From Group Sequential to Adaptive Designs

A two‐stage group‐sequential design for delayed treatment responses with the possibility of trial restart

“…To address these properties, a modified procedure that shares the same underlying logic has been proposed. 11 In this modified procedure, the boundary determination focuses on a non-binding futility boundary and does not permit a final decision for rejection after halting recruitment due to crossing l 1 .…”

“…5 However, in practice, nonbinding stopping boundaries are generally preferred over binding ones as they allow for more flexibility in the conduct of a trial, even though this typically causes some power loss. 4 As an adjusted version for the determination of {u 1 , l 1 , d1 , d 2 }, Jennison and Hampson proposed the following constraints 11 :…”

“…Given {u 1 , l 1 , d1 , d 2 }, there are three equations with four unknown parameters. Thus, Jennison and Hampson 11 suggest fixing d1 = Φ −1 (1 − 𝛼) and successively solve the equations above for {l 1 , u 1 , d 2 }. With these boundaries, it is not possible to reject at interim after observing Z 1 < l 1 since all type I error rate is spent with respect to the upper boundaries.…”

“…Moreover, required notation, the error-spending concept and some important distributional results follow. In Section 3, two methods proposed by Hampson and Jennison 5,11 will be introduced. The new method will be introduced in Section 4.…”

Randomized Phase II Cancer Clinical Trials by JUNG, SIN‐HO

“…Especially in Section 4.5, the author discusses using the exact binomial testing on t ‐year survival probability but the patients whose event times are censored before t should be excluded from the analysis. Several researchers have proposed using Kaplan–Meier estimator or Nelson–Aalen estimator of the survival probability during interim analyses without necessitating trial suspension (cf., Jennison and Turnbull, ;, ). Although the author mentions the use of asymptotic testing based on the Kaplan–Meier estimator in this section, more discussion on these methods in the next edition of this book would help the readers achieve the broad sense of phase II cancer clinical trials with survival endpoint.…”