二階段的試驗。但就實務的考量而言，考慮到此終止期間所耗損的成本，因而在 1998 年發展出廣為美國西南腫瘤集團 (Southwest Oncology Group，簡稱 SWOG) 所使用的 Hybrid 設計，
允許在檢視第一階段結果時繼續收取患者，倘若結果顯示新藥並不顯著具有療效時，便暫時停止收取患者，此時再次檢視所有患者的結果，以決定是否進行第二階段的試驗還是永久終止試驗。Fleming 和 Hybrid 設計在當時均是主觀的預先設定每一階段的樣本大小，
以便在給定型一誤和型二誤之下求得相對應的臨界值。但主觀的判定樣本大小在實務考量上並非十分恰當，故本論文中針對此特點進行修正，分就 Simon 所提出的兩種選取樣本數的準則來選取樣本數。

最後再就每一兩階段設計所對應的期望樣本數 EN(p)、試驗在第一階段早期終止的機率 PET(p)、總樣本數 n=n1+n2，以及真實的型一誤和型二誤來比較第二期臨床試驗中各種不同兩階段設計的優劣，以及探討各設計所適用的時機、解決的問題與實際上的使用現況等等。

關鍵字

臨床試驗；兩階段設計；早期終止機率；期望樣本大小

並列摘要

The purpose of a phase II cancer clinical trial is to understand the effectiveness and safety of a new drug and also to determine the dosage and the form of a drug. For enthical consideration, a trial should include minimal test patients that are sufficient to validate the effectiveness and safety of a drug and also reduce the cost and time of a clinical trial. A two-stage clinical trial is designed to fullfill such purposes.

The earliest two-stage desig was proposed by Gehan (1960), but the precise executing process was not given.Fleming (1982) proposed a multiple-stage design. The most commonly used design for the Phase II cancer clinical trial is the two-stage design proposed by Simon (1989). However, designs proposed by Fleming (1982) and Simon (1989) have to stop when evaluating the effectiveness at the end of the first stage. This can result in longer trial duration. Southwast Oncology group suggested a hybrid design that allows
recruiting some patients while evaluating the effectivenes at the end of the first stage.

To use Fleming's design and the hybrid design, one need to specify the total sample size of the trials to compute the critical values. This is not very practical. This thesis adapts the two criteria suggested by Simon (1989) and implements into Fleming's design and the hybrid design to avoid prespecifying the total sample size. Under various settings, the performance of the four proposed designs as well as Simon's desgins as measured by the total sample size, the expected sample size, the probability of early termination,Type I error and Type II error is disscussed.

並列關鍵字

clinical trials；two stage designs；early termination probability；expected sample size

參考文獻

林盛峰(2088),"二元反應臨床試驗的二階段設計之實務考量",國立台北大學統計研究所碩士論文.

Fleming, T. R. (1982),"One-sample multiple testing procedure for phase II clinical trials", Biometrics, 38(1), 143-151.

Gehan, E. A. (1961),"The determination of number of patients in a follow-up trial of a new chemotherapeutic agent", Chronic Diseases, 13(4), 346-353.

Green, S. J. and Dahlberg, S. (1992),"Planned versus attained design in phase II clinical trials",Statistics in medicine, 11(7), 853-862.

Herndon, James E. II (1998),"A design alternative for two-stage, phase II, multicenter cancer clinical trials", Controlled clinical trials, 19(5), 440-450.