A phase I‐II design based on periodic and continuous monitoring of disease status and the times to toxicity and death

Abstract

A Bayesian phase I-II dose-finding design is presented for a clinical trial with four coprimary outcomes that reflect the actual clinical observation process. During a prespecified fixed follow-up period, the times to disease progression, toxicity, and death are monitored continuously, and an ordinal disease status variable, including progressive disease (PD) as one level, is evaluated repeatedly by scheduled imaging. We assume a proportional hazards model with piecewise constant baseline hazard for each continuous variable and a longitudinal multinomial probit model for the ordinal disease status process and include multivariate patient frailties to induce association among the outcomes. A finite partition of the nonfatal outcome combinations during the follow-up period is constructed, and the utility of each set in the partition is elicited. Posterior mean utility is used to optimize each patient's dose, subject to a safety rule excluding doses with an unacceptably high rate of PD, severe toxicity, or death. A simulation study shows that, compared with the proposed design, a simpler design based on commonly used efficacy and toxicity outcomes obtained by combining the four variables described above performs poorly and has substantially smaller probabilities of correctly choosing truly optimal doses and excluding truly unsafe doses.

Risk–benefit trade-offs and precision utilities in phase I-II clinical trials

“…32,33 Toxicity can also be defined as a time-to-event variable monitored continuously over a prespecified follow-up period. 2,4,6,33 The effect of patient prognosis on a timeto-event outcome may manifest differently based on the estimand employed to characterize a treatment or dose effect, leading to potentially conflicting treatment evaluations. 3,13,34 As a simple illustration, we assume a proportional hazards model with an exponential distribution for the baseline hazard function in the numerical example below.…”

“…Although, for practical reasons, most phase I-II designs use early efficacy endpoints such as objective response evaluated within 30 days, modern dose-finding designs are beginning to incorporate more long-term clinical outcomes, such as overall survival or PFS time. 32 , 33 Toxicity can also be defined as a time-to-event variable monitored continuously over a prespecified follow-up period. 2 , 4 , 6 , 33 The effect of patient prognosis on a time-to-event outcome may manifest differently based on the estimand employed to characterize a treatment or dose effect, leading to potentially conflicting treatment evaluations.…”

Confirmatory adaptive group sequential designs for single‐arm phase II studies with multiple time‐to‐event endpoints

“…As in Lee et al. (2020) we assume that we can observe toxicities ($$T^{\lbrace 1\rbrace }$$), progressions ($$T^{\lbrace 2\rbrace }$$) and death ($$T^{\lbrace 3\rbrace }$$). In the spirit of a competing risk analysis (see, e.g., chapter 3 of Beyersmann et al., 2011) we define two events characterized by the sets $$E_1=\lbrace 1,3\rbrace$$ for toxicity and $$E_2=\lbrace 2,3\rbrace$$ for efficacy.…”

Generalized phase I‐II designs to increase long term therapeutic success rate

“…Many phase I designs 6–9 and phase I‐II designs 3,10–16 have been proposed. Most of these designs use binary outcomes, although phase I‐II designs have been proposed for ordinal outcomes, 17 event times, 18,19 or more than two outcomes 5,20 …”

“…Most of these designs use binary outcomes, although phase I-II designs have been proposed for ordinal outcomes, 17 event times, 18,19 or more than two outcomes. 5,20 A practical requirement in dose finding trials is that the outcomes, Y T or Y = Y R , Y T ð Þ, must be evaluated over a time period, 0, t 1 ½ , short enough to avoid delaying accrual to evaluate previous patients' outcomes to make adaptive decisions. For designs based on event times, such as the phase I time-to-event continual reassessment method (TITE-CRM) 21 or late-onset efficacy-toxicity phase I-II design, 22 follow up intervals must be short enough so that outcomeadaptive decisions can be made without unduly suspending accrual.…”