Dose escalation with over-dose and under-dose controls in Phase I/II clinical trials

Abstract

To save valuable time and resources in new drug development, Phase I/II clinical trials with toxicity control and drug efficacy as dual primary endpoints have become increasingly popular. Escalation with over-dose control (the EWOC) is a Bayesian adaptive Phase I clinical trial design that can accurately estimate the maximum tolerated dose (MTD) level and control the probability of overdosing patients during the dose allocation phase. In this paper, we extend EWOC to Phase I/II clinical trials by controlling for under-dosing with a Gumbel Copula model to provide patients with at least minimum drug efficacy. We propose a utility function to measure the composite effect of toxicity and efficacy and select the optimal dose. To deal with the common issue that the efficacy endpoint often cannot be quickly ascertained, we employ Bayesian data augmentation to handle delayed efficacy and allow for flexible patient accrual without a waiting period. Extensive simulations demonstrate that the proposed new design not only provides better therapeutic effect by reducing the probability of treating patients at under-dose levels while protecting patients from being overdosed, but also improves trial efficiency and increases the accuracy of dose recommendation for subsequent clinical trials. We apply the proposed design to a Phase I/II solid tumor trial.

Two-Stage Design for Phase I–II Cancer Clinical Trials Using Continuous Dose Combinations of Cytotoxic Agents

“…For treatments where efficacy is ascertained in a relatively short period of time such as one or two cycles of therapy, sequential designs for updating the joint probability of toxicity and efficacy and estimating the optimal dose have been studied extensively in the literature for single‐agent trials; see for example Murtaugh and Fisher (), Thall and Russell (), Braun (), Ivanova (), Thall and Cook (), Chen et al . () and Sato et al . ().…”

A Bayesian hierarchal modeling approach to shortening phase I/II trials of anticancer drug combinations

“…Assuming r > 0 and c > 0, the gamma process with mean rt and variance rt / c [ GP ( c × rt , c )] is set as the prior stochastic process of the cumulative hazard function Λ ( t ). Under these conditions, the posterior stochastic process for Λ ( t ) becomes a gamma process, which has Ga [ c × r ( τ s − τ s‐ 1 ) + N + ( τ s ), c + R + ( τ s )] with an independent increment in the interval [ τ s‐ 1 , τ s ], where N + ( τ s ) is the sum total of N mi ( t ) for all patients up to τ s , and R + ( τ s ) is the sum total of Y mi ( t ) for all patients up to τ s .…”

“…Because dose combinations assigned to subsequent cohorts are determined after confirming evaluations of toxicity and efficacy, complete evaluation of overall response is another potential factor resulting in extension of trial duration. In trial designs where toxicity and efficacy have been evaluated concurrently, methods to shorten trial duration have also been studied . Chen et al applied this design to a trial evaluating combination treatment with BKM120 and everolimus, targeting patients diagnosed with advanced solid malignancies.…”

“…In trial designs where toxicity and efficacy have been evaluated concurrently, methods to shorten trial duration have also been studied . Chen et al applied this design to a trial evaluating combination treatment with BKM120 and everolimus, targeting patients diagnosed with advanced solid malignancies. Unobserved efficacy data were treated as missing data, with a piecewise exponential model.…”

“…Using the piecewise exponential model makes it necessary to determine a valid number of intervals based on the assumption that the hazard function is constant at each interval. The design used by Chen et al did not allow changes in randomization probability for each treatment group within the trial duration.…”